article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
many observables in nuclear collisions are difficult to calculate analytically because the number of particles is neither large enough to justify rigorously the application of statistical mechanics nor small enough to justify impulse approximations . in most cases , numerical simulations of transport equations
are required to compare theory with experiment . however , due to the complexity of the algorithms employed and the number of untested dynamical assumptions @xcite , it is not straight forward to check or even reproduce the numerical results from parton cascade codes .
recently , several analytic tests were proposed to help check the accuracy and identify limitations of such codes . in the context of earlier non - relativistic transport models tests such as comparing with the analytic krook - wu model @xcite
have been useful . in the context of hydrodynamics ,
numerical hydrodynamic codes have been tested in cases of expansion of baryon free matter into vacuum @xcite and slab - on - slab " collision @xcite . to address the new generation of parton cascade , a new open standard for codes and routines ( oscar)@xcite has been developed to enable objective testing of essential components of algorithms and ensure reproducibility of numerical results . in this context
, we have proposed in ref.@xcite a dynamical test of the evolution of the transverse energy in @xmath5 dimensional expansion for cascade models .
another test of the frame and scattering scheme dependence of cascade models was proposed in @xcite . in this paper , we propose two further tests for the oscar standard : the equation of state and the collision rate test of one - component relativistic classical gases .
the tests provide information about the nature of the code s initial momentum distribution and check its evolution algorithm in free expansion in periodic spatial inhomogeneities against nontrivial analytic expressions . the transverse to parallel momentum flux ratio tests the evolution of space - time and momentum space correlations .
the asymptotic homogeneous equilibrated state tests the equation of state of the model against the expected ideal gas laws .
the collision frequency test checks the basic collision algorithm .
we apply these tests to the newly developed zpc parton cascade model @xcite . in order to pass these tests
, we uncovered and fixed a problem with one of the random number generators used in zpc .
we discuss that example in detail to illustrate the importance of applying such numerical tests .
we propose to add these tests to the group already in the oscar standard@xcite .
the paper is organized as follows : in section 2 , we recall basic statistical mechanics relations of ideal gas and calculate the free streaming evolution of periodic slab initial spatial distribution .
in addition the collision rate is discussed .
numerical tests of zpc are compared to analytical predictions in section 3 .
we conclude by emphasizing the importance of testing cascade codes .
the partition function for an ideal relativistic classical gas with mass @xmath6 and degeneracy @xmath7 is given by @xcite ( @xmath8 ) @xmath9 in the @xmath10 limit , the free energy , @xmath11 is well approximated by @xmath12 the pressure , @xmath13 is given by the well known ideal gas law @xmath14 in terms of the density @xmath15 , and the energy density is given by @xmath16 where @xmath17 .
for fixed @xmath18 an interesting quantity that reflects the softness of the equation of state is @xmath19 in the @xmath20 relativistic limit , @xmath21 , while in the nonrelativistic limit , @xmath22 , in a cascade simulation , these thermodynamic quantities can be measured as a function of time by computing the spatially averaged energy - momentum tensor @xmath23 then @xmath24 and @xmath25 . in section 3
, we will compare the analytic results of the pressure - energy density ratio as a function of @xmath26 , and the energy density as a function of @xmath27 .
a simple test of equilibration in a _ periodic _ box of volume @xmath28 and fixed particle number is provided by taking the initial spatial distribution to be confined to one half of the box ( say with @xmath29 ) : @xmath30 ) f(p ) .
\label{f0}\ ] ] in the above , @xmath31 is the equilibrium particle density , @xmath32 is the length of the box , and @xmath33 for the free streaming case , @xmath34 . for the interesting case of massless partons , the momentum flux component
, @xmath35 evolves from its initial value @xmath36
\label{txx0}\ ] ] to its final value @xmath37 via @xmath38 here @xmath39 , @xmath40 , and @xmath41 .
the infinite sum can be expressed in terms of the lerch @xmath42 functions , but in practice the series converges very rapidly .
for example the evolution at the `` midway '' points @xmath43 involves damped oscillations with maxima and minima occuring at times @xmath44 . at @xmath45 , the @xmath46 " momentum flux oscillates within the envelops @xmath47 reaching the envelop at times @xmath44 .
note that free streaming in this geometry leads to global homogeneous equilibrium through mixing from neighboring slabs .
the deviation of the parton cascade evolution from this result when collisions terms are turned on are of physical interest as tests of collective hydrodynamic behavior .
another interesting probe of the evolution is the transverse - longitudinal momentum flux anisotropy , @xmath48 initially in the @xmath49 region we set @xmath50 while in the @xmath51 region @xmath52 . in the case of free streaming ,
this anisotropy is given by @xmath53 we note that of course other non - equilibrium initial conditions , e.g. , homogeneous spatial distribution with anisotropic momentum distribution , can be set up to test dynamical relaxation toward global equilibrium . in this paper
we will study only the evolution of spatial inhomogeneous conditions .
the collision rate per unit volume @xmath54 is another quantity that can be easily monitored .
it is related to the number of collisions @xmath55 in a period of time @xmath56 via : @xmath57 we take the integrated parton elastic cross section to be : @xmath58 the total collision rate per unit volume can be calculated from : @xmath59 and it is expected to be independent of the microscopic form of the differential cross section .
the @xmath60 on the right hand side of equation ( [ rate2 ] ) takes into account the two identical incoming particles .
the identity of two incoming gluons has been taken into account in the scattering cross section . in appendix
a , we show that the above 6-dimensional integral can be reduced to a 1-dimensional integral : @xmath61 in which @xmath62 .
for the cascade calculations , we take the temperature @xmath63 gev , and set the total number of particles in the periodic box to be @xmath64 .
the volume is related to the density of particle at zero - chemical potential , @xmath65 via @xmath66 . in zpc , there are @xmath67 computation cells of volume @xmath68 .
we specify the cell size to the cascade code . to study the mass dependence , we choose @xmath69 different masses : @xmath70 , @xmath71 , and @xmath72 gev .
for the dependence on the scattering cross section , the interaction length @xmath73 is set to be @xmath74 , @xmath75 , and @xmath76 . in the above
@xmath75 is a rough estimate for the mean free path , @xmath77 we specify the screening mass @xmath78 to the cascade code .
the parameters for the cascade measurement are shown in table ( [ para1 ] ) .
1 shows the pressure - energy density ratio @xmath79 as a function of @xmath26 , and fig . 2 shows the energy density as a function of temperature @xmath27 for different particle masses .
we see very good agreement between the predictions and the cascade results .
this indicates that the initial momentum distributions are correctly generated .
there is no time dependence of the pressure and energy density over a time period of @xmath80 fm during which each particle has experienced @xmath81 collisions on average . for periodic slab initial conditions , @xmath35 evolution at position @xmath82 in the free streaming case
is compared with prediction in fig .
3 . also shown in fig .
3 is the result ( filled circles ) with interactions turned on .
we see that interactions reduce the rate of collective expansion compared to free streaming and also the oscillations damp faster than in the free streaming case . for the non - interacting free streaming case , good agreement of prediction (
dashed line ) and cascade results ( pluses ) is shown for the @xmath35 evolution at @xmath82 in fig . 3 and time evolution of free streaming @xmath35 spatial distribution in fig . 4 .
time evolution of free streaming @xmath83 spatial distribution ( fig .
5 ) also agrees well with the prediction .
fig . 6 ,
7 give scaled collision rate per unit volume as a function of @xmath26 .
the data points with the same @xmath26 , and same @xmath78 overlap .
the three data sets at the same @xmath26 correspond to three different screening masses @xmath78 .
the smaller the screening mass , the lower the collision rate .
the collision rate is the same for the zpc default yukawa type scattering differential cross section and the straight line propagation .
this indicates the collision number loss is not due to particle shielding . with small screening mass
, the interaction range is large .
when the interaction range is much larger than the mean free path , non - causal collisions become more abundant . to process the non - causal collision
, we pick up one collision out of several collisions for one particle according to the ordering time ( see ref .
@xcite for details ) .
this process neglects other collisions in the same non - causal collision set .
some of them will not be recovered later .
the larger the interaction range , the larger percentage of non - causal collisions . hence ,
more collisions are neglected and the collision rate is lower than expected . in the dilute limit ,
the percentage of non - causal collisions out of total number of collisions for massless particles is proportional to the number of particles inside the causal sphere in the two colliding particle center of mass frame .
the radius of the sphere is proportional to the impact parameter .
so @xmath84 @xmath85 is the probability of having impact parameter @xmath86 . here
, @xmath87 is for the averaged boost factor from lab frame to the two colliding particle center of mass frame . in the case of massless particles , @xmath88 .
@xmath89 gives the number of particles in the sphere with radius @xmath86 in the two colliding particle center of mass frame .
the exact radius of the sphere depends on the definition of non - causal collisions @xcite and the collision prescription of cascade .
the non - causal to total ratio is closely related to the ratio of interaction length to the mean free path : @xmath90 we see @xmath91 decreases linearly with @xmath92 as @xmath93 .
this motivates the algorithm of reducing the non - causal collision percentage @xcite by subdividing the particles by a factor @xmath94 so that @xmath95 while decreasing @xmath96 .
this preserves the mean free path @xmath97 while @xmath98 .
the number of non - causal collisions , and total number of collisions for @xmath99 particles with @xmath100 , @xmath101 gev , and a time period of @xmath80 fm is summarized in table ( [ para2 ] ) . in the table
, @xmath78 is screening mass .
@xmath102 gives the ratio of interaction length to the estimate of mean free path before rescaling .
@xmath94 is the scaling parameter , i.e. , the total number of particles is increased by a factor of @xmath94 and the cross section is decreased by a factor of @xmath94 ( but the number of collisions is still for @xmath99 particles ) .
@xmath103 is the ratio of interaction length to the estimate of mean free path including the rescaling , i.e. , @xmath104 .
@xmath105 is the ratio of interaction length to the measured mean free path .
the mean free path is measured through the formula , @xmath106 , in which @xmath54 is the collision rate per unit volume and the particles are moving at the speed of light .
the ratio of the number of non - causal collisions to the total number of collisions is also plotted in fig . 8 against the ratio of interaction length to the mean free path . in fig . 8 , the open circles are data against @xmath103 and the filled circles are against @xmath105
it shows that when the density increases , i.e. , when @xmath92 increases , the difference between @xmath105 and @xmath103 increases .
this tells us that when density is high , the naive formula for the estimate of mean free path , @xmath107 , needs to be corrected .
also we see the data deviate from linear formula and has a tendency of saturation when density is large .
this is consistent with the fact that the ratio should always be less than @xmath108 from the definition of the non - causal collision . when we fix @xmath78 and increase @xmath94 from @xmath109 to @xmath72 , the total number of collisions goes up from @xmath110 for @xmath111 to @xmath112 for @xmath113 .
for @xmath114 , the total number of collisions is @xmath115 .
we see clearly the trend toward a constant value of total number of collisions when @xmath94 is increased .
the collision rate with @xmath114 is shown in fig .
6 as an open triangle ( the @xmath111 data is shown as an open circle ) .
the collision rate with @xmath114 is within @xmath116 of the analytic prediction . during the preliminary study of the collision rate
, we found when @xmath117 , the collision rate is higher than the predicted rate . by looking more carefully into the code , we found out that it was caused by the larger than statistical fluctuations in the position distribution . when the fugacity @xmath75 is not uniform in space , the collision rate per unit volume
: @xmath118 in which : @xmath119 by using the inequality , @xmath120 and the fact that the system we prepared has zero chemical potential and hence averaged fugacity is one , we arrive at : @xmath121 the equality holds only when the fugacity has no spatial dependence .
this shows that when there are space clusters existing for some time period , the collision rate is higher than that expected for the uniform system .
it was found only in @xmath117 and not in other cases because in the @xmath117 case , particles are moving very slowly , and they stay in clusters for much longer time .
we traced the origin of the non - uniform distribution .
it was caused by some correlation of random number generators ( see appendix b ) . when we use ran1 from @xcite , and generate first the momenta for all the particles and then generate the positions for all the particles , there are no abnormal fluctuations .
when we generate momentum and position together , we found abnormal fluctuations .
this does not occur when ran3 from @xcite is used .
we correct the generation by separating the generation of particle momentum and particle position .
from the above study , we show that the equation of state and the collision rate can be used to test the initial conditions and collision mechanisms of relativistic parton cascade . for massless particles , when the interaction range is much larger than the mean free path , the cascade collision rate is lower than the theoretical value .
other methods , e.g. , particle partition , have to be used to correct the collision rate .
the comparison of free streaming and interacting cascade approach to equilibrium indicate qualitative similarities of the two cases . however , the damping and speed of collective motion are quite different
. a detailed comparison of free streaming , ideal hydrodynamics , and cascade approach to global equilibrium in the case of half filled periodic box initial conditions will be addresses in another paper @xcite . as discussed in this paper , spatial distribution with larger than statistical fluctuations gives higher than thermal reaction rate .
hijing @xcite predicts initial spatial clusters of partons for nucleus - nucleus collisions at collider energies . this implies higher than thermal collision rates
@xcite and many other interesting physical phenomena beyond the widely used hot gluon scenario @xcite predictions .
we emphasize the importance of using analytic tests in debugging numerical simulation codes .
the spatial distribution with abnormal fluctuations illustrates well the usefulness of the analytic collision rate test .
as more components are added to the cascade code , more tests will be needed to ensure the consistency of different parts of the cascade code and to enable disentangling of the actual physical assumptions that define the model . + * acknowledgments * + we thank s. a. chin , p. danielewicz , v. koch , b. a. li , s. pratt , j. randrup for useful discussions .
we also thank brookhaven national laboratory and lawrence berkeley laboratory for providing computing facilities .
to calculate the collision rate per unit volume , equation ( [ rate2 ] ) , we first replace @xmath122 by @xmath123 and get : @xmath124 the delta function can be used to integrate out one angle by using : @xmath125 after carrying out the angular integrals , the collision rate becomes @xmath126 the @xmath127 function constraint can be written as : @xmath128 now we change integration variables from @xmath129 , @xmath130 to @xmath131 , and @xmath132 .
the collision rate becomes : @xmath133 the integration over @xmath134 can be carried out first .
the absolute lower bound of @xmath131 is @xmath135 , but @xmath134 is also restricted by the @xmath127 function . by noticing @xmath136 , and hence @xmath137 ,
we see the lower bound should be @xmath138 .
the upper bound is determined from the square root in the integrand to be @xmath139 .
the result is : @xmath140 now the collision rate per unit volume is reduced to a 2-dimensional integral : @xmath141 the @xmath142 integral can be readily carried out to be : @xmath143 in the above , @xmath144 is the modified bessel function .
taking @xmath145 , we arrive at equation ( [ rate3 ] ) .
the following is the a sample program that generates the particle momentum and position together .
ran1 is taken from @xcite and declared as real*8 function .
program dist implicit real*8 ( a - h , o - z ) parameter ( mul = 40000 ) parameter ( size = 5.19653401d0 ) external ran1 common /para/
xmp temp = 0.5d0 xmp = 5d0 do i = 1 , mul call energy(e , temp ) call momentum(px , py , pz , e ) x = 2d0 * ran1(iseed ) - 1d0 x = x * 5d0 * size y = 2d0 * ran1(iseed ) - 1d0 y = y * 5d0 * size z = 2d0 * ran1(iseed ) - 1d0 z = z * 5d0 * size end do stop end subroutine energy(e , temp ) implicit real*8 ( a - h , o - z ) external ran1 common /para/
xmp 1000 continue e = ran1(iseed ) e = e * ran1(iseed ) e = e * ran1(iseed ) if ( e .le .
0d0 ) goto 1000 e = - temp * log(e ) if ( ran1(iseed ) .gt . &
exp((e - sqrt(e * * 2 + xmp * * 2))/temp ) ) then goto 1000 end if return end subroutine momentum(px , py , pz , e ) implicit real*8 ( a - h , o - z ) external ran1 parameter ( pi = 3.14159265358979d0 ) cost = 2d0 * ran1(iseed ) - 1d0 sint = sqrt(1d0 - cost * * 2 ) phi = 2d0 * pi * ran1(iseed ) px = e * sint * cos(phi ) py = e * sint * sin(phi )
pz = e * cost return end .... fig .
10 gives the position @xmath134 distribution from the above program using ran1 and ran3 .
ran1 result has much larger fluctuations than the expected statistical fluctuations , while ran3 result is consistent with the expected statistical fluctuations .
when the generation of momentum and position are separated , the distributions all have reasonable fluctuations . another possible ( and more efficient ) way to solve the problem is to generate gaussian momentum distribution for large @xmath26 values . since with the above program , the total number of random number used is @xmath146 , while we can easily show the period of ran1 with the given parameters is beyond @xmath147 , it is not clear that the large fluctuations are due to the period of the random number generator ran1 .
but this is a concrete example of ill - generated distribution .
.input parameters to zpc : cell size and screening mass for different temperatures and masses of particles .
@xmath148 , @xmath149 , @xmath150 correspond to three decreasing screening masses as described in the text.[para1 ] [ cols="^,^,^,^",options="header " , ] | we develop two further numerical tests of parton cascade models : the ideal gas equation of state and the collision frequency tests . the equation of state test checks the initial momentum distribution generator and free expansion dynamics in periodic inhomogeneous geometries .
the collision rate test is sensitive to spatial inhomogeneities and the collision algorithm .
these tests are applied to the recently developed zpc parton cascade model .
the tests helped uncover unphysical correlations induced by one of the commonly used random number generators and showed the necessity of particle subdivisions for convergence to the exact analytic limit .
cu - tp-875 * equation of state and collision rate + tests of parton cascade models@xmath0 * + bin zhang@xmath1 , miklos gyulassy@xmath1 , and yang pang@xmath2 + @xmath1__physics department , columbia university , _ _ + _ new york , ny 10027 , usa _ + @xmath3__brookhaven national laboratory , _ _ + _ upton , new york , 11973 , usa _ + ( 500,5)(0,0 ) ( 1,2)(1,0)180 @xmath4this research was supported by the u.s .
department of energy under contract no .
de - fg02 - 93er40764 .
y. pang was partially supported by the u.s .
department of energy grant de - fg02 - 92er40699 , de - ac02 - 76ch00016 , and alfred p. sloan foundation . |
discovery of coherent oscillations with frequencies in the 300600 hz range during x - ray bursts from a number of low - mass x - ray binaries ( see strohmayer & bildsten 2003 for a review ) has triggered the efforts to use the information on the amplitude of variability and the folded pulse shape to put constraints on the compactness of the neutron star and thus its equation of state as well as the emission pattern from the neutron star surface ( e.g. miller & lamb 1998 ) .
however , a rather limited photon statistics does not allow to reach a high accuracy ( strohmayer et al . 1997 ; nath , strohmayer & swank 2002 ; muno , zel & chakrabarty 2002 ) .
the millisecond coherent pulsations discovered in the persistent emission of the four sources : ( with the period @xmath9 ms ; wijnands & van der klis 1998a ) , xte j1751@xmath8305 ( @xmath10 ms ; markwardt et al .
2002 ) , xte j0929@xmath8314 ( @xmath11 ms ; galloway et al . 2002 ) and xte j1807@xmath8294 ( @xmath12 ms ; markwardt , smith & swank 2003 ) allow one to increase the statistics by folding the profile over a longer observational period ( days rather than seconds in the x - ray burst oscillations ) .
thus , for example , showed coherent pulsations with rms amplitude of 57 per cent during its april 1998 outburst ( cui et al .
1998 ) with almost constant shape of its energy spectrum ( gilfanov et al . 1998 ) and very similar pulse profiles .
the pulse profiles of show strong energy dependence with soft photons lagging behind the hard ones ( soft phase lags , see cui et al .
1998 ) . by analysing the phase - resolved spectra , gierliski , done & barret ( 2002 ) showed that this results from the fact that the two main spectral components , a soft black body and a hard comptonized , vary out of phase with the first lagging the last one .
while the variations of the black body flux can be described by a single sine wave , the pulse of the hard comptonized component is strongly skewed .
gierliski et al .
( 2002 ) suggested that the doppler boosting can play a role in changing the shape of the profile .
the radiation pattern of standard accreting x - ray pulsars is influenced by strong magnetic field @xmath13 g. on the other hand , the magnetic field of the accreting millisecond pulsars in low mass x - ray binaries is much weaker , @xmath14@xmath15 g ( see e.g. wijnands & van der klis 1998a ; psaltis & chakrabarty 1999 ) , and does not affect significantly properties of the emitted radiation . thus these sources can also serve as laboratories for studying radiative processes at the surface of weakly - magnetized neutron stars . in this paper , we construct a detailed model for the x - ray emission from the surface of a rapidly rotating neutron star accounting for relativistic effects .
we compare the model with the data on and put constraints on the inclination of the system , the position of the emitting region relative to the rotational pole and its emission pattern as well as the stellar radius .
the data used for the analysis are described in [ sec : data ] .
the model and useful analytical formulae for light curve calculations are presented in [ sec : model ] .
the main results are given in
[ sec : results ] .
the discussion and the summary are given in
[ sec : disc ] and [ sec : summary ] , respectively .
we study the data obtained by the _ rossi x - ray timing explorer _ ( ) during the april 1998 outburst of , using ftools 5.2 . to improve the statistics we average the data between 1998 april 11 and 29 ( resulting in 118 ks of data ) . for spectral
fitting we extract the proportional counter array ( pca ) spectra from all five detectors , top layer only , and use the data in 320 kev energy band .
we also extract spectra from high - energy x - ray timing experiment ( hexte ) in the 20150 kev band , from clusters 0 and 1 .
we constructed the energy - dependent pulse profiles from the pca data ( all detectors , all layers ) by following the procedure described in cui et al .
( 1998 ) and gierliski et al .
( 2002 ) correcting the photon arrival times for orbital motions of the pulsar and the spacecraft . for pulse profile
fitting we created two folded light curves in 16 phase bins , for energies 34 kev and 1218 kev .
the statistical uncertainties are about the same as uncertainties in the background , resulting in errors of 0.2 per cent and 0.3 per cent , respectively , in the count rate in the two considered energy bands .
accreting matter following the magnetic field lines close to the neutron star is stopped in the very vicinity of the surface by radiation producing radiation dominated shock ( basko & sunyaev 1976 ; lyubarskii & sunyaev 1982 ) . for the source luminosity of a few percent of the eddington luminosity ( gilfanov et al .
1998 ; gierliski et al . 2002 ) , the vertical ( i.e. along the radial direction ) extension of the shocked plasma is smaller than a characteristic horizontal size and certainly smaller than the stellar radius . the hard x - rays produced in the shock can irradiate the surrounding stellar surface , so that the black body emission region can cover a somewhat larger area . for our calculations
we assume that all photons originate at the stellar surface ( i.e. the height of the emitting region is set to zero ) .
we assume a circular spot and consider two extreme cases for the relative positions of the black body and the comptonizing regions : ( 1 ) a homogeneous slab , i.e. the hot comptonizing layer covering the whole black body spot , ( 2 ) a point like comptonizing region in the centre of a black body region . in order to compute the light curve as observed by a distant observer
, we first specify the radiation spectrum and angular dependence in the frame co - rotating with the star .
we then make lorentz transformation to obtain the radiation intensity in the non - rotating frame close to the stellar surface . at the final step ,
we follow photon trajectories in the schwarzschild spacetime to infinity .
let us now consider a star with an azimuthally symmetric surface radiation intensity @xmath16 that can vary over the surface , where @xmath17 is the angle between an emitted photon and the local normal to the stellar surface . let @xmath18 be a surface element ( spot ) at colatitude @xmath19 ( see fig .
[ fig : geom ] for the geometry ) .
the primed quantities correspond to the frame co - rotating with the spot .
let @xmath20 and @xmath21 be unit vectors pointing from the star centre towards the observer and the spot , respectively , and @xmath22 be the inclination of the spin axis .
the inclination of the spot varies periodically [ eq : psi ] = = i + i , where the phase @xmath23 , with @xmath24 being the pulsar frequency and @xmath25 is chosen when the spot is closest to the observer .
we compute the original direction of the photon @xmath26 near the stellar surface ( which is transformed to @xmath20 at large distance from the star ) assuming schwarzschild geometry where photon orbits are planar : [ eq : k0 ] _ 0= [ + ( - ) ] / , where @xmath27 is the angle between @xmath26 and @xmath21 , i.e. @xmath28 . the relation between @xmath27 and @xmath29 ( i.e. light bending ) can be obtained by standard techniques ( pechenick , ftaclas & cohen 1983 ) : [ eq : bend ] = _
r^ ^-1/2 , where the impact parameter [ eq : impact ] b= , @xmath30 is the schwarzschild radius , @xmath31 is the mass and @xmath32 is the radius of the compact star .
the observed flux at energy @xmath33 is @xmath34 , where @xmath35 is the radiation intensity at the infinity and @xmath36 is the solid angle occupied by @xmath18 on the observer s sky .
the solid angle can be expressed through the impact parameter [ eq : omegab ] = b b /d^2 , where @xmath37 is the distance to the source and @xmath38 is the azimuthal angle corresponding to rotation around @xmath20 .
the impact parameter @xmath39 depends on @xmath29 only , but not on @xmath38 .
the apparent area of the spot as measured by photon beams in the non - rotating frame near the stellar surface is @xmath40 ( see terrell 1959 ; lightman et al .
1975 ; ghisellini 1999 ) and the relation between @xmath27 and @xmath17 is described by the relativistic aberration formula ( for motions parallel to the spot surface ) [ eq : aberr ] = , where @xmath41 is the doppler factor .
thus the spot area projected on to the plane perpendicular to the photon propagation direction , i.e. a photon beam cross - section , is lorentz invariant ( see e.g. lightman et al . 1975 ; lind & blandford 1985 ) : [ eq : salpha ] s = s .
representing @xmath42 and using equations ( [ eq : impact ] ) and ( [ eq : salpha ] ) we get from ( [ eq : omegab ] ) [ eq : omega ] = .
the doppler factor can be expressed as follows [ eq : dop ] = 1/[(1- ) ] , where @xmath43 and @xmath44 is the spot velocity as measured in the non - rotating frame , [ eq : beta ] = = . here
@xmath45 is the velocity at the equator and @xmath46 is the angle between the spot velocity and @xmath26 .
the pulsar frequency is corrected here for the redshift @xmath47 . using equation ( [ eq : k0 ] )
it is easy to show that [ eq : cosxi ] = _ 0 = = - i . in the case of a weak field and negligible bending the solid angle occupied by the spot
is @xmath48 . in reality
, this formula can be applied even when bending is significant since the relation between @xmath49 and @xmath50 is close to linear ( beloborodov 2002 ) , [ eq : cosbend ] /r + ( 1-/r ) , for a star with @xmath51 , so that @xmath52 . the radiation intensity observed at
the infinity is @xmath53 , where @xmath54 is the redshift ( misner , thorn & wheeler 1973 ) and the energy @xmath55 is measured in the non - rotating frame near the stellar surface .
( in the actual calculations , one can neglect the redshift factor in all formulae for photon energies , since it is the same for the photons originating at the stellar surface . )
now we can make lorentz transformation to the spot rest frame where we know the angular and energy distribution of the radiation field @xmath16 .
the intensities are related via @xmath56 with the doppler shift given by @xmath57 .
combining equations above we get the expression for the observed flux : [ eq : fluxspot ] f_e= ^3 i ( e, ) s , with the visibility condition @xmath58 .
we account thus here for the special relativistic effects ( doppler boosting , relativistic aberration ) , the gravitational redshift and light bending in schwarzschild geometry . in order to evaluate the flux at a given phase @xmath59 , one first computes @xmath29 using equation ( [ eq : psi ] ) , then inverting ( [ eq : bend ] ) gets @xmath27 and substitutes the results into equations ( [ eq : cosxi ] ) and ( [ eq : dop ] ) to obtain the doppler factor which is used later to compute @xmath17 from equation ( [ eq : aberr ] ) .
the light curve from the antipodal spot can be easily obtained by replacing in all formulae @xmath60 and @xmath61 .
we note here that the formalism outlined above neglects the time delays due to different photon paths ( see pechenick et al .
1983 ) which can be easily incorporated
. however , they are negligible comparing to the pulsar period above a few ms . for a spot of the finite size
, the observed flux can be obtained by integrating over the ( visible ) spot area . on the other hand ,
if we know from the observations the black body flux @xmath62 , its temperature @xmath63 and the distance to the source @xmath37 , one can estimate the size of the ( black body ) emitting region .
the `` observed '' radius of the spot @xmath64 should be corrected for the effects of gravitation and orientation ( using a procedure described in
[ sec : appr ] , see also psaltis , zel & dedeo 2000 ) to obtain the actual spot area .
in our calculations of the model light curves for we use the distance @xmath65 kpc ( int zand et al .
2001 ) , @xmath66 kev , and @xmath67 km ( see fig . [
fig : speave ] and gierliski et al . 2002 ) .
since the observed profiles from have almost sinusoidal shapes , they can not be produced by two antipodal spots .
therefore , we compute the light curves from the primary spot ( closest to the observer ) only
. we also neglect multiple images since neutron or strange stars have sizes larger than @xmath68 ( see e.g. lattimer & prakash 2001 ) . in order to compute a light curve @xmath69 at a given energy @xmath33
, one needs to specify the intrinsic radiation pattern in the spot rest frame @xmath16 . in a slab geometry of the emitting region ,
the black body photons come from the bottom of the hot comptonizing slab .
the black body intensity transmitted through the slab of thomson optical thickness @xmath70 is @xmath71 , where @xmath72 is the zenith ( polar ) angle measured from the slab normal , @xmath73 is the electron concentration , @xmath74 is thomson cross - section , and @xmath75 is the height of the slab .
the radiation flux from a unit area is then strongly peaked along the normal direction ( a `` knife''like emission pattern , see curves marked by 0 at fig .
[ fig : ray ] ) .
the scattered radiation is expected to have a very different angular distribution , which depends on the scattering order ( and thus on photon energy ) .
the radiation flux does not peak anymore in the normal direction , but instead peaks at some intermediate angle ( a `` fan''like emission pattern , see solid curves marked by @xmath76 in fig . [
fig : ray ] ) .
it approaches the asymptotic distribution @xmath77 in a few scatterings ( the exact number is a function of the optical depth ) .
we compute the angular distribution of radiation following procedure described in sunyaev & titarchuk ( 1985 ) . the intensity of the black body radiation in the frame co - moving with the spot
can be represented in the following way : [ eq : bb ] i^bb_e()(-/ ) f_bb(e ) , where @xmath78 is the observed time - averaged black body flux ( i.e. the best - fitting bbodyrad model spectrum , see fig .
[ fig : speave ] ) .
we assume that the energy and angular dependences of the high energy ( scattered many times ) radiation can be separated as [ eq : sc ] i^sc_e( ) i_(,)f_sc(e ) . here
@xmath79 is the observed , time - averaged energy dependence of the comptonized flux ( i.e. the best - fitting thcomp model spectrum , see fig . [ fig : speave ] ) . since the main spectral components of the time - averaged spectrum of are a power - law like comptonized component ( with weak compton reflection ) extending at least to 200 kev and a black body contributing about 30 per cent to the flux in the 35 kev region , ( see fig .
[ fig : speave ] and gilfanov et al . 1998 ; gierliski et al . 2002 ) , we compute the total observed flux as a function of phase @xmath69 at every energy @xmath33 summing up the contributions from the black body and the comptonized components .
we compute them independently and renormalise so that their phase - averaged values are equal to the observed fluxes @xmath80 ( see fig .
[ fig : speave ] ) .
if the black body spot is not covered completely by a comptonizing layer , some black body photons do not enter the comptonizing region and escape to the observer directly . in that case , the `` effective '' @xmath81 that describes the mean ( i.e. averaged over the spot ) angular dependence of the black body radiation can be smaller than the actual optical depth in the comptonizing region @xmath82 that describes the angular dependence of the comptonized radiation . since the exact geometry is unknown , we use @xmath81 and @xmath82 as two independent parameters for the light curve computations .
the emission pattern for the black body radiation is given by equation ( [ eq : bb ] ) while the comptonized radiation is described by equation ( [ eq : sc ] ) where instead of @xmath81 we use @xmath82 as a parameter for the data fitting .
these patterns are referred to as model 1 in
[ sec : results ] and table [ table : fit ] . in order to check the robustness of the results and their dependence on the assumed angular distribution of radiation , we consider a second model where instead of @xmath83 we use a linear function @xmath84 with @xmath85 as a free parameter ( limited by @xmath86 from below to ensure positiveness of the function ) .
the corresponding fits are referred below as model 2 .
thus , the model parameters are the neutron star mass @xmath31 , its radius @xmath32 , rotational frequency @xmath24 ( @xmath87 hz for ) , inclination @xmath22 , colatitude of the spot centre @xmath19 , the optical depth @xmath81 and a parameter determining the angular distribution of the comptonized radiation @xmath82 or @xmath85 ( plus a free phase factor ) . using the formalism developed by beloborodov ( 2002 )
we derive here simple analytical formulae for the light curves and the oscillation amplitudes that could be useful for understanding the main physical effects .
let us first consider a small spot and make approximation to the light bending formulae
. the bolometric flux can be obtained from equation ( [ eq : fluxspot ] ) by integrating over @xmath33 and using the cosine relation ( [ eq : cosbend ] ) which describes well the light bending near a star with @xmath88 : [ eq : df1 ] f&= & ( 1-/r)^2 ^4 i( ) s/d^2 + & = & ( 1- ) ^2 ^5 i( ) , with the visibility condition @xmath89 .
( for the antipodal spot , one should substitute @xmath90 and @xmath91 when computing the doppler factor . )
it is clear from equation ( [ eq : df1 ] ) that the ( bolometric ) black body flux observed from a spot at a rapidly spinning star is a factor @xmath92 times the flux from a slowly rotating star ( see fig . [
fig : relateff]a ) .
two powers of @xmath41 come from the solid angle transformation , one from the energy , one from the arrival time contraction , and the fifth from the change in the projected area due to aberration .
the scattered radiation , on the other hand , has an angular distribution which differs from that of the black body . in the optically thin case , it is closer to @xmath93 ( see fig . [
fig : ray ] ) .
equation ( [ eq : df1 ] ) for the observed flux can be rewritten then as [ eq : fsc ] f_sc=(1-/r)^2 ^4 fs/d^2 .
one sees that the flux is modified by approximately factor @xmath94 ( see fig .
[ fig : relateff]b ) .
the angular distribution of the hard radiation ( defined in our model by parameters @xmath82 or @xmath85 ) is reflected in the shape of the light curve at high photon energies , where comptonizing emission dominates .
since the comptonization spectrum is close to a power - law @xmath95 , with the same approximations as above , one can easily get the monochromatic flux at a given energy : f^sc_e=(1-/r)^3/2 ^2 + f(e_0 ) s/d^2 , which varies in a way similar to the bolometric flux if @xmath96 . neglecting the doppler effect and using equation ( [ eq : df1 ] )
one can relate the variability amplitude of the black body flux to the angles @xmath22 and @xmath19 and the neutron star radius ( beloborodov 2002 ) : [ eq : ampl ] a ( f_-f_)/(f_+f_)= u / q , where we defined [ eq : uq ] u&=&(1-/r ) i , + q&=&/r+(1-/r ) i .
this peak - to - peak amplitude is @xmath97 larger than sometimes quoted rms amplitude if the light curve is a pure sine wave . for a circular spot of an angular radius @xmath98 emitting as a black body
, the observed flux can be obtained from equation ( [ eq : df1 ] ) : [ eq : fluxrho ] f_= ( 1-)^2 r^2 ^2 ( q+ ^2 + u ) , if the spot is always visible , i.e. @xmath99 . here
the intensity is related to the local temperature @xmath100 .
we can easily now obtain the oscillation amplitude : [ eq : amplspot ] a_=. equation ( [ eq : fluxrho ] ) allows us to estimate the size of the spot @xmath101 from the equation : ^2=r^2 ^2 ( q+ ^2 ) , if the distance @xmath37 , the temperature @xmath102 and the observed phase - averaged black body flux @xmath103 ( and thus @xmath104 ) are known from observations . for small @xmath98 ,
the spot radius is @xmath105 .
we fit the observed pulse profiles in the energy bands 34 kev and 1218 kev simultaneously .
the second band is completely dominated by the comptonized radiation and thus gives direct information about its angular distribution . in the lower energy band
, the black body affects the observed pulse profile .
since the mass of the compact star in is not known , we consider masses between @xmath106 and @xmath2 .
masses in the range of @xmath107@xmath2 are expected for neutron stars in low mass x - ray binaries that have accreted material from their binary companions , lower masses are possible , for example , for strange stars .
we restrict the inclination to @xmath108 to be consistent with the absence of the x - ray eclipses ( bildsten & chakrabarty 2001 ) .
the best - fitting parameters are presented in table [ table : fit ] .
the two sets of errors correspond to the 90 and 99 per cent confidence intervals with @xmath109 and @xmath110 , respectively .
cllllccc @xmath31 & @xmath32 & @xmath111 & @xmath22 & @xmath19 & @xmath81 & @xmath82 or @xmath85 & @xmath112dof + ( @xmath113 ) & ( km ) & & ( deg ) & ( deg ) & & & + + @xmath114 & @xmath115 & @xmath116 & @xmath117 & @xmath118 & @xmath119 & @xmath120 & @xmath121 + @xmath1 & @xmath122 & @xmath123 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & @xmath128 + @xmath107 & @xmath129 & @xmath130 & @xmath131 & @xmath132 & @xmath133 & @xmath134 & @xmath135 + @xmath136 & @xmath137 & @xmath138 & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath143 + + @xmath114 & @xmath144 & @xmath145 & @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 + @xmath1 & @xmath151 & @xmath152 & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath157 + @xmath107 & @xmath158 & @xmath159 & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164 + @xmath136 & @xmath165 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 & @xmath171 + [ table : fit ] we see from table [ table : fit ] that the results obtained with the two models 1 and 2 are very consistent with each other . for both models , a large inclination is preferred and the lower limit is @xmath172 ( at 99 per cent confidence level ) for @xmath173 .
the fits become much worse at masses higher than @xmath2 .
the estimated stellar radius ( even in @xmath174 units ) grows with the mass and varies between @xmath175 and @xmath176 ( @xmath177 for model 1 ) .
using simple analytical estimations we explain the physical reasons behind this behaviour in [ sec : constr ] .
the radii are consistent with some of the neutron star equations of state if @xmath178 . for a smaller
assumed mass of @xmath179 , the radius is more consistent with the equations of state for strange stars , while for even smaller mass of @xmath180 , the obtained stellar radius of @xmath181 km is too small even for a strange star ( see [ sec : massrad ] and fig .
[ fig : eos ] for details ) .
the colatitude @xmath19 of the spot centre varies approximately as @xmath182 .
a larger @xmath19 is needed for smaller @xmath32 in order to keep the spot velocity ( and thus the doppler factor ) at approximately constant level .
the estimated radius of the spot ( not a fitting parameter ) corrected for orientation and gravitational effects varies between 3.0 and 3.7 km for @xmath106 and @xmath2 , respectively .
one can get some estimates of the size of the comptonizing region from a rather large difference between the best - fitting @xmath82 and @xmath81 obtained in model 1 .
the values of @xmath82 are consistent with the optical depths obtained from the fits of the time - averaged spectrum with a comptonization model for a slab geometry . for example
, model compps ( poutanen & svensson 1996 ) , gives @xmath183 for the electron temperature @xmath184 kev ( which is not well constrained because of the low signal above 100 kev ) .
this leads us to a conclusion that a large fraction of the black body photons reaches the observer without interaction with the hot comptonizing medium ( contradicting assumed homogeneous slab geometry ) . in order to get the `` effective '' @xmath185 ( see eq . [ [ eq : bb ] ] ) , about 40 per cent of the black body area should be covered by the comptonizing medium with @xmath186 , while the remaining 60 per cent should emit as a pure black body ( i.e. with no @xmath187 reduction due to scattering ) .
thus , we conclude that the radius of the comptonizing region @xmath188 is about 2 km . the angular distributions of the scattered radiation ( see dashed curve in fig .
[ fig : all]b ) determined by parameters @xmath82 and @xmath85 is also very similar for models 1 and 2 .
model 2 is somewhat less restrictive and therefore gives generally better fits .
the required emission pattern is certainly very much different from that given by the lambert law and more radiation is escaping at intermediate angles @xmath189 than along the normal .
it is this pattern which is responsible for the origin of a strong first harmonic ( i.e. at double spin frequency ) in the pulsar light curve at high energies ( modified also by doppler effects , see fig .
[ fig : relateff ] and [ sec : appr ] ) . in order to check the dependence of the results on the assumed geometry
, we repeated the fitting procedure considering an extreme case with a point - like comptonizing source at the centre of the black body spot .
the best - fitting parameters are similar to the situation where both emitting regions coincide .
this is expected since the size of the emission region should not play a large role since @xmath190 . however , when estimating the size of the spot , we assumed the soft emission is a black body , i.e. the observed color temperature is equal to the effective temperature . for the hydrogen and helium atmospheres of weakly magnetized neutron stars
one expects that the color correction is @xmath191@xmath192 ( zavlin , pavlov & shibanov 1996 ) .
the shift of the peak of the emission towards higher energies from the corresponding black body emission results from the fact that the outer layers of the atmosphere are cooler than the deeper layers .
since the bound - free and free - free cross - sections rapidly decrease with increasing photon energy , one sees deeper and hotter layers at high energies .
however , when the accretion takes place and the radiative shock forms ( its existence is supported by the presence of the hard x - ray tail ) , the soft radiation can be produced by reprocessing the hard photons from the shock .
the temperature gradient of the atmosphere then can be opposite to that of standard neutron star atmospheres .
now the hot layers are at the top , softer photons are coming from the hotter region ( since absorption cross - section is large for small energies ) , while hard photons are coming from the cooler region below and the color correction can be smaller than unity .
the actual vertical temperature dependence is however unknown . to investigate the influence of different possible color corrections on our best - fitting parameters , we consider two extreme cases of @xmath193 and @xmath194 and fitted the data with model 2 .
the area of the `` black body '' spot now changes by a factor @xmath195 . for @xmath193 ,
the spot centre colatitude @xmath19 increases by about 20 per cent comparing to the case of no color correction .
since now the spot is larger , @xmath19 has to increase to keep the same oscillation amplitude ( see eq . [ [ eq : amplspot ] ] ) .
the best - fitting stellar radius increases significantly only for the smallest considered mass of @xmath196 reaching @xmath197 km , for @xmath5 the radius is @xmath198 km , while for higher masses the change is negligible . in the case of @xmath199
the spot becomes smaller and @xmath19 also slightly decreases . at @xmath196 the best - fitting radius is now @xmath200 km , while all other parameters remain the same within the errors .
there is no change in the best - fitting parameters for higher stellar masses .
this results from the fact that the light curve depends weakly on the spot size if it small comparing to the stellar radius ( see eqs [ [ eq : fluxrho ] , [ eq : amplspot ] ] ) .
one of the important assumptions that can affect our results is that the two emission components are assumed to be co - spatial or are at least co - centred . in our interpretation the comptonized emission peaks earlier than the black body due to a different angular distribution more affected by the doppler effect .
however , an additional phase shift is possible if the comptonized emission region is physically shifted on the stellar surface .
we fitted the data with model 2 and the phase shift @xmath201 of the comptonized emission as an additional free parameter .
a major effect of adding a new parameter is that the confidence interval for the stellar radius increases .
thus , for example , for @xmath202 , the best fit with @xmath203/dof@xmath204 was obtained for @xmath205 rad ( a positive shift corresponds to an earlier arrival time ) and other parameters very similar to those presented in table [ table : fit ] , while the constraints on the radius weaken : @xmath206 ( 90 per cent confidence limits ) .
an @xmath207-test gives @xmath208 which corresponds to a 7 per cent chance that the decrease in @xmath203 was random ( i.e. the improvement is not highly statistically significant ) .
we show one of the fits to the pulse profiles with the fixed @xmath31 , @xmath22 , and @xmath32 in fig .
[ fig : all]a . the corresponding intrinsic angular distribution of the radiation flux in the two spectral components is shown in fig .
[ fig : all]b .
one sees a dramatic difference between the dependences of the black body and comptonized components .
the observed energy dependence of the pulse profiles and the phase lags at the pulsar frequency between different energies ( fig .
[ fig : all]c ) are reproduced reasonably well .
a natural consequence of the model is that the phase lags change significantly at smaller energies together with the black body contribution to the total flux , while they reach a constant value at about 68 kev where the black body flux becomes negligible ( see also gierliski et al .
2002 ) . since in reality
the angular dependence of comptonization radiation _ is _ a function of energy , one would expect a weaker dependence of lags on energy even above 10 kev .
let us consider first for simplicity the black body emission from a single spot . in that case , the doppler effect does not change significantly the light curve ( compare dashed and solid curve in fig .
[ fig : relateff]a ) and thus it is mainly characterized by the amplitude @xmath209 that can be estimated for a slowly rotating star using equation ( [ eq : ampl ] ) . for a given @xmath111 ,
the amplitude in its turn depends on @xmath22 and @xmath19 only .
thus , at the @xmath210 plane we can determine a curve that would correspond to a given amplitude @xmath209 ( see solid curves in fig . [
fig : dopamrin]a ) .
one notes that this curve is symmetric around the @xmath211 line since one can exchange @xmath22 and @xmath19 in the expression for @xmath209 ( see eqs [ [ eq : ampl][eq : uq ] ] ) without affecting the result .
if the black body emission is more beamed towards the spot normal because of a non - zero @xmath81 , this curve will be somewhat shifted down and left since @xmath209 then increases .
for example , the black body variability amplitude observed in is @xmath212 per cent ( gierliski et al .
2002 ) and , correcting for @xmath81 , we can determine the region at the @xmath210 plane where it can be reproduced for a given @xmath111 .
the situation is different when we consider a less beamed emission pattern @xmath213 with @xmath214 ( i.e. in our case radiation above @xmath215 kev ) .
the oscillation amplitude is then strongly affected by the doppler effect ( compare dashed and solid curve in fig .
[ fig : relateff]b ) .
the maximum of the doppler factor and the spot projected area are shifted in phase by @xmath0 0.25 of the period and the peak of the emission is then shifted towards the phase where @xmath41 has the maximum
. however , the amplitude and the shape of the profile carry the information needed to determine @xmath85 and the maximum doppler factor @xmath216 .
again at the @xmath210 plane we can determine the curve of constant @xmath216 ( see dashed curves in fig .
[ fig : dopamrin]a ) .
note , that since the doppler factor ( to the first order in @xmath217 ) depends on the product @xmath218 ( see eqs [ [ eq : dop][eq : cosxi ] ] ) , but not on @xmath22 and @xmath19 individually , this curve is almost symmetric around the @xmath211 line .
a very important point here is that the curves of constant @xmath209 and @xmath216 do not coincide ( see fig .
[ fig : dopamrin]a ) . at the @xmath210 plane
thus there are at most two points ( for a given star mass @xmath31 and radius @xmath32 ) with the given values of these two parameters : points of the crossing of the curves of constant @xmath209 and @xmath216 .
this leaves some ambiguity regarding the exchange of @xmath22 and @xmath19 .
it is possible , however , that the number of crossing points is one or even zero which would then mean that there is no physical solution with a given doppler factor and the variability amplitude within the considered model .
let us now apply our simple arguments to the data on .
the observed light curve implies a maximum doppler factor about 1.015 and the variability amplitude @xmath209 for a black body spot given by equation ( [ eq : ampl ] ) of @xmath219 ( actually observed higher value results from a sharper emission pattern with @xmath185 , see eq .
[ [ eq : bb ] ] ) .
the contours of constant @xmath209 do not depend on the stellar mass but only on the radius measured in @xmath174 ( see fig . [
fig : dopamrin]b ) .
we can now find a point at the @xmath210 plane that satisfy both constraints .
the curves of the constant @xmath209 and @xmath216 intercept in two points : at a large inclination and small @xmath19 as well as at a small inclination @xmath220 and large @xmath19 .
the second solution is much less probable .
first , we do not see any emission from the secondary antipodal spot .
this implies that either it is blocked by the star itself ( which in turn requires the stellar radius to be @xmath221 for @xmath222 and @xmath223 ) , or it is blocked by the accretion disc . in the later case
we have additional constraints . at inclinations smaller than @xmath224
an upper limit on the inner disc radius ( i.e. the magnetospheric radius ) required to block the secondary spot is @xmath225 ( see lower dotted curve in fig . [
fig : dopamrin]a ) .
with such a small inner radius , the expected amplitude of compton reflection component should be probably larger than the observed @xmath226 ( unless the inner disc is highly ionised ) and the shock should appear at uncomfortably large angle from the rotational pole if the magnetic field is a central dipole . on the other hand , at inclinations @xmath227 the antipodal spot is blocked by the disc if the magnetospheric radius is smaller than the corotation radius @xmath228 km which is also required for accretion to take place .
second , small @xmath22 is inconsistent with the x - ray ( chakrabarty & morgan 1998 ) and optical ( giles , hill & greenhill 1999 ; homer et al .
2001 ) modulations at the binary period , and finally , it is also inconsistent with the lower limit @xmath229 obtained by wang et al .
( 2001 ) based on the modelling of the optical / ir emission with the x - ray irradiated disc .
we conclude that large inclinations @xmath230 and small @xmath231@xmath232 are preferred to small @xmath22 and large @xmath19 . for small stellar masses
@xmath106 , the interception point for given @xmath209 and @xmath216 fixes the possible inclination @xmath22 between @xmath233 and @xmath234 depending on the assumed radius . the minimum possible radius still satisfying constraints on the inclination angle @xmath108 is @xmath235 . for higher masses , the minimum possible radius ( in @xmath174 units ) increases , reaching @xmath236 for @xmath202 , and for @xmath4
there is no solution whatsoever .
thus it is quite natural that the @xmath203 is worse for the high mass star ( see table [ table : fit ] ) .
this simple analysis also explains why the higher is the stellar mass , the higher the inclination should be in order to satisfy constraints on both @xmath209 and @xmath216 . our results ( see table [ table : fit ] and fig .
[ fig : eos ] ) put constraints on mass of the compact object . for @xmath196
the obtained stellar radius of @xmath181 km is too small to be consistent with any existing equations of state for neutron or strange stars .
for masses higher than @xmath2 , the fits are much worse than for smaller masses .
thus we conclude that the stellar mass is bound between 1.2 and 1.6 solar masses .
we can also put constraints on the equations of state of compact stars ( which of course could be much stronger if the mass of compact object were known ) .
thus , if @xmath5 the only possible solution is a strange star ( chakrabarty 1991 ; li et al .
1999 ; ray et al .
2000 ; gondek - rosiska et al . 2000 ) with the equation of state derived by dey et al . (
1998 ) . for the mass about @xmath237 ,
almost all known neutron star equations of state predict larger radii than our best - fitting value .
the exception are the models similar to bpal12 from prakash et al .
( 1997 ) and the models with a kaon condensate ( glendenning & schaffner - bielich 1999 ; see lattimer & prakash 2001 for more details ) . however , the maximum mass for these models is @xmath238 which is smaller than the best estimate of the neutron star mass in vela x-1 of @xmath239 ( barziv et al .
thus we are inclined to conclude that our constraints together with the minimum neutron star mass in vela x-1 put a lower limit on to the neutron star mass in of about @xmath240 .
strange stars ( mit bag models ) with maximum masses in the range of @xmath241@xmath242 would also fit into our constraints . at larger mass of @xmath4
a strange star is still a possible solution as well as a neutron star described by equations of state similar to that of baldo et al .
( 1997 ) , while the most stiff equations of state , such as for example that by sahu et al .
( 1993 ) , can be ruled out ( see fig . [
fig : eos ] ) . assuming alignment of the orbital angular momentum and that of the compact object , we can estimate the mass of the companion star @xmath243 from our lower limits on the inclination of the rotational axis ( see table [ table : fit ] ) . using the measured mass function @xmath244
( chakrabarty & morgan 1998 ) we find that @xmath245 for @xmath246 , respectively .
the companion radius is about @xmath247 ( bildsten & chakrabarty 2001 ) .
the mass measurements of radio pulsars in binary systems and their neutron star companions revealed their very narrow distribution with the mean @xmath248 and a very small dispersion ( thorsett & chakrabarty 1999 ) .
no evidence for extensive mass accretion was found .
in order to spin up a star with the moment of inertia @xmath249 g @xmath250 to the frequency of 400 hz one needs to accrete just about @xmath251 .
thus if the initial mass is within the same distribution one does not expect the mass of the compact object to exceed @xmath241@xmath2 . for a mass of @xmath237 , our best - fitting radius ( with model 2 )
is @xmath252 km ( 90 per cent confidence interval ) .
such a radius is consistent with the upper limit @xmath253 km obtained from the fact that pulsations are observed at fluxes differing by a factor of 100 ( burderi & king 1998 ; li et al .
1999 ) . here
one assumes that the magnetic and the ram pressure from the accreting material are equal at the magnetospheric radius which is larger than the stellar radius @xmath32 and smaller than the corotation radius @xmath254 . for @xmath255 ,
our best fit @xmath256 km is already inconsistent with the given upper limit .
this constraint , however , is more relaxed if the magnetic field does not have a dipole structure and/or the accretion disc is not gas - pressure dominated ( psaltis & chakrabarty 1999 ) . the angular dependence of the hard radiation that is needed to produce the observed pulse profiles above 8 kev is consistent with that expected from the electron scattering dominated slab with thomson optical depth @xmath257@xmath258 .
it would be natural to assume that the emission originates in a plane - parallel shock at the surface of the compact star ( basko & sunyaev 1976 ) .
the observed spectral slope @xmath259 ( the energy spectrum @xmath260 ) is expected from bulk motion comptonization in a radiation dominated strong shock ( blandford & payne 1981 ) . most of the papers on this subject neglect generation of new soft photons by reprocessing the hard radiation at the neutron star surface which would significantly soften the emergent spectra . thus , in order to reproduce @xmath259 slope
, thermal comptonization should play an important role ( lyubarskii & sunyaev 1982 ) .
as discussed in gilfanov et al .
( 1998 ) , there should exist a mechanism that adjust the spectrum so that it does not change much when the accretion rate varies by a factor of 100 .
reprocessing of the hard photons produced in a shocked region into soft ones that cool the shock could be such a mechanism .
this two - phase structure in the energy balance is known to stabilize the spectrum at the observed values when the optical depth is of the order of unity ( haardt & maraschi 1993 ; stern et al .
the electron temperature adjusts itself to the variations of the optical depth ( and luminosity ) so that the product of the electron temperature and optical depth is approximately constant
. variation of the accretion rate that is the main candidate for the aperiodic variability observed in ( wijnands & van der klis 1998b ) thus can cause the changes in the optical depth , but not in the spectral shape .
however , the relative amplitude of the black body and comptonized components can vary .
indeed , at small luminosities the black body seems to the more prominent ( gilfanov et al .
1998 ) which would correspond to a smaller optical depth of the comptonizing layer .
a larger relative normalization of the black body component should result also in a higher variability amplitude which in fact was increasing from 4 to 7 per cent ( rms amplitude ) over the duration of the 1998 april outburst ( cui et al .
. there could be additional effects related to the electron - positron pair production ( by photon - photon interactions ) which becomes important when the compactness parameter @xmath261 is larger than unity .
for the luminosity of @xmath262 erg s@xmath263 ( which is a characteristic luminosity during the april 1998 outburst of , see gierliski et al .
2002 ) and the emission region size of @xmath264 km , the compactness is about @xmath265 and it could be much larger if the region has a smaller height . thus , the data are consistent with the presence of pairs which could be responsible for producing the optical depth in the range @xmath266 depending on the compactness and for fixing the electron temperature in the range @xmath267 kev working as a thermostat ( stern et al .
1995 ; poutanen & svensson 1996 ; malzac , beloborodov & poutanen 2001 ) .
if the hard radiation is slightly beamed towards the neutron star surface due to the bulk motion , the amount of soft photons produced would be larger than from an isotropic source and there will be no problem in reproducing the observed spectrum that satisfies the energy balance ( cf .
gierliski et al .
2002 ) .
the radiation escaping from the shock can be affected by cold plasma in the accretion column .
assuming the area of the emission region of @xmath268 km@xmath269 and the accretion rate @xmath270 g s@xmath263 ( needed to produce the observed luminosity at efficiency of 0.15 ) one can estimate from the continuity equation the characteristic optical depth of the plasma over the distance equal to the spot radius @xmath271 .
( at higher distances from the star , the magnetic field lines diverge and photons can escape freely . )
this means that the flow can affect the radiation escaping along the normal to the spot and the radiation can be blocked from the observer at certain phases . since the area emitting the black body radiation is larger than the cross - section of the accretion column , one would expect that the effect of shadowing is smaller for the black body . possibly , this effect is observed in the light curve where the hard radiation has a break at @xmath272 ( see fig .
[ fig : all]a ) .
we may also see this effect from the best - fitting angular distribution ( see fig .
[ fig : all]b ) which shows a drop along the normal at @xmath273 .
our results confirm the proposal by gierliski et al .
( 2002 ) that the main physical reason behind the soft lags is a two - component nature of the spectrum .
a natural consequence of this model is a break in the phase lag at the energy where the black body component vanishes , i.e. around 8 kev ( see fig . [
fig : all]c ) .
we showed here that the black body and the comptonized components having different angular distribution generate light curves which are affected by the doppler effect in a different way .
the doppler shift is an important ingredient but not the major cause of the soft lags . in our model ,
the doppler effect comes into play only through its influence on the variability amplitude of the _ bolometric _ flux .
on the other hand , ford ( 2000 ) and weinberg et al .
( 2001 ) proposed a model where the lags in are produced by the doppler effect alone through its influence on the _ monochromatic _ flux . in both papers a black body emission from the spot
is assumed .
when the spot is moving towards the observer , the doppler effect shift the spectrum towards higher energies , while motions away from the observer softens the spectrum and generate a deficit at high energies .
thus , higher energy photons , in the tail of the black body distribution , arrive _ only _ at phases when motion is predominantly towards the observer , while the flux of photons at the peak of the black body distribution has a maximum together with the projected area , i.e. quarter of the period later .
this results in hard leads ( or soft lags ) as well .
as discussed in gierliski et al .
( 2002 ) , there are a number of problems with this proposal .
first , the photons in the black body tail contribute very little to the total observed spectrum above 8 kev which is dominated by the hard comptonized component .
thus , any soft lags in the black body flux would be completely invisible in the total flux .
second , a break in the phase lag energy dependence does not have any physical explanation in that model . even stronger doppler boosting is expected when photons from the spot are scattered off a much more rapidly rotating inner disc ( sazonov & sunyaev 2001 ) .
the scattered radiation will be leading or lagging the incident spot radiation depending on whether the disc is co - rotating or counter - rotating . for a black body radiation from the spot
, this would translate into soft lags or leads , respectively .
however , since the spectrum from the spot is closer to a power law , no phase lags should appear ( chen & shaham 1989 ) .
a different angular distribution of the black body and comptonized components as well as a non - coherent nature of scattering can affect this result .
it could be an interesting problem for future studies .
we believe that this model is not directly related to , since the spot is probably situated close to the rotational pole and the fraction of the spot photons scattered in the disc is negligibly small .
cui et al .
( 1998 ) interpreted soft lags as due to compton down - scattering of intrinsic hard photons with energies above 10 kev in a cold electron cloud of thomson optical depth @xmath274 .
the softer photons ( scattered more times ) escape from the medium later producing soft lags . as the authors pointed out
, the hard photons should escape to the observer directly through some kind of a hole in the cloud since otherwise a spectral cutoff should appear at @xmath275 kev ( e.g. sunyaev & titarchuk 1980 ; lightman , lamb & rybicki 1981 ) . this model , however , predicts a significantly reduced variability in the softer energy band ( chang & kylafis 1983 ; brainerd & lamb 1987 ) contrary to what is observed .
the two main spectral components , black body and comptonized , are clearly identified in the time - averaged spectrum of .
the pulse profiles show significant energy dependence and the softer photons lag the harder ones .
such a behaviour can be reproduced if the hard comptonized emission peaks earlier than the black body emission ( gierliski et al . 2002 ) .
the black body component shows about 25 per cent ( peak - to - peak ) variability amplitude and the doppler effect ( producing a @xmath0 7 per cent variability ) does not change its pulse profile much ( fig .
[ fig : relateff ] ) .
the comptonized emission from an optically thin slab with a broader ( `` fan''like ) angular distribution has an intrinsically smaller variability amplitude .
the doppler effect then shifts significantly the peak of the pulse producing soft lags . by modelling the observed pulse profiles
, we obtained an upper limit on the mass of the compact object @xmath276 .
we also put the lower limit of @xmath277 on the stellar mass since for smaller masses the obtained radii are too small to be consistent with any published equation of state for neutron or strange stars .
this also constrains the inclination of the system to be larger than @xmath7 . for the masses @xmath278 ,
the best - fitting stellar radii @xmath279@xmath280 km are consistent with those given by the neutron star equations of states as well as some strange star models , while for @xmath281 with that given by the equations of state for strange stars only .
we thank andrei beloborodov , tomasz bulik , marat gilfanov , george pavlov and boris stern for useful discussions .
we are grateful to marat gilfanov for providing the pulse profiles of for comparison and to dorota gondek - rosiska for the data on the equations of state for the neutron and strange stars .
this research has been supported by the academy of finland and the jenny and antti wihuri foundation .
jp is grateful to nordita ( copenhagen ) and to didier barret at centre detude spatiale des rayonnements ( toulouse ) for the hospitality during his visits . 99 baldo m. , bombaci i. , burgio g. f. , 1997 , a&a , 328 , 274 barziv o. , kaper l. , van kerkwijk m. h. , telting j. h. , van paradijs j. , 2001 a&a , 377 , 925 basko m. m. , sunyaev r. a. , 1976 , mnras , 175 , 395 beloborodov a. m. , 2002 , apj , 566 , l85 bildsten l. , chakrabarty d. , 2001 , apj , 557 , 292 blandford r. , payne d. , 1981 , mnras , 194 , 1033 brainerd j. , lamb f. k. , 1987 , apj , 317 , l33 burderi l. , king a. , 1998 , apj , 505 , l135 chakrabarty d. , morgan e. h. , 1998 , nature , 394 , 346 chakrabarty s. , 1991 , phys . rev .
d , 43 , 627 chang k. m. , kylafis n. d. , 1983 , apj , 265 , 1005 chen k. , shaham j. , 1989 , apj , 339 , 279 cui w. , morgan e. h. , titarchuk l. g. , 1998 , apj , 504 , l27 dey m. , bombaci i. , dey j. , ray s. , samanta b. c. , 1998 , phys . lett .
b , 438 , 123 ford e. c. , 2000 , apj , 535 , l119 galloway d. k. , chakrabarty d. , morgan e. h. , remillard r. a. , 2002 , apj , 576 , l137 ghisellini g. , 1999 , in casciaro b. , fortunato d. , francaviglia m. & masiello a. , eds , xiii national meeting on general relativity of the sigrav , recent developments in general relativity .
springer - verlag ( astro - ph/9905181 ) gierliski m. , done c. , barret d. , 2002 , mnras , 331 , 141 giles a. b. , hill k. m. , greenhill j. g. , 1999 , mnras , 304 , 47 gilfanov m. , revnivtsev m. , sunyaev r. , churazov e. , 1998 , a&a , 338 , l83 glendenning n. k. , schaffner - bielich j. , 1999 , phys .
c , 60 , 025803 gondek - rosiska d. , bulik t. , zdunik l. , gourgoulhon e. , ray s. , dey j. , dey m. , 2000 , a&a , 363 , 1005 gondek - rosiska d. , kluniak w. , stergioulas n. , 2003 , a&a , submitted ( astro - ph/0206470 ) haardt f. , maraschi l. , 1993 , apj , 413 , 507 homer l. , charles p. a. , chakrabarty d. , van zyl l. , 2001 , mnras , 325 , 1471 in t zand j. j. m. et al . , 2001 , a&a , 372 , 916 lattimer j. m. , prakash m. , 2001 , apj , 550 , 426 li x. d. , bombaci i. , dey m. , dey j. , van der heuvel e. p. j. , 1999 , phys . rev .
83 , 3776 lightman a. p. , lamb d. q. , rybicki g. b. , 1981 , apj , 166 , 301 lightman a. p. , press w. h. , price r. h. , teukolsky s. a. , 1975 , problem book in relativity and gravitation .
princeton university press , princeton lind k. r. , blandford r. d. , 1985 , apj , 295 , 358 lyubarskii yu .
e. , sunyaev r. a. , 1982 , sva lett . , 8
, 330 nath n. r. , strohmayer t. e. , swank j. h. , 2002 , apj , 564 , 353 malzac j. , beloborodov a. m. , poutanen j. , 2001 , mnras , 326 , 417 markwardt c. b. , swank j. h. , strohmayer t. e. , in t zand j. j. m. , marshall f. e. , 2002 , apj , 575 , l21 markwardt c. b. , smith e. , swank j. h. , 2003 , iau circ .
8080 miller m. c. , lamb f. k. , 1998 , apj , 499 , l37 misner c. w. , thorn k. s. , wheeler j. a. , 1973 , gravitation .
freeman , san francisco muno m. p. , zel f. , chakrabarty d. , 2002 , apj , 581 , 550 pechenick k. r. , ftaclas c. , cohen j. m. , 1983 , apj , 274 , 846 poutanen j. , svensson r. , 1996 , apj , 470 , 249 prakash m. , bombaci i. , prakash m. , ellis p. j. , lattimer j. m. , knorren r. , 1997 , phys . rep . , 280 , 1 psaltis d. , chakrabarty d. , 1999 , apj , 521 , 332 psaltis d. , zel f. , dedeo s. , 2000 , apj , 544 , 390 ray s. , dey j. , dey m. , ray k. , samanta b. c. , 2000 , a&a , 364 , l89 sahu p. k. , basu r. , datta b. , 1993 , apj , 416 , 267 sazonov s. y. , sunyaev r. a. , 2001 , a&a , 373 , 241 stern b. e. , poutanen j. , svensson r. , sikora m. , begelman m. c. , 1995 , apj , 449 , l13 strohmayer t. e. , jahoda k. , giles a. b. , lee u. , 1997 , apj , 486 , 355 strohmayer t. , bildsten l. , 2003 , in lewin w. h. g , van der klis m. , eds , compact stellar x - ray sources . cambridge university press , cambridge , in press ( astro - ph/0301544 ) sunyaev r. a. , titarchuk l. g. , 1980 , a&a , 86 , 121 sunyaev r. a. , titarchuk l. g. , 1985 , a&a , 143 , 374 terrell j. , 1959 , phys .
, 116 , 1041 thorsett s. e. , chakrabarty d. , 1999 , apj , 512 , 288 wang z. et al . , 2001 ,
apj , 563 , l61 weinberg n. , miller m. c. , lamb d. q. , 2001 , apj , 546 , 1098 wijnands r. , van der klis m. , 1998a , nature , 394 , 344 wijnands r. , van der klis m. , 1998b , apj , 507 , l63 zavlin v. e. , pavlov g. g. , shibanov yu .
a. , 1996 , a&a , 315 , 141 zdziarski a. a. , johnson w. n. , magdziarz p. , 1996 ,
mnras , 283 , 193 | the pulse profiles of the accreting x - ray millisecond pulsar at different energies are studied .
the two main emission component , the black body and the comptonized tail that are clearly identified in the time - averaged spectrum , show strong variability with the first component lagging the second one . the observed variability can be explained if the emission is produced by comptonization in a hot slab ( radiative shock ) of thomson optical depth @xmath0 0.31 at the neutron star surface .
the emission patterns of the black body and the comptonized radiation are different : a knife" and a fan " like , respectively .
we construct a detailed model of the x - ray production accounting for the doppler boosting , relativistic aberration and gravitational light bending in the schwarzschild spacetime .
we present also accurate analytical formulae for computations of the light curves from rapidly rotating neutron stars using formalism recently developed by beloborodov ( 2002 ) .
our model reproduces well the pulse profiles at different energies simultaneously , corresponding phase lags , as well as the time - averaged spectrum .
we constrain the compact star mass to be bounded between @xmath1 and @xmath2 . by fitting the observed profiles
, we determine the radius of the compact object to be @xmath3 km if @xmath4 , while for @xmath5 the best - fitting radius is @xmath6 km , indicating that the compact object in can be a strange star .
we obtain a lower limit on the inclination of the system of @xmath7 .
[ firstpage ] accretion , accretion discs methods : data analysis pulsars : individual ( sax j1808.4@xmath83658 ) x - rays : binaries |
the simplest model of breaking the electroweak gauge symmetry @xmath0 spontaneously is the standard higgs model .
it consists of the spin - zero higgs boson and three massless goldstone bosons , the latter ultimately being absorbed by the weak gauge bosons .
the standard higgs model is regarded to be an effective theory , only valid up to a cutoff energy @xmath1 .
the maximal allowed value of @xmath1 depends on the higgs value of the higgs mass @xmath2 , and it is connected to the renormalization group ( rg ) behaviour of the higgs sector . in the absence of a more complete theory ,
it is important to understand the perturbative limitations of the standard higgs model : if @xmath2 is too large , perturbation theory ceases to be a useful tool for calculating physical observables of the higgs sector , such as cross sections and higgs decay width .
here we review the derivation of upper perturbative bound on @xmath2 . in sect .
[ basics ] we briefly present the framework of the higgs sector , introducing the lagrangian and the higgs running coupling .
[ pertlim ] lists the various criteria which can be used to judge breakdown of perturbation theory , and the results for perturbative @xmath2 upper bounds are derived .
a short discussion and comparison with lattice results are given in sect .
the lagrangian of the standard higgs sector is @xmath3 where @xmath4 the doublet @xmath5 has a nonzero expectation value in the physical vacuum , @xmath6 to facilitate perturbative calculations , the field @xmath7 is expanded around the physical vacuum , absorbing the vacuum expectation value by the shift @xmath8 .
hence the field @xmath9 has zero vacuum expectation value . rewriting eq .
( [ l0higgs ] ) , the higgs lagrangian has the form @xmath10 with the three - point and four - point interactions of the fields given by @xmath11 here @xmath12 is the so(3 ) vector of goldstone scalars , @xmath13 , with @xmath14 .
the tadpole term and an additive constant have been dropped .
looking at eq .
( [ lagrhiggs ] ) , the @xmath15 and @xmath16 bosons are massless , in agreement with the goldstone theorem @xcite .
the higgs mass @xmath2 and the higgs quartic coupling @xmath17 are related by @xmath18 where @xmath19 is the fermi constant , and @xmath20 is the physical vacuum expectation value , @xmath21 gev .
the lagrangian @xmath22 corresponds to taking the limit of vanishing gauge and yukawa couplings of the full standard model lagrangian , and it is the starting point for carrying out calculations using the equivalence theorem @xcite . the equivalence theorem can also be used to calculate radiative corrections without having to use the full sm lagrangian @xcite , but the use of proper renormalization conditions is crucial @xcite .
the standard higgs model lagrangian contains only one degree of freedom . once @xmath2 is determined experimentally , the coupling @xmath17 is fixed according to eq .
( [ lambda ] ) .
this will be one of the many predictions to be tested experimentally when a higgs particle has been discovered .
calculating physical observables with energy scales larger than @xmath2 , renormalization group methods suggest the use of the running coupling @xmath23 .
the evolution of @xmath23 as a function of @xmath24 is given by the renormalization group equation @xmath25 with the initial condition at scale @xmath26 imposed by eq .
( [ lambda ] ) . at one loop ,
the beta function is @xmath27 and we find @xmath28 for @xmath29 gev , assuming @xmath30 gev . for even larger values of @xmath2 , the beta function is dominated by the contributions from the higgs coupling @xmath31 , and we can neglect the top yukawa coupling , @xmath32 , as well as other contributions . in this limit
, the three - loop beta function for the higgs coupling is @xcite : @xmath33\;. \label{beta}\ ] ] neglecting the appropriate powers of @xmath17 , the above equations determine the @xmath34-loop running coupling for @xmath35 .
explicitly , the one - loop heavy - higgs running coupling is @xmath36
there are two scenarios in which perturbation theory can break down .
first , the higgs coupling can be nonperturbative for large values of @xmath2 ; see eq .
( [ lambda ] ) .
secondly , the higgs _ running _ coupling can be nonperturbative if @xmath28 and @xmath24 is large . increasing the scale @xmath24 in eq .
( [ run1lp ] ) , @xmath23 eventually becomes infinite , indicating the one - loop landau pole .
, increasing @xmath24 eventually leads to a negative higgs running coupling .
this observation is related to the problem of vacuum stability .
see @xcite for a review .
] perturbation theory ceases to be reliable long before reaching the location of this pole . to judge for which values of the ( running ) higgs coupling the interactions become strong ,
one calculates perturbative amplitudes , assuming small perturbative couplings . increasing @xmath37 , the perturbative expansion will eventually break down , giving upper perturbative bounds on the ( running ) higgs coupling .
one has to distinguish two different kinds of physical observables .
low - energy quantities such as higgs decay amplitudes do not require rg methods , hence no running coupling is involved .
therefore they immediately yield an upper perturbative bound on @xmath2 via eq .
( [ lambda ] ) . on the other hand ,
cross sections with large center - of - mass energy , @xmath38 , do require the use of the running coupling .
hence the onset of nonperturbative effects in such processes is a function of both @xmath2 and @xmath39 .
there are several criteria we can use to judge the convergence of perturbation theory : * the size of radiative corrections should be small , with higher order contributions being suppressed .
* the renormalization scheme dependence ( _ e.g. , _ @xmath40 _ vs.
_ oms ) of physical amplitudes should diminish when including higher order corrections . *
the dependence of an amplitude on the scale @xmath24 should decrease with increasing order in perturbation theory . *
scattering amplitudes should not violate perturbative unitarity by a large amount . _
exact _ amplitudes , of course , have _ no _ scheme dependence , _ no _ scale dependence , and feature _ no _ unitarity violations . converging perturbative amplitudes should therefore approach this limit reasonably fast .
yet there are also limitations to nonperturbative calculations in the sm higgs sector .
since the heavy - higgs running coupling has a positive beta function even in the nonperturbative approach @xcite , there remains a cutoff scale @xmath1 beyond which the theory fails . in return , this cutoff scale bounds the possible higgs mass range from above .
we will return to this aspect in sect .
[ disc ] .
a heavy higgs particle ( @xmath41 ) decays mostly into pairs of ( longitudinally polarized ) gauge bosons .
the branching ratio for decays into a pair of top quarks is about 10% . neglecting the subleading corrections due to gauge and yukawa couplings , the two - loop @xmath42 corrections to these decay channels
have been calculated .
using the oms renormalization scheme , the bosonic decay channel receives corrections@xcite @xmath43 and the two - loop result for the fermionic decay width is @xcite @xmath44 comparing the magnitude of the one - loop and two - loop corrections , perturbation theory seems to work up to values of @xmath45 , that is , higgs masses of about 1 tev .
however , conversion of the results into the @xmath40 scheme as well as an analysis of the scale dependence using rg methods reveals that higher - order terms may spoil perturbation theory already for values of @xmath46 , or equivalently , @xmath47 gev @xcite .
the principles of such analyses will be discussed in the following section in connection with cross sections rather than decay widths .
a typical scattering process related to the sm higgs sector is @xmath48 .
this process is of interest for future colliders such as the lhc or linear @xmath49 and @xmath50 colliders . at tree level ,
the amplitude consists of a gauge contribution @xmath51 and a higgs sector contribution @xmath52 .
the @xmath52 term is connected to the longitudinally polarized component of the gauge bosons .
it can be calculated using @xmath53 , receiving contributions from both three - point and four - point interactions , see fig . [ feynman ] .
= 2.5 in at high center - of - mass energy , @xmath38 , the three - point interactions contributing to the amplitude @xmath54 are suppressed by powers of @xmath55 : @xmath56 this is also true when including higher order corrections , as can be shown using dimensional arguments .
hence the entire high - energy @xmath57-dependence originates from scattering graphs related to @xmath58 , the part of the higgs lagrangian which is identical to a massless @xmath59 theory with n=4 .
the resulting high - energy scattering graphs up to two loops are shown in fig .
[ feyn2lp ] . calculating these leading corrections up to two loops ,
the renormalized oms amplitude is @xcite @xmath60 ^ 2\ , \biggl[\,1\,+ \left ( 24 \ln \frac{s}{\mu^2 } - \ , 42.65 \right ) \,\frac{\lambda(\mu)}{16\pi^2}\ , \\ & & \phantom{\frac{1}{8\pi s } [ \lambda(\mu)]^2\ , } + \left ( 432 \ln^2 \frac{s}{\mu^2 } - 1823.3\ln \frac{s}{\mu^2 } + \;2457.9\,\right ) \frac{\lambda^2(\mu)}{(16\pi^2)^2}\ , + \ ; { \rm o}\left(\lambda^3(\mu)\right)\,\biggr]\,,\nonumber\end{aligned}\ ] ] where the explicit @xmath24 dependence has been kept using the results in @xcite .
an anomalous dimension prefactor has been neglected since it is close to unity for the values of @xmath61 and @xmath2 considered here . using the three - loop running coupling in connection with the above two - loop result ,
we obtain the rg improved next - to - next - to - leading - log ( nnll ) cross section .
the nnll result contains a complete summation of the leading logarithms ( ll ) , @xmath62 , the next - to - leading logarithms , @xmath63 , and the nnlls , @xmath64 , @xmath65 . just using the one - loop cross section together with the two - loop running coupling yields the nll cross section , and the ll cross section corresponds to the tree - level result with @xmath23 given by eq .
( [ run1lp ] ) .
= 2.5 in the product @xmath66 depends only on the running coupling and the ratio @xmath67 , making the above cross section a useful quantity in determining perturbative upper bounds on @xmath23 .
taking @xmath68 , all rg logarithms are completely resummed into @xmath69 and we obtain @xmath70 it is striking that the coefficients appearing in the cross section are much larger than the corresponding coefficients of the decay widths , eqs .
( [ hwwwidth ] ) and ( [ hffwidth ] ) .
peculiarly , a value @xmath71 leads to a _ negative _ nll cross section , indicating the complete failure of perturbation theory for such a coupling . corresponds to , for example , @xmath72 gev and @xmath73 tev . ] using the various criteria listed earlier , the upper perturbative bound is about @xmath74 @xcite.this upper bound on a perturbative higgs running coupling represents a strong restriction on the higgs mass , stronger than the bound obtained in the case of the decay widths , @xmath46 .
choosing , for example , @xmath75 to be about 2 tev , the higgs mass has to be less than about 400 gev to allow for a perturbative calculation of the cross section .
recent results @xcite , however , show that this bound can be relaxed ; see the following section .
the perturbative behaviour of the scattering cross sections can be improved by performing a partial summation of higher - order contributions @xcite . realizing that chains of bubble diagrams , @xmath76 , and diagrams with bubble substructures ( see fig . [ feyn2lp ] for examples up to two loops ) are the dominant contribution to the @xmath57-dependent part of the scattering amplitude , a class of nonlogarithmic terms accompanying the logarithmic @xmath57-dependent terms can be partially summed .
this summation can be formulated in terms of a modified scale entering the running coupling .
instead of using the standard choice @xmath68 , the improved choice is @xcite @xmath77 the vast improvement due to this choice can be seen when plotting the scale dependence of the cross section . in fig .
[ figsigmu ] the quantity @xmath66 is given as a function of @xmath78 , fixing the running coupling at @xmath79 to be 1.5 ( which corresponds to @xmath80 ) .
= 2.5 in taking @xmath68 , the cross section receives large radiative corrections and has a strong scale dependence . in particular , the scale dependence remains large even when performing a nnll resummation .
in contrast , the improved scale @xmath79 leads to much smaller corrections , and the scale dependence almost vanishes when going beyond the ll approximation .
of course , increasing the value of the running coupling , perturbation theory also breaks down for the summed cross section .
this happens for @xmath81 of about 4.0 , where perturbative unitarity is violated @xcite .
such a running couplings corresponds to , for example , @xmath82 gev and @xmath75 close to 2 tev .
experimental measurement of the higgs quartic coupling ( for example , the measurement of the @xmath52 contribution to high - energy @xmath83 scattering ) and verifying its theoretical prediction is challenging . for small values of @xmath17 , the electroweak @xmath51 terms are dominant .
larger values of @xmath17 increase the @xmath52 contributions of the cross sections ( see fig .
[ figsigall ] ) , but accumulation of a sufficient amount of data and elimination of background is a difficult task .
the perturbative upper bounds on the higgs ( running ) coupling indicate that perturbation theory is reliable for higgs masses up to 700 gev if the process considered has a center - of - mass energy of less than 2 tev . going to higher cms energy , the value of @xmath2 has to be reduced to satisfy @xmath84 . of course ,
these bounds are process dependent .
however , all @xmath85 high - energy scattering processes involving @xmath86 , @xmath87 , @xmath88 and @xmath9 particles are related by the underlying so(4 ) symmetry of @xmath58 and therefore have similar amplitudes .
in particular the unitarity argument , first used by @xcite , involves the different channels and supports the above bound .
the perturbative upper bounds can be compared with results obtained from lattice calculations @xcite .
such results depend somewhat on the lattice action and regularization used .
interestingly , they obtain similar bounds on the higgs mass , excluding the existence of a heavier standard model higgs boson .
together with the results reviewed here , this excludes the possibility of a strongly interacting higgs boson @xcite .
the author thanks the organizers of the _ xxxvi cracow school of theoretical physics _ for the invitation to this informative and well - organized school . the pleasant atmosphere as well as
the picturesque surroundings provided a great setting for lectures and discussions .
j. cornwall , d. levin , and g. tiktopoulos , phys .
d * 10 * , 1145 ( 1974 ) , appendix a ; c.e .
vayonakis , lett .
nuovo cimento * 17 * , 383 ( 1976 ) ; g. gounaris , r. kgeler , and h. neufeld , phys .
d * 34 * , 3257 ( 1986 ) ; y. yao and c. yuan , _
ibid_. * 38 * , 2237 ( 1988 ) .
j. bagger and c. schmidt , phys .
d * 41 * , 264 ( 1990 ) ; a. aoki , kyoto university report no .
rifp-705 , 1987 ( unpublished ) : in _ physics at tev scale _ , proceedings of the meeting , tsukuba , japan , 1988 , edited by k. hidaka and k. hikasa ( kek , tsukuba , 1988 ) ; h. veltman , phys .
d * 41 * , 2294 ( 1990 ) . | this lecture reviews various higgs - sector amplitudes which have been calculated to two loops in the higgs quartic coupling . after explaining the framework of these calculations ,
the perturbative behaviour of the amplitudes is discussed , and perturbative upper bounds on the higgs boson mass are given .
16.cm 22.cm # 1 # 2 # 3 phys .
rev .
* # 1 * , # 2 ( # 3 ) # 1 # 2 # 3 phys .
rev .
lett . *
# 1 * , # 2 ( # 3 ) # 1 # 2 # 3 phys .
rev .
d * # 1 * , # 2 ( # 3 ) # 1 # 2 # 3 phys .
lett .
* b#1 * , # 2 ( # 3 ) # 1 # 2 # 3 nucl .
phys .
* b#1 * , # 2 ( # 3 ) # 1 # 2 # 3 rev . mod .
phys . *
# 1 * , # 2 ( # 3 ) # 1 # 2 # 3 z. phys .
c * # 1 * , # 2 ( # 3 ) |
the motivation to install roman pot detectors at 220 m within atlas is quite clear .
it extends nicely the project of measuring the total cross sections using roman pots at 240 m @xcite by measuring hard diffraction at high luminosity in atlas in the lhc . as we will see in the following , it is also complementary to the fp420 project which aims at tagging protons at 420 m. the physics motivation of this project corresponds to different domains of diffraction : * a better understanding of the inclusive diffraction mechanism at the lhc by studying in detail the structure of pomeron in terms of quarks and gluons as it was done at hera @xcite . of great importance
is also the measurement of the exclusive production of diffractive events @xcite and its cross section in the jet channel as a function of jet transverse momentum .
its understanding is necessary to control the background to higgs signal .
* looking for higgs boson diffractive production in double pomeron exchange in the standard model or supersymmetric extensions of the standard model @xcite .
this is clearly a challenging topic especially at low higgs boson masses where the higgs boson decays in @xmath0 and the standard non - diffractive search is possible .
we will detail in the following the trigger strategy .
* sensitivity to the anomalous coupling of the photon by measuring the qed production cross section of @xmath1 boson pairs . this might be the best way to access the anomalous coupling before the start of the ilc .
* photoproduction of jets * other topics such as looking for stop events or measuring the top mass using the threshold scan method @xcite which will depend strongly on the production cross section .
we propose to install roman pots in atlas at 216 and 224 m on each side of the main atlas detectors .
the project is a collaboration between the physics institutes and universities of prague , cracow , stony brook , michigan state university , lpnhe ( paris 6 ) , giessen , and in addition the university of chicago and the argonne national laboratory for the timing detectors .
the roman pot design follows as close as possible the design which is currently used by the totem collaboration and the luminosity group of the atlas collaboration which aims at measuring the total cross section using roman pots at 240 m. the only difference is that we only need the horizontal arms and not the vertical arms since hard diffractive protons are scattered horizontally .
we will follow the totem experience to build the roman pots in vakuum praha and to use the same technics for the step motors and the lvdt system .
r0.5 assuming one can go down to 10 ( resp . 15 ) @xmath2 from the beam center , it is possible to measure protons with @xmath3 , and @xmath4 on each side of atlas ( resp .
@xmath5 , @xmath6 ) where @xmath7 is the momentum fraction of the initial proton carried away by the pomeron @xcite .
this can be translated in missing mass acceptance as illustrated in fig 1 .
the missing mass acceptance using only the 220 m pots starts at 135 gev , but increases slowly as a function of missing mass .
it is clear that one needs both fp420 and rp220 projects , or in other words the possibility to detect scattered protons at 220 and 420 m to obtain a good acceptance on a wide range of masses since most events are asymmetric ( one tag at 220 m and another one at 420 m ) .
the precision on mass reconstruction using either two tags at 220 m or one tag at 220 m and another one at 420 m is of the order of 2 - 3 % on the full mass range .
this shows the advantage of this measurement which allows to give a very good mass resolution on a wide range of masses , and thus to detect higgs bosons at low masses decaying into @xmath0 .
the idea is to enhance the signal over background ratio by benefitting from the good resolution of the detectors and the suppression of the @xmath8 jet background due to the @xmath9 suppression rule for @xmath8 jet exclusive production .
we propose to put inside the roman pots two kinds of detectors , namely silicon detectors to measure precisely the position of the diffracted protons , and the mass of the produced object , such as the higgs boson , and @xmath7 , and precise timing detectors .
the position detectors will consist in either five layers of silicon strips of 50 @xmath10 m and two additional layers used for triggering , or 3d silicon detectors if they are available industrially by the time we need to instal the roman pots .
if the silicon strip option is chosen , there will be four different orientations , namely x , y , u , and v ( u and v being orientated within 45 degrees with respect to x and y ) .
the strip size will be 50 @xmath10 m and the detector size about 2 cm , which allows a measurement up to @xmath11 .
the silicon strip detectors will be edgeless which means that the dead edge will be of the order of 30 - 50 @xmath10 m so that we can move the detector as close to the beam as possible without losing some acceptance due to the dead edge .
the detectors will be read out by the standard abcnext chip being developped in cracow for the silicon detector of atlas .
the latency time of the abcnext chip is of the order of 3.5 @xmath10s which gives enough time to send back the local l1 decision from the roman pots to atlas ( see the next paragraph about trigger for more detail ) , and to receive the l1 decision from atlas , which means a distance of about 440 m. it is also foressen to perform a slight modification of the abcnext chip to include the trigger possibilities into the chip .
the other option is to use 3d silicon detectors using the same readout system as before ( abcnext chip ) .
these detectors use a lateral electric field , instead of vertical in conventional planar techniques .
holes of the order of 10 @xmath10 m crossing the full thickness of the detector are filled with a conductive medium in order to collect the ionisation ( electrons or holes ) depending on the applied bias .
both kinds of options will be tested in prague and in saclay using the full electronics chain ( including the abcnext chip ) and a laser or a radioactive source .
beam tests at desy or cern are also foreseen .
it is planed to install the roman pot together with the silicon detectors during a shut down of the lhc in 2009 - 2010 .
the timing detectors are necessary at the highest luminosity of the lhc to identify from which vertex the protons are coming from .
it is expected that up to 35 interactions occur at the same bunch crossing and we need to identify from which interaction , or from which vertex the protons are coming from .
a precision of the order of a few mm or 5 - 10 ps is required to distinguish between the different vertices and to make sure that the diffracted protons come from the hard interactions .
picosecond timing detectors are still a challenge and are developped in a collaboration between saclay , stony brook , the university of chicago and argonne national laboratory for medical and particle physics applications .
the proton timings will be measured in a crystal of about 2.5 cm located inside the roman pots , and the signal will be read out by micro - channel plates photomultipliers developped by photonis .
the space resolution of those detectors should be of the order of a few mm since at most two protons will be detected in those detectors for one given bunch crossing at the highest luminosity .
the detectors are read out with a constant fraction discriminator which allows to improve the timing resolution significantly compared to usual electronics . a first version of the timing detectors is expected to be ready in 2009 - 2010 with a worse resolution of 40 - 50 ps , and the final version by 2012 with a resolution of 5 - 10 ps .
in this section , we would like to give the principle of the trigger using the roman pots at 220 m as well as the rates obtained using a simulation of the atlas detector and trigger framework .
the principle of the trigger is shown in fig . 2 in the case of a higgs boson decaying into @xmath0 as an example .
the first level trigger comes directly from two different silicon strip layers in each roman pot detector .
it is more practical to use two dedicated planes for triggering only since it allows to use different signal thresholds for trigger and readout .
the idea is to send at most five strip addresses which are hit at level 1 .
a local trigger is defined at the roman pot level on each side of the atlas experiment by combining the two trigger planes in each roman pot and the roman pots as well .
if the hits are found to be compatible ( not issued by noise but by real protons ) , the strip addresses are sent to atlas , which allows to compute the @xmath7 of each proton , and the diffractive mass .
this information is then combined with the information coming from the central atlas detector , requesting for instance two jets above 40 gev in the case shown in fig .
2 . at l2 ,
the information coming from the timing detectors for each diffracted proton can be used and combined with the position of the main vertex of atlas to check for compatibility . once
a positive atlas trigger decision is taken ( even without any diffracted proton ) , the readout informations coming from the roman pot detectors are sent to atlas as any subdetector . *
* trigger on dpe events at 220 m : * this is the easiest situation since two protons can be requested at level 1 at 220 m. three different options are considered : * - trigger on high mass higgs * ( @xmath12 gev ) given by atlas directly ( decay in @xmath13 , @xmath14 ) , by requesting two high @xmath15 jets and two positive tags in roman pots , ( high @xmath15 jets given directly by atlas , and low @xmath15 jet special trigger for qcd studies highly prescaled ) .
this configuration will not rise any problem concerning the l1 rate since most of the events will be triggered by atlas anyway , and the special diffractive triggers will be for qcd measurements and can be highly prescaled .
* * trigger on dpe events at 220 and 420 m * this is the most delicate scenario since the information from the 420 m pots can not be included at l1 .
the strategy is the following ( see table 1 ) : ( higgs ... ) decaying in @xmath0 by requesting a positive tag ( one side only ) at 220 m with @xmath16 ( due to the @xmath17 rp acceptance in @xmath7 , the proton momentum fractional loss in the @xmath18 roman pot can not be too high if the higgs mass is smaller than @xmath19 ) , and topological cuts on jets such as the exclusiveness of the process ( @xmath20 , @xmath21 , where @xmath22 are the pseudorapidities of the two l1 jets , and @xmath23 the pseudorapidity of the proton in the @xmath18 roman pots ) .
this trigger can hold without prescales to a luminosity up to 2.10@xmath24 @xmath25s@xmath26 , ( single diffraction , or double pomeron exchange ) for qcd studies : can be heavily prescaled , given by atlas with lepton triggers .
let us note that the rate will be of the order of 1 hz at l2 by adding a cut on a presence of a tag in the 420 pots , on timing , and also on the compatibility of the rapidity of the central object computed using the jets or the protons in roman pots .
& rp200 & @xmath16 & jet @xmath27 & bunch & rate @xmath28 $ ] & reduction & reduction & prop . & crossing & @xmath29 $ ] & factor & factor & @xmath30 & 0.35 & 2.6 & 120 & 300 & 1200 @xmath31 & 3.5 & 26 & 8.9 & 22 & 88 @xmath32 & 7 & 52 & 4.2 & 9.8 & 39.2 @xmath33 & 17.5 & 130 & 1.9 & 3.9 & 15.6 @xmath34 & 35 & 260 & 1.3 & 2.2 & 8.8 in this short report , we described the main aspects of the project to install roman pots at 220 m within atlas : silicon detectors , measurement of the proton timings , and the trigger properties .
this project is aimed to be proposed to atlas and the lhcc together with the fp420 one .
m. boonekamp , r. peschanski and c. royon , phys .
* b598 * ( 2004 ) 243 ; m. boonekamp , a. de roeck , r. peschanski and c. royon , phys
* b 550 * ( 2002 ) 93 ; v. a. khoze , a. d. martin and m. g. ryskin , eur .
j. * c23 * ( 2002 ) 311 , eur .
j. * c24 * ( 2002 ) 581 .
m. boonekamp , j. cammin , s. lavignac , r. peschanski and c. royon , phys . rev . *
d 73 * ( 2006 ) 115011 atlas note about rp220 project , to be submitted | we give a short description of the project to install roman pot detectors at 220 m from the interaction point in atlas .
this project is dedicated to hard diffractive measurements at high luminosity . |
understanding the physical and chemical structure of the inner hundred au of the envelopes around protostellar objects is important because part of this material will be included in circumstellar disks from which planets may form eventually .
the innermost envelopes are characterized by high temperatures ( @xmath6 100 k ) and densities ( @xmath11 @xmath12 ) , making them eminently suited for observations at submillimeter wavelengths .
these regions are , however , heavily diluted in single - dish beams ( @xmath13 size compared to typical single - dish beam sizes of 1020@xmath1 ) and interpretation of their line and continuum emission relies on extrapolation of the density and temperature distribution from observations on larger scales .
furthermore , studies based on observations of lower excitation lines are complicated by the fact that a large fraction of their emission may arise in the outer cold regions of the envelopes , where the lines can also become optically thick .
interferometry with the submillimeter array ( sma ) avoids most of these complications and thus provides a unique possibility to probe the warm and dense material in the innermost regions of protostellar systems . in this paper
we present sma observations for the low - mass protostar ngc 1333-iras2a .
this study will serve as an illustration of the type of data and analysis techniques that are possible for these objects and will be applied in the future to a much larger sample of low - mass protostars currently being observed at the sma .
the first goal of this project is to constrain the physical structure on scales of a few hundred au from the continuum data .
the larger ( @xmath14 au ) scale temperature and density distribution is well determined from analysis of single - dish submillimeter continuum images ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) but holes , cavities and departures from spherical symmetry are likely at smaller radii . in addition , any disk will start to contribute significantly to the emission on scales of a few hundred au .
since the dust continuum flux scales with frequency as @xmath15 or steeper , submillimeter interferometry is particularly well suited to probe the disks and distinguish them from the envelope .
very little is known about the physical properties such as mass and sizes of disks in these deeply embedded stages , parameters which the sma data can constrain .
a second goal of this project is to study the chemistry in the innermost region , where large changes in abundances are expected due to evaporation of icy grain mantles that inject species such as h@xmath4o and ch@xmath2oh into the gas - phase .
reactions between evaporated molecules result in a distinct `` hot core chemistry '' containing even more complex organic molecules . in recent years
the existence of such hot cores has been suggested in a number of low - mass young stellar objects ( e.g. , * ? ? ?
* ) , most strongly in the protostellar binary iras 162932422 where warm gas and high abundances of a number of species are thought to be present in the innermost ( @xmath16 au ) regions @xcite .
detections of various complex organic molecules , e.g. , hcooh , ch@xmath2och@xmath2 , ch@xmath2cho , ch@xmath2cn , and c@xmath4h@xmath17cn have been reported from single - dish observations toward iras 162932422 @xcite and another low - mass protostar ngc 1333-iras4a @xcite .
still , it is not clear how many low - mass protostars show complex organic molecules and whether these reflect the heating of ices or other processes such as the action of the outflows ( see , e.g. , discussion in * ? ? ?
yet another fascinating question is whether any of the organic molecules reside in the circumstellar disks .
high angular resolution interferometer observations of iras
16293 - 2422 for example @xcite show that line emission of the organic molecules peaks close to the position of two compact dust continuum sources thought to mark the location of circumstellar disks around each of the binary components .
this paper presents high angular resolution submillimeter observations of dust continuum emission and high excitation transitions of organic and other molecules toward the class 0 protostar ngc 1333-iras2a using the sma .
the data are interpreted in the context of physical and chemical models previously constrained from single - dish observations and coupled with detailed radiative transfer calculations .
the paper is laid out as follows : 2 describes the details of the observations and 3 shows the qualitative results in terms of the detected continuum and line emission - including the detections of high excitation lines of , e.g. , ch@xmath2oh . 4 presents a detailed analysis based on the available continuum and line radiative transfer models and 5 discusses the origin of the observed emission .
ngc 1333-iras2a ( @xmath18 , @xmath19 ) @xcite was observed on 2004 october 17 with the submillimeter array @xcite at 350 ghz .
the array had 7 antennae in the `` compact - north '' configuration covering projected baselines from 18 to 164 k@xmath20 corresponding to a beam size of [email protected] ( p.a .
83@xmath22 ) with natural weighting .
the weather conditions were excellent with typical 225 ghz zenith sky opacities of 0.0250.03 .
the 2 ghz bandwidth in each sideband of the sma digital correlator was configured with 1 chunk of 512 channels ( 345.853345.957 ghz and 354.453354.557 ghz ) and 3 chunks of 256 channels ( 345.765345.869 , 345.279345.383 , 344.295344.399 ghz and 354.541354.645 , 355.027355.131 , 356.011 - 356.115 ghz ) . with the width of each chunk of 104 mhz ,
the resulting spectral resolutions are 0.20.4 mhz ( 0.150.30 km s@xmath23 ) .
the remaining bandwidth was covered by 20 chunks each of 32 channels with a resulting resolution of 3.25 mhz ( 2.8 km s@xmath23 ) ; no bright lines are expected in these chunks , which are used to determine the continuum .
the separation between the upper and lower sideband is 10 ghz .
the complex gains were calibrated through observations of the two nearby quasars , 3c84 and 0359 + 509 , the flux through observations of uranus and the passband through observations of uranus and venus .
the initial reduction was performed in the `` mir '' package @xcite and subsequently maps created and cleaned in miriad @xcite .
the resulting rms sensitivity was 15 mjy beam@xmath23 over the 2 ghz bandwidth ( continuum ) and 0.2 - 0.35 jy beam@xmath23 per channel in the narrow band setups .
continuum emission was clearly detected toward iras2a .
the nearby companion iras2b @xcite is located about 30 away , i.e. , outside the primary beam of the sma 350 ghz observations ( half power beam width of 35 ) and is not picked up by the observations presented in this paper . a map of the continuum emission is shown in fig .
[ zerothorder ] . for
iras2a gaussian fits to the slightly extended emission ( 2.3@[email protected]@xmath1 fwhm ) give a total integrated flux of approximately 1 jy ( a peak flux of 0.46 jy beam@xmath23 ) . for comparison ,
@xcite report a flux of [email protected] jy at 850 @xmath25 m in an aperture with a radius of 45@xmath1 from the jcmt / scuba whereas @xcite report a `` disk '' flux of @xmath26 jy from interferometer observations with the jcmt and cso interferometer covering baselines of @xmath27 70190 k@xmath20 .
these differences illustrate that a large fraction of the emission is resolved out due to the interferometer s lack of short spacings . in order to make any ( qualitative or quantitative ) statements about the origin and properties of the compact
, emitting component , the extended envelope therefore needs to be subtracted , and the interpretation of the interferometric data requires a model . in total 10 lines were detected at the 3@xmath28 level in this single track including the optically thick co 32 and hcn 43 lines , optically thin lines of h@xmath3cn , so , and so@xmath4 and high excitation lines of organic molecules including ch@xmath2oh , ch@xmath2od and ch@xmath2och@xmath2 with a tentative detection of ch@xmath2ocho .
the lines detected are listed in table [ linelist ] and their spectra in the central beam shown in fig .
[ spectra_overview ] .
results of gaussian fits to the lines observed in the central beam toward the continuum position are given in table [ linefits ] . besides the co 32 and hcn 43 lines that clearly show structure due to resolved out extended emission at the systemic velocity , and possible self - absorption , the lines are all well represented by gaussian profiles with line widths of 34 km s@xmath23 ( fwhm ) .
so@xmath4 is the only exception with a narrower line width of @xmath27 2 km s@xmath23 .
llll + co 32 & 345.7960 & @xmath29 & @xmath29 + hcn 43 & 354.5055 & @xmath29 & @xmath29 + h@xmath3cn 43 & 345.3398 & 1.4 ( 0.06 ) & 4.2 ( 0.2 ) + so @xmath30 & 344.3108 & 1.9 ( 0.07 ) & 3.9 ( 0.2 ) + so@xmath4 @xmath31 & 354.0455 & 0.54 ( 0.07 ) & 2.4 ( 0.4 ) + ch@xmath2oh @xmath32 ( a - type ) & 345.9040 & 0.99 ( 0.07 ) & 3.6 ( 0.3 ) + ch@xmath2oh @xmath33 ( e - type ) & 345.9192 & 0.79 ( 0.08 ) & 3.4 ( 0.4 ) + ch@xmath2od @xmath34 ( a - type ) & 345.3196 & 0.58 ( 0.13 ) & 3.1 ( 0.8 ) + ch@xmath2och@xmath2 @xmath35 & 344.3579 & 0.32 ( 0.15 ) & 3.5 ( 2.1 ) + ch@xmath2ocho @xmath36 & 354.6078 & 0.27 ( 0.08 ) & 2.8 ( 1.0 ) maps of the observed line emission for selected species are shown in fig .
[ zerothorder ] .
most of the species ( except co 32 ) show compact emission around the continuum position and only co , hcn , h@xmath3cn and so are clearly resolved and show velocity gradients associated with the outflow .
the widths of the gaussian fits to these lines are not significantly larger than those of the remaining species .
the organic species and the so@xmath4 are largely unresolved with the observing resolution corresponding to scales of @xmath37 au .
this suggests that the observed emission either probes the innermost region of the protostellar envelope ( where the temperatures are 5060 k or higher ) or is associated with the compact component observed in the continuum data , possibly the circumstellar disk .
the development of detailed radiative transfer models has made it possible to place constraints on the radial distribution of molecular abundances throughout protostellar envelopes . in short , these models use dust continuum observations to constrain the physical conditions in the envelopes ( e.g. temperature and density ) , which subsequently can be used for non - lte molecular excitation and line radiative transfer calculations for comparison to both multi - line single - dish and interferometer data .
interferometer observations of continuum emission can be used to constrain the properties of the envelope on small ( @xmath38 au ) scales and infer the presence of circumstellar disks ( e.g , * ? ? ?
* ; * ? ? ?
* ) if emission from the envelope is taken into account .
images of the continuum and line emission from the protostar can be computed and then fourier transformed for direct comparison to the interferometer observations .
this comparison in the fourier plane has two direct advantages : i ) no image deconvolution and restoration of the interferometer data is necessary and ii ) the comparisons between the models and interferometer data implicitly take the problem of missing short spacings into account @xcite .
the short spacings are included through the single - dish observations used to constrain the physical and chemical structure of the envelope in the first place . as basis for a more detailed analysis we adopt the structure of the ngc 1333-iras2 envelope derived from detailed dust and line radiative transfer modeling @xcite .
this model assumes a spherically symmetric envelope heated by a central source and the temperature profile is calculated self - consistently through the dust radiative transfer .
the single - dish observations are well - fitted by a power - law density distribution , @xmath39 , with @xmath40 .
the mass of the envelope is 1.7 @xmath8 within a radius of 12,000 au .
this model has previously been compared to 3 mm interferometer data from bima and ovro and was found to explain well the envelope structure down to scales of about 600 - 1000 au ( 3@xmath1 - 5@xmath1 ) , the smallest scales probed @xcite .
[ cont_uvamp ] compares the observed visibilities from the sma observations with the envelope models @xcite .
consistent with the 3 mm interferometer observations @xcite a compact continuum source is required to explain the observed emission in addition to the extended envelope picked up at the shorter baselines . in the 3 mm observations continuum emission
was seen out to baselines of @xmath41 k@xmath20 which did not resolve the central source . at the longer baselines
observed with the sma the visibility amplitudes continue to decline indicating that the compact component is resolved at scales of about 1@xmath1 ( @xmath42 au ) .
this is consistent with the result of @xcite who suggested that the jcmt - cso observations with baselines of @xmath27 70190 k@xmath20 resolved the disk around iras2a . to constrain the properties of this central component ,
the data were fitted with the prediction from the envelope model explaining the larger scale emission but adding a gaussian brightness distribution simulating the disk .
the gaussian brightness distribution was taken to be spherically symmetric and its best fit flux and width constrained through estimates of the @xmath43 statistics .
for an inner radius of the envelope of 23 au ( corresponding to a temperature of 250 k ) , a gaussian distribution with a fwhm of 1.1@xmath1 ( a diameter of @xmath44 au ) and a total flux of 0.6 jy provides the best fit .
this is somewhat larger than the size inferred by @xcite of @xmath45 au ( scaled to a distance of 220 pc ) .
this difference reflects i ) the shorter baselines included in the observations in this paper constraining the disk contribution on larger scales and ii ) the fact that @xcite did not include the contribution from the envelope in their estimates .
if an inner cavity exists in the envelope such as suggested for iras 162932422 ( * ? ? ?
* jrgensen et al . , in prep . )
the flux of the compact component would be higher .
for example , for a cavity with a radius of @xmath27 285 au ( corresponding to an inner temperature of 75 k ) the flux of the central component increases to 0.8 jy and its diameter decreases to 0.9@xmath1 ( @xmath46 au ) . note that this is just an example : the interferometer does not place exact constraints on the cavity size and therefore some degeneracy exists between the envelope and disk parameters .
observations of the mid - infrared source with the spitzer space telescope would place exact constraints on the dust column density toward the source center and could thereby resolve this degeneracy . no matter
which inner radius is assumed for the model ( 23 au or 285 au ) , the flux at the longest baselines is underestimated .
varying the slope of the density profile within the uncertainties allowed by the scuba maps does not improve the best fit at the longer baselines as the emission there is dominated by the gaussian disk component .
naturally the gaussian profile is probably an oversimplification of the disk structure on the smallest scales .
this point is illustrated by the solid line in the rightmost panel of fig .
[ cont_uvamp ] which includes an envelope ( with an inner cavity of radius 285 au ) , a disk with gaussian brightness distribution ( flux of 0.5 jy and diameter of 300 au ) and an unresolved 0.3 jy point source component .
this two component model improves the fit statistically even taking the extra degree of freedom , introduced with the unknown point source flux , into account . @xcite
reached a similar conclusion for the disk around the t tauri star hl tau : 2.7 mm and 0.87 mm continuum emission toward this source was found to be more centrally peaked than a single gaussian .
@xcite attributed this central condensation in emission to temperature and density gradients in a simple power - law disk model .
what this discussion emphasizes is the need for good constraints on the continuum emission on a wide variety of interferometer baselines and from single - dish observations coupled with reliable models for both the envelope and disk components .
thereby it becomes possible to simultaneously constrain the flux of both the extended envelope , and of the disk on intermediate ( few hundred au ) and small scales ( @xmath9 au ) but naturally these estimates rely on our assumptions about the disk and envelope structure .
the line observations are compared to the predictions based on models of the envelope constrained by the single - dish observations of @xcite . in those papers abundances of h@xmath3cn , so and ch@xmath2oh were determined and an upper limit of the so@xmath4 line reported . in this paper , we focus on the chemistry of the optically thin species and defer the discussion of co and hcn to a future paper . the aim is to constrain the abundance profiles in the innermost regions which were difficult to probe with single - dish telescopes or low - frequency interferometers . for h@xmath3cn , so and so@xmath4 collisional rate coefficients from the _ leiden atomic and molecular database _ @xcite were used to calculate the nlte molecular excitation . for the remaining species no data exist for the observed high excitation transitions and lte was assumed . for ch@xmath2oh one of the observed transitions probes a - type and the other e - type ( the distinction reflecting the rotation of the oh group with respect to the ch@xmath2 group ) , and the abundance of ch@xmath2oh quoted in this paper refers to the sum of the abundance of these two species .
the abundance structures for all species are summarized in table [ abundstruct ] .
each model is consistent with both single - dish and interferometer observations with @xmath47 - except for ch@xmath2oh for which the single - dish data indicate a hot core abundance an order of magnitude lower .
[ so_uvplot]a compares the model prediction for the so @xmath30 from @xcite with the sma results .
a good fit is obtained with a constant abundance of @xmath27 2.53@xmath48 in agreement with the results of @xcite .
it is noteworthy how well the sma data are described by this constant abundance model .
the single - dish observations of @xcite are not sensitive to the abundance in the innermost ( @xmath49 k ) region . in iras
162932422 @xcite an enhancement by two orders of magnitude in the innermost , @xmath50 k , region was inferred on the basis of single - dish observations . in the case of iras2a
, such enhancements appear to be ruled out by the interferometer observations presented in this paper .
in contrast to so , so@xmath4 only shows unresolved emission close to the central protostar .
@xcite did not detect so@xmath4 in the ngc 1333-iras2 envelope and the detection in this paper suggests a compact origin of the emission from this species with a jump in abundance of at least two orders of magnitude .
the absence of a clear so abundance enhancement at small scales is interesting in this context . simply comparing our observed [ so@xmath4]/[so ] ratio @xmath52 to the models of the sulphur chemistry by @xcite
would indicate that ngc 1333-iras2 is more evolved than iras 162932422 .
in terms of absolute abundances , however , the upper limit of @xmath53 of so is in contradiction with any of the hot core models that predict so abundances @xmath54 ( e.g. * ? ? ?
* ; * ? ? ?
these results therefore suggest that so does not probe a hot core around iras2a - or that its chemistry is significantly different from other known hot cores . for h@xmath3cn
no good fit to the single - dish observations is obtained with a constant abundance model .
this is similar to the case for co and hco@xmath55 where constant abundance models underestimate the intensity of the lowest excitation 10 transitions .
@xcite suggested that the abundances vary radially with a so - called `` drop '' profile : in these models freeze - out occurs in the region of the envelope where the temperature is low enough to prevent immediate desorption , but where the density is high enough that the freeze - out timescales are shorter than the lifetime of the core .
similar type abundance profiles have been confirmed by high angular resolution observations of h@xmath4co and co toward other protostars @xcite and found in detailed chemical - dynamical models @xcite . for the h@xmath3cn 43 observations in this paper ( fig .
[ h13cn_uvplot]b ) and the single - dish observations @xcite a combined best fit is obtained with an abundance profile where depletion occurs at the same radius as it does for co and hco@xmath55 at a density of @xmath56 @xmath12 and evaporation at small radii unresolved by the interferometer observations where the temperatures exceeds 7090 k ( radii smaller than @xmath27 100 au ) .
this is expected if hcn comes off dust grains at roughly the same temperatures as h@xmath4o ice mantles @xcite . in terms of tracing the dense gas in the envelope , hcn and h@xmath3cn appear to be most useful probes of the inner envelope and/or disk with high abundances , strong transitions and only weakly affected by outflows ( as it is also the case for the lower excitation @xmath57 lines ( e.g. , * ? ? ?
* ) ) . for the organic molecules the emission appears compact and mostly unresolved . for these species the abundances in the region where @xmath58 k ( diameter @xmath59 au or scales of 0.6@xmath1 ; 300 k@xmath20 ) are constrained by the observed high excitation lines , whereas their line intensities are largely independent of the abundance in the region outside this radius . for ch@xmath2oh
the abundance in the inner envelope derived from the interferometer data is an order of magnitude higher than that derived on the basis of multi - transition single - dish observations @xcite .
the ch@xmath2oh observations shown in figs . [ ch3oh_uvplot]c and [ ch3oh_uvplot]d do show a slowly decreasing visibility amplitude with increasing baseline lengths , but the data are consistent with an unresolved component at the 2@xmath28 level ( dashed lines in fig .
[ ch3oh_uvplot]c and d ) .
an even better fit is obtained with a slightly extended ( @xmath27150 - 200 au ) gaussian component ( dashed - dotted lines ) .
[ h13cn_uvplot][ch3oh_uvplot ]
as explained above and seen for other low - mass protostars an additional source of continuum emission has to be present in addition to the envelope to explain the interferometer data .
the sma observes the brightness distribution of this source and can thereby constrain its physical parameters . taking the compact flux from the continuum data and adopting a dust opacity of 1.82 @xmath60 g@xmath23 at 850 @xmath25 m ( * ?
* used in the envelope models ) , the mass ( gas+dust ) of a disk with a temperature of 150 k is 0.02 @xmath8 .
this in turn corresponds to an average h@xmath4 column density of 1.6@xmath61 @xmath62 for a circular disk filling the beam with a diameter of 200 au . even in models where the envelope does not have an inner cavity ,
a disk column density of @xmath63 @xmath62 is non - negligible compared to the hot core column density ( i.e. , pencil beam column density of the material in the envelope with a temperature higher than 90 k ) of 3@xmath64 @xmath62 .
these estimates of the disk column densities are simplistic and most likely a lower limit to the dust mass .
first , the temperature of the dust may not be coupled to that of the gas . lowering the dust temperature ,
however , increases the dust mass required to produce the same flux at the given wavelength .
for example , adopting a disk temperature of 30 k increases the dust mass by a factor 5 to 0.1 @xmath8 .
second , fig .
[ cont_specindex ] illustrates that the flux is consistent with a spectral index @xmath65 ( @xmath66 ) from cm @xcite through mm @xcite and submm wavelengths ( this paper ) . the data in fig .
[ cont_specindex ] are all from interferometer observations and cover an unprecendented range up to 350 ghz .
the power - law exponent of @xmath27 2 suggests that the compact source is caused by optically thick thermal dust emission from dust with opacity @xmath67 .
it would therefore be more appropriate to estimate the dust mass on the basis of the longer wavelength points .
taking the 3.6 cm measurements of @xcite , rather than the 850 @xmath25 m measurements from this paper , results in a dust mass of 0.10.5 @xmath8 ( for temperatures of 15030 k ) .
of course at the longer wavelengths it is less clear whether the dust opacity law is valid and if sources other than dust could be contributing to the observed compact emission . as discussed in sect .
[ continstruct ] , the continuum interferometer data resolve the compact structure into at least two components , which we have modeled as an extended gaussian structure ( diameter of 300 au ) and an unresolved component .
the size of the gaussian component is similar to that of disks around more evolved t tauri stars ( e.g. , * ? ? ?
similarly - sized disks are found around other deeply embedded protostars ( e.g. , l1157 ; * ? ? ?
* ) but smaller disks are found around other protostars ( e.g. , iras 16293 - 2422b ; * ? ? ?
an interesting question is whether the smaller disks result from truncation in binary systems ( e.g. , * ? ? ? * ) or whether there is an evolution of disk size from the earliest stages and whether the smaller disks simply reflect the youngest systems ( e.g. , * ? ? ?
the fairly extended gaussian component in ngc 1333-iras2a would argue in favor of the first suggestion , i.e. , that a circumstellar disk quickly builds up around low - mass protostars unless truncated in a binary system .
@xcite did not resolve the compact continuum source seen at 3.6 cm with a beam size of about 0.3@xmath1 ( 66 au ) .
if the 3.6 cm source is related to the submillimeter / millimeter source as the sed suggests , this would imply that the centimeter observations only pick up the peak , unresolved component of the submillimeter structure .
the sensitivity of the centimeter observations by @xcite suggests that they would not detect an extended ( 300 au ) , lower surface brightness component such as seen in the submillimeter observations .
in fact ngc 1333-iras2a may be a close ( @xmath68 au ) binary itself such as suggested by its quadrupolar outflow probed down to small scales ( e.g. * ? ? ? * ) .
it is possible that the compact component seen in these data traces the disks around each of these binary components with the resolved gaussian component reflecting an extended circumbinary disk .
the unresolved source is responsible for about 30% of the total flux of the compact component suggesting that it contains a similar fraction of the mass . in the discussion of the hot core scenario in [ lineabund ] we assumed that the envelope extends to small radii ( 23 au ; 0.1@xmath1 ) where the temperature increases to 250 k. the analysis of the continuum emission indicates that a disk is present around ngc 1333-iras2 , however , and an open question remains how much , if any , of the observed line emission can be attributed to such a disk . in particular , this is relevant if an inner cavity exists in the envelope around ngc 1333-iras2 : as shown in fig .
[ cont_uvamp ] the continuum data are also consistent with an envelope where the temperature never exceeds 75 k. such an inner cavity might be expected if the extended gaussian component discussed above reflects the disk formed as a result of the rotation of the core .
the size of the cavity ( radius of 285 au for the example with an envelope with an inner temperature of 75 k ) would then be the centrifugal radius of the envelope within which the disk is formed ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in this scenario the disk would also have to be warm as indicated by the observed high excitation lines .
observations of compact ch@xmath2oh emission around another class 0 object , l1157 , have suggested that the ch@xmath2oh emission is related to an accretion shock close to the central disk @xcite .
@xcite suggest that a disk with a heated accretion layer with a temperature of 170250 k exists around the class i young stellar object elias 29 from high @xmath69 fir / sub millimeter co observations .
lll|ll molecule & & envelope w. @xmath70 au & envelope w. @xmath71 au + & @xmath72@xmath73 & @xmath74@xmath75 & hot core abundance , @xmath76@xmath77 & disk abundance , @xmath78@xmath79 + hcn@xmath80 & 2@xmath81 & 2@xmath48 & 7@xmath81 & 2@xmath48 + so & 3@xmath48 & @xmath29 & @xmath828@xmath48 & 5@xmath83 + so@xmath4 & @xmath822.5@xmath83 & @xmath29 & 2@xmath81 & 9@xmath83 + ch@xmath2oh & 1.0@xmath48 & @xmath29 & 2@xmath84 & 3@xmath81 + ch@xmath2och@xmath2 & @xmath29 & @xmath29 & 3@xmath81 & 2@xmath48 + @xmath73envelope abundance where @xmath85 k ( for hcn : where furthermore @xmath86 @xmath12 ) .
so@xmath4 and ch@xmath2oh abundances from single - dish observations by @xcite .
+ @xmath75envelope abundance where @xmath87 @xmath12 and @xmath85 k. + @xmath77envelope abundance where @xmath50 k in model extending to @xmath88 k ( inner radius of 23.4 au ) .
+ @xmath79disk abundance in model where envelope extends to @xmath89 k ( inner radius of 285 au ) .
+ @xmath80calculated from h@xmath3cn abundance adopting @xmath90c:@xmath3c isotope ratio of 70 . + assuming that the compact line emission comes from a homogeneous medium with a density of @xmath91 @xmath12 ( constant density in a 100 au thick disk given the estimated column density ) and a temperature of 150 k ( see above ) ,
an estimate of the column density of each molecular species can be made using the non - lte escape probability code _
_ radex__moldata ] @xcite .
these can then be compared to the column densities inferred from the dust emission above . with these assumptions , the so and so@xmath4 abundances would be 5@xmath83 and 9@xmath83 , respectively , and the h@xmath3cn abundance 3@xmath2110@xmath92 ( corresponding to an hcn abundance of 2@xmath93 assuming standard isotopic ratios ) .
table [ abundstruct ] summarizes the derived abundance structures .
the value for the so abundance is consistent with the upper limit by @xcite toward the pre - main sequence disks around lkca15 .
the disk column densities should be considered as lower limits ( see sect .
[ continorigin ] ) and therefore any abundances derived using the dust column densities are upper limits . a more detailed treatment of the disk physical structure ( including its temperature and density variations ) together with more observations of the emission from a larger number of molecules is needed for a complete picture of the disk chemistry in these early deeply embedded stages .
possibly the best proof for either the `` hot core '' or `` disk '' scenario can come from resolving the molecular emission and addressing their intrinsic velocity fields .
@xcite suggested that the discovery of organic molecules was evidence for hot cores due to the alignment with the dust emission .
this is , however , not the case as their dust emission observed at smaller scales more likely arises in the disks around the two binary components in iras 162932422 .
in fact , if the reservoirs of organic molecules observed in iras 162932422 are related to the larger scale envelopes seen for example by single - dish continuum observations , molecular species should be seen on intermediate scales between the `` compact disk sources '' and the `` larger scale envelope '' .
observations of more complex , asymmetric sources such as iras 162932422 might therefore be the best way to distinguish between the `` accretion disk '' and `` hot core envelope '' scenarios .
we have presented a detailed analysis of the physical and chemical structure of the low - mass protostar , ngc 1333-iras2a based on high angular resolution ( 1 - 2@xmath1 ; 200 - 400 au ) observations with the submillimeter array .
the data are compared to detailed continuum and line radiative transfer models to zoom in on the inner few hundred au unhindered by emission from the lower density outer envelope .
the conclusions are as follows : * compact continuum emission is detected at 850@xmath25 m : the sma observations resolve the compact continuum component , which can be fitted by two structures in addition to the large scale envelope constrained by single - dish observations : a resolved , 300 au sized , gaussian brightness distribution and an unresolved ( @xmath94 au ) structure , the latter contributing about 30% of the flux of the compact continuum component .
* the spectrum of the compact continuum component follows a power - law , @xmath95 , from centimeter through submillimeter wavelengths .
the favored explanation is that its origin is optically thick thermal dust emission from a circumstellar disk .
the size of the disk around this young object suggests that the build - up of circumstellar disks proceeds rapidly in the protostellar stages .
smaller disks inferred around other protostars may be a consequence of tidal truncation by nearby binary companions .
* compact emission from lines of complex organic molecules including ch@xmath2och@xmath2 and ch@xmath2ocho , high excitation ch@xmath2oh transitions , deuterated methanol ch@xmath2od as well as lines of co , hcn , h@xmath3cn , so and so@xmath4 are detected .
the line data are interpreted in the context of previously published single - dish studies of the chemical structure of the protostellar envelope .
the so data are consistent with a low constant abundance throughout the envelope - an so abundance enhancement at small scales is ruled out by the current observations .
the complex organic species and so@xmath4 in contrast are enhanced on small scales .
h@xmath3cn follows the drop abundance structure also traced by co and hco@xmath55 with a high abundance in the outer region where the density is low and no depletion has occurred , an intermediate region with low temperatures and higher densities where depletion can occur and an inner region where hcn either evaporates off dust grains due to high temperatures or resides in the circumstellar disk . * a disk of the size suggested by the continuum emission
can not be neglected in the interpretation of line data .
if this disk has a warm ( @xmath6100 k ) layer , this could be the origin of the high excitation tracers and complex organic species on the smallest scales .
this paper illustrates the potential of the sma for studying the detailed variations in the physical and chemical structure of low - mass protostellar envelopes , probing their innermost regions and constraining the properties of circumstellar disks .
the sma data and presented analysis will form a template for a large survey of deeply embedded low - mass protostars currently ongoing at the sma .
we are grateful to the sma staff and in particular charlie qi for useful discussions about reduction of sma data .
we extend special thanks to those of hawaiian ancestry on whose sacred mountain we are privileged to be guests .
we thank the referee for a prompt and well - considered report .
this research was supported by nasa origins of solar systems grant nag5 - 13050 .
fls acknowledges financial support from the swedish research council .
astrochemistry research in leiden is supported by a nwo spinoza grant and a nova grant . | the submillimeter array has opened a new window to study the innermost warm and dense regions of the envelopes and disks around deeply embedded protostars .
this paper presents high - angular resolution ( @xmath0@xmath1 ) submillimeter observations of the class 0 young stellar object ngc 1333-iras2a .
dust continuum emission and lines of complex organic molecules such as ch@xmath2och@xmath2 and ch@xmath2ocho , high excitation ch@xmath2oh transitions , deuterated methanol ch@xmath2od as well as lines of co , hcn , hc@xmath3cn , so and so@xmath4 are detected on @xmath5 au scales .
the observations are interpreted using detailed radiative transfer models of the physical and chemical structure , consistent with both single - dish and interferometer data .
the continuum emission is explained by an extended envelope and a compact but resolved component , presumably a circumstellar disk with a diameter of 200300 au and a mass of @xmath6 a few@xmath7 @xmath8 .
if related to the rotation of the envelope , then the size of this disk suggests a centrifugal barrier of 200300 au , which implies that the temperature in the envelope does not increase above 100 k. its large size also suggests that the build - up of disks proceeds rapidly throughout the early protostellar stages .
the smaller ( @xmath9 au ) disks found around other deeply embedded protostars may be a result of tidal truncation .
the high - resolution observations of so can be explained with a simple constant abundance , @xmath6 @xmath10 , constrained through single - dish observations , whereas those of h@xmath3cn and the organic species require high abundances , increased by one to two orders of magnitude , or an additional compact source of emission at small scales .
the compact molecular emission could originate in a hot core region of the inner envelope , but a more likely reservoir is the circumstellar disk . |
in @xcite , the authors analyzed the average capacity for a point - to - point channel model where slow changes result in varying severity of noise . in this work
, we apply a similar concept to interference networks by assuming that slowly changing deep fading conditions result in link erasures .
we consider the linear interference network introduced by wyner @xcite , with the consideration of two fading effects .
long - term fluctuations that result in link erasures over a complete block of time slots , and short - term fluctuations that allow us to assume that any specific joint realization for the non - zero channel coefficients , will take place with zero probability .
we study the problem of associating receivers with transmitters and setting up the backhaul links for coordinated multi - point ( comp ) transmission , in order to achieve the optimal average degrees of freedom ( dof ) .
this problem was studied in @xcite for the case of no erasures . here ,
we extend the schemes in @xcite to consider the occurrence of link erasures , and propose new schemes that lead to achieving better average dof at high probabilities of erasure .
we use the standard model for the @xmath4user interference channel with single - antenna transmitters and receivers , @xmath5 where @xmath6 is the time index , @xmath7 is the transmitted signal of transmitter @xmath8 , @xmath9 is the received signal at receiver @xmath10 , @xmath11 is the zero mean unit variance gaussian noise at receiver @xmath10 , and @xmath12 is the channel coefficient from transmitter @xmath8 to receiver @xmath10 over the time slot @xmath6 . we remove the time index in the rest of the paper for brevity unless it is needed . for any set
@xmath13 $ ] , we use the abbreviations @xmath14 , @xmath15 , and @xmath16 to denote the sets @xmath17 , @xmath18 , and @xmath19 , respectively .
finally , we use @xmath20 $ ] to denote the set @xmath21 .
each transmitter can only be connected to its corresponding receiver as well as one following receiver , and the last transmitter can only be connected to its corresponding receiver .
more precisely , @xmath22.\ ] ] in order to consider the effect of long - term fluctuations ( shadowing ) , we assume that communication takes place over blocks of time slots , and let @xmath0 be the probability of block erasure . in each block , we assume that for each @xmath8 , and each @xmath23 , @xmath24 with probability @xmath0 .
moreover , short - term channel fluctuations allow us to assume that in each time slot , all non - zero channel coefficients are drawn independently from a continuous distribution .
finally , we assume that global channel state information is available at all transmitters and receivers .
for each @xmath25 $ ] , let @xmath26 be the message intended for receiver @xmath10 , and @xmath27 $ ] be the transmit set of receiver @xmath10 , i.e. , those transmitters with the knowledge of @xmath26 .
the transmitters in @xmath28 cooperatively transmit the message @xmath26 to the receiver @xmath10 .
the messages @xmath29 are assumed to be independent of each other .
the _ cooperation order _
@xmath1 is defined to be the maximum transmit set size : @xmath30 a message assignment strategy is defined by a sequence of transmit sets @xmath31 , k\in\{1,2,\ldots\}$ ] . for each positive integer @xmath32 and
@xmath33 $ ] , @xmath34 , |{\cal t}_{i , k}| \leq m$ ] .
we use message assignment strategies to define the transmit sets for a sequence of @xmath4user channels .
the @xmath35 channel in the sequence has @xmath36 users , and the transmit sets for this channel are defined as follows .
the transmit set of receiver @xmath10 in the @xmath35 channel in the sequence is the transmit set @xmath37 of the message assignment strategy .
the average power constraint at each transmitter is @xmath38 . in each block of time slots , the rates @xmath39 are achievable
if the decoding error probabilities of all messages can be simultaneously made arbitrarily small as the block length goes to infinity , and this holds for almost all realizations of non - zero channel coefficients .
the sum capacity @xmath40 is the maximum value of the sum of the achievable rates .
the total number of degrees of freedom ( @xmath41 ) is defined as @xmath42 . for a probability of block erasure @xmath0 ,
we let @xmath43 be the average value of @xmath41 over possible choices of non - zero channel coefficients . for a @xmath32-user channel , we define @xmath44 as the best achievable @xmath43 over all choices of transmit sets satisfying the cooperation order constraint in . in order to simplify our analysis
, we define the asymptotic average per user dof @xmath45 to measure how @xmath44 scale with @xmath32 , @xmath46 we call a message assignment strategy _ optimal _ for a given erasure probability @xmath0 , if there exists a sequence of coding schemes achieving @xmath45 using the transmit sets defined by the message assignment strategy .
a message assignment strategy is @xmath47 if it is optimal for all values of @xmath0 .
we first consider the case where each receiver can be served by only one transmitter .
this reflects the problem of associating mobile users with cells in a cellular downlink scenario .
we start by discussing orthogonal schemes ( tdma - based ) for this problem , and then show that the proposed schemes are optimal
. it will be useful in the rest of this section to view each realization of the network where some links are erased , as a series of subnetworks that do not interfere .
we say that a set of @xmath36 users with successive indices @xmath48 form a subnetwork if the following two conditions hold : the first condition is that @xmath49 or it is the case that message @xmath50 does not cause interference at @xmath51 , either because the direct link between the transmitter carrying @xmath50 and receiver @xmath52 is erased , or the transmitter carrying @xmath50 is not connected to the @xmath53 receiver .
secondly , @xmath54 or it is the case that message @xmath55 does not cause interference at @xmath56 , because the carrying transmitter is not connected to one of the receivers @xmath57 and @xmath58 .
we say that the subnetwork is @xmath59 if the transmitters carrying messages for users in the subnetwork have successive indices and for any transmitter @xmath6 carrying a message for a user in the subnetwork , and receiver @xmath60 such that @xmath61 and @xmath62 , the channel coefficient @xmath63 .
for @xmath64 $ ] , let @xmath65 be the number of messages available at the @xmath53 transmitter , and let @xmath66 .
it is clear that the sequence @xmath67 can be obtained from the transmit sets @xmath68 $ ] ; it is also true , as stated in the following lemma , that the converse holds .
we borrow the notion of _ irreducible _ message assignments from @xcite . for @xmath2 , an irreducible message assignment will have each message assigned to one of the two transmitters connected to its designated receiver .
[ lem : equiv ] for any irreducible message assignment where each message is assigned to exactly one transmitter , i.e. , @xmath69 $ ] , the transmit sets @xmath28 , @xmath64,$ ] are uniquely characterized by the sequence @xmath67 .
since each message can only be available at one transmitter , then this transmitter has to be connected to the designated receiver .
more precisely , @xmath70 , and @xmath71 .
it follows that each transmitter carries at most two messages and the first transmitter carries at least the message @xmath72 , i.e. , @xmath73 , and @xmath74 .
assume that @xmath75 $ ] , then @xmath76 $ ] .
for the remaining case , we know that there exists @xmath77 such that @xmath78 , since @xmath79 ; we handle this case in the rest of the proof .
let @xmath80 be the smallest index of a transmitter that carries no messages , i.e. , @xmath81 .
we now show how to reconstruct the transmit sets @xmath82 from the sequence @xmath83 .
we note that @xmath84 , \forall i\in[x]$ ] , and since @xmath85 , it follows that @xmath86 , \forall i\notin [ x]$ ] .
it follows that @xmath87 . since @xmath88 , we know that at most one transmitter in the first @xmath89 transmitters carries two messages . since @xmath87 , and
@xmath90 $ ] , it follows that there exists an index @xmath91 $ ] such that @xmath92 , and @xmath93\backslash\{y\}$ ] .
it is now clear that the @xmath94 transmitter carries messages @xmath95 and @xmath96 , and each transmitter with an index @xmath97 is carrying message @xmath98 , and each transmitter with an index @xmath99 is carrying message @xmath100 .
the transmit sets are then determined as follows .
@xmath101 $ ] and @xmath102 .
we view the network as a series of subnetworks , where the last transmitter in each subnetwork is either inactive or the last transmitter in the network . if the last transmitter in a subnetwork is inactive , then the transmit sets in the subnetwork are determined in a similar fashion to the transmit sets @xmath103 $ ] , in the above scenario . if the last transmitter in the subnetwork is the @xmath104 transmitter , and @xmath105 , then each message in this subnetwork is available at the transmitter with the same index .
we use lemma [ lem : equiv ] to describe message assignment strategies for large networks through repeating patterns of short ternary strings . given a ternary string @xmath106 of fixed length @xmath107 such that @xmath108 , we define @xmath67 , @xmath109 as follows : * @xmath110 if @xmath111 , * @xmath112 if @xmath113 .
we now evaluate all possible message assignment strategies satisfying the cell association constraint using ternary strings through the above representation .
we only restrict our attention to irreducible message assignments , and note that if there are two transmitters with indices @xmath10 and @xmath8 such that @xmath114 and each is carrying two messages , then there is a third transmitter with index @xmath36 such that @xmath115 that carries no messages .
it follows that any string defining message assignment strategies that satisfy the cell association constraint , has to have one of the following forms : * @xmath116 , * @xmath117 , * @xmath118 , * @xmath119 .
we now introduce the three candidate message assignment strategies illustrated in figure [ fig : msgassignment ] , and characterize the tdma per user dof achieved through each of them ; we will show later that the optimal message assignment strategy at any value of @xmath0 is given by one of the three introduced strategies .
we first consider the message assignment strategy defined by the string having the form @xmath116 . here , each message is available at the transmitter having the same index .
[ lem : highp ] under the restriction to the message assignment strategy @xmath120,$ ] and orthogonal tdma schemes , the average per user dof is given by , @xmath121 we will first explain a transmission scheme where @xmath122 dof is achieved , and then modify it to show how to achieve @xmath123 . for each user with and odd index @xmath10 , message @xmath26
is transmitted whenever the channel coefficient @xmath124 ; the rate achieved by these users contributes to the average per user dof by @xmath125 . for each user with an even index @xmath10 , message @xmath26
is transmitted whenever the following holds : @xmath126 , @xmath50 does not cause interference at @xmath51 , and the transmission of @xmath26 will not disrupt the communication of @xmath127 to its designated receiver ; we note that this happens if and only if @xmath128@xmath129 @xmath130 .
it follows that the rate achieved by users with even indices contributes to the average per user dof by @xmath131 .
we now show a modification of the above scheme to achieve @xmath123 . as
above , users with odd indices have priority , i.e. , their messages are delivered whenever their direct links exist , and users with even indices deliver their messages whenever their direct links exist and the channel connectivity allows for avoiding conflict with priority users .
however , we make an exception to the priority setting in atomic subnetworks consisting of an odd number of users , and the first and last users have even indices ; in these subnetworks , one extra dof is achieved by allowing users with even indices to have priority and deliver their messages .
the resulting extra term in the average per user dof is calculated as follows . fixing a user with an even index , the probability that this user is the first user in a subnetwork consisting of an odd number of users in a large network is @xmath132 ; for each of these events , the sum dof is increased by @xmath133 , and hence the added term to the average per user dof is equal to half this value , since every other user has an even index .
the optimality of the above scheme within the class of orthogonal tdma - based schemes follows directly from ( * ? ? ?
* theorem @xmath133 ) for each realization of the network
. we will show later that the above scheme is optimal at high erasure probabilities . in @xcite ,
the optimal message assignment for the case of no erasures was characterized .
the per user dof was shown to be @xmath134 , and was achieved by deactivating every third transmitter and achieving @xmath133 dof for each transmitted message .
we now consider the extension of this message assignment illustrated in figure [ fig : lowp ] , which will be shown later to be optimal for low erasure probabilities .
[ lem : lowp ] under the restriction to the message assignment strategy defined by the string @xmath135 , and orthogonal tdma schemes , the average per user dof is given by , @xmath136 for each user with an index @xmath10 such that @xmath137 or @xmath138 , message @xmath26 is transmitted whenever the link between the transmitter carrying @xmath26 and the @xmath53 receiver is not erased ; these users contribute to the average per user dof by a factor of @xmath139 . for each user with an index @xmath10 such that @xmath140 , message @xmath26 is transmitted through @xmath141 whenever the following holds : @xmath142
, message @xmath50 is not transmitted because @xmath143 , and the transmission of @xmath26 will not be disrupted by the communication of @xmath127 through @xmath144 because @xmath145 ; these users contribute to the average per user dof by a factor of @xmath146 . using the considered message assignment strategy
, the tdma optimality of this scheme follows from ( * ? ? ?
* theorem @xmath133 ) for each network realization .
we now consider the message assignment strategy illustrated in figure [ fig : middlep ] .
we will show later that this strategy is optimal for a middle regime of erasure probabilities .
[ lem : middlep ] under the restriction to the message assignment strategy defined by the string @xmath147 , and orthogonal tdma schemes , the average per user dof is given by , @xmath148 as in the proof of lemma [ lem : highp ] , we first explain a transmission scheme achieving part of the desired rate , and then modify it to show how the extra term can be achieved .
let each message with an odd index be delivered whenever the link between the transmitter carrying the message and the designated receiver is not erased ; these users contribute to the average per user dof by a factor of @xmath149 . for each user with an even index @xmath10 ,
if @xmath150 , then @xmath26 is transmitted through @xmath144 whenever the following holds : @xmath124 , message @xmath127 is not transmitted through @xmath144 because @xmath151 , and the transmission of @xmath26 can will not be disrupted by the communication of @xmath50 through @xmath141 because either @xmath152 or @xmath153 ; these users contribute to the average per user dof by a factor of @xmath154 . for each user with an even index @xmath10 such that @xmath10 is a multiple of @xmath155 ,
@xmath26 is transmitted through @xmath141 whenever @xmath156 , and the transmission of @xmath26 will not disrupt the communication of @xmath50 through @xmath157 because either @xmath143 or @xmath158 ; these users contribute to the average per user dof by a factor of @xmath159 .
we now modify the above scheme to show how @xmath160 can be achieved .
since the @xmath53 transmitter is inactive for every @xmath10 that is a multiple of @xmath155 , users @xmath161 are separated from the rest of the network for every @xmath10 that is a multiple of @xmath155 , i.e. , these users form a subnetwork .
we explain the modification for the first four users , and it will be clear how to apply a similar modification for every following set of four users . consider the event where message @xmath72 does not cause interference at @xmath162 , because either @xmath163 or @xmath164 , and it is the case that @xmath165 , @xmath166 , @xmath167 , and @xmath168 ; this is the event that users @xmath169 form an atomic subnetwork , and it happens with probability @xmath170 . in this case , we let messages @xmath171 and @xmath172 have priority instead of message @xmath173 , and hence the sum dof for messages @xmath174 is increased by @xmath133 .
it follows that an extra term of @xmath175 is added to the average per user dof .
the tdma optimality of the illustrated scheme follows from ( * ? ? ?
* theorem @xmath133 ) for each network realization . in figure
[ fig : monenorm ] , we plot the values of @xmath176 , @xmath177 , and @xmath178 , and note that @xmath179 equals @xmath123 at high probabilities of erasure , and equals @xmath180 at low probabilities of erasure , and equals @xmath160 in a middle regime . , [ lem : lowp ] , and [ lem : middlep ] , normalized by @xmath181 . ]
we now show that under the restriction to tdma schemes , one of the message assignment strategies illustrated in lemmas [ lem : highp ] , [ lem : lowp ] , and [ lem : middlep ] is optimal at any value of @xmath0 .
[ thm : tdma ] for a given erasure probability @xmath0 , let @xmath182 be the average per user dof under the restriction to orthogonal tdma schemes , then at any value @xmath183 the following holds , @xmath184 where @xmath123 , @xmath180 , and @xmath160 are given in , , and , respectively .
the inner bound follows from lemmas [ lem : highp ] , [ lem : lowp ] , and [ lem : middlep ] . in order to prove the converse
, we need to consider all irreducible message assignment strategies where each message is assigned to a single transmitter .
we know from lemma [ lem : highp ] that the tdma average per user dof achieved through the strategy defined by the string of all ones having the form @xmath116 equals @xmath123 , and hence the upper bound holds in this case .
we now show that the tdma average per user dof achieved through strategies defined by strings of the form @xmath185 is upper bounded by a convex combination of @xmath123 and @xmath180 , and hence , is upper bounded by @xmath186 .
the considered message assignment strategy splits each network into subnetworks consisting of a transmitter carrying two messages followed by a number of transmitters , each is carrying one message , and the last transmitter in the subnetwork carries no messages .
we first consider the case where the number of transmitters carrying single messages is odd .
we consider the simple scenario of the message assignment strategy defined by the string @xmath187 , and then the proof will be clear for strategies defined by strings of the form @xmath188 that have an arbitrary odd number of ones . in this case , it suffices to show that the average per user dof in the first subnetwork is upper bounded by a convex combination of @xmath123 and @xmath180 .
the first subnetwork consists of the first five users ; @xmath72 and @xmath171 can be transmitted through @xmath189 .
@xmath173 , @xmath172 and @xmath190 can be transmitted through @xmath191 , @xmath192 , and @xmath193 , respectively , and the transmit signal @xmath194 is inactive .
we now explain the optimal tdma scheme for the considered subnetwork .
we first explain a simple scheme and then modify it to get the optimal scheme .
each of the messages @xmath72 , @xmath173 , and @xmath190 is delivered whenever the direct link between its carrying transmitter and its designated receiver is not erased .
message @xmath171 is delivered whenever message @xmath72 is not transmitted , and message @xmath173 is not causing interference at @xmath162 .
message @xmath172 is transmitted whenever @xmath190 is not causing interference at @xmath195 , and the transmission of @xmath172 through @xmath192 will not disrupt the communication of @xmath173 .
we now explain the modification ; if there is an atomic subnetwork consisting of users @xmath169 , then we switch the priority setting within this subnetwork , and messages @xmath171 and @xmath172 will be delivered instead of message @xmath173 .
the tdma optimality of this scheme for each realization of the network follows from ( * ? ? ?
* theorem @xmath133 ) .
now , we note that the average sum dof for messages @xmath196 is equal to their sum dof in the original scheme plus an extra term due to the modification .
the average sum dof for messages @xmath197 in the original scheme equals @xmath198 , and the sum of the average sum dof for messages @xmath199 and the extra term is upper bounded by @xmath200 .
it follows that the average per user dof is upper bounded by @xmath201 .
the proof can be generalized to show that the average tdma per user dof for message assignment strategies defined by strings of the form @xmath202 with an odd number of ones @xmath107 , is upper bounded by @xmath203 . for message assignment strategies defined by a string of the form @xmath202 with an even number of ones @xmath107 ,
it can be shown in a similar fashion as above that the tdma average per user dof is upper bounded by @xmath204 . also , for strategies defined by a string of the form @xmath205 with a number of ones @xmath107 , the tdma average per user dof is the same as that of a strategy defined by a string of the form @xmath202 with the same number of ones , and hence , is upper bounded by a convex combination of @xmath123 and @xmath180 . finally , for strategies defined by a string of the form @xmath206 with a number of ones @xmath107 , it can be shown in a similar fashion as above that the average per user dof is upper bounded by @xmath207 .
we now characterize the average per user dof for the cell association problem by proving that tdma schemes are optimal for any candidate message assignment strategy . in order to prove an information theoretic upper bound on the per user dof for each network realization
, we use lemma @xmath155 from @xcite , which we restate below . for any set of receiver indices
@xmath13 $ ] , define @xmath208 as the set of indices of transmitters that exclusively carry the messages for the receivers in @xmath209 , and the complement set is @xmath210 .
more precisely , @xmath211\backslash\cup_{i \notin { \cal a } } { \cal t}_i$ ] .
[ @xcite ] [ lem : dofouterbound ] if there exists a set @xmath212 $ ] , a function @xmath213 , and a function @xmath214 whose definition does not depend on the transmit power constraint @xmath38 , and @xmath215 , then the sum dof @xmath216 .
[ thm : mone ] the average per user dof for the cell association problem is given by , @xmath217 where @xmath123 , @xmath180 , and @xmath160 are given in , , and , respectively . in order to prove the statement , we need to show that @xmath218 ; we do so by using lemma [ lem : dofouterbound ] to show that for any irreducible message assignment strategy satisfying the cell association constraint , and any network realization , the asymptotic per user dof is given by that achieved through the optimal tdma scheme .
consider message assignment strategies defined by strings having one of the forms @xmath116 , @xmath219 , and @xmath205 .
we view each network realization as a series of atomic subnetworks , and show that for each atomic subnetwork , the sum dof is achieved by the optimal tdma scheme . for an atomic subnetwork consisting of a number of users @xmath107 , we note that @xmath220 users are active in the optimal tdma scheme ; we now show in this case using lemma [ lem : dofouterbound ] that the sum dof for users in the subnetwork is bounded by @xmath220 .
let the users in the atomic subnetwork have the indices @xmath221 , then we use lemma [ lem : dofouterbound ] with the set @xmath222 , except the cases of message assignment strategies defined by strings having one of the forms @xmath116 and @xmath205 with an even number of ones , where we use the set @xmath223 .
we now note that each transmitter that carries a message for a user in the atomic subnetwork and has an index in @xmath210 , is connected to a receiver in @xmath209 , and this receiver is connected to one more transmitter with an index in @xmath208 , and hence , the missing transmit signals @xmath224 can be recovered from @xmath225 and @xmath226 .
the condition in the statement of lemma [ lem : dofouterbound ] is then satisfied ; allowing us to prove that the sum dof for users in the atomic subnetwork is upper bounded by @xmath227 .
the proof is similar for message assignment strategies defined by strings that have the form @xmath228 .
however , there is a difference in selecting the set @xmath209 for atomic subnetworks consisting of users with indices @xmath229 , where @xmath230 , and messages @xmath231 and @xmath232 are both available at transmitter @xmath233 . in this case
, we apply lemma [ lem : dofouterbound ] with the set @xmath209 defined as above , but including indices @xmath234 and excluding indices @xmath235 .
it can be seen that the condition in lemma [ lem : dofouterbound ] will be satisfied in this case , and the proved upper bound on the sum dof for each atomic subnetwork , is achievable through tdma . in figure
[ fig : monemax ] , we plot @xmath236 at each value of @xmath0 . the result of theorem [ thm : mone ] implies that the message assignment strategies considered in lemmas [ lem : highp ] , [ lem : lowp ] , [ lem : middlep ] are optimal at high , low , and middle values of the erasure probability @xmath0 , respectively .
we note that in densely connected networks at a low probability of erasrue , the _ interference - aware _ message assignment strategy in figure [ fig : lowp ] is optimal ; through this assignment , the maximum number of interference free communication links can be created for the case of no erasures . on the other hand
, the linear nature of the channel connectivity does not affect the choice of optimal message assignment at high probability of ersure . as the effect of interference diminishes at high probability of erasure , assigning each message to a unique transmitter , as in the strategy in figure [ fig : highp ] , becomes the only criterion of optimality . at middle values of @xmath0 , the message assignment strategy in figure [ fig : middlep ]
is optimal ; in this assignment , the network is split into four user subnetworks . in the first subnetwork ,
the assignment is optimal as the maximum number of interference free communication links can be created for the two events where there is an atomic subnetwork consisting of users @xmath237 or users @xmath169 .
we have shown that there is no message assignment strategy for the cell association problem that is optimal for all values of @xmath0 .
we show in this section that this statement is true even for the case where each message can be available at more than one transmitter ( @xmath3 ) .
recall that for a given value of @xmath1 , we say that a message assignment strategy is universally optimal if it can be used to achieve @xmath45 for all values of @xmath0 .
[ thm : comp ] for any value of the cooperation constraint @xmath238 , there does not exist a universally optimal message assignment strategy the proof follows from theorem [ thm : mone ] for the case where @xmath2 .
we show that for any value of @xmath3 , any message assignment strategy that enables the achievability of @xmath45 at high probabilities of erasure , is not optimal for the case of no erasures , i.e. , can not be used to achieve @xmath45 for @xmath239 . for any message assignment strategy , consider the value of @xmath240 and note this value equals the average number of transmitters in a transmit set that can be connected to the designated receiver . more precisely , @xmath241 where @xmath28 in corresponds to an optimal message assignment strategy at high probabilities of erasure .
it follows that there exists a value @xmath242 such that for any message assignment strategy that enables the achievability of @xmath45 for @xmath243 , almost all messages are assigned to the two transmitters that can be connected to the designated receiver , i.e. , if we let @xmath244 , then @xmath245 .
we recall from @xcite that for the case of no erasures , the average per user dof equals @xmath246 .
we also note that following the same footsteps as in the proof of ( * ? ? ?
* theorem @xmath247 ) , we can show that for any message assignment strategy such that @xmath245 , the per user dof for the case of no erasures is upper bounded by @xmath248 ; we do so by using lemma [ lem : dofouterbound ] for each @xmath4user channel with the set @xmath209 defined such that the complement set @xmath249 , i=(2m-1)(j-1)+m , j\in{\bf z}^+\}$ ] .
the condition of optimality identified in the proof of theorem [ thm : comp ] for message assignment strategies at high probabilities of erasure suggest a new role for cooperation in dynamic interference networks .
the availability of a message at more than one transmitter may not only be used to cancel its interference at other receivers , but to increase the chances of connecting the message to its designated receiver .
this new role leads to three effects at high erasure probability .
the achieved dof in the considered linear interference network becomes larger than that of @xmath32 parallel channels , in particular , @xmath250 .
secondly , as the effect of interference diminishes at high probabilities of erasures , all messages can simply be assigned to the two transmitters that may be connected to their designated receiver , and a simple interference avoidance scheme can be used in each network realization , as we show below in the scheme of theorem [ thm : mtwoic ] .
it follows that channel state information is no longer needed at transmitters , and only information about the slow changes in the network topology is needed to achieve the optimal average dof . finally , unlike the optimal scheme of (
* theorem @xmath155 ) for the case of no erasures , where some transmitters are always inactive , achieving the optimal dof at high probabilities of erasure requires all transmitters to be used in at least one network realization .
we now restrict our attention to the case where @xmath251 . here , each message can be available at two transmitters , and transmitted jointly by both of them .
we study two message assignment strategies that are optimal in the limits of @xmath252 and @xmath253 , and derive inner bounds on the average per user dof @xmath254 based on the considered strategies . in @xcite , the message assignment of figure [ fig : mtwojone ]
was shown to be dof optimal for the case of no erasures ( @xmath239 ) .
the network is split into subnetworks , each with five consecutive users .
the last transmitter of each subnetwork is deactivated to eliminate inter - subnetwork interfeerence . in the first subnetwork ,
message @xmath173 is not transmitted , and each other message is received without interference at its designated receiver .
note that the transmit beams for messages @xmath72 and @xmath190 contributing to the transmit signals @xmath191 and @xmath194 , respectively , are designed to cancel the interference at receivers @xmath162 and @xmath195 , respectively .
an analog scheme is used in each following subnetwork .
the value of @xmath254 is thus @xmath255 for the case where @xmath239 . in order to prove the following result
, we extend the message assignment of figure [ fig : mtwojone ] to consider the possible presence of block erasures .
[ thm : mtwoicaware ] for @xmath251 , the following average per user dof is achievable , @xmath256 where , @xmath257 and is asymptotically optimal as @xmath252 . we know from @xcite that @xmath258 , and hence , it suffices to show that the inner bound in is valid . for each @xmath25 $ ] .
message @xmath26 is assigned as follows , @xmath259 we illustrate this message assignment in figure [ fig : mtwojonenew ] .
we note that the transmit signals @xmath260 are inactive , and hence , we split the network into five user subnetworks with no interference between successive subnetworks .
we explain the transmission scheme in the first subnetwork and note that a similar scheme applies to each following subnetwork . in the proposed transmission scheme ,
any receiver is either inactive or receives its desired message without interference , and any transmitter will not transmit more than one message for any network realization .
it follows that 1 dof is achieved for each message that is transmitted .
messages @xmath72 , @xmath171 , @xmath172 , and @xmath190 are transmitted through @xmath189 , @xmath191 , @xmath192 , and @xmath193 , respectively , whenever the coefficients @xmath261 , @xmath165 , @xmath262 , and @xmath263 , respectively .
note that the transmit beam for message @xmath72 contributing to @xmath191 can be designed to cancel its interference at @xmath162 .
similarly , the interference caused by @xmath190 at @xmath195 can be cancelled through @xmath192 .
it follows that @xmath181 dof is achieved for each of @xmath264 , and hence , @xmath265 .
also , message @xmath171 is transmitted through @xmath189 if it can not be transmitted through @xmath191 and message @xmath72 is not transmitted through @xmath189 .
more precisely , message @xmath171 is transmitted through @xmath189 if @xmath266 and @xmath267 and @xmath163 , thereby achieving an extra @xmath268 dof .
similarly , message @xmath172 can be transmitted through @xmath193 if @xmath269 and @xmath270 and @xmath271 .
it follows that , @xmath272 finally , message @xmath173 will be transmitted through @xmath192 if message @xmath172 is not transmitted through @xmath192 , and message @xmath171 is not causing interference at @xmath273 .
message @xmath172 is not transmitted through @xmath192 whenever the coefficient @xmath269 , and message @xmath171 does not cause interference at @xmath273 whenever the coefficient @xmath274 or the coefficient @xmath275 or @xmath171 can be transmitted through @xmath189 .
more precisely , message @xmath173 is transmitted through @xmath192 if and only if all the following is true : * @xmath167 , and @xmath276 * @xmath274 , or @xmath275 , or it is the case that @xmath277 and @xmath267 .
it follows that @xmath278 dof is achieved for message @xmath173 , where , @xmath279 similary , @xmath173 can be transmitted through @xmath191 if and only if message @xmath171 is not transmitted through @xmath191 and message @xmath172 is either not transmitted or can be transmitted without causing interference at @xmath273 , i.e. , if and only if all the following is true : * @xmath166 , and @xmath274 * @xmath276 , or @xmath280 , or it is the case that @xmath281 and @xmath270 . the above conditions are satisfied with probability @xmath278 . since we have counted twice the event that @xmath167 and @xmath282 and @xmath166 and @xmath274 , it follows that @xmath283 dof is achieved for @xmath173 . summing the dof achieved for other messages in ,
we conclude that , @xmath284 which is the same inequality as in .
although the scheme of theorem [ thm : mtwoicaware ] is optimal for the case of no erasures ( @xmath239 ) , we know from theorem [ thm : comp ] that better schemes exist at high erasure probabilities . since in each five user subnet in the scheme of theorem
[ thm : mtwoicaware ] , only three users have their messages assigned to the two transmitters that can be connected to their receivers , and two users have only one of these transmitters carrying their messages , we get the asymptotic limit of @xmath285 for the achieved average per user dof normalized by @xmath181 as @xmath253 .
this leads us to consider an alternative message assignment where the two transmitters carrying each message @xmath10 are the two transmitters @xmath286 that can be connected to its designated receiver .
such assignment would lead the ratio @xmath287 as @xmath253 . in the following theorem
, we analyze a transmission scheme based on this assignment .
[ thm : mtwoic ] for @xmath251 , the following average per user dof is achievable , @xmath288 where , @xmath289 and , @xmath290 for any message assignment , no message can be transmitted if the links from both transmitters carrying the message to its designated receiver are absent , and hence , the average dof achieved for each message is at most @xmath291 .
it follows that @xmath292 .
we then need only to prove that the inner bound in is valid . in the achieving scheme ,
each message is assigned to the two transmitters that may be connected to its designated receiver , i.e. , @xmath293 $ ] .
also , in each network realization , each transmitter will transmit at most one message and any transmitted message will be received at its designated receiver without interference .
it follows that 1 dof is achieved for any message that is transmitted , and hence , the probability of transmission is the same as the average dof achieved for each message . each message @xmath26 such that @xmath294 is transmitted through @xmath141 whenever @xmath142 , and is transmitted through @xmath144 whenever @xmath152 and @xmath124 .
it follows that @xmath295 dof is achieved for each of these messages , where , @xmath296 we now consider messages @xmath26 such that @xmath297 .
any such message is transmitted through @xmath141 whenever @xmath142 and @xmath298 .
we note that whenever the channel coefficient @xmath299 , message @xmath26 can not be transmitted through @xmath141 as the transmission of @xmath26 through @xmath141 in this case will prevent @xmath50 from being transmitted due to either interference at @xmath300 or sharing the transmitter @xmath141 .
it follows that @xmath301 dof is achieved for transmission of @xmath26 through @xmath141 .
also , message @xmath26 is transmitted through @xmath144 whenever it is not transmitted through @xmath141 and @xmath124 and either @xmath152 or message @xmath50 is transmitted through @xmath157 .
more precisely , @xmath26 is transmitted through @xmath144 whenever all the following is true : @xmath124 , and either @xmath302 or it is the case that @xmath156 and @xmath299 and @xmath303 .
it follows that @xmath304 is achieved for transmission of @xmath26 through @xmath305 , and hence , @xmath306 dof is achieved for each message @xmath26 such that @xmath297 , where , @xmath307 we now consider messages @xmath26 such that @xmath308 .
any such message is transmitted through @xmath141 whenever all the following is true : * @xmath142 . * either @xmath309 or @xmath50 is not transmitted .
* @xmath127 is not causing interference at @xmath51 .
the first condition is satisfied with probability @xmath181 . in order to compute the probability of satisfying the second condition
, we note that @xmath50 is not transmitted for the case when @xmath299 only if @xmath310 is transmitted through @xmath157 and causing interference at @xmath300 , i.e. , only if @xmath311 and @xmath312 and @xmath303 .
it follows that the second condition is satisfied with probability @xmath313 .
the third condition is not satisfied only if @xmath124 and @xmath314 , and hence , will be satisfied with probability at least @xmath315 . moreover ,
even if if @xmath124 and @xmath314 , the third condition can be satisfied if message @xmath127 can be transmitted through @xmath316 without causing interference at @xmath317 , i.e. , if @xmath318 and @xmath319 .
it follows that the third condition will be satisfied with probability @xmath320 , and @xmath321 dof is achieved by transmission of @xmath26 through @xmath141 , where , @xmath322 message @xmath26 such that @xmath308 is transmitted through @xmath144 whenever @xmath124 , and @xmath323 , and either @xmath324 or @xmath50 is transmitted through @xmath157 .
it follows that @xmath325 dof is achieved by transmission of @xmath26 through @xmath144 , where , @xmath326 and hence , @xmath327 dof is achieved for each message @xmath26 such that @xmath308 . we finally get , @xmath328 which is the same inequality as in .
we plot the inner bounds of and in figure [ fig : mtwo ] .
we note that below a threshold erasure probability @xmath329 , the scheme of theorem [ thm : mtwoicaware ] is better , and hence is proposed to be used in this case . for higher probabilities of erasure , the scheme of theorem [ thm : mtwoic ] should be used .
it is worth mentioning that we also studied a scheme based on the message assignment @xmath330 $ ] , that is introduced in @xcite .
however , we did not include it here as it does not increase the maximum of the bounds derived in and at any value of @xmath0 .
finally , although the considered channel model allows for using the interference alignment scheme of @xcite over multiple channel realizations ( symbol extensions ) , all the proposed schemes require only coding over one channel realization because of the sparsity of the linear network .
we considered the problem of assigning messages to transmitters in a linear interference network with link erasure probability @xmath0 , under a constraint that limits the number of transmitters @xmath1 at which each message can be available .
for the case where @xmath2 , we identified the optimal message assignment strategies at different values of @xmath0 , and characterized the average per user dof @xmath236 . for general values of @xmath331
, we proved that there is no message assignment strategy that is optimal for all values of @xmath0 .
we finally introduced message assignment strategies for the case where @xmath251 , and derived inner bounds on @xmath254 that are asymptotically optimal as @xmath252 and as @xmath253 . , v. s. annapureddy , and v. v. veervalli , `` interference channels with coordinated multi - point transmission : degrees of freedom , message assignment , and fractional reuse , '' _ submitted to ieee trans .
theory , available at http://arxiv.org/abs/1211.2897._ nov . | a linear interference network is considered .
long - term fluctuations ( shadow fading ) in the wireless channel can lead to any link being erased with probability @xmath0 .
each receiver is interested in one unique message that can be available at @xmath1 transmitters . in a cellular downlink scenario ,
the case where @xmath2 reflects the cell association problem , and the case where @xmath3 reflects the problem of setting up the backhaul links for coordinated multi - point ( comp ) transmission . in both cases ,
we analyze degrees of freedom ( dof ) optimal schemes for the case of no erasures , and propose new schemes with better average dof performance at high probabilities of erasure . for @xmath2 , we characterize the average per user dof , and identify the optimal assignment of messages to transmitters at each value of @xmath0 . for general values of @xmath1 , we show that there is no strategy for assigning messages to transmitters in large networks that is optimal for all values of @xmath0 . |
ad acknowledges travel grants from iau and australian institute of physics . | we have used the wide field spectrograph on the australian national university 2.3-m telescope to perform the integral field spectroscopy for a sample of the galactic planetary nebulae .
the spatially resolved velocity distributions of the h@xmath0 emission line were used to determine the kinematic features and nebular orientations .
our findings show that some bulge planetary nebulae toward the galactic center have a particular orientation .
the majority of planetary nebulae ( pne ) show predominantly axisymmetric morphologies , i.e. elliptical and bipolar .
there is a strong argument in favor of the axisymmetric morphologies of pne being related to the angular momentum of their central engines ( e.g. , ( * ? ? ?
* miszalski et al . 2009 ) ; ( * ? ? ?
* nordhaus et al . 2010 ) ) .
recently , it has been found that the nebular inclinations are closely related to the orbital inclinations of their close - binary central stars ( see e.g. , ( * ? ? ?
* jones et al .
2012 ) ; ( * ? ? ?
* tyndall et al . 2012 ) ) .
pn as a test particle can be used to trace the dynamics of the galaxy .
the major axes of bipolar pne are assumed to be perfectly aligned with the angular momentum vectors of their central stars , so a population of bipolar galactic pne can be used to determine the angular momentum distribution of the local stars within the galaxy .
recent studies showed that some bipolar pne within the bulge near the galactic center have a homogeneous orientation ( ( * ? ? ? * weidmann & daz 2008 ) ; ( * ? ?
? * rees & zijlstra 2013 ) ) . ,
title="fig:",width=115 ] , title="fig:",width=115 ] , title="fig:",width=115 ] , title="fig:",width=115 ] , title="fig:",width=115 ] , title="fig:",width=432 ] in this study , we aim to disentangle 3d gaseous structures of some galactic pne by means of integral field spectroscopy .
we used the wide field spectrograph ( ( * ? ? ?
* wifes ; dopita et al .
2010 ) ) on the 2.3-m anu telescope at siding spring observatory .
wifes is an image - slicing integral field unit ( ifu ) feeding a double - beam spectrograph .
it has a field - of - view of @xmath1 and a spatial resolution of @xmath2 .
our observations were carried out with a spectral resolution of @xmath3 .
the spectra were reduced using the iraf pipeline wifes ( ( * ? ? ?
* described in detail by danehkar et al .
2013 ) ) .
we obtain the kinematic features by applying a gaussian curve fit to each spaxel of ifu datacube ( see fig .
[ fig1 ] ) .
the observed kinematic maps are then compared to 3-d kinematic models in order to find the orientations of the nebular symmetric axes onto the plane of the sky , i.e. the position angle ( pa ) measured from the north toward the east in the equatorial coordinate system ( ( * ? ? ?
* see danehkar et al .
2014 ) ) .
the pa is transferred into the galactic position angle ( gpa ) in the galactic coordinate system .
fig .
[ fig1 ] ( bottom ) shows orientations of the pne based on their symmetric axes projected onto the sky plane .
the spatial orientations of the galactic pne have been studied by few authors ( ( * ? ? ?
* phillips 1997 ) ; ( * ? ?
? * corradi et al . 1998 ) ; ( * ? ? ?
* weidmann & daz 2008 ) ; ( * ? ? ?
* rees & zijlstra 2013 ) ) .
although the orientations of nearby galactic pne are found to be completely arbitrary in the galactic disc , some distant pne seem to have a particular orientation within the galactic bulge , as seen in fig .
[ fig1 ] ( in agreement with ( * ? ? ?
* weidmann & daz 2008 ) ; ( * ? ? ? * rees & zijlstra 2013 ) ) .
galactic bulge pne are affected by the bugle dynamics and the supermassive black hole located at the galactic center , and are required to be observed and investigated further . |
in the pase two decades , the lattice boltzmann ( lb ) method has received great success in modeling various fluid systems , and the capability in simulating multiphase flows can be recognized as its unique advantage that distinguishes from the traditional numerical methods @xcite .
one of most fundamental problems in lb method for multiphase flows is how to describe the interfacial dynamics , which is a natural consequence of intermolecular interactions between different phases . up to now , based on different physical pictures of the interactions , several types of lb multiphase models has been established , which are substantially divided into four categories : the colour model @xcite , the pseudo - potential model @xcite , the free - energy model @xcite and the phase - field based model @xcite . in most of these lb multiphase models
@xcite , the interface is not tracked explicitly , and the region with non - zero density gradient is identified as the interface . therefore the physics of interface tracking equation is unknown .
fortunately , the phase - field theory provides a firm foundation on the interface physics , in which the interface is tracked by an order parameter that mimics the cahn - hilliard equation ( che ) @xcite , @xmath1 where @xmath2 represents an order variable , @xmath3 is the mobility coefficient , @xmath4 is the chemical potential and a function of @xmath2 , @xmath5 where @xmath6 and @xmath7 are related to the interface thickness @xmath8 and surface tension @xmath9 by the relationships @xmath10 and @xmath11 .
@xmath12 in eq.(1 ) is the fluid velocity and governed by the incompressible navier - stokes equations @xcite @xmath13 @xmath14 + { { \bf{f}}_s } + { \bf{g}},\ ] ] where @xmath15 is the fluid density , @xmath16 is the pressure , @xmath17 is the kinematic viscosity , @xmath18 is the surface tension , and @xmath19 is the external force . some researchers have constructed some lb multiphase models based on the phase - field theory , where the interface is needed to be tracked explicitly by an index or order distribution function @xcite .
@xcite proposed a lb model for incompressible multiphase flows , in which they adopted an index distribution function to track the interface and a pressure distribution function for solving the flow field . based on this model , they successfully simulated the two - dimensional rayleigh - taylor instability , and later the three - dimensional case using d3q15 lattice structure in lb equations for both the interface capturing and flow field @xcite .
although this model is rather robust , it suffers from some limitations , one of which is that the recovered interface equation is inconsistent with the che noticed by zheng _
_ @xcite . to this end
, they developed a lb model for the che , in which a source term on a spatial difference of the distribution function is introduced @xcite .
the model is subsequently extended to the three - dimension using d3q7 lattice model @xcite .
recently , zu _ et al . _
@xcite introduced another similar lb model for the che where a spatial difference term on the equilibrium distribution function was included .
they also modified the equilibrium distribution function in lb equation for flow field such that the continuity equation ( 3a ) can be derived .
however , the computation of the macroscopic pressure and velocity in their scheme is implicit .
more recently , we proposed a novel lb model for two - dimensional multiphase flows @xcite .
on one hand , a simpler time - dependent source term is incorporated in lb equation for the interface capturing . as a result
, the two - dimensional che can be recovered correctly . on the other hand , a equilibrium distribution function
is delicately designed for flow field to derive the correct continuity equation while the hydrodynamic properties can be computed explicitly .
this improved model is also extended to study axisymmetric multiphase flows @xcite as a continuing work , in this paper an efficient three - dimensional lb model for incompressible multiphase flow systems is developed based on the multiple - relaxation - time ( mrt ) method .
this model has some distinct advantages .
firstly , the model for the che requires only seven discrete velocities in three dimensions ( d3q7 ) , therefore the expenditure in data storage and computational time is smaller than that of other models using d3q15 lattice @xcite .
secondly , the mrt collision model is adopted , which has a better accuracy and stability than the single - relaxation - time ( srt ) model used commonly in other lb multiphase models @xcite .
finally , the present model is able to deal with fluid flows at a large peclet number or a high reynolds number .
the rest of this paper is organized as follows .
[ sec : method ] presents our three - dimensional mrt lb model for multiphase flows .
the model then is verified by several classical numerical experiments in sec .
[ sec : results ] .
finally , we made a brief summary in sec .
[ sec : sum ] .
the lb equation with a mrt collision model for the che can be written as , @xmath20 + { \delta_t}{f_i}({\bf{x}},t),\ ] ] where @xmath21 is the order distribution function used to track the interface , @xmath22 is the equilibrium distribution function defined as @xcite @xmath23 where the discrete velocities @xmath24 , the weighting coefficients @xmath25 and the sound speed @xmath26 depend on the choice of the discrete - velocity model , @xmath27 is a parameter related to the mobility . in this work , an efficient d3q7 discrete - velocity model is adopted for the che , and @xmath24 then can be given by @xmath28 and further according to the following equations ( @xmath29 is the unit matrix ) @xmath30 @xmath25 can be obtained as @xmath31 where @xmath32 is a free parameter , @xmath33 . to ensure the positive weighting coefficients ,
the parameter @xmath32 should satisfy @xmath34 .
the transformation matrix @xmath35 of the d3q7 model is defined by @xcite @xmath36 which is constructed based on the polynomial set of the discrete velocities .
@xmath37 in eq .
( 1 ) is a diagonal relaxation matrix , @xmath38 where @xmath39 , and if the parameters @xmath40 equal to each other , the mrt model can reduce to the srt model . to recover the correct che , the source term @xmath41 in eq .
( 9 ) should be defined as @xmath42_{ij}r_j,\ ] ] where @xmath43 is given by @xcite @xmath44 in the present model , the order parameter is computed by @xmath45 and the density is taken as a linear function of the order parameter , @xmath46 where @xmath47 and @xmath48 represent the densities of liquid and gas phases , respectively . the evolution of lb equation ( 4 ) can be commonly divided into two steps , i.e. , the collision process , @xmath49 + { \delta_t}{f_i}({\bf{x}},t),\ ] ] and the propagation process , @xmath50 to reduce the matrix operations , it is wise that the collision process of mrt model is implemented in the moment space . by premultiplying the transformation matrix , we can easily derive the equilibrium distribution function in moment space , @xmath51 where @xmath52 , @xmath53 and @xmath54 are the x- , y- and z- components of macroscopic velocity @xmath12 , respectively .
similarly , the source term @xmath55 in the moment space can be presented as @xmath56 the chapman - enskog analysis is carried out on the lb evolution equation ( 4 ) , and the results demonstrate that the che can be derived correctly from the present mrt model , and the relationship between the mobility @xmath3 and the relaxation parameter is also derived as @xmath57 where @xmath58 , and @xmath59 .
the mrt lb equation with a source term for the nses reads as @xcite @xmath60 + { \delta _ t}g_i,\ ] ] where @xmath61 is the density distribution function , @xmath62 is the equilibrium distribution function and is defined by @xmath63 with @xmath64.\ ] ] to simulate the fluid flows in three dimension , we can choice several types of lattice velocity models , such as d3q15 or d3q19 @xcite
. the d3q15 lattice model is used in this work due to its smaller data storage and higher computational efficiency . following the work of qian @xmath65 @xcite ,
the discrete velocities @xmath66 of the d3q15 lattice model can be given by @xmath67 where @xmath68 .
according to the ordering of @xmath66 , the weight coefficients are presented as @xmath69 the transformation matrix @xmath70 in eq .
( 20 ) can be given by @xcite @xmath71 and the corresponding diagonal relaxation matrix @xmath72 is denoted by @xmath73 where @xmath74 .
the source term @xmath75 in eq . is defined as @xcite @xmath76_{ij}{t_j}\ ] ] where @xmath77,\ ] ] in which @xmath78 is an interfacial force , and @xmath79 is the surface tension taken the potential form @xmath80 . in the present model
, the macroscopic pressure and velocity can be obtained from @xmath81 @xmath82.\ ] ] the collision process of mrt lb equation for nses is also carried out in the moment space .
after some algebraic manipulations , the equilibrium distribution function @xmath62 and source term @xmath83 in the moment space can be respectively derived as @xmath84 and @xmath85^{\rm t}\nonumber\\\end{aligned}\ ] ] we also conduct the chapman - enskog analysis on lb equation ( 20 ) , and the results show that the present model can correctly recover to the nses with the kinematic viscosity determined by @xmath86 where @xmath87 , and @xmath88 . in practical applications , the derivative terms in present model
should be discretized by suitable difference schemes . as widely adopted in the references @xcite , an explicit difference scheme @xmath89 is used for computing the time derivative in eq .
( 12 ) , and the isotropic central schemes @xcite @xmath90 and @xmath91}{c_s^2 \delta_t^2}}\ ] ] are employed for calculating the gradient and the laplacian operator , respectively . in above equations , @xmath92 represents an arbitrary function .
it should be noted that the schemes ( 35 ) and ( 36 ) not only can preserve a secondary - order accuracy in space , but also can ensure the global mass conservation of a multiphase system @xcite .
in this section , we will validate the present three - dimensional mrt lb model with several numerical examples .
we first test the performance of the three - dimensional lb model in interface capturing by simulating two classical benchmark problems : rotation of the zalesak s sphere @xcite and deformation field flow @xcite . in these simulations , the evolution equation ( 4 )
is only adopted in that the velocity distribution has been specified in advance .
next , we simulated the three - dimensional rayleigh - taylor instability to show the ability of the present model for multiphase flows , where a comparison between the numerical results and some available results is also conducted . : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] + ( a ) + : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the previous three - dimensional lb model @xcite.,title="fig:",width=115,height=116 ] ( b ) + : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] + ( a ) + : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] : ( a ) the present mrt model ; ( b ) the present srt model.,title="fig:",width=115,height=116 ] ( b ) + rotation has been widely used in the literatures to test interface tracking methods @xcite .
here we consider the rotation of the zalesak s sphere which has a slot of 16 lattice units width in a @xmath93 domain .
this sphere is initially centered at @xmath94 and the radius @xmath95 occupies 40 lattice units .
the revolution of the sphere is driven by a constant vorticity velocity field , @xmath96 where @xmath97 takes a value of @xmath98 so that the sphere complete one cycle every @xmath99 time steps . in our simulations ,
the interface thickness @xmath8 and surface tension @xmath9 are fixed as 2.0 and 0.04 , respectively ; the relaxation matrix is set as @xmath100 the periodic boundary condition is applied at all boundaries . to examine the mobility effect ,
we introduce the dimensionless peclet number , which is defined as , @xmath101 where @xmath102 is the characteristic length , and set as the value of @xmath95 .
different mobilities can be obtained by changing the value of @xmath103 .
figure 1 shows the revolution processes of the sphere during one period at @xmath104 .
it is seen from fig .
1 that the present model can precisely capture the interface shape in one period , and the sphere returns to its initial configuration at time @xmath105 , which is accordance with the expected results . for a comparison ,
the results obtained by the previous three - dimensional lb model @xcite are also presented in fig .
it is observed that the previous model not only produces some obvious sawteeth on the sphere surface , but also induces some unphysical disturbances around the computational domain .
this implies that the present model can obtain a more accurate and stable interface .
it has been mentioned above that the srt model can be deemed as a special case of the mrt model when all the relaxation parameters equal to each other . due to the more adjustable relaxation factors
, the mrt model should show more potential to achieve a better numerical stability against the srt model .
to illustrate this point , we present a comparison between them using this case .
figure 2 depicts snapshots of rotating sphere during one period at @xmath106 by the mrt and srt models .
it is clearly seen that the results of the srt model are unstable .
the slot of the sphere is slightly distorted and there produces some extra jetsam at the corners of the computational domain .
in contrast , the present mrt model can capture the moving interface of the sphere correctly . to further quantitatively describe the accuracy of the interface capturing , the global relative error on the order parameter
is introduced as @xmath107 where @xmath108 is the value of the order parameter at position @xmath109 and time @xmath105 , and @xmath110 is the exact solution at initial time .
we computed the relative errors generated by the lb previous model @xcite and the present mrt and bgk models at various peclet numbers , and show the results in fig .
3 . as seen from this figure , one can find that the mrt model is more accurate than the srt model , which is further more accurate than the previous model , especially at large pelcet number .
the effect of the model parameter @xmath32 is also examined , and it is found from fig .
3 that the parameter @xmath32 has almost no effect on the accuracy of the mrt model . without loss of generality , in the following simulations we fix the parameter @xmath32 at @xmath111 . : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] + ( a ) + : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite .
the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] ( b ) + : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite .
the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] : ( a ) the present mrt model ; ( b ) the present srt model ; ( c ) the previous lb model @xcite . the times from the left pattern to the right are @xmath112 , @xmath113 , @xmath114 , @xmath115 and @xmath105.,title="fig:",width=115,height=119 ] ( c ) + the above test does not induce large changes of the interface . to show the capacity of the present mrt model , in this subsection we will consider a rather challenging problem of deformation field flow , in which the interface could undergo a large deformation @xcite .
the initial setup of this problem is described as follows .
a sphere with a radius of @xmath116 is placed in a @xmath93 computational domain centered at @xmath117 .
the velocity field of this flow is strongly nonlinear and is given by @xmath118 and we postmultiply a time - dependent function @xmath119 to make the flow periodic , where @xmath120 is the iteration step divided by @xmath121 . in our simulations , @xmath97 and @xmath105 are given as @xmath98 and @xmath122 ; the other physical parameters and boundary condition are set as those in the last test . in theory , the sphere will undergo the deformation continuously until time @xmath114 , and the velocity field then is revised in time , the sphere goes back and returns to its original position .
figure 4 shows the evolution of the interface pattern at @xmath123 obtained by the use of the mrt model , the srt model and the previous lb model @xcite .
it can be clearly observed that the present mrt model can give an accurate prediction in the evolution of the interface : it has a largest deformation at time @xmath114 and moves back to the initial configuration at time @xmath105 .
the behaviors of the interface conform to the theoretical results .
in contrast , the results of the srt model and the previous three - dimensional lb model are unstable . at time @xmath113 or @xmath105 ,
some jagged shapes are produced by the srt model in the vicinity of the interface . at the same time
, the previous lb model is worst in tracking the interface .
a massive amount of unphysical disturbances can be clearly observed in the system .
we also perform simulations with different peclet numbers , and compare the relative errors generated by these lb models .
the results are presented in fig .
5 . from this figure
, one can find that the present mrt model is more accurate than the srt model and the previous lb model . at last
, we simulated a benchmark problem of the rayleigh - taylor instability ( rti ) .
rti is a classical and common instability phenomenon , which occurs at perturbed interface between two different fluids , where the gravity force is applied .
rti studies are useful since it has particular relevance and importance in fields including inertial confinement fusion @xcite and astrophysics @xcite . for this reason
, rti has become a subject of intensive researches in the past 60 years using theoretical analysis @xcite , experimental methods @xcite and also numerical approaches @xcite . however , to our best knowledge , most of previous numerical studies are limited to the two - dimensional case @xcite , and relatively few attention has been paid on the three - dimensional rti @xcite . on the other hand ,
the simulated reynolds numbers considered in previous studies are small in general . in this subsection
, we applied the present mrt model to study the three - dimensional rayleigh - taylor instability , and the effect of reynolds number on the evolution of the interface was examined .
.,title="fig:",width=67,height=230 ] .,title="fig:",width=67,height=230 ] .,title="fig:",width=67,height=230 ] .,title="fig:",width=67,height=230 ] .,title="fig:",width=67,height=230 ] .,title="fig:",width=67,height=230 ] .,title="fig:",width=67,height=230 ] .,title="fig:",width=67,height=230 ] + ( a)(b ) + the physical problem we considered here is a rectangular box with an aspect ratio of @xmath124 , where @xmath125 is the box width .
the initial interface is located at the midplane ( @xmath126 ) , with an imposed square - mode perturbation @xmath127,\ ] ] and the initial order distribution then can be given by @xmath128 according to @xcite , the reynolds numbers ( re ) characterizing rti can be defined as @xmath129 where @xmath130 is a gravitational acceleration , and @xmath17 is the kinematic viscosity . in our simulations , we take the densities of liquid and gas phases as 3.0 and 1.0 , corresponding to an atwood number of 0.5 ; some physical parameters are given as : @xmath131 , @xmath132 , @xmath131 @xmath133 , @xmath134 ; the peclet number defined in [ 10,11 ] is fixed at 50.0 ; the relaxation matrix is set to be @xmath135 and the relaxation factors in @xmath72 are chosen to be unity expect for the ones related to the kinematic viscosity . the periodic boundary conditions in the lateral directions and no - slip boundary condition in the vertical direction are applied in our studies .
figure 6(a ) shows the evolution of the density contours in immiscible rti at a reynolds number of 1024 .
it can be seen that , due to the gravity effect , a heavy fluid and a light one penetrate into each other at early time , and then forms the spike and bubble , respectively .
after that , the heavy fluid rolls up along the flank of the spike , and a mushroom - like structure appears ( t=3.0 ) , which can be attributed to kelvin - helmholtz instabilities providing the rolling motion of the interface .
finally , the mushroom develops further and becomes much bigger .
the patterns of the liquid interface obtained by the present model compare well with previous results @xcite .
we also simulated the three - dimensional rti at a high reynolds number of 4000 , which has not been considered in ref .
[ 15 ] , and presented the results in fig .
it can be observed that the interface takes on almost similar behaviors before time 3.0 .
however , it subsequently presents a distinct manner . due to larger shear interaction between different layers ,
the interface becomes more complex , which eventually undergoes a breakup inducing some tiny dissociative drops in the system . to observe the evolution of the interface more clearly , we plotted in fig .
7 the interface patterns at the diagonal vertical plane ( @xmath136 ) with above two reynolds numbers .
it is found that for both @xmath137 and @xmath138 , the development of the initial mode follows the pattern similar from two - dimensional simulations @xcite . in the following ,
two pairs of counter - rotating vortices are formed at the spike tip and the saddle point ( see time 3.0 ) , which is significantly different from the two - dimensional results @xcite .
the vortices grow with time for @xmath137 , while they become unstable for @xmath138 , resulting in the mixing of two different fluids at the vicinity of the spike tip .
we also gave a quantitative study on the reynolds number effect .
figure 8 depicts the evolution of the positions of the bubble front and spike tip obtained by the present mrt model and the previous numerical results in ref . [
it is shown that the results of @xmath137 obtained by our model agree well with those in ref . [ 15 ] , which verifies the numerical accuracy of the present mrt model in dealing with complex interfacial flows . from fig .
8 , one can also find that there is no evident difference in trajectory of bubble front at two different reynolds numbers , while the spike tip moves slightly faster at a larger reynolds number . ) with different reynolds numbers : ( a ) re=1024 ; ( b ) re=4000 .
the corresponding times from the left pattern to the right are t=1.0 , 2.0 , 3.0 , 4.0 , where t is normalized by the characteristic time @xmath139 . , title="fig:",width=67,height=187 ] ) with different reynolds numbers : ( a ) re=1024 ; ( b ) re=4000 .
the corresponding times from the left pattern to the right are t=1.0 , 2.0 , 3.0 , 4.0 , where t is normalized by the characteristic time @xmath139 . , title="fig:",width=67,height=187 ] ) with different reynolds numbers : ( a ) re=1024 ; ( b ) re=4000 .
the corresponding times from the left pattern to the right are t=1.0 , 2.0 , 3.0 , 4.0 , where t is normalized by the characteristic time @xmath139 .
, title="fig:",width=67,height=187 ] ) with different reynolds numbers : ( a ) re=1024 ; ( b ) re=4000 .
the corresponding times from the left pattern to the right are t=1.0 , 2.0 , 3.0 , 4.0 , where t is normalized by the characteristic time @xmath139 . , title="fig:",width=67,height=187 ] ) with different reynolds numbers : ( a ) re=1024 ; ( b ) re=4000 .
the corresponding times from the left pattern to the right are t=1.0 , 2.0 , 3.0 , 4.0 , where t is normalized by the characteristic time @xmath139 .
, title="fig:",width=67,height=187 ] ) with different reynolds numbers : ( a ) re=1024 ; ( b ) re=4000 .
the corresponding times from the left pattern to the right are t=1.0 , 2.0 , 3.0 , 4.0 , where t is normalized by the characteristic time @xmath139 . , title="fig:",width=67,height=187 ] ) with different reynolds numbers : ( a ) re=1024 ; ( b ) re=4000 .
the corresponding times from the left pattern to the right are t=1.0 , 2.0 , 3.0 , 4.0 , where t is normalized by the characteristic time @xmath139 . , title="fig:",width=67,height=187 ] ) with different reynolds numbers : ( a ) re=1024 ; ( b ) re=4000 .
the corresponding times from the left pattern to the right are t=1.0 , 2.0 , 3.0 , 4.0 , where t is normalized by the characteristic time @xmath139 .
, title="fig:",width=67,height=187 ] + ( a)(b ) + and @xmath139 , respectively.,width=384,height=288 ]
in this work , an efficient three - dimensional lattice boltzmann model based on the mrt collision method is proposed for multiphase flow systems .
this model is a straightforward extension of our previous two - dimensional model @xcite to the three - dimensions .
the present model for the che only utilizes seven discrete velocities in three dimensions , while most of previous lb models need at least fifteen discrete velocities . as a result
, the computational efficiency of the present model can be greatly improved in simulating three - dimensional multiphase flows .
in addition , the advanced mrt collision model is adopted in lb equations for both the che and the nses , which has a better stability than the srt model .
two classical interface - capturing problems including rotation of the zalesak s sphere and deformation field flow were conducted to test the model , and the results show that the present mrt model is more stable and accurate than the srt model and the previous lb model in tracking the interface . finally , the present mrt model is applied to study the three - dimensional rayleigh - taylor instability at various reynolds numbers .
it is found that the numerical results at low reynolds numbers agree well with previous data , while the instability at a high reynolds numbers induces a more complex structure of the interface .
this work is financially supported by the national natural science foundation of china(grant nos .
11272132 ) , and the fundamental research funds for the central universities ( grant no .
2014ts065 ) . | in this paper , an efficient three - dimensional lattice boltzmann ( lb ) model with multiple - relaxation - time ( mrt ) collision operator is developed for the simulation of multiphase flows .
this model is an extension of our previous two - dimensional model ( h. liang , b. c. shi , z. l. guo , and z. h. chai , phys .
rev .
e. 89 , 053320 ( 2014 ) ) to the three dimensions using the d3q7 ( seven discrete velocities in three dimensions ) lattice for the chan - hilliard equation ( che ) and the d3q15 lattice for the navier - stokes equations ( nses ) . due to the smaller lattice - velocity numbers used
, the computional efficiency can be significantly improved in simulating real three - dimensional flows , and simultaneously the present model can recover to the che and nses correctly through the chapman - enskog procedure .
we compare the present mrt model with the single - relaxation - time model and the previous three - dimensional lb model using two benchmark interface - tracking problems , and numerical results show that the present mrt model can achieve a significant improvement in the accuracy and stability of the interface capturing .
the developed model is also able to deal with multiphase fluids with very low viscosities due to the using of the mrt collision model , which is demonstrated by the simulation of the classical rayleigh - taylor instability at various reynolds numbers .
the maximum reynolds number considered in this work reaches up to @xmath0 , which is larger than those of almost previous simulations .
it is found that the instabilty induces a more complex structure of the interface at a high reynolds number . |
the launch of the _ chandra x - ray observatory _ on 1999 july 23 opened a new era in our study of the x - ray universe . with its revolutionary mirror assembly design ,
_ chandra _ provides @xmath14 resolution imaging with @xmath15 1 astrometry over the @xmath16 kev range ( weisskopf , odell , & van speybroeck 1996 ) .
furthermore , the deepest _ chandra _ observations to date are @xmath17 times more sensitive than the deepest pre-_chandra _ observations in the @xmath18 kev soft x - ray band ( hasinger et al .
1998 ) and more than two orders of magnitude more sensitive than the deepest pre-_chandra _ observations in the @xmath19 kev hard x - ray band ( ueda et al .
1999 ; fiore et al .
1999 ) . at this resolution and depth ,
_ chandra _ has resolved @xmath20 % of the soft x - ray background and @xmath15 80% of the hard x - ray background into discrete sources ( mushotzky et al .
2000 ; barger et al . 2001 ; giacconi et al .
2001 ; hornschemeier et al . 2001 ; tozzi et al .
_ chandra _ s imaging resolution is also well - matched to the typical deep optical / near - infrared imaging resolutions , providing for the first time largely unambiguous identifications of the x - ray hosts .
this affords an exciting new opportunity to study x - ray host identifications , thereby aiding our understanding of what powers the x - ray background . in particular , even heavily obscured active galactic nuclei ( agn ) are expected to be luminous and largely unattenuated in the hard x - ray band : the great hope is that studying the resolved hard x - ray background can provide an unbiased picture of the history of accretion - driven energy production in the universe . with the goal of simultaneously studying both the x - ray background and the properties of intracluster media at high redshift
, we obtained a deep , 184.7 ks exposure with the advanced ccd imaging spectrometer ( acis - i ) on board _ chandra _ of three spatially - proximate high - redshift galaxy clusters in the lynx field of the spices survey ( eisenhardt et al .
2001 , see 4 below ) .
all three clusters were identified in the _ rosat _ deep cluster survey ( rdcs ; rosati et al .
1995 , 1998 ) and reside within a single acis - i field .
two are at @xmath21 , among the most distant x - ray - emitting galaxy clusters known : rx 0848 + 4453 at @xmath22 ( stanford et al .
1997 ) and rx 0849 + 4452 at @xmath23 ( rosati et al .
the third cluster is at @xmath24 ( rx 0848 + 4456 ; rosati et al .
1998 ) . at 184.7
ks , our _ chandra _ observation is among the deepest observations of the x - ray sky taken to date
. analysis of the diffuse x - ray emission from the intracluster media of the two @xmath25 clusters is presented in stanford et al .
( 2001 ) and suggests little evolution in the x - ray luminosity @xmath3 temperature ( @xmath26 ) relation to @xmath27 .
analysis of the diffuse x - ray emission from the @xmath28 cluster is presented in holden et al .
( 2001 ) and suggests that 10@xmath320% of the x - ray emission actually comes from a group at @xmath29 along the line of sight , showing the importance of detailed and multiwavelength studies of galaxy clusters when using them as probes of cosmological mass functions .
observations of the 153 x - ray point sources are the subject of the remainder of this paper . to aid in follow - up observations , this manuscript details and provides a catalog of the unresolved x - ray sources we identify from the deep 17 @xmath2 17 acis - i image ( 2 ) .
in 3 we discuss results from the x - ray data alone , including the hardness ratio and the fraction of the x - ray background which is resolved from these observations .
spices optical ( @xmath30 ) images cover the central 16@xmath2 16 of this field and spices near - infrared ( @xmath31 ) images cover the central 56@xmath2 56 of this field : we present the results of our cross - correlation of the optical / near - infrared imaging data with the x - ray catalog in 4 .
the 18 redshifts we have obtained for lynx _ chandra _ sources are also discussed in 4 .
section 5 presents our first spectroscopic results and 6 highlights some initial results from this work .
we summarize our conclusions in 7 .
an additional , interesting x - ray source in this field , cxo52 , is the subject of stern ( 2002 ) .
unless otherwise stated , all magnitudes presented are in the vega system .
cosmology - dependent parameters are calculated for two models : an einstein - de sitter universe consistent with previous work in this field ( @xmath32 , @xmath33 , and @xmath34 ) and the dark energy universe favored by recent work on high - redshift supernovae and fluctuations in the cosmic microwave background ( @xmath35 , @xmath36 , and @xmath37 ; riess et al .
acis - i observations of the lynx field were obtained on 2000 may 3 ( 65 ks ; obs - id 1708 ) and 2000 may 4 ( 125 ks ; obs - id 927 ) .
the aimpoint for the observations was @xmath38 , @xmath39 + 4454329 and both exposures were obtained in the faint mode when acis was at a temperature of @xmath40 c. the galactic absorbing column for this field is @xmath41 .
the position angle of the observations was 258.45 .
the data were reduced and analyzed using the _ chandra _ interactive analysis of observations ( ciao ) software ( release v1.1 , see http://asc.harvard.edu/ciao ) . the level 1 data were processed using the quantum efficiency uniformity file for @xmath40 c , which corrects the effective area for loss due to charge transfer inefficiency .
the correction compensates for the loss of events far from the readout , especially at high energies , and is relevant when fitting the total spectrum in the largest energy range .
the data were filtered to include only the standard event grades 0 , 2 , 3 , 4 , and 6 .
all bad columns were removed by eye , as were the columns close to the border of each node , since the grade filtering is not efficient in these columns .
we also identified flickering pixels in the images in the chip coordinates : for each chip , we checked photon arrival times for all pixels on which more than two photons were detected . if two or more photons were detected within an interval shorter than 7 seconds , they were considered flickering pixels and excluded from further analysis .
as the brightest source ( cxo39 ) has a count rate of only 1 photon every @xmath42 seconds , this procedure should not affect our source catalog .
the removal of columns and pixels reduces slightly the effective area of the detector , the result of which has been included when calculating the soft and hard exposure maps .
time intervals with background rates larger than @xmath43 over the quiescent value ( @xmath44 counts s@xmath45 per chip in the @xmath46 kev band ) were removed .
this procedure gave 60.7 ks of effective exposure out of the first observation , and 124 ks out of the second , for a total of 184.7 ks .
the two observations are almost coincident on the sky , so that the total coverage is 298 arcmin@xmath47 .
the two event files were merged into a single file , which was used in the reduction and analysis process .
the astrometry of the two observations , contiguous in time , was perfectly consistent , as verified by registering the positions of the brightest sources on top of each other .
we did not include data from the chips s-2 and s-3 in the following analysis , since the point spread function ( psf ) is very broadened ( @xmath48 arcsec ) and they would add a very little effective area even for high fluxes .
[ xcolorfig ] shows the final , combined x - ray data , with a false - color mapping derived from the _ chandra _ energy bands . a soft band from 0.5@xmath32 kev and a hard band from 2@xmath37 kev were used in the detection process to select a soft and hard source lists , as detailed below
. the hard band detection window was cut at 7 kev since above this energy the effective area of _ chandra _ is decreasing , while the instrumental background is rising , giving very inefficient detection of sky photons .
x - ray count rates in the above bands were used to calculate x - ray fluxes in a 0.5@xmath32 kev soft band and 2@xmath310 kev hard band , assuming an average x - ray spectrum .
we ran source extractor ( bertin & arnouts 1996 ) on the summed 0.5@xmath37 kev image binned by a factor of two ( one binned pixel corresponds to 0.982@xmath2 0.982 ) .
since we are far from being background limited , the detection efficiency proved to be much higher for the 0.5@xmath37 kev image rather than the soft and hard bands separately .
giacconi ( 2001 ) use the same procedure for their analysis of the _ chandra _ deep field - south ( cdf - s ) .
source extractor parameters were chosen after extensive simulations , and we adopted an internal detection threshold signal - to - noise ( @xmath49 ) ratio of 2.4 , with a gaussian filter fwhm of 15 and a minimum area of 5 pixels . since source extractor is not tailored to work with a very low background , it gives a wrong estimate of the internal background . to tackle this problem
, we built a smoothed map of the background computed from the data themselves in the corresponding band , after the removal of the sources down to a very low threshold .
this background map was used as a weight ( variance ) map to define the detection threshold .
this modified detection algorithm is several orders of magnitude faster than the wavelet detection algorithm of rosati ( 1995 ) or the wavdetect algorithm in the ciao software ( freeman 2002 ) .
table 1 presents the catalog of @xmath50 kev point sources identified in the combined image , selected as follows .
we first measured the signal - to - noise ( @xmath49 ) ratio of all the candidate detections in circular extraction apertures . a radius @xmath51 was used , where @xmath52 is the modeled psf full - width at half - maximum , determined as a function of off - axis angle to reproduce the broadening of the psf .
a minimum @xmath53 was used in the central regions .
sources with @xmath54 within the extraction area were considered detections .
this limit was chosen to proved a census of x - ray sources to faint flux levels , where the choice of source extractor parameters has already significantly reduced the number of spurious detections .
we stress as a compromise between providing a census of x - ray sources to faint levels and limiting the contamination from false detections .
we stress that the condition of having @xmath54 in the extraction area corresponds to a high significance detection the faintest detected sources have more than 10 counts .
these catalogs , obtained from the source extractor run and the a - posteriori @xmath49 cut , typically include less than five spurious sources as tested with simulations with the model of _ axaf _ response to x - rays ( marx ) software ( for a detailed description of the simulations , see tozzi et al .
2001 ) . after performing this procedure on the soft and hard images
separately , we obtained the soft and hard catalogs .
a combined catalog was then produced by matching the two .
source counts were measured with simple aperture photometry within @xmath55 .
the background for each source was calculated locally within an annulus of inner radius @xmath56 and an outer radius of @xmath57 .
numerical simulations show that our aperture photometry systematically underestimates the net counts by 4% . for the following analysis
, we have applied this aperture correction . finally , we transformed the net count rate into energy flux in the soft 0.5@xmath32 kev band and in the hard 2@xmath310 kev band .
the conversion factors used were @xmath58 in the soft band and @xmath59 in the hard band , assuming an absorbing column of @xmath60 @xmath61 ( galactic value ) and a photon index @xmath62 .
the uncertainties in the conversion factors reflect the range of possible values for the photon index , @xmath63 . as suggested by the spectral analysis of the stacked spectra , these values are representative of our sample ( see below ) . since the conversion factors were computed at the aimpoint , where the effective area of _
chandra_/acis is at its maximum , the count rate of each source was corrected for vignetting .
namely , the net count rate was multiplied by the ratio of the exposure map at the aimpoint to the value of the exposure map averaged within the extraction region .
such a correction was done in the soft and hard bands separately .
we detect 132 sources in the 0.5@xmath32 kev ( soft ) x - ray band down to a 2.1@xmath64 limiting flux of @xmath65 and 111 sources in the 2@xmath310 kev ( hard ) x - ray band down to a 2.1@xmath64 limiting flux of @xmath66 .
only the point sources are considered here .
as mentioned above , the diffuse galaxy cluster emissions are explored by stanford et al .
( 2001 ) and holden et al .
( 2001 ) . of the total catalog of 153 sources , 42 ( 27% )
are detected only in the soft band and 21 ( 14% ) are detected only in the hard band .
these results are very similar to the fractions derived by tozzi et al .
( 2001 ) for their 300 ks acis map of the cdf - s : referred to the total sample , they find 26% of the sources are detected only in the soft band and 11% of the sources are detected only in the hard band .
astrometric positions were initially determined from the acis - i aspect solution .
following comparison of x - ray source positions to the deep , ground - based imaging discussed in 4 , we refined the absolute pointing of the x - ray data with an offset of 109 west , 135 south , and no rotation .
these offsets were determined by matching x - ray sources with greater than 25 total counts ( @xmath50 kev ) to @xmath67 sources .
the positions presented in table 2 therefore match the astrometric frame of the spices survey .
to compute the number counts in the soft and hard bands , we first calculate the sky coverage as a function of the flux .
the effective sky coverage of a given flux is defined as the area on the sky where a soft ( or hard ) source can be detected with a @xmath54 in the extraction region defined above .
we have performed extensive simulations with marx to show this procedure is accurate within a few percent ( see tozzi et al .
the same procedure has been used for the cdf - s ( giacconi et al .
2001 ) . in fig .
[ fig_nssoftxrb ] we present the soft x - ray source counts derived from our observations . for comparison , we also show source counts from the deep _ rosat _ observations of the lockman hole ( hasinger et al .
1998 ) , extrapolation of the source counts to fainter fluxes from fluctuation analysis of the _ rosat _ data ( hasinger et al . 1993 ) , and recent , deep _
chandra _ soft x - ray source counts from mushotzky et al .
we find the _ chandra _ results in excellent agreement with _
rosat _ in the region of overlap @xmath68 2s .
the _ chandra _
data extend the results to @xmath69 2s
. a maximum - likelihood power - law fit to the soft x - ray @xmath70 data finds @xmath71 confidence contours for the slope and the normalization of the soft counts are shown in the upper right corner of fig .
[ fig_nssoftxrb ] , compared with the results of mushotzky et al .
note that here we used an average slope to be consistent with the average spectrum of our sample of sources . in fig .
[ fig_nshardxrb ] we present the hard x - ray source counts derived from our observations down to a flux limit of @xmath72 2s . at
bright levels source counts from the _ advanced satellite for cosmology and astrophysics _
( _ asca _ ) and _ bepposax _ are shown ( giommi et al .
1998 ; della ceca et al .
1999 ; ueda et al .
1999 ) . at faint levels
we show recent , deep _
chandra _ hard x - ray source counts from mushotzky et al .
. a maximum - likelihood power - law fit to the hard x - ray @xmath70 data finds @xmath73 confidence contours are again provided , as per fig .
[ fig_nssoftxrb ] .
our hard - band @xmath70 results are consistent with deep observations of the lockman hole obtained with the _ x - ray multi - mirror - newton _ satellite ( _ xmm - newton _ ; hasinger et al .
our normalization of the hard - band @xmath74 relation is 1@xmath64 higher than the results for the cdf - s who find a normalization to the above relation of @xmath75 where @xmath76 ( tozzi et al . 2001 ) . in both the soft and the hard x - ray bands ,
the lynx - field source counts are slightly lower at each x - ray flux than those determined for ssa13 ( mushotzky et al .
the difference likely reflects cosmic variance due to large scale structure on size scales similar to the _ chandra _ field - of - view .
in particular , the ssa13 results analyze only one acis chip , thus sampling one - quarter the area of the results discussed here .
we note that four of the deepest 16 @xmath2 16 _ chandra _ fields ( cdf - s , hdf - n , the lockman hole , and lynx ) find consistent normalizations to the @xmath70 relation , while ssa13 and the ms1137 fields give 40% higher normalizations ( tozzi , in preparation ) what fraction of the hard x - ray background is resolved by the deep lynx _
[ fig_hardxrb2 ] presents the integrated contribution of resolved hard x - ray sources as a function of x - ray flux . in the range @xmath77 , the integrated hard x - ray flux density in the 2@xmath310 kev band is @xmath78 , depending on the assumed average power law ( @xmath79 , respectively ) .
including the hard - band @xmath70 relation derived by ( 1999 ) for a wider - area , brighter _ asca _ survey , we find that @xmath80 of the hard x - ray background is resolved to a flux limit of @xmath81 . as seen in fig .
[ fig_hardxrb2 ] , this is close to the _ high energy astronomy observatory-1 _ ( _ heao-1 _ ) value of @xmath82 ( marshall et al .
1980 ) , but @xmath83% lower than the more recent determinations from _ bepposax _ and _ asca _ ( vecchi et al .
1999 ; ishisaki et al .
1999 ) . for comparison , from the first 300 ks observation of the cdf - s , tozzi et al .
( 2001 ) resolve @xmath84 of the hard x - ray background while mushotzky et al .
( 2000 ) resolve @xmath85 in their 101 ks observation of ssa13 .
we conclude that , given the uncertainty on the value of the total background , a fraction between @xmath86% and @xmath87% of the hard x - ray background is still unresolved in our data set .
a long - standing problem in x - ray astrophysics has been that the spectral slope of the x - ray background is significantly steeper than the spectral slopes of the first ( brightest ) soft x - ray sources which were identified .
the majority of these bright sources are identified with unobscured agn ( schmidt et al .
this problem , the so - called `` spectral paradox , '' implies that a new population of fainter sources with harder x - ray indices exists : obscured agn are the likely culprit .
_ chandra _ has now identified this population .
[ fig_softhard ] illustrates the hardness ratio @xmath88 for the lynx sample , where @xmath89 and @xmath90 are the net counts in the hard ( @xmath91 kev ) and soft ( @xmath92 kev ) bands , respectively .
the hardness ratio is plotted against the hard x - ray flux .
there are 21 sources , corresponding to 14% of the total combined sample , which are detected only in the hard band ( plotted at @xmath93 ) and 42 which are detected only in the soft band ( not plotted ) .
a population of hard sources , largely absent at bright fluxes , clearly emerges at faint fluxes .
these results are analogous with the ones found in the cdf - s ( giacconi et al .
we note that in the lynx field there are three bright outliers in the hardness ratio flux plane , in the upper right corner .
these are very absorbed sources ( @xmath94 cm@xmath47 ) .
in particular , the brightest one , cxo12 , dominates the spectrum of the bright end of our sample , giving an effective @xmath96 already for fluxes @xmath97 erg @xmath61 s@xmath45 .
cxo12 is identified with the @xmath98 type 2 quasar ax j08494 + 4454 reported by ohta ( 1996 ) and nakanishi ( 2000 ) from _ asca _ observations of the lynx field . of the other two hard , bright x - ray sources which stand out in fig .
[ fig_softhard ] , only one ( cxo36 ) is within the area imaged optically in the current work .
optical spectra from these other two sources are not discussed in the current work .
[ fig_allsources ] presents the average spectral index of the population , determined from the stacked spectrum of all 154 sources in the combined sample . as a background
, we used the stacked spectrum of all the background regions extracted around each source .
the resulting photon files were scaled by the ratio of the total area of extraction of the sources and the corresponding area for the background .
this procedure guarantees a correct background subtraction despite the non - uniformities of the instrumental background across acis - i .
the response matrices for the stacked spectra were obtained from the counts - weighted average of the matrices for individual sources .
xspec , the x - ray spectral fitting package ( arnaud 1996 ) , was used to compute the slope of a power law spectrum with local @xmath99 absorption in the energy range @xmath100 kev .
we excluded bins below @xmath101 kev because the calibration is still uncertain below this energy .
we obtained a photon index @xmath102 of @xmath103 and a column density @xmath104 @xmath61 which is consistent with the galactic value ( errors refer to the 90% confidence level ) .
the results of the spectral fits are shown in fig .
[ fig_allsources ] .
we find an average spectrum for the detected sources is consistent with the average shape of the hard background ( @xmath105 ) .
the lynx field is one of four fields which comprise spices ( eisenhardt et al .
this long - term project was conceived as the next generation analog to the @xmath106-selected canada - france redshift survey ( cfrs ; lilly et al . 1996 ) : by studying a sample of @xmath107-selected faint field galaxies over a larger , 100 arcmin@xmath47 field , we are attempting to address the history of galaxy formation and evolution out to @xmath108 in a manner which is less - biased by large scale structure and recent star formation .
in particular , selecting galaxies at @xmath109 m should essentially translate to selecting by mass , thereby providing a powerful discriminant of theories of galaxy assembly ( kauffmann & charlot 1998 ) .
four unremarkable but well - studied fields were chosen spread across the sky for efficient follow - up throughout the year .
details of the motivation and imaging observations are presented in eisenhardt et al .
we briefly summarize the lynx observations here .
optical @xmath30 imaging was obtained using the kitt peak national observatory 4 m mayall telescope with its prime focus ccd imager ( pfccd ) equipped with a thinned ar coated @xmath110 tektronics ccd .
this configuration gives a 16@xmath2 16 field of view with 047 pix@xmath45 .
the filters used were a harris @xmath111-band ( @xmath112 ; @xmath113 ) , harris @xmath114-band ( @xmath115 ; @xmath116 ) , harris @xmath106-band ( @xmath117 ; @xmath118 ) , and an rg850 long - pass @xmath119-band filter .
for the @xmath106 and @xmath119 imaging , the ccd was operated using `` short scan , '' where the ccd was mechanically displaced while its charge is shifted in the opposite direction to reduce fringing to very low levels .
the combined , processed images reach limiting vega magnitudes of 26.9 ( @xmath111 ) , 25.4 ( @xmath114 ) , 24.6 ( @xmath106 ) , and 24.2 ( @xmath119 ) , where these numbers represent 3@xmath64 limits in 3 diameter apertures .
the corresponding ab magnitude limits are 26.8 , 25.6 , 25.1 , and 24.7 , respectively .
the seeing for the summed images ranges from 12 ( @xmath111 ) to 10 ( @xmath119 ) .
data reduction followed standard techniques ( for details , see eisenhardt et al .
2001 ) . near - infrared @xmath31 imaging was obtained at the kpno 4 m with its infrared imager ( irim ; fowler et al .
1988 ) equipped with a rockwell international nicmos 3 256 @xmath2 256 hecdte array giving 06 pix@xmath45 .
four pointings were obtained to cover a @xmath120 area .
the average exposure time in the fully sampled regions of the resulting mosaic was 10.14 ks in @xmath121 ( @xmath122 ) and 8.49 ks in @xmath123 ( @xmath124 ) .
the seeing in the mosaics is 12 @xmath3 13 .
data reduction followed standard techniques ( for details , see eisenhardt et al .
calibrations of the optical and near - infrared images onto the landolt and cit systems were obtained using observations of landolt ( 1992 ) and ukirt standard stars ( hawarden et al .
2001 ) , respectively .
the @xmath121-band images reach a 3@xmath64 limiting magnitude of 22.9 ( vega ) in a 3 diameter aperture .
the corresponding depth for the @xmath123 images is @xmath125 .
we used source extractor ( release v2.1.6 ; bertin & arnouts 1996 ) to create independent catalogs of sources selected from the @xmath111- , @xmath106- , and @xmath123-band images .
photometry was generated for 30 diameter apertures and evaluated in the same apertures for all six coaligned bands .
[ fig_offsets ] shows the x - ray - to - optical positional offsets after the x - ray positions have been shifted by 109 west and 135 south ( see 2.2 ) .
based on these results , we matched the optical / near - infrared and x - ray source catalogs using a 15 radius aperture for host identification .
table 2 presents the optical / near - infrared properties of the _ chandra _ sources listed in table 1 .
as indicated in the table , for some sources , primarily fainter x - ray sources for which the x - ray position is more uncertain , we have used a larger ( @xmath126 ) match radius .
non - detections are listed with an ellipsis ( `` ... '' ) while sources with bad photometry , due either to source saturation or contamination from the stellar bleed trail of a nearby , bright source , are listed as 99.00 .
only those sources within the smaller near - infrared field are tabulated in the near - infrared photometry columns .
sources from table 1 not within the optical field are not included in table 2 .
only 104 of the x - ray sources are within the optical image field of view . of these , 80 , or 77% , have @xmath127 ( vega ) identifications .
the infrared field of view covers 24 of the x - ray sources , of which 17 , or 71% , have @xmath128 ( vega ) identifications . when we restrict ourselves to _ chandra _ sources with signal - to - noise ratios greater than 3 , the absolute numbers of x - ray sources in each region drops by @xmath129% but the fraction of sources with @xmath106 and/or @xmath123 identifications does not change substantially . from the spices imaging we find that the surface density of sources to @xmath7 is @xmath15 15 arcmin@xmath130 , so the probability of a chance coincidence of an x - ray source within 15 of a @xmath131 source is @xmath132% .
we therefore expect none of the 17 @xmath133 identifications to be spurious .
the surface density of @xmath127 ( vega ) sources is @xmath15 23.5 arcmin@xmath130 so the probability of a chance juxtaposition of an x - ray source with an @xmath134 source with our 15 match radius is @xmath135% .
we therefore expect that up to five of the faint host identifications in table 2 to be spurious .
note that most of the optical identifications are in fact much brighter than @xmath6 , and are thus less likely to be spurious . in fig .
[ fig_ircolor ] we present color - magnitude diagrams for sources detected in our near - infrared imaging field .
x - ray sources are indicated with larger symbols . we see that for a given optical magnitude , the x - ray source hosts tend to be red in optical - to - near - ir color , hugging the so - called _ red envelope_. several of these sources meet the standard identity criteria of extremely red objects ( eros ; graham & dey 1996 ) : one source has @xmath136 and four sources have @xmath137 . we know from previous x - ray surveys that many bright , early - type galaxies at moderate redshift are x - ray luminous ;
_ chandra_-identified eros may represent the higher - redshift tail of that population .
indeed , cowie et al .
( 2001 ) suggest that identifying optically - faint , hard x - ray sources may prove an efficient strategy to locate luminous , evolved galaxies at high redshift ( @xmath138 ) .
we note that optically - faint , _ soft _ x - ray sources are also occasionally identified with high - redshift , early - type galaxies ( cxo138 ; 6.4.2 ) .
highly - obscured sources are also expected to have red colors and may represent some fraction of extremely red x - ray sources .
contrary to that expectation , the @xmath139 type ii quasar in the cdf - s identified by norman et al .
( 2001 ) is relatively blue in optical - to - near - ir color : @xmath140 and @xmath141 .
a smaller subset of the x - ray sources have identifications which have blue optical - to - near - ir colors .
these sources tend to be associated with broad - lined , bright quasars .
thirteen of the _ chandra _ sources in the lynx field are located in regions for which we have images from the wide field planetary camera 2 ( wfpc2 ; trauger et al .
1994 ) on the _ hubble space telescope _
( _ hst _ ; fig .
[ fig_hst ] ) .
these observations , sampling several independent observing programs , primarily target galaxy clusters in the lynx field and range from 2 to 10 orbits ( 4.8 to 27 ks ) through the f702w ( 12 ks ; rx j0848 + 4456 region ) and f814w ( all other imaging ) filters .
the observations are described in van dokkum et al .
( 2001 ) , holden et al .
( 2001 ) , and wu ( in preparation ) . only nine ( @xmath142% ) of the sources are detected in the _ hst _ images , lower than expected for a larger sample given that @xmath143% of the full catalog have @xmath127 identifications .
this fraction is also lower than the 92% of cdf - s x - ray sources with optical counterparts in the deeper _ hst_/wfpc2 study of schreier et al .
( 2001 ) , which reaches a detection limit of @xmath144 .
of the four lynx non - detections ( cxo46 , cxo127 , cxo135 , and cxo136 ) , three are detected in the x - ray with signal - to - noise ratios @xmath145 . the nine detections show a range of morphology .
cxo42 and cxo49 are unresolved spatially , though both show faint , surrounding diffuse emission .
cxo41 also has a very strong core , embedded in extended emission suggestive of a tidal tail or spiral arm .
cxo44 and cxo124 also appear to be associated with merging systems , while the remaining sources ( cxo52 , cxo60 , cxo122 , and cxo205 ) are spatially extended .
the faint , soft x - ray source cxo205 appears to be associated with an early - type galaxy .
[ fig_xopt ] illustrates the x - ray - to - optical flux ratio of soft and hard x - ray sources .
following hornschemeier et al .
( 2001 ) , we plot lines of constant x - ray - to - optical flux ratio using the relation @xmath146 derived from the kron - cousins @xmath114-band filter transmission function . at bright x - ray fluxes ,
we plot the results of shallow , wide - area surveys : for soft x - rays , we show agn from the _ rosat _ survey reported by schmidt et al .
( 1998 ) and for hard x - rays , we show sources from the _ asca _ survey reported by akiyama et al .
( 2000 ) . at faint x - ray fluxes
we show the results of the _ chandra _ lynx survey reported here and results from the 221.9 ks _ chandra _ study of the _ hubble _ deep field north ( hdf - n ) reported in hornschemeier et al .
the majority of sources have @xmath147 , which , as we see from the shallow surveys , is typical of agn .
however , at the fainter fluxes probed by _
chandra _ , new populations become apparent . at faint , soft x - ray fluxes ,
several sources are seen which are x - ray underluminous for their optical magnitudes .
optical spectroscopy ( see 5 and hornschemeier et al .
2001 ; tozzi et al .
2001 ) reveal many of these sources to be apparently normal galaxies .
we discuss these sources further in 6.4 .
several galactic m dwarfs are also identified with @xmath148 , discussed further in 6.3 .
we note that one late - type galactic dwarf has been identified in the lynx field more than two orders of magnitude brighter in the x - ray bands than the other galactic x - ray emitters : we discuss this interesting source further in 6.3 . at
faint , hard x - ray fluxes , a population of sources appears which are x - ray _ overluminous _ for their optical magnitudes .
they potentially represent an astronomically interesting new population . unfortunately , most are extremely faint optically and lack spectroscopic identification .
extremely high - redshift ( @xmath149 ) agn might have bright x - ray fluxes but faint @xmath114-band fluxes due to absorption from the lyman transitions of hydrogen along our line of sight . obscured agn , for which high column density absorption near the central black hole shields both the broad - line region and soft x - ray emission , might also be overluminous in the hard x - ray relative to their optical magnitudes .
several examples of these _ type ii quasars _ have now been identified in the deepest _ chandra _ surveys ( norman et al .
2001 ; dawson et al .
to date , the spices survey has obtained spectroscopic redshifts for 219 of the 485 @xmath128 sources in the lynx field .
this work started in february 1997 and has all been done in multislit mode with the low resolution imaging spectrometer ( lris ; oke et al .
1995 ) on the keck telescopes .
observations typically used the 150 lines mm@xmath45 grating ( @xmath150 ; @xmath151 ) and sample the wavelength range 4000 to 1@xmath152 m .
typical slitlet lengths were @xmath15 20 and we performed @xmath15 3 spatial offsets between each 1200 s @xmath153 s exposure in order to facilitate removal of fringing at long wavelength ( @xmath154 ) .
masks were designed to contain sources of comparable @xmath106-band magnitudes . for the masks containing the brightest sources , two 1200 s exposures were sufficient to obtain redshifts for most sources on the mask . for the masks containing the faintest sources , five 1800 s exposures
were not uncommon and some sources have been on multiple faint source masks so that they now have @xmath155 10 hr of spectroscopic integration .
our infrared selection led to the unintentional spectroscopic targeting of a handful of x - ray sources prior to our obtaining the _ chandra _ map .
on ut 2000 october 01 we observed a mask dedicated to bright _ chandra _ sources for 2400 s in photometric conditions . in total , we have spectroscopic redshifts for 18 of the lynx _ chandra _ sources currently .
table 3 summarizes the spectroscopic results and fig .
[ spectra ] presents the spectra .
[ zhist ] shows the histogram of redshifts for spectroscopically - identified x - ray sources in this field .
all data reductions were performed using iraf and followed standard slit spectroscopy procedures .
we calculated the pixel - to - wavelength transformation using a near lamp spectrum , generally observed immediately subsequent to the science observations ( rms variations of 0.6 are typical for the 150 lines mm@xmath45 grating ) , and employed telluric emission lines to adjust the wavelength zero - point .
the spectra on photometric nights were flux - calibrated using observations of standard stars from massey & gronwall ( 1990 ) .
consistent with previous spectroscopic studies of faint x - ray hosts ( fiore et al .
2000 ; barger et al . 2001 ; giacconi et al .
2001 ) , the lynx _ chandra _ sources are associated with a wide range of astronomical objects .
we find two galactic dwarfs , seven obvious agn , and nine apparently normal galaxies . of the galaxies , four are early - type and five show evidence of star formation .
table 3 presents the optical and x - ray luminosities for the sample of sources for which we have spectroscopic redshifts .
luminosities have been calculated for two cosmologies : an einstein - de sitter ( eds ) universe with @xmath156 , consistent with previous work in this field ( tozzi et al .
2001 ) and the dark energy ( @xmath157 ) universe favored by recent supernovae and microwave background observations ( riess et al . 2001 ) .
we calculate rest - frame @xmath111-band luminosities @xmath158 by interpolating rest - frame 4400 apparent ab magnitude , @xmath159 , from our broad - band imaging .
magnitudes in the ab system ( oke 1974 ) are defined by @xmath160 , where @xmath161 is measured in @xmath152jy . for the quasar cxo50 , our highest - redshift source
, we do not have an imaging data point long - ward of rest - frame 4400 : for this source we instead assume a standard quasar optical spectral index , @xmath162 ( richstone & schmidt 1980 ; schneider et al .
1992 ) and interpolate the flux density from the @xmath106-band magnitude .
a cosmology - dependent luminosity distance is used to relate apparent magnitude @xmath163 to absolute magnitude @xmath164 , and the same optical spectral index is used to calculate the offset between vega - based @xmath158 and ab - system @xmath165 : @xmath166 ( kennefick , djorgovski , & de calvalho 1995 ) . as seen in table 3 ,
the differences in absolute magnitude are not large between the two cosmologies considered , @xmath167 . for the figures below we
plot only the einstein - de sitter values . for @xmath168 and @xmath169 , the luminosity in the rest frame energy band @xmath170 , @xmath171 ,
is then related to the flux observed in that same energy band , @xmath172 , by @xmath173 ( see hogg 1999 ) , where we adopt an x - ray spectral index @xmath62 for our sample .
[ fig_xlum ] plots the x - ray luminosity of our spectroscopic sample against the rest - frame @xmath111-band luminosity , @xmath158 . fig .
[ fig_zlum ] plots x - ray luminosity against redshift for our sample , fortified with the hdf - n _ chandra _ sources from hornschemeier et al . ( 2001 ) .
we have recalculated their luminosities for our cosmology and @xmath62 .
we have classified the optical spectra in three broad categories : agn , galaxies , and stars , following a scheme similar to that of schmidt et al .
( 1998 ) and hornschemeier et al .
a description of the categories follows : * * agn : * we classify sources as agn that have either ( 1 ) broadened emission lines ( fwhm @xmath174 ) or ( 2 ) high - ionization state emission lines such as [ ] .
unfortunately the coarse spectral resolution at which many of these data were taken hinders us from separating broad - lined agn from narrow - lined agn .
seven of the eighteen sources in the current sample are classified as agn .
they are predominately identified with the most x - ray - luminous sources , @xmath175 .
many are also luminous at optical wavelengths ( fig .
[ fig_xlum ] ) . * * emission - line and early - type galaxies : * extragalactic sources without obvious agn features in their optical spectra are classified as galaxies , though this does not rule out the presence of an active nucleus .
we distinguish between galaxies showing emission lines and early - type galaxies , where the latter have redder continua marked by continuum breaks at 2640 ( b2640 ) , 2900 ( b2900 ) , and 4000 ( d4000 ) .
many of the emission - line galaxies show early - type features as well .
nine of the eighteen sources in the current sample are classified as galaxies .
for the lower x - ray luminosity sources , stellar processes such as binaries , winds , and supernovae can produce the x - ray emission .
for the higher x - ray luminosity sources , a buried active nucleus is likely present ( see 6.4 ) . *
* stars : * two of the eighteen sources in the current sample are classified as low - mass galactic dwarfs .
hornschemeier et al .
( 2001 ) also report two x - ray - emitting m4 dwarfs in the _ chandra _ field encompassing the hdf - n .
* cxo39 ( @xmath176 , agn ) : * this x - ray - luminous ( @xmath177 , @xmath178 ) source shows many high equivalent width emission features with deconvolved velocity widths fwhm @xmath179 .
high - ionization species of ne are also detected .
the spectra were taken without an order - blocking filter , making the continuum unreliable long - ward of 7000 .
* cxo40 ( @xmath180 , emission - line galaxy ) : * this galaxy shows a spectrum typical of young , star - forming galaxies : emission from , the doublet , and absorptions from both cah@xmath181k and the hydrogen balmer lines .
the x - ray luminosity is approximately an order of magnitude lower than that typical of sources with obvious agn features in their optical spectra ( @xmath182 ) , though this does not rule out some fraction of the x - ray emission deriving from a nuclear supermassive black hole .
* cxo41 ( @xmath183 , agn ) : * this source shows broad ( fwhm @xmath184 ) and emission .
the optical spectrum long - ward of 7000 is suggestive of a normal star - forming galaxy , a higher redshift analog of cxo40 .
however , the broad , high x - ray luminosity ( @xmath185 ) , and strong core in the _ hst _ image conclusively show that an agn plays a prominent role in this galaxy .
* cxo42 ( @xmath186 , agn ) : * this source shows broad emission ( fwhm @xmath187 ) and narrow emission lines from high - ionization species of ne .
the lines are individually unresolved at our coarse spectral resolution ( fwhm @xmath188 ) .
no stellar features are seen .
we classify the source as an agn .
the _ hst _
image shows the source to be unresolved .
* cxo44 ( @xmath189 , emission - line galaxy ) : * the optical spectrum of this source shows weak , spectrally unresolved emission superposed on an early - type galaxy spectrum .
no obvious agn features are evident in the optical spectrum .
the galaxy is moderately x - ray luminous ( @xmath190 ) .
the _ hst _
image shows a complicated , merging system , suggestive of merger - induced ( nuclear ? ) activity powering the x - ray emission in this system .
* cxo49 ( @xmath191 , emission - line galaxy ) : * this source shows a single emission line straddled by a faint , blue continuum .
the most likely identification of such a line is ( stern et al . 2000 ) .
the morphology from _ hst _ image shows an unresolved source . * cxo50 ( @xmath192 , agn ) : * this broad - lined agn is the highest - redshift source in the current sample .
it shows the classic features of a broad absorption line quasar ( balqso ) .
note that the absorption blue - ward of is telluric in origin , from the atmospheric a - band .
* cxo54 ( @xmath193 , agn ) : * this source is similar to cxo41 and cxo42 , showing broad ( fwhm @xmath194 ) and narrow emission .
the [ ] strength is considerably weaker in this source compared to cxo41 .
we also detect emission from this source .
the redshift places this source at the same redshift of the @xmath195 cluster .
* cxo57 ( @xmath196 , agn ) : * similar to many of the other galaxies , this source shows broad ( fwhm @xmath197 ) and strong , narrow emission .
* cxo63 ( @xmath198 , agn ) : * this source shows strong and emission of moderate width , fwhm @xmath199 .
emission from and are also detected .
again , the spectra were taken without an order - blocking filter , making the continuum unreliable long - ward of 7000 .
* cxo65 ( @xmath200 , emission - line galaxy ) : * this source shows an unremarkable galaxy spectrum with a prominent 4000 break ( d4000 ) and unresolved emission .
no emission is detected .
the source is moderately x - ray luminous , @xmath201 and @xmath202 . * cxo66 ( @xmath203 , emission - line galaxy ) : * this is the least x - ray luminous galaxy in the current spectroscopic sample ( @xmath204 ) .
the spectrum shows classic early - type galaxy features ( ca h@xmath181k absorption , d4000 , h@xmath205 absorption ) with weak and emission evident .
* cxo72 ( m7 v ) : * cxo72 has spectral characteristics intermediate between those of late type dwarfs and giants , suggestive of an intermediate gravity object . the spectral energy distribution between 6500 and 9400 and lack of cah absorption near 6950 are suggestive of an m7 iii giant , while the strengths of the doublet near 7700 and the feature near 8400 are more suggestive of an m7 v dwarf .
we favor the latter interpretation as it provides a more realistic distance for a low - extinction , high galactic latitude source : 700 pc , using the kirkpatrick & mccarthy ( 1994 ) optical absolute magnitude tabulation for late - type dwarfs . unlike the other galactic dwarf identified in this paper or the two x - ray - emitting dwarfs discussed in hornschemeier et al .
( 2001 ) , we detect hard x - rays from cxo72 . as seen in fig .
[ fig_xopt ] , the soft x - ray - to - optical flux ratio is also quite different than the other x - ray dwarfs . assuming the dwarf interpretation , the corresponding hard x - ray luminosity is @xmath206 .
we discuss this source further in 6.3 .
* cxo128 ( @xmath207 , early - type galaxy ) : * this galaxy , detected only in the soft x - ray band , has an early - type spectrum lacking any emission features .
the ca h+k absorption doublet , 4000 break ( d4000 ) , and g - band are very prominent .
the @xmath106-band magnitude is approximately @xmath208 for a @xmath209 giant elliptical galaxy .
the redshift of this source is similar to the group at @xmath210 , slightly foreground and south of the @xmath28 cluster ( see holden et al .
2001 ) .
* cxo138 ( @xmath211 , early - type galaxy ) : * this extremely red galaxy ( @xmath212 , @xmath213 ) has a spectrum similar to that of the old , dead , and red radio galaxy lbds53w091 ( dunlop et al .
1996 ; spinrad et al .
1997 shown for comparison in fig .
11 ) : the optical spectrum is devoid of emission lines and shows continuum breaks at rest - frame 2640 ( b2640 ) , 2900 ( b2900 ) , and 4000 ( d4000 ) .
these features are seen in _ international ultraviolet explorer _ ( _ iue _ ) spectra of f - type main - sequence stars ( fanelli et al .
1992 ) , implying this is an old ( @xmath214 gyr ) galaxy at high - redshift .
the redshift matches that of the two high - redshift clusters targeted by the _ chandra _ observations .
the source , detected in the soft x - ray band near our detection limit , has an x - ray luminosity higher than the other spectroscopically non - active galaxies .
* cxo139 ( m4 v ) : * this m4 dwarf is detected only in the soft band , similar but slightly fainter than the two m4 dwarfs detected in _
observations of the field encompassing the hdf - n ( hornschemeier et al .
the x - ray - to - optical flux ratio is similar to those stars . using optical / near - infrared absolute magnitudes of an m4 v dwarf calculated by kirkpatrick & mccarthy ( 1994 ) ,
we derive a distance of @xmath215 pc for cxo139 .
the corresponding x - ray luminosity is @xmath216 .
again , this luminosity is similar to the two soft x - ray - emitting m4 dwarfs reported in hornschemeier et al .
we discuss x - ray emission from galactic sources further in 6.3 .
* cxo164 ( @xmath217 , early - type galaxy ) : * the spectrum of this source is devoid of emission lines and shows absorption from ca h@xmath181k , a d4000 break , and g - band absorption .
the source is detected only in the hard x - ray band , suggesting that the source may harbor an obscured agn from which no feature is evident in the optical data .
* cxo167 ( @xmath218 , early - type galaxy ) : * similar to cxo164 , cxo167 is detected only in the hard x - ray band and shows an early - type galaxy spectrum in the optical devoid of emission lines .
the feature near the redshifted line is spurious , the residual from a cosmic ray .
two of the sources in our sample are spectroscopically identified with late - type galactic dwarfs .
hornschemeier et al .
( 2001 ) also find two m4 dwarfs in their survey of the field encompassing the hdf - n .
three of these four sources are very similar : mid - m dwarfs detected only in the soft x - ray band with @xmath219 .
the remaining source , cxo72 , is a late - m dwarf , detected out to 7 kev and has @xmath220 .
we briefly review x - ray emission from late - type stars followed by a discussion of this intriguing new source .
phenomenologically , the fraction of stars showing stellar activity is found to increase with lower stellar mass , reaching 100% at spectral type m7 , then to decrease such that few objects later than type l5 show activity ( hawley , gizis , & reid 1996 ; gizis et al .
physically , the observed activity , seen in the form of elevated and/or x - ray emission , is believed to result from collisional heating of ions and electrons along magnetic field lines in stellar chromospheres . for typical chromosphere temperatures ,
such stars are only detected in the soft x - ray band .
stars more massive than @xmath221 have both convective and radiative zones which allow the formation of a stable internal dynamo through the @xmath222 mechanism ( parker 1955 ) .
more rapidly spinning stars have stronger magnetic fields and thus show enhanced activity .
angular momentum loss from stellar winds causes stars to spin - down as they age , with the more massive stars exhibiting stronger winds and thus spinning down more rapidly .
the rise in stellar activity from spectral type k5 to spectral type @xmath15 m9 thus simply tracks the fraction of stars sufficiently young for their given spectral type and associated spin - down rate that the internal dynamo can power stellar activity . for spectral types m7
m9 , the spin - down rate is apparently longer than the hubble time .
this scenario is supported by the observed correlation between activity and rotation rate in late type stars ( kraft 1967 ; basri 1987 ) .
the soft - x - ray emitting m4 dwarfs detected by _
chandra _ in the deep surveys is consistent with this scenario .
curiously , however , none of these sources show elevated h@xmath223 emission .
cxo72 is a more unusual source . as seen in fig .
[ fig_xopt ] , for its optical @xmath114-band magnitude , this star is much brighter in the soft x - ray than the other galactic x - ray sources discussed here .
some of the difference is likely due to the redder spectral energy distribution of its later spectral class .
however , cxo72 is also detected in the hard x - ray band implying a more energetic energy production mechanism than in the soft - band - only galactic sources .
rutledge et al .
( 2000 ) report the _
detection of an x - ray flare from the lithium - bearing m9 brown dwarf lp944@xmath340 , with _ soft _ x - ray emission detected for 1@xmath32 hr during a 12.1 hr observation . unlike that source ,
the x - ray emission from cxo72 is non - transient ( fig .
[ fig_lightcurve ] ) .
we posit that cxo72 is most likely associated with a very low mass binary system , among the lowest mass such systems observed , with accretion powering the x - ray emission .
burgasser et al .
( 2000 ) recently reported strong h@xmath223 emission from a t dwarf ( methane brown dwarf ) discovered in the 2mass survey .
they suggest a close [ @xmath224 interacting binary system with roche lobe overflow powering the observed emission . both the h@xmath223 emission in the t dwarf system and the x - ray emission in cxo72
are not seen to temporally vary in brightness , again suggesting that flaring is not the source of observed stellar activity .
alternatively , most of the hard x - rays from cxo72 come from the second observing period ( see fig . [ fig_lightcurve ] ) , suggesting that time - variable phenomena could be involved .
nine of the x - ray sources discussed here show optical spectra devoid of strong or obvious agn features .
as seen in fig .
[ fig_zlum ] , these sources tend to lie at low redshift ( @xmath225 ) with @xmath226 and @xmath227 .
many of the sources are detected in only one x - ray band .
similar sources are reported in mushotzky et al .
( 2000 ) , barger et al .
( 2001 ) , hornschemeier et al .
( 2001 ) , and tozzi et al .
we briefly review physical processes which might produce such emission . with
a typical @xmath228 thought to derive from shocks developing in unsteady wind outflows ( pallavicini et al .
1987 ) , isolated late - type stars are not prodigious x - ray emitters .
however , a galaxy with a significant young stellar population will produce soft and hard x - rays from several astrophysical mechanisms , primarily associated with the beginning and end stages of massive star evolution ( see helfand & moran 2001 ) .
pre - main sequence stars will emit x - rays in their t tauri phase ( koyama et al .
shocks and hot gas associated with stellar winds , galactic winds , and supernovae will also produce x - ray emission , though the dominant source of hard x - ray emission in a starburst galaxy will be accretion - driven from high - mass x - ray binaries ( hmxbs ) . from a sample of major local group galaxies and galactic ob stars within 3 kpc of the sun ,
helfand & moran ( 2001 ) infer a specific @xmath229 kev x - ray luminosity per o star of @xmath230 . using the models of leitherer , carmelle , & heckman ( 1995 ) we can then relate hard x - ray luminosity to the galaxy star formation rate , @xmath231 . for a salpeter initial mass function ( imf ) with slope 2.35 , an upper mass cutoff 100 @xmath232 , and solar metallicity , the models of leitherer et al .
( 1995 ) show that a region producing @xmath233 for at least @xmath234 yr will have @xmath235 o stars
. therefore , @xmath236 where the hard x - ray luminosity @xmath237 is measured in .
all of the galaxies in our sample showing evidence of star formation ( sources classified as emission - line galaxies ) are indeed detected in the hard x - ray band , with @xmath238 , implying a wide range of possible star formation rates .
however , luminosity , @xmath239}$ ] , also provides a crude estimate of star formation rate ( kennicutt 1992 ) : @xmath240}~ m_\odot~ { \rm yr}^{-1},\ ] ] where @xmath239}$ ] is measured in units of . for our sample of emission - line galaxies , the sfr derived from the x - ray luminosity is typically more than three orders of magnitude greater than that derived from the luminosity .
the implication is that a buried agn is likely producing much of the x - ray emission ( see 6.4.3 ) though dust - enshrouded star formation can not be ruled out .
the _ einstein observatory _ found that some early - type galaxies are powerful x - ray emitters , with x - ray luminosities correlated with optical luminosities ( forman , jones , & tucker 1985 ) , albeit it with a significant dispersion ( @xmath241 ; brown & bregman 1998 ) .
for the x - ray luminous galaxies , hot ( @xmath242 k ) interstellar gas lost by stars during normal stellar evolution is thought responsible for producing the x - ray emission .
the gas thermalizes in the galactic potential well and then cools as it flows to the center of the galaxy .
the x - ray spectra of these galaxies are dominated by thermal emission at @xmath243 kev .
it is thought that the x - ray - faint galaxies have lost much of their interstellar gas through galactic winds or from ram pressure stripping by ambient intercluster or intergroup gas .
these x - ray - faint galaxies show a harder x - ray spectral component , @xmath244 kev ( matsumoto et al . 1997 ) , with a very soft ( @xmath245 kev ) component .
recent _ chandra _ observations of the x - ray - faint elliptical galaxy ngc 4697 resolves most of the soft and hard x - ray emission into point sources , the majority of which are low - mass x - ray binaries ( lmxbs ; sarazin , irwin , & bregman 2000 ) . in fig .
[ fig_xlum ] we plot rest - frame @xmath92 kev luminosity , @xmath246 , against rest - frame @xmath111-band absolute magnitude , @xmath158 , for the spectroscopically - observed _ chandra _ sources from our survey .
we also plot 34 early - type galaxies observed by _
rosat _ by brown & bregman ( 1998 ) .
their sample comprises the 34 optically - brightest , early - type galaxies from the optically - selected , flux - limited sample of faber et al .
( 1989 ) , excluding sources with @xmath247 , dwarfs , and m87 , known to contain an x - ray bright agn .
only two sources in our sample classified as early - type galaxies cxo128 and cxo138 are detected in the soft x - ray band .
both reside within portions of the @xmath248 plot ( fig .
[ fig_xlum ] ) consistent with the early - type galaxies studied by brown & bregman ( 1998 ) .
several of the sources classified as having emission - line galaxy spectra also have soft x - ray luminosities consistent with the bright early - type sample .
the soft x - ray luminosities for these sources can presumably be explained by hot gas and stellar sources . on the other hand , sources cxo164 and cxo167 are both classified as early - type galaxies from their optical spectra but are detected only in the _ hard _ x - ray band .
the x - ray emission in these two sources presumably is powered by a buried agn .
interestingly , the two soft - band - only early - type galaxies are at redshifts where the lynx field has known clusters while the two hard - band - only early - type galaxies are at redshifts lacking large - scale structure in the lynx field .
unification models of agn invoke orientation as one of the dominant parameters that affect the observed optical properties of an active nucleus .
for each broad - lined quasar whose optical emission is doppler - boosted along our line of sight , a population of _ misdirected _ quasars should exist whose agn optical emission is severely attenuated .
norman et al .
( 2001 ) report a clear example of one such source , where the optical spectrum shows strong , narrow - lined emission .
the well - studied high - redshift radio galaxies ( hzrgs ) are the radio - loud component of this population : they also typically only exhibit narrow emission lines ( mccarthy 1993 ; stern et al .
presumably the more centrally - condensed broad - lined region ( blr ) of these sources has been obscured while the spatially - extended narrow - lined region ( nlr ) has not been .
these spatial scales are consistent with reverberation mapping studies of agn . for a less luminous active nucleus
, the stellar emission from the galaxy might dominate the optical spectral energy distribution , providing an optical spectrum devoid of agn features .
the effects of obscuration decrease at high energies , allowing x - ray detection of obscured agn in apparently normal galaxies .
how do the x - ray - emitting galaxies compare to other galaxy populations ? in fig .
[ fig_kz ] we show the @xmath107-band brightness of sources from our survey plotted against spectroscopic redshift .
high - redshift radio galaxies from several radio surveys are also plotted ( see de breuck et al . 2001 , and references therein ) . at each redshift ,
hzrgs are the most luminous galaxies known at observed 2@xmath152 m ( see fig . 10 in de breuck et al .
2001 ) . since 2@xmath152
m samples stellar emission from the low - mass stars which dominate the baryon content of a galaxy , this is interpreted as hzrgs being the most massive systems at each cosmic epoch ( dey 1999 , and references therein ) . in fig .
[ fig_kz ] we also plot 219 redshifts from our followup of @xmath128 sources in the lynx field of spices .
as we can see from this figure , the _ chandra _ sources are among the brighter 2@xmath152 m sources at each redshift , though they are not as luminous as the hzrgs identified from larger area surveys .
we present first results from our deep , 184.7 ks _ chandra _ map of the lynx field .
this is the third - deepest x - ray field observed thus far by _
, exceeded in depth only by the megasecond campaigns in the hdf - n and cdf - s ( hornschemeier et al . 2001 ; tozzi et al . 2001 ) .
we present a catalog of all 153 unresolved field x - ray sources detected in this survey which will provide essential information on the sources responsible for the soft ( @xmath92 kev ) and hard ( @xmath229 kev ) x - ray backgrounds .
we discuss basic results from the x - ray data alone , showing the fainter x - ray sources predominantly have steeper x - ray spectra .
we also present multiband optical and near - ir ( @xmath249 ) identifications for the @xmath250% of lynx _ chandra _ sources which are within the area studied by spices , the @xmath107-selected imaging and redshift survey we are pursuing ( see eisenhardt et al .
we find @xmath251% of the spectroscopically - studied subsample are identified with galaxies showing obvious agn features ( broad lines and high - ionization emission ) , @xmath17% are identified with galaxies showing no obvious agn features , and 2 ( @xmath252% ) of the sources are identified with late - type galactic dwarfs .
one galactic source , cxo72 , is unusual , having an m7 spectral classification and an unusually hard x - ray spectrum .
we discuss this sparse spectroscopic sample in relation to other recently reported deep _
fields and briefly review x - ray emission from sources not showing obvious agn features in their optical spectra .
we conclude indirectly that many such sources in our survey must indeed harbor an obscured agn .
the catalogs presented here will prove useful for more complete spectroscopic followup allowing direct comparison to models of the x - ray background and studies of the fraction of massive galaxies harboring active , enshrouded nuclei .
we are indebted to numerous colleagues for stimulating conversation and insight regarding this work .
in particular , we thank john gizis , davy kirkpatrick , john stauffer , jim liebert , and bob rutledge for their input regarding cxo72 , the hard x - ray - emitting , late - type galactic dwarf . we also thank fiona harrison , peter mao , david helfand , and ed moran for contributing discussion regarding the whole breadth of this project .
we thank roberto della ceca for providing us with the hard - band @xmath70 results from _
asca_. the authors wish to extend special thanks to those of hawaiian ancestry on whose sacred mountain we are privileged to be guests . without their generous hospitality , many of the observations presented herein would not have been possible .
the work of ds and pe were carried out at the jet propulsion laboratory , california institute of technology , under a contract with nasa .
the work of sd was supported by igpp / llnl university collaborative research program grant # 02@xmath3ap@xmath3015 , and was performed under the auspices of the u.s .
department of energy by university of california lawrence livermore national laboratory under contract no . w@xmath37405@xmath3eng@xmath348 .
ad s research is supported by noao which is operated by aura under cooperative agreement with the nsf .
this work has been supported by the following grants : nsf grant ast 95@xmath328536 ( hs ) and nsf career grant ast 9875448 ( re ) . | we present our first results on field x - ray sources detected in a deep , 184.7 ks observation with the advanced ccd imaging spectrometer ( acis - i ) on the _ chandra x - ray observatory_. the observations target the lynx field ( @xmath0 + 4454 ) of spices , the spectroscopic photometric infrared - chosen extragalactic survey , which contains three known x - ray - emitting clusters at redshifts of @xmath1 , and 1.27 . not including the known clusters , in the 17@xmath217 acis - i field we detect 132 sources in the 0.5@xmath32 kev ( soft ) x - ray band down to a limiting flux of @xmath4 and 111 sources in the 2@xmath310 kev ( hard ) x - ray band down to a limiting flux of @xmath5 .
the combined catalog contains a total of 153 sources , of which 42 are detected only in the soft band and 21 are detected only in the hard band . confirming previous _ chandra _ results , we find that the fainter sources have harder x - ray spectra , providing a consistent solution to the long - standing ` spectral paradox ' . from deep optical and near - infrared follow - up data , 77% of the x - ray sources have optical counterparts to @xmath6 and 71% of the x - ray sources have near - infrared counterparts to @xmath7 .
four of the 24 sources in the near - ir field are associated with extremely red objects ( eros ; @xmath8 ) .
we have obtained spectroscopic redshifts with the keck telescopes of 18 of the lynx _ chandra _ sources .
these sources comprise a mix of broad - lined active galaxies / quasars , apparently normal galaxies , and two late - type galactic dwarfs .
intriguingly , one galactic source is identified with an m7 dwarf exhibiting non - transient , hard x - ray emission .
thirteen of the _ chandra _ sources are located within regions for which we have _ hubble space telescope _ imaging .
nine of the sources are detected , showing a range of morphologies : several show compact cores embedded within diffuse emission , while others are spatially extended showing typical galaxy morphologies .
two of the _ chandra _ sources in this subsample appear to be associated with mergers .
we briefly review non - agn mechanisms to produce x - ray emission and discuss properties of the lynx _ chandra _ sample in relation to other samples of x - ray and non - x - ray sources .
2 @xmath9 2 @xmath10 2s @xmath11 2 @xmath12 2hz @xmath13 # 1to 0pt#1 # 1= # 1#2=.45 = .45 # 1#2#3#4#5#6#7 to#2 ' '' '' |
the @xmath2li nucleus is a prominent example of the so - called borromean nuclei which can be described as a three body system ( @xmath3li plus two neutrons ) where none of the internal two body subsystems is bound @xcite .
few - body models and techniques have been very successful @xcite but they must necessarily rely heavily on the underlying two - body interactions .
phenomenological interaction parameters for the neutron-@xmath3li system have usually been obtained by constraints of the properties of @xmath2li . however , comparison between measurements and these theories suffered from the fact that essentially all theoretical investigations assumed spin zero for both @xmath3li and @xmath2li .
this is very unsatisfactory and also rather surprising especially considering the huge amounts of work invested in @xmath2li .
substantially more complicated models are required but the available techniques are also very developed .
the new experimental results have prompted us to reinvestigate the present @xmath1li-@xmath2li connection .
the experimental data about the properties of @xmath1li , and therefore also of @xmath2li , still remain incomplete .
the quantum numbers of the @xmath1li ground state are not firmly established . the first attempt to measure the @xmath1li spectrum was made by wilcox et al .
@xcite placing the ground state of the unbound nucleus @xmath1li at [email protected] mev .
more recent experiments @xcite concluded the existence of a very low - lying state in @xmath1li , although its quantum numbers were not determined . in @xcite
the authors identified the 1@xmath9state as the most probable @xmath1li ground state .
this parity inversion was already discussed much earlier for neighboring nuclei in @xcite and for @xmath1li in @xcite . a probable @xmath5state as ground state of @xmath1li was experimentally suggested in @xcite . in @xcite
a @xmath7resonance was observed at an energy of 538@xmath862 kev , together with a weak evidence of a barely unbound @xmath5 or @xmath7resonance in @xmath1li .
a bit later , abramovich et al .
@xcite and zinser et al .
@xcite claimed that the ground state of @xmath1li corresponds to a very low lying @xmath10=0 state . in the first case
the energies of the doublet 2@xmath11/1@xmath11 were given to be 27 kev and 88 kev , respectively , while in @xcite the energy of the low - lying @xmath5state was established to be below 50 kev .
recent experiments have been performed in order to clarify this puzzle . in @xcite the existence of a @xmath7resonance ,
probably the 2@xmath9 state , at [email protected] mev has been confirmed , and the existence of a lower @xmath7resonance , the 1@xmath9 state , has been found at [email protected] mev . although the existence of a low lying @xmath5state is not confirmed in this work , they emphasize that part of the observed strength at the threshold might represent @xmath10=0 strength . more recently , thoennessen et al .
@xcite have repeated the experiment of kryger et al .
@xcite with improved energy resolution in order to establish more stringent limits of the @xmath5wave parameters .
they concluded that a low lying @xmath5state was observed with a scattering length of at least 20 fm corresponding to a peak energy of less than 50 kev .
they also found indications of a @xmath7state around 540 kev , and they did not rule out another @xmath7-state at @xmath12240 kev . in @xcite the experimental spectrum was fitted with a @xmath7state at @xmath13 kev and width @xmath14 kev .
substitution of this resonance by combination of one @xmath5 and one @xmath7 or combination of two @xmath7 s provided fits which were less significant due to the increase of fit parameters .
inclusion of the extra @xmath5wave or @xmath7wave moved the 500 kev state up by 15 and 25 kev , respectively .
no other states were identified , either because they were not populated , or because the experimental resolution was insufficient .
although there is no evidence for a state at @xmath12250 kev there is a small enhancement , also seen in @xcite .
finally the last experiment by chartier et al .
@xcite produced @xmath1li by proton knockout of @xmath2be .
the measured resulting @xmath3li nucleus in the ground state is predominantly produced from @xmath5-wave neutron emission of @xmath1li which is concluded to resemble the neutron configuration in the @xmath2be ground state .
thus the ground state of the n-@xmath3li system is an @xmath5-wave below 50 kev . as a consequence of all this
it is well established experimentally that the ground state of @xmath1li corresponds to a very low lying @xmath5state .
furthermore a @xmath7resonance at an energy of around 0.5 mev is present in the @xmath1li spectrum .
the other quantum numbers specifying the couplings of these states remain so far unknown .
also the presence of a second @xmath7resonance at around 0.25 mev is not well established and another @xmath5-state , which necessarily also must be present in the low energy spectrum , is entirely unknown .
the observed resonances or virtual states in the neutron-@xmath3li spectrum are related to the second @xmath15 and the first @xmath16-states .
the first @xmath15 and the lowest @xmath17-states are occupied in the @xmath3li - nucleus .
the lowest @xmath15-state is strongly bound and the energy of the @xmath17-state can be estimated as the one - neutron separation energy of @xmath3li , i.e. 4.1 mev @xcite . strictly speaking
this energy applies to @xmath3li whereas both the @xmath17 and @xmath16-states discussed in the present context are related to the @xmath1li - system . from the theoretical side , initial calculations @xcite considered the ground state of @xmath1li to be the 1@xmath9 state at an energy of 810 kev , as given in @xcite .
however in @xcite is already mentioned the possibility that the ground state of @xmath1li is 2@xmath11 , as discussed in @xcite .
the first three - body computation of the @xmath2li properties @xcite predicted that , to be consistent with the available data , @xmath1li should have an @xmath5-state below 300 kev in conflict with the 800 kev at that time believed to be the ground state resonance energy . in @xcite
it was confirmed that core momentum distributions after fragmentation of @xmath2li were consistent with a low lying @xmath5state in @xmath1li , and it was estimated that @xmath2li has nearly 50% of @xmath5wave motion between one neutron and the core .
this fact of a negative parity state was contradicted by subsequent microscopic calculations @xcite that predicted a 1@xmath9 assignment for the ground state of @xmath1li . on the other hand , in @xcite
an @xmath5state was predicted to be the ground state of @xmath1li . in @xcite a method to solve faddeev equations in coordinate space is developed in detail .
this method is especially suitable to describe the large distance behavior of weakly bound three body systems .
in @xcite we have shown that combination of this method to describe @xmath2li and the simple sudden approximation as fragmentation reaction model are sufficient to reproduce the shape of the experimental neutron and core momentum distributions after fast fragmentation of @xmath2li on a light target .
we have shown that the agreement with the experiment is obtained only when the final interaction between fragments is included , and also when a low lying @xmath5state is present in @xmath1li .
inclusion of only @xmath7states in the @xmath1li spectrum necessarily leads to too broad neutron momentum distributions .
in @xcite we have investigated the angular correlations after breakup of @xmath18he and @xmath2li , and we have shown that the presence of a low lying @xmath5state gives rise to a very different angular distribution compared to nuclei as @xmath18he where the @xmath7resonances dominate in the neutron
core system .
the experimental data for the angular distribution arising from @xmath2li was later published @xcite and found to be in agreement with the distribution predicted in @xcite .
this is imperative evidence of strong @xmath5wave configurations in both @xmath1li and @xmath2li .
more detailed calculations of momentum distributions after fragmentation of @xmath2li on light targets are reported in @xcite .
the agreement with the experiments is found to be rather good when around 40% of @xmath7wave is present in the relative neutron@xmath0li motion .
nevertheless the precise energies and quantum numbers of the @xmath1li states are not well established , since the momentum distributions are mainly sensitive to the average energies of the doublets 1@xmath11/2@xmath11 and @xmath19/2@xmath9 @xcite .
for @xmath1li these states arise from couplings of the @xmath3li - spin of 3/2 to the available @xmath15 and @xmath16 neutron states . in @xmath2li
the pauli principle prevents the neutrons from both choosing the lowest state .
parity then requires an equal distribution within each doublet .
thus , most observables are not sensitive to the precise positions of these doublets , only to their average values .
this is the reason why so many theoretical computations agree fairly well with experiments even with the obviously wrong assumptions of spin zero for both @xmath3li and @xmath2li .
the invariant mass spectrum is an exception which could be sensitive to the precise position of the resonances and virtual states .
very recently the fragmentation model used in @xcite has been implemented to take into account the interaction between each of the halo constituents and the target , as well as to include coulomb interaction , making the method valid not only for light targets , but also for intermediate and heavy targets .
this is the participant spectator method described in details in @xcite .
the method works well for @xmath18he , where essentially no uncertainties in the neutron
core interaction are present .
this success for the case of @xmath18he leads us to turn the problem around and investigate in details which structure of @xmath1li is consistent with the firmly established properties of both @xmath3li , @xmath1li and @xmath2li .
first the neutron-@xmath3li potential must be consistent with the structure of @xmath3li and the energies and quantum numbers of the established two - body resonances and virtual states must be reproduced
. then also the known @xmath2li three - body structure must be reproduced .
the unavoidable facts are one and only one particle bound state with angular momentum and parity @xmath20 , with binding of 295@xmath835 kev @xcite , with large interaction cross sections on all targets and with narrow fragment momentum distributions after breakup .
the complicated model with finite core spin is then inescapable .
the measured differential or total @xmath2li reaction cross sections are interpreted as evidence for an exceptionally large spatial extension of the neutron matter and a correlated wave function mixing roughly equally relative neutron - core @xmath5 and @xmath7-states .
the quantitative details must necessarily be extracted by use of model computations .
apart from the binding energy of about 0.30 mev the crucial parameter is the @xmath7-wave content determined to be about 40% in a rather successful model @xcite , which is able to describe essentially all measured breakup cross sections .
the breakup cross sections are first of all sensitive to this @xmath5-wave content and we shall therefore aim at reproducing these 60% in the @xmath2li structure .
the model dependence of the extracted results has to be minimized as far as possible .
this important issue can be illustrated by two different examples .
the first is the interpretation of the neutron momentum distribution observed in breakup reactions .
narrower distributions result from inclusion of both final state interactions and from removal of those parts of the three - body wave function where only one or two of the three halo particles appear after the reaction , i.e. true three - body breakup @xcite
. a larger fraction of neutron - core @xmath5-wave would compensate and delusively reproduce the measurement .
the second example is the interaction cross section often attempted to be expressed in terms of a corresponding value of the @xmath2li root mean square radius .
this interpretation is also model dependent as evidenced by the different results , i.e. radii of [email protected] fm @xcite , [email protected] fm @xcite and [email protected] fm @xcite from glauber models with parametrized densities , three - body densities and from elastic proton scattering with gaussian densities , respectively .
the first of these , off hand least sophisticated , analyses @xcite used parametrized static densities of various kinds . the second analysis used a reaction model accounting for the few - body correlations obviously present in the initial wave function @xcite
the essential input is then the three - body ground state wave function and the resulting radius relies heavily on the properties of this choice .
the result of about 3.53 fm quoted above arises from two - body interactions where the virtual states and resonances can not all agree with the established experimental information .
there average positions are unrealistically low and there is no spin splitting in the calculation .
the @xmath7-wave content is almost 60% and the presently accepted use of a three - body interaction for fine tuning is not applied .
the third analysis is from elastic proton scattering resulting in an even larger radius @xcite .
again this is one numerical value and corresponding results for @xmath21he are in conflict with other information .
avoiding the model dependence entirely is not possible . minimizing the uncertainties
at least require that all effects influencing the observables in question must be accounted for in the model .
the safest is to use one consistent model with all anticipated physical effects included and with one set of parameters reproducing all measured data from the two - body input data to the differential breakup cross sections .
we rely on the three - body structure model @xcite and the reaction models developed for systems precisely like @xmath2li @xcite .
the paper is organized as follows : in section [ sec2 ] we briefly sketch the method used to construct the three body wave function and the model to compute the breakup cross sections .
we continue in section [ sec3 ] with the details of the interactions involved in the particular case of @xmath2li . in section [ sec4 ]
we show results for the @xmath2li wave function and for different relevant observables after fragmentation reactions .
finally section [ sec5 ] contains the summary and the conclusions .
the wave function of the three body halo system is obtained by solving the faddeev equations in coordinate space @xcite .
this is done by writing each of the three faddeev equations in terms of each of the three sets of hyperspherical coordinates @xmath22 , @xmath23 , where the index @xmath24 is related to a given jacobi system .
the three
body wave function @xmath25 is then a sum of the three faddeev components , each of them expanded for a given @xmath26 in terms of a complete sets of generalized angular functions @xmath27
@xmath28 the angular wave functions are solutions of the angular part of the faddeev equations @xmath29 where @xmath30 is a cyclic permutation of @xmath31 , @xmath32 is a normalization mass , @xmath33 is the two - body interaction between particles @xmath34 and @xmath35 and @xmath36 is the @xmath26-independent part of the kinetic energy operator .
the expression for @xmath37 is given in @xcite .
the radial functions @xmath38 are the solutions of the coupled set of `` radial '' differential equations @xmath39 where @xmath40 is an anticipated three - body potential and the eigenvalues of the angular part , @xmath41 , enter as effective potentials .
the functions @xmath42 and @xmath43 are defined as the angular integrals @xmath44 the observables obtained after fragmentation of @xmath2li are computed by using the participant spectator method @xcite .
the breakup cross sections are obtained by adding the incoherent contributions from processes where one , two , or three halo particles simultaneously interact with the target .
the interaction with the target of one of them ( participant ) is described by a phenomenological optical potential and the black disk model is used for the other two particles ( spectators ) .
let us assume that the halo constituent @xmath24 is the participant and the final state can be described as two independent two body systems , one made by particle @xmath24 and the target @xmath45 , and the second made by the remaining halo constituents @xmath34 and @xmath35 .
this final state description is appropriate for large momentum transfer processes in which the constituent @xmath24 is violently removed from the projectile . under these assumptions , and working in the frame of the halo projectile
, the differential cross sections can be written as @xcite @xmath46 @xmath47 where eq.([elas ] ) and ( [ abs ] ) refer to the cases in which @xmath24 is elastically scattered and absorbed by the target , respectively .
the cross sections @xmath48 and @xmath49 are the differential elastic and absorption cross sections for the participant target scattering .
the momenta @xmath50 , @xmath51 , and @xmath52 are the relative momenta in the final state between the center of mass of the two final two body systems , between the halo constituents @xmath34 and @xmath35 , and between @xmath24 and the target , respectively . the momentum @xmath53 is the momentum transfer in the process .
the quantum numbers @xmath54 are the total spin and its third component of the halo nucleus , @xmath55 are the spin and third component of the two body final state made by particles @xmath34 and @xmath35 , and @xmath56 is the final state third component of the spin of particle @xmath24 . the target is assumed to have spin zero .
the expressions are then correct when the participant @xmath24 has spin 0 or 1/2 , but we shall employ them even for the spin of 3/2 corresponding to the @xmath3li - core . we expect this is a very accurate approximation .
the function @xmath57 in eq.([elas ] ) is included to remove the non zero component of the three body bound state plus target in the final state , that represents elastic scattering of the halo nucleus as a whole . finally , @xmath58 is the overlap between the @xmath59wave function in the final state and the wave function of the three body projectile . in @xcite
all the details about the participant spectator model are given , as well as the optical potentials used to describe the interaction between participant @xmath24 and the target .
to obtain the @xmath2li wave function by use of the procedure sketched above one needs to specify the two body interactions entering in eq.([eq1 ] ) , i.e. the neutron neutron and the neutron-@xmath0li interactions .
the neutron - neutron interaction is specified in @xcite .
the neutron@xmath3li interaction is obviously decisive in order to establish the @xmath1li spectrum .
we therefore must be very careful in the choices of both form and parameters .
results from previous work can not be used directly since we want , and perhaps even need , a better correspondence between core - density and potentials of both core and valence neutrons .
we begin with the core density and potential of the core neutrons .
the @xmath0li - core consists of two neutrons and two protons in the @xmath15shell and four neutrons and one proton in the @xmath17shell . the mean square radius of @xmath0li is then given by @xmath60 , where @xmath61 and @xmath62 are the mean square radii of the @xmath15 and @xmath17orbits , respectively .
if we use gaussian potentials for @xmath5 and @xmath7waves with equal range , @xmath63 , we can estimate the range of the interaction by using that i ) the root mean square radius of @xmath0li is 2.32 fm , ii ) the lowest @xmath5state must be bound and the second must correspond to a low lying neutron virtual state in @xmath1li , and iii ) the @xmath7potential must have a bound state with a neutron separation energy around 4.1 mev @xcite .
the precise position of the @xmath17-state is not important in the present work and we shall use the value 4.1 mev obtained for @xmath3li even though the potential applies to the neutron-@xmath3li system .
the @xmath15 and @xmath17 bound states in this potential then produce a @xmath0li density distribution which strictly speaking is that of the core particles in @xmath1li .
this is lack of selfconsistency in line with an inert core approximation , but only applied to estimate the range of the potential .
this is only of marginal importance provided the resonances and virtual states remain unchanged . in table
[ tab1 ] we show these potentials fulfilling conditions ii ) and iii ) for three different ranges of the gaussian . in all the three cases
the @xmath5potentials have only a bound state and a virtual state around 300 kev , and the @xmath7potential has only a bound state at @xmath64 mev .
as we can see in the table , only a range of around 2.0 fm is giving rise to a root mean square radius for @xmath0li in agreement with the experimental value of 2.32 fm .
.parameters of the trial potentials @xmath65 used for the occupied @xmath15 and @xmath17levels in @xmath0li .
the @xmath5potential has a single bound state and a virtual state around 300 kev .
the @xmath7potential has a single bound state at @xmath64 mev .
the root mean square radius @xmath66 of @xmath0li is obtained from the mean square radii @xmath67 and @xmath68 of the @xmath5 and @xmath7-orbits , respectively .
[ cols="^,^,^,^",options="header " , ] [ tab3 ] columns ( a ) and ( b ) of these spectra can also be consistent with a second @xmath16resonance around 0.24 mev , although now this resonance corresponds to the 2@xmath9level instead of the 1@xmath9level as given in @xcite . in the spectra ( c ) and ( d ) in the central part of fig .
[ fig6 ] the lowest @xmath16resonance is very close to threshold ( the crossing between @xmath69 kev and @xmath7wave content of 40% in fig .
[ fig5 ] is very close to zero @xmath19energy ) .
as previously indicated there is no experimental evidence of such a low @xmath16resonance , and it is established that the ground state corresponds to an @xmath15state . finally , in the lower part of fig . [ fig6 ]
we show the @xmath1li spectrum by assuming @xmath70 kev . again , since the three
body force is providing less binding the average energy of the @xmath5states and the @xmath7resonances is smaller than in the previous cases .
the role played by the three
body force is now not very important .
the crossing between the lines corresponding to constant @xmath71=50 kev and constant @xmath7wave content of 40% in fig .
[ fig4 ] gives rise to the two spectra shown in the figure . in this case
also a @xmath16resonance ( the 2@xmath9state ) appears at an energy below 100 kev . from fig .
[ fig5 ] the same procedure gives negative energies of the 1@xmath9level , meaning that this level is bound .
this is of course against the well known fact that @xmath1li is not bound , and the corresponding @xmath1li spectrum can certainly not correspond to the real one . actually , from the spectra ( c ) and ( d ) in the middle part of fig .
[ fig6 ] , we already see that a value of @xmath71 slightly below 200 kev is already binding the @xmath19level . for completeness
we give in table [ tab3 ] the two and three body potentials corresponding to all the spectra shown in fig .
[ fig6 ] . according to the experimental and theoretical information collected about @xmath1li and @xmath2li
we can conclude that the spectra shown in the upper part of fig .
[ fig6 ] , especially in ( c ) and ( d ) , and spectra ( a ) and ( b ) shown in the middle part , are the main candidates for the @xmath1li spectrum . from the different observables obtained experimentally after fragmentation of @xmath2li the invariant mass spectrum is the observable most sensitive to the energies of the different resonances and virtual states .
investigation of the invariant mass spectrum could then help to establish which spectrum is most probable for @xmath1li . in fig .
[ fig7 ] we show the invariant mass spectrum for all the @xmath1li spectra shown in fig . [ fig6 ] .
the observables after fragmentation of @xmath2li are computed following the participant spectator method described in @xcite .
the external parts of each graph in the figure show the computed invariant mass spectrum ( thin lines ) and the result after convoluting with the experimental beam profile ( thick lines ) .
this curve is compared with the experimental data taken from @xcite and @xcite .
these sets of experimental data do not agree in the absolute numbers , and we have scaled those of @xcite to the same maximum as in @xcite .
note that the only difference between the @xmath1li spectra ( a ) and ( b ) and between ( c ) and
( d ) is the ordering of the @xmath72 and the @xmath73levels . therefore the compared solid lines and dashed lines in fig .
[ fig7 ] differ only in the ordering of the 1@xmath11/2@xmath11doublet .
for the solid lines the @xmath72state is at 50 kev , while for the dashed lines the 2@xmath11state is at 50 kev . as we can see in all the panels in fig .
[ fig7 ] the difference between solid and dashed lines is very small , especially after convoluting with the experimental beam profile .
the only difference appears before convoluting and at very small energies , where the cases with the 2@xmath11state at 50 kev ( dashed lines ) have a higher peak .
this is due to the larger statistical weight of the @xmath73level compared to the 1@xmath11level ( 5/3 times bigger ) . in any case , from the invariant mass distributions and its comparison with the present experimental data it is not possible to determine which of the two states in the doublet @xmath72/@xmath73 corresponds to the low
lying @xmath5state found experimentally .
let us concentrate now on the @xmath1li spectra ( a ) and ( b ) .
in the three cases , upper , middle , and lower parts , one of the @xmath5states is placed at 50 kev , and the @xmath19resonance is fixed at 0.54 mev .
basically the only difference between the three cases is the energy of the 2@xmath9level , that goes from almost 0.4 mev to a small value below 100 kev .
looking at the corresponding invariant mass distributions in the left part of fig .
[ fig7 ] we observe that calculations before convoluting with the experimental beam profile show two peaks corresponding to the low @xmath5state at 50 kev , and the @xmath74resonance .
these two peaks approach each other when the energy of these two levels becomes closer , and the one corresponding to the @xmath74resonance becomes higher when its energy approaches the threshold .
in fact , in the lower part of fig .
[ fig7 ] both peaks can not be distinguished .
the peak corresponding to the @xmath19state at 0.54 mev can not be seen .
looking at the @xmath7wave contribution ( inner panels ) we note that only in the lower part this peak can be noticed in the small bump produced in the total invariant mass distribution , although this bump disappears after convolution with the beam profile .
the main effect produced by the different energies of the @xmath74resonance is that the lower the energy of the state the narrower the computed invariant mass spectrum .
this fact is preserved after convoluting the computed distributions with the experimental beam profile .
the main effect of the convolution is that the separated peaks associated to the low lying @xmath5state and the @xmath74resonance merge into one . in any case
the comparison with the experimental distribution shows that the computed ones are too narrow , except perhaps for values of the @xmath74level close or higher than 0.4 mev .
this would mean that the two states in the @xmath19/@xmath74doublet would be almost degenerate .
therefore a @xmath74level below the @xmath19level is not likely .
let us focus now on the @xmath1li spectra ( c ) and ( d ) .
again one of the @xmath5levels is placed at 50 kev , but now the @xmath74level is fixed at 0.54 mev , while the energy of the @xmath19level takes different values . in the upper part
the 1@xmath9resonance has an energy of 0.25 mev .
the computed invariant mass spectrum shows two shoulders produced by the two @xmath7resonances .
these two states are clearly seen in the @xmath7wave contribution shown in the inset of the upper part of fig .
nevertheless the structure observed in the computed invariant mass spectrum disappears after convolution .
the agreement with the experimental data is quite satisfactory , especially with the one given in @xcite . if we reduce the energy of the @xmath19level its contribution appears in a higher and narrower peak . in the middle part of fig .
[ fig7 ] we show the extreme case in which the @xmath19level is very close to the threshold ( its energy is around 10 kev ) .
as we see in the inset , the @xmath7wave contribution has a wide peak at 0.5 mev and a sharp and very high peak at a very low energy corresponding to the 1@xmath9resonance .
this is obviously producing a similar high and narrow peak in the computed invariant mass spectrum .
nevertheless after convolution this peak is not visible anymore , and the comparison with the experiment can also be considered rather good .
this is because although convolution is diluting the presence of the sharp peak at low energies there is still a resonance at 0.54 mev , with high statistical weight , that makes the tail of the total distribution behave accordingly .
therefore a @xmath19level below the @xmath74 is the most likely structure for the @xmath1li spectrum .
this result agrees with @xcite . from the simultaneous analysis of figs .
[ fig6 ] and [ fig7 ] we can then summarize the results in the following points : i ) we can not establish which of the two states in the doublet @xmath72/@xmath73 is the one at a very low energy , ii ) a @xmath74 resonance below the @xmath19level is not likely , producing too narrow invariant mass spectra , unless the @xmath19 and the @xmath74states are close to degeneracy , iii ) the most likely situation corresponds to a @xmath19level below the @xmath74state , although it can not be extracted by direct comparison with the experimental invariant mass spectra , ( iv ) experiments with a better beam energy resolution are desirable , since the predicted structure in the invariant mass spectra then could be detected and the results from the different spectra directly distinguished . as a final test , we show in fig .
[ fig8 ] some additional observables after fragmentation of @xmath2li on carbon .
the calculations have been performed assuming the best candidates for the @xmath1li spectrum , i.e. the spectra ( c ) and ( d ) in the upper part of fig [ fig6 ] .
the computed results for both spectra are completely indistinguishable . in part
( a ) of fig .
[ fig8 ] we show the core momentum distribution .
the thin line is the pure calculation while the thick line is the result obtained after convoluting with the experimental beam profile @xcite . in part
( b ) we show the neutron momentum distribution , for which experimental data are not available yet .
in part ( c ) we show the angular distribution , where @xmath75 is the angle between the @xmath1li momentum and the relative momentum between neutron and core after the fragmentation .
the curve is very sensitive to the @xmath5 and @xmath7-wave mixing and the agreement is therefore a strong indication of a @xmath7-wave content of about 40% .
for both the distributions in ( a ) and ( c ) the agreement with the experimental data is very good .
finally we want to discuss the radius of @xmath2li .
it has been extracted by model computations and fitting of measured interaction cross sections .
the results depend somewhat on the assumed density distributions and the reaction models .
we obtain interaction cross sections almost constant between 1030 and 1045 mb for a beam energy varying between 300 and 700 mev / nucleon on a carbon target .
this should be compared to the experimental values of @xmath76 mb at 800 mev / nucleon and @xmath77 mb at 400 mev / nucleon .
also the measured interaction cross sections for heavier targets like copper and lead agree quite well with our model computations using the new neutron- @xmath3li interaction .
our resulting root mean square radius is about 3.2 fm .
it should perhaps be emphasized that a larger range and in particular another choice of radial shape of the three - body interaction could be designed to increase the resulting radius by fractions of a fm possibly without changing anything else .
the recent measurements of the properties of @xmath1li confine the neutron-@xmath3li interaction which is decisive for the elaborate three - body computations of @xmath2li . to use the new information
the @xmath2li model must be realistic and in particular the hyperfine structure arising from spin - spin couplings must be included when the measured energy spectrum of @xmath1li is used for comparison .
we first sketch our method of adiabatic hyperspherical expansion of the faddeev components to determine the structure of @xmath2li from a given set of two - body interactions .
second we sketch the participant - spectator reaction model used for all the fragmentation cross sections .
the neutron - neutron interaction is well known and not crucial as long as the scattering length and effective ranges are correct .
the same could be said about the neutron-@xmath3li interaction but in this case we do not have sufficient information .
we believe that compelling experimental evidence show that an @xmath5-state is present at around 50 kev and a @xmath7-state around 0.5 mev .
we proceed by constructing an @xmath5-state neutron-@xmath3li gaussian potential of range @xmath78 with a deeply bound state and one very low - lying virtual state .
we also construct a @xmath17 gaussian potential with the same range @xmath78 with a bound state at -4.1 mev .
these two potentials produce a density distribution of @xmath3li with the measured root mean square radius of 2.32 fm when @xmath79 fm .
the core density and potential radius is now defined . the three - body computation for @xmath2li with these interactions supplemented with an appropriate spin - orbit term would lead to a disaster since the additional two neutrons would occupy the low - lying pauli forbidden states of the core - neutrons .
therefore we introduce the phase equivalent potentials for both @xmath15 and @xmath17 states .
these potentials have precisely the same low - energy scattering properties but without the pauli forbidden bound states . for weakly bound three - body systems
this has proven to be an accurate procedure .
although technically difficult this is possible and above all this is a physically appropriate prescription to account for the most important many - body effect in the present context .
the radial shapes of the potentials are not important provided the low - energy scattering properties are correct .
therefore we are content with using gaussians .
the result is that we have two free strength parameters , one for @xmath15 and one for @xmath16-states .
the last part of the neutron-@xmath3li interaction arises from the @xmath3li - spin of @xmath4 coupling to the neutron angular momenta of @xmath15 and @xmath16 resulting in two sets of spin - split states of opposite parity but the same spins of 1 and 2 .
again we maintain the same range parameter @xmath78 of a gaussian potential leaving us with two new strength parameters .
these four parameters may be understood as related to the four different @xmath1li - states .
in addition the properties of @xmath2li can now be used to constrain the parameters .
unfortunately we must also use the three - body potential introduced to fine tune the three - body energy which is necessary even when using a set of two - body interactions reproducing all low - energy scattering properties .
this must be a short range potential and we make it as neutral as possible under that assumption , i.e. we only use it to adjust the binding energy to the measured value and not to upset other unrelated properties . we know that fragmentation reaction cross sections rather strongly select a neutron-@xmath3li relative @xmath7-wave content within @xmath2li of about 40% .
we also know that most of these cross sections as well as the @xmath2li binding energy and size are essentially independent of the spin - splitting because the pauli principle for the outer neutrons requires at most one identical particle in each of these spin - split states .
if one is occupied the other is as well .
then the average is the most important quantity .
thus a given @xmath7-wave content selects a one dimensional relation between the average energies of the @xmath15 and @xmath16 states .
if the three - body potential had been known or determined from independent sources both the statistically averaged energies could have been found from @xmath7-wave content and three - body binding energy .
still an educated guess using experience provide a reasonable estimate .
we continue with spin - splitting both @xmath5 and @xmath7-waves and demanding that one @xmath5-state must be at 50 kev and one @xmath7-state at 0.54 mev . with the average positions fixed
this provides the spectrum of the four @xmath1li states except that the ordering is yet undetermined and the inaccuracy from the three - body potential is also still present .
therefore we compute sensitive fragmentation data like the invariant mass spectrum which should be able to distinguish between the various @xmath1li spin assignments .
unfortunately the precision of this experimental data is not sufficient to exclude more than two of the four different , but essentially discrete , spin structures . in conclusion ,
the strong connections between the properties of @xmath1li and @xmath2li are used to produce a consistent description .
experimental constraints from both nuclei are essential .
we maintain the previous level of accuracy in the description of both structure and breakup reactions of @xmath2li .
the spectrum of @xmath1li is confined to have average energies of @xmath5 and @xmath7-states at @xmath80 mev .
the individual @xmath7-states should appear at about 0.54 mev ( most likely the 2@xmath9resonance ) and at @xmath81 mev ( probably the 1@xmath9resonance ) .
the @xmath5-states are at about 50 kev and at @xmath82 mev , where the present knowledge is insufficient or too inaccurate to establish the ordering of the 1@xmath11/2@xmath11 doublet .
furthermore , the three - body potential energy , accounting for three - body polarization , off - shell effects , core - excited states beyond two - body phenomenology or other cluster configurations , must contribute by about @xmath83 mev to the binding energy . | the neutron@xmath0li interaction and the corresponding low - energy @xmath1li spectrum are decisive for the properties of @xmath2li described as a three - body system ( n+n+@xmath3li ) .
we compute structure and breakup properties of @xmath2li as function of this interaction .
the hyperfine structure due to the spin @xmath4 of both @xmath3li and @xmath2li is needed and treated with special care .
we use the hyperspherical adiabatic expansion of the faddeev equations for the structure and the participant spectator model for the breakup reactions of @xmath2li .
we use established experimental constraints of both @xmath2li ( binding energy , size and differential breakup cross sections ) and of @xmath1li ( a virtual @xmath5state below @xmath6 kev and a @xmath7resonance around 0.54 mev)-state energies should strictly speaking be negative , but we shall here and throughout the paper omit the minus sign and use the absolute value . ] .
another @xmath7resonance must then be present below 0.54 mev and another @xmath5level must be present between about 0.5 mev and 1 mev depending on unknown spin assignments .
all established facts are in agreement with our predictions obtained within the same consistent model . 21.45.+v , 25.10.+s , 25.60.-t , 25.60.gc |
recent experiments on thin film growth on well characterized substrates using molecular beam epitaxy have provided detailed information about growth kinetics and morphology . in particular , growth in the submonolayer regime
has been studied extensively using both scanning tunneling microscopy ( stm ) @xcite and diffraction methods such as helium atom beam scattering @xcite and low energy electron diffraction @xcite .
it was observed that for a variety of systems and a broad temperature range , island nucleation is the dominant mechanism for crystal growth .
a variety of island morphologies has been found .
fractal - like islands , resembling diffusion limited aggregation ( dla ) clusters , were observed in au on ru(0001 ) @xcite , cu on ru(0001 ) @xcite and pt on pt(111 ) @xcite . on the other hand ,
compact islands were observed in ni on ni(001 ) system @xcite . scaling properties of the island density as a function of deposition rate and coverage , as well as the island size distribution
have been studied experimentally by several groups @xcite .
field ion microscopy ( fim ) experiments that provide direct access to diffusion processes at the atomic scale , have also been performed @xcite .
this technique was used to identify the diffusion modes of adatoms @xcite as well as small islands @xcite on fcc(001 ) metal surfaces and to measure their diffusion coefficients .
theoretical studies aimed at providing a better understanding of the relation between key processes at the atomic scale and the resulting morphologies have been done using monte carlo ( mc ) simulations @xcite . in simulations of island growth during deposition ,
atoms are deposited randomly on the substrate at rate @xmath1 [ given in monolayer ( ml ) per second ] and then hop according to a microscopic model .
the hopping rate @xmath9 ( in units of hops per second ) of a given atom to each unoccupied nearest neighbor ( nn ) site is given by @xmath10 where @xmath11 @xmath12 is the commonly used attempt rate , @xmath13 is the activation energy barrier , @xmath14 is the boltzmann constant and @xmath15 is the temperature .
the activation energy barrier @xmath13 depends on the local environment of the hopping atom , namely the configuration of occupied and unoccupied adjacent sites .
two approaches have been taken in the construction of the energy barriers for hopping .
one approach was to construct simple models that include the desired features , such as stability and mobility of small islands , and that take into account properties such as bond energies @xcite .
the advantage of this approach is that the models are well defined and use only a few parameters .
these models are useful for studies of scaling and morphology but can not provide a quantitative description of diffusion on a particular substrate .
a second approach is based on the use of an approximate many - body energy functional to calculate the hopping energy barriers for a complete set of relevant configurations @xcite .
this approach provides a good description of diffusion processes on the given substrate but only limited understanding due to the large number of parameters . in this work
we propose a procedure that combines the advantages of both approaches . using sensible assumptions about the bond energies and diffusion paths we obtain a simple formula for the activation energy barriers .
we then optimize the parameters of this formula by using energy barriers obtained from the embedded atom method for cu diffusion on cu(001 ) .
this procedure provides two simple models which combine the best features of both approaches .
model i , which has three parameters , provides good quantitative description of the landscape of hopping energy barriers .
model ii , which has only two parameters and is as simple as other minimal models , still incorporates the essential physics of diffusion of adatoms on fcc(001 ) metal surfaces .
we find that on the cu(001 ) substrate , edge mobility ( i.e. , the mobility of an adatom at the perimeter of an island ) is much higher than the single adatom mobility , giving rise to compact island shapes .
dimer mobility is comparable to the single adatom mobility while trimers are also mobile @xcite .
mobility of small islands has a significant effect on the asymptotic scaling properties and is thus particularly important .
we also find that vacancies have higher mobility than adatoms .
the paper is organized as follows . in section
ii we introduce the models .
simulation and results are presented in section iii with emphasis on scaling and morphology .
the results and their implications are discussed in section iv and a summary is given in section v.
in this work we use a set of energy barriers for cu on cu(001 ) @xcite obtained using the embedded atom method ( eam ) @xcite .
this method uses semiempirical potentials and provides a good description of self diffusion of cu on cu(001 ) and similar surfaces @xcite .
specifically , we use the eam functions of cu developed by adams , foiles , and wolfer @xcite which are fitted to a similar data base as the one employed by foiles , baskes , and daw @xcite .
the hopping energy barriers are calculated for all local environments as shown in fig .
1 , where seven adjacent sites , @xmath16 are taken into account @xcite . each one of these sites can be either occupied ( @xmath17 ) or vacant ( @xmath18 ) , giving rise to @xmath19 barriers .
a binary representation is used to assign indices to these barriers . for each configuration @xmath20 the barrier
is given by @xmath21 , where @xmath22 takes the values @xmath23 .
the full set of hopping energy barriers ( given in ev ) is presented in table i. to show these values in a compact form , each barrier in table i corresponds to a configuration in which the occupied sites are the union of the occupied sites in the picture on top of the given column and on the left hand side of the given row .
the column in table i in which a given configuration appears is determined by the occupancy of sites @xmath24 while the row is determined by sites @xmath25 .
one can define @xmath26 such that for each configuration @xmath27 . to demonstrate the use of table i
, we will check the barrier for the configuration in which sites @xmath28 and @xmath29 are occupied and all other sites adjacent to the hopping atom are vacant . for this configuration , according to eq .
( [ n1n2 ] ) , @xmath30 and @xmath31 ( @xmath32 ) .
the barrier , which is found in the column with the index @xmath33 and the row with index @xmath34 , is @xmath35 ev . in table
i we use the symmetries of the configurations in the @xmath36 cell to reduce the number of entries .
there is a mirror symmetry plane perpendicular to the surface and containing the arrow of the hopping atom .
consequently , the columns of @xmath37 and @xmath38 , in which site @xmath39 is occupied , stand also for the symmetric configurations in which @xmath40 is occupied . in the other four columns , there are some configurations that , due to symmetry , appear twice . in such cases ,
the barrier for the configuration with larger @xmath41 appears in _
italics_. here we consider only hopping moves in which a single atom hops each time .
however , it turns out that in some cases the molecular statics calculations , used to obtain the barriers , give rise to concerted moves . in such moves the atom at site @xmath42 follows the hopping atom and takes the place vacated by the hopping atom
this fact significantly reduces the barrier .
we have found that for all configurations in which concerted moves appear , these concerted moves can be suppressed by adding one more row of atoms on the left hand sides of sites @xmath43 and @xmath29 . in table
i , the energy values for those configuration in which a concerted move was found , are shown in parenthesis .
the barrier obtained when the concerted move was suppressed is shown to the left of the parenthesis .
concerted moves may affect film growth by modifying either the stability and mobility of small islands or the morphology of large islands .
the effect of the shear move ( @xmath44 ) , in which concerted motion reduces the barrier from @xmath45 ev to @xmath46 ev , was studied in ref .
it was claimed that for experiments with very low deposition rates this move effectively raises @xmath4 from 3 to 8 , where @xmath4 is the size of the critical island nucleus .
simple calculations of the hopping rates , using eq .
( [ hoppingrate ] ) indicate that , since the reduced barrier is still rather large , the effect of this concerted move is negligible for the temperatures and deposition rates studied in this paper .
we find that there are other concerted moves which , unlike the shear move discussed above , cause a dramatic reduction in the barrier for the given move ( table i ) .
if the atoms inside the @xmath36 square represent an isolated island and sites around are vacant , it is easy to see that moves such as @xmath47 , and @xmath48 simply lead to a more stable configuration of the pentamer and heptamer islands , respectively .
once this more stable configuration is obtained , the concerted move can not take place again .
since these more stable configurations can be obtained through other fast moves , the concerted move does not have a significant effect on the overall morphology of the film .
other moves such as @xmath49 , @xmath50 , and @xmath51 can occur only as long as the islands of 4 , 6 and 8 atoms respectively , have not reached their most stable configuration .
if the above configurations , in which concerted moves are possible , occur in a denser environment , and in particular if the sites on the left hand side of sites @xmath43 and @xmath29 are occupied , the concerted moves are suppressed .
thus , we conclude that in both cases the concerted moves do not have a significant effect in the simulations presented here and they will be ignored . to gain a better understanding of the barrier energy landscape we discuss the barrier height distribution ( without concerted moves ) in fig .
we observe that this distribution exhibits four peaks .
this feature is in agreement with ref .
@xcite where a different method @xcite was used to calculate the barriers .
each peak , corresponds to a single or a double column in table i ( there is a little ambiguity in the region between peaks iii and iv due to a slight overlap ) . in general
, peak i includes very fast moves toward island edges , peak ii includes moves along the edge , peak iii includes , most notably , the single atom move , while peak iv includes detachment moves . when an atom on the surface hops into a vacant nearest neighbor site it has to cross the energy barrier between the initial and final sites .
we have used molecular statics in conjunction with eam functions to find the diffusion path of an adatom .
not surprisingly , we have found that the top of the energy barrier to go from a four - fold coordinate site to an adjacent site is at the bridge site . by slowly moving the adatom within the molecular statics calculation
, it turns out that the hopping energy barrier is simply the difference between the energy at the bridge site and in the initial site .
the occupancy of the seven adjacent sites ( fig .
1 ) affects both energies .
we will now express the energy of the hopping atom in its initial site as : @xmath52 where @xmath17 if site @xmath53 is occupied and 0 otherwise .
the energy of an isolated atom is @xmath54 while the reduction in its energy due the presence of an atom in a nearest ( next nearest ) neighbor site is given by @xmath55 ( @xmath56 ) . here
we assume that the contributions of nearest neighbor and next nearest neighbor ( nnn ) atoms to the energy are additive .
the energy of the hopping atom when it is on the bridge site is given by : @xmath57 where @xmath58 is the energy of an isolated atom on top of a bridge site , while @xmath59 is the reduction in the energy due to the presence of an atom in one of the four sites adjacent to the bridge site .
we do not include here nnn type contributions since their effect is small .
therefore , for a given configuration the barrier @xmath60 for an atom to hop into an adjacent vacant site is given by : @xmath61 where @xmath62 and @xmath41 is given by eq .
( [ barindex ] ) . to examine the formula above
, we found the parameters that best describe the 128 eam barriers by minimizing the sum of squares @xmath63 ^ 2 .
\label{optimize}\ ] ] the values obtained for the parameters in eq .
( [ longform ] ) are @xmath64 , @xmath65 , @xmath66 and @xmath67 ev .
thus , we find that to within about @xmath68 ev , @xmath69 . replacing both @xmath55 and @xmath59 by @xmath70 ev we obtain a three parameter model ( model i ) in which the energy barriers are : @xmath71 where @xmath41 is given by eq .
( [ barindex ] ) . the distribution of energy barriers obtained from eq .
( [ barrier3 ] ) is shown in fig .
2(b ) . a very good agreement in the location of the four peaks with the eam energy barriers is found , but eam peaks are significantly broader .
this can be due to the fact that during the hopping move adjacent atoms can relax within their potential well .
model i accounts for these effects only on average and therefore gives rise to narrower peaks .
one can further simplify the model by ignoring the nnn interactions which are relatively small , namely choosing @xmath72 .
this model ( model ii ) has only two parameters and is described by : @xmath73 where @xmath41 is given by eq .
( [ barindex ] ) .
repeating the optimization procedure of eq .
( [ optimize ] ) for the barriers given by eq .
( [ barrier4 ] ) , we obtain @xmath74 and @xmath75 . in this model , all the energy barriers in each peak collapse into a single value . in spite of its simplicity , this model provides a good description of self diffusion of cu on the cu(001 ) substrate . to find out which island sizes are mobile and to obtain the diffusion coefficients of mobile islands on cu(001 ) , we have done simulations of single cluster diffusion . to obtain the statistics required for a precise determination of the diffusion coefficients we performed 1000 runs for diffusion of monomers , dimers , trimers , pentamers ( @xmath76 ) and heptamers ( @xmath77 ) .
each run was carried out for a time equal to @xmath78 seconds , which is more than 100 times larger than the time scale for hopping of an isolated adatom at the given temperature ( @xmath79k ) .
the diffusion coefficients were obtained from the relation @xmath80 , where @xmath81 is the number of atoms in the cluster , @xmath82 is the distance between the initial position of the center of mass of the cluster and its position after time @xmath83 , @xmath84 is the diffusion coefficient for a cluster of size @xmath85 and @xmath86 represents an average over the 1000 runs .
the diffusion coefficients for monomers , dimers , trimers , pentamers and heptamers are shown in table ii .
it is also found that under the conditions studied here the diffusion coefficients for islands of size @xmath87 , @xmath88 and @xmath89 are negligible .
to assess the validity of our models , we have performed mc simulations of island growth for a range of deposition rates using the eam as well as model i and model ii barriers @xcite .
we used a continuous time mc technique @xcite in which moves are selected randomly from the list of all possible moves at the given time with the appropriate weights .
the time is advanced after each move according to the inverse of the sum of all rates .
the existence of four peaks in the spectrum of energy barriers indicates that there are four typical time scales of hopping .
the two lowest peaks include very fast moves towards and along island edges and motion of vacancies .
the single atom move is in the third peak while moves in which atoms detach from islands are in the highest peak . from statistics collected during the simulations we find that most of the computer time is consumed by moves @xmath906 ( 96 ) and 7 ( 112 ) .
the first move occurs in trimers while both moves occur for atoms hopping along straight island edges .
the reason that these moves consume so much time is that the reverse move typically has the same low barrier so the atom can continuously hop back and forth for a long time . to make the simulation feasible
, we had to somewhat suppress these moves by artificially raising the barriers in the first and second peaks .
this was done for all barriers lower than 0.4 ev according to : @xmath91 where @xmath92 . since the moves associated with these barriers are still orders of magnitude faster than for the two higher peaks we expect that this modification will have only small effect on the island morphology .
we tried various values of @xmath93 and found that up to @xmath94 this is indeed the case . in fig .
3 we show the island morphologies obtained in mc simulations using the energy barriers from eam [ fig .
3(a ) ] , model i [ fig .
3(b ) ] and model ii [ fig .
these simulations were done on a @xmath95 lattice at @xmath79k and deposition rate @xmath96 ml / s .
the snapshots presented here are for a coverage @xmath97 ml .
these snapshots indicate that models i and ii provide a good description of the model with the full eam barriers as far as the island morphology is concerned .
the islands are rather compact , dominated by overall square and rectangular shapes with a small number of kinks . the island density vs. coverage is presented in fig . 4 for the eam ( @xmath98 ) , model i ( o ) and model ii ( + ) energy barriers
in all cases the island density quickly increases at low coverage , then saturates and remains nearly constant thereafter .
it starts to decrease at higher coverage , when coalescence sets in .
for model i the island density is slightly higher than for the eam barriers while for model ii it is about @xmath99 higher .
these differences are not intrinsic to the models .
they can be traced to the fact that the single atom hopping energy barriers obtained from the optimization procedure that we use are @xmath64 ev for model i and @xmath74 ev for model ii ; these values are higher than the value of @xmath100 ev from the eam calculations .
these discrepancies could be fixed by an overall rescaling of the barriers in each model so that the single atom barrier would be exactly equal to the eam value . since there is some arbitrariness in this procedure we chose not to apply it in the simulations presented here .
the important result of this work is that our models reproduce the essential features of diffusion , island growth and scaling .
the overall factor in island density to make the models agree with the eam calculations could be obtained by rescaling .
while model i reproduces the spectrum of eam energy barriers more accurately , the importance of model ii is that it captures the essential features of adatom diffusion on the cu(001 ) in spite of its simplicity .
snapshots of the island morphology using the eam barriers for three deposition rates , @xmath101 , @xmath102 and @xmath103 ml / s are presented in fig .
5(a)-(c ) , respectively .
the simulations were done on a @xmath95 lattice at @xmath79k and the snapshots were taken at @xmath97 ml . it is observed that decreasing the deposition rate gives rise to fewer and larger islands . in fig .
6 we present the island density vs. coverage for the eam barriers , at deposition rates ( from top to bottom ) @xmath104 and @xmath105 ml / s .
it is observed that in addition to the decrease in island density as the deposition rate decreases , the plateau becomes broader and flatter .
the island density vs. deposition rate is presented in fig . 7 for the eam ( @xmath98 ) ,
model i ( o ) and model ii ( + ) energy barriers for @xmath79k and @xmath97 ml .
this coverage is well within the quasi - steady state regime in which the island density is nearly constant ( see fig . 6 ) . in this regime scaling arguments and rate equations indicate that in the asymptotic limit of low deposition rate the island density exhibits power law behavior of the form @xmath106 where @xmath107 is the adatom diffusion coefficient ( the hopping rate for an isolated atom is @xmath108 due to the four possible directions for hopping ) .
the exponent @xmath109 depends on microscopic properties of the system such as stability @xcite and mobility @xcite of small islands , anisotropic diffusion @xcite and magic islands ( islands which are stable while larger islands are unstable ) @xcite .
it was found that for systems in which the smallest stable island is of size @xmath110 ( where islands of size @xmath111 dissociate ) , @xmath112 . in this case
@xmath4 is the critical island size @xcite
. under the conditions studied here , detachment of atoms from islands is negligible , while mobility of small islands is a significant effect .
therefore , we define here the critical island size @xmath4 as the size for which all islands of size @xmath113 are mobile while islands of size @xmath114 are immobile . in this case
the asymptotic value of @xmath109 in the limit where @xmath115 is given by @xcite : @xmath116 this result is exact if all mobile islands have the same diffusion coefficient , namely @xmath117 , @xmath118 .
still , it is a good approximation if all the diffusion coefficients are of the same order of magnitude .
table ii shows that the diffusion coefficients of pentamers and heptamers are practically negligible . for eam
the diffusion coefficient of the trimer is also rather small and we expect the system to be described by @xmath119 .
models i and ii exhibit small but not negligible mobility of trimers and therefore we expect them to be described by either @xmath119 or @xmath120 . in fig .
7 we observe nearly two and a half decades of scaling behavior , in a range of deposition rates between @xmath101 and @xmath105 ml / s .
the best fit through the six data points provides @xmath121 for the eam , @xmath6 for model i and @xmath7 for model ii barriers .
these results are significantly lower than the asymptotic values for @xmath119 ( @xmath122 ) or @xmath120 ( @xmath123 ) . in ref .
@xcite a similar deviation was found between simulation results and the asymptotic predictions of the rate equations .
it was argued that this deviation is due to the slow convergence of @xmath109 to its asymptotic value , when the deposition rate @xmath1 is lowered .
this conclusion is strengthened by the fact that if only the four left - most data points are used in the linear fit , the resulting @xmath109 equals @xmath124 for the eam , @xmath125 for model i and @xmath126 for model ii .
these results are somewhat closer to the asymptotic value . moreover , within the error bars they are larger than the asymptotic value that arises from adatom mobility alone , namely @xmath127 .
therefore , the possibility that @xmath109 converges to @xmath128 should be discarded .
since the conditions used here are experimentally relevant , our results indicate that one should be careful in drawing conclusions about processes at the atomic scale from scaling results ( see section iv d ) .
the scaling properties of the island size distribution have been studied both experimentally @xcite and theoretically @xcite .
these studies indicated that the island size distribution depends on the stability and mobility of small islands and is modified in the case of magic island sizes @xcite .
the scaled island size distributions are presented for the three models in fig .
for all three models the deposition rates are @xmath129 ( @xmath98 ) , @xmath130 ( o ) and @xmath131 ml / s ( + ) .
these results are based on statistics collected from 20 runs on a @xmath95 size lattice at @xmath79k and @xmath97 ml .
the general shape of the distributions seems to resemble previous results @xcite for @xmath132 where the peaks rise more slowly on the left hand size ( small @xmath133 ) due to the depletion of the mobile islands .
this trend is qualitatively similar to the results for the case where small islands are unstable , as shown in ref .
note also that the peak height increases considerably as @xmath1 decreases
. this may be due to coalescence which is found to become more pronounced as the deposition rate decreases .
coalescence causes @xmath134 to increase , pushing up the scaled island size distribution which includes the factor @xmath135 .
for both the eam barriers and models i and ii we obtain significant dimer mobility .
the trimer mobility for the eam barriers is rather small , and somewhat larger in models i and ii . in general
, we find that the energy barriers relevant for dimer and trimer mobility are in the same peak as the single atom hopping .
the differences in hopping rates between monomers , dimers and trimers are due to very small differences in the energy barriers , which at @xmath79k are significant .
these differences in hopping rates decrease as the temperature increases . in model ii , which has only one activation barrier for each peak , the activation barriers for monomer , dimer and trimer mobilities are all equal .
combinatorial summation of paths then shows that for model ii the diffusion coefficient of dimers ( trimers ) is equal to one half ( one quarter ) of the monomer diffusion coefficient , at all temperatures .
therefore , in model ii , at any temperature in which atoms are mobile , dimers and trimers are also mobile .
the mobility of small islands has a significant effect on the asymptotic scaling of the island density vs. deposition rate . in principle , this can be used to extract these diffusion properties at the atomic scale from stm and diffraction results at larger scales .
however , our simulations indicate that for experimentally relevant parameters the system is likely to be away from the asymptotic regime , and therefore one should be careful in drawing conclusions from the empirically determined scaling exponents .
the relation between mobility of small islands and edge mobility was examined in ref .
it was shown that both types of mobilities are generated by the same hopping moves and are therefore related , namely , small island mobility implies edge mobility .
we have found that for cu(001 ) atom mobility along straight island edges ( for which the barriers belong to peak ii ) is much higher than the single atom mobility ( with a barrier in peak iii ) . this indicates , using an argument from ref .
@xcite , that islands should form compact shapes for a broad range of temperatures .
this conclusion is confirmed by our mc simulations and is in agreement with stm results @xcite .
for the temperature studied here detachment of atoms from islands is negligible and therefore edge mobility is the dominant process which shapes the islands .
it is found that for cu(001 ) the mobility of a single vacancy is higher than the single adatom mobility .
we observe that the barrier for the diffusion of a single vacancy is sensitive to the environment beyond the @xmath36 square used to calculate energies in this paper .
the energy barrier for diffusion of a single vacancy is @xmath136 ev if all sites around the @xmath36 square are vacant .
adding three more atoms on the left it increases it to @xmath137 ev .
embedding the hopping vacancy in an occupied @xmath138 square increases the barrier to @xmath139 ev while the barrier for a single vacancy in the substrate is found to be @xmath140 ev ( using a slab of 20 layers , 121 atoms / layer ) .
we conclude that , although the vacancy mobility is higher than the adatom mobility , at high coverage the difference may not be dramatic @xcite .
for example , at low temperature a difference of @xmath141 ev in the barrier may significantly affect the hopping rate .
a more general conclusion is that models i and ii best describe diffusion and nucleation at relatively low coverage . at high coverage , far beyond the percolation threshold hopping can be considered primarily as motion of vacancies .
when the adlayer is already dense , relaxation effects during the moves are more important than at low coverage , and the configuration of occupied sites beyond the @xmath36 square used here may affect the energy barriers . in spite of the prediction that a single vacancy is more mobile than an adatom ,
note that for ag(001 ) there is evidence that vacancy clusters have lower diffusivity than islands @xcite .
the simulations presented here were carried out with parameters typical to homoepitaxial growth experiments on cu(001 ) surfaces , namely , linear lattice size of 250 sites ( about 650 ) , @xmath142k and @xmath1 between @xmath143 and @xmath144 ml / s @xcite . in ref .
@xmath145 was measured for @xmath146k , @xmath1 between @xmath147 and @xmath131 ml / s , and coverage @xmath148 .
this value of @xmath109 is in a very good agreement with our simulation results .
the authors of ref .
@xcite claimed that this indicates that in this system and parameter range , @xmath149 .
our results indicate that in this parameter range the system is away from the asymptotic regime .
thus , the value @xmath145 can be obtained eventhough @xmath150 or @xmath151 .
a value of @xmath152 was reported ( for @xmath146k , and @xmath1 between @xmath131 and @xmath153 ) and @xmath154 ( for @xmath155k , and @xmath1 between @xmath147 and @xmath131 ) in the same coverage , @xmath148 @xcite .
these values are close to those reported in ref .
@xcite , namely @xmath156 ( for @xmath157k , and @xmath1 between @xmath158 and @xmath159 ) , and @xmath160 ( for @xmath161k , and @xmath1 between @xmath143 and @xmath130 ) .
a possible interpretation of the higher values of @xmath109 observed in these experiments , is that processes such as coalescence become significant as the coverage increases .
we expect our approach to apply also to other fcc(001 ) metal surfaces such as ni(001 ) and ag(001 ) in which diffusion occurs through hopping rather than by the exchange mechanism .
eam calculations for ni(001 ) and ag(001 ) are consistent with our models @xcite . for surfaces in which the exchange mechanism is favorable , as it is believed to be the case for al @xcite , pt @xcite , pd and au @xcite modifications of the models are required .
however , we believe that the approach presented here should still provide a simple and useful model for diffusion on these surfaces .
we have studied adatom self diffusion and island growth on cu(001 ) using mc simulations at the atomic scale . as input to the mc simulation we used a complete set of energy barriers obtained from the embedded atom method . to reduce the number of parameters and to obtain better understanding of the diffusion processes
we have constructed two simple models .
model i , with three parameters , provides a good quantitative description of the full landscape of hopping energy barriers .
model ii , with only two parameters , is a minimal model which still incorporates the essential features of adatom diffusion on the cu(001 ) surface .
we examined the diffusion properties of small islands and found that the mobility of dimers is comparable to the single adatom mobility while trimers are also somewhat mobile .
the mobility of adatoms along island edges was found to be much higher than the single adatom mobility .
since atoms detachment from islands is negligible , edge mobility is the dominant process that shapes islands into the compact forms .
a further conclusion from the eam calculations is that vacancy mobility , which is dominant at very high coverages , is higher than the single adatom mobility , which is dominant at low and intermediate coverages .
mc simulations of island growth show similar morphologies for the eam barriers and models i and ii . in all cases
the islands generally form square or rectangular shapes with a small number of kinks .
studies of the scaling of the island density @xmath0 vs. deposition rate @xmath1 show that @xmath2 where @xmath121 for the eam , @xmath6 for model i and @xmath7 for model ii barriers .
these results are lower than the asymptotic value of @xmath8 anticipated for systems that exhibit dimer ( and even trimer ) mobility .
it indicates that for the experimentally relevant conditions studied here the system has not reached the asymptotic regime . this result call for caution in using scaling results to draw conclusions about diffusion properties at the atomic scale . since both models
i and ii have only few parameters , one needs a very small set of calculated energy barriers to determine these parameters for a given fcc(001 ) substrate , and a few more barriers to verify that the model applies to that substrate
. this may open the way for a more effective use of first principle calculations of energy barriers on the surface as an input to the kinetics calculations .
we believe that the procedure described in this paper would prove useful much beyond the fcc(001 ) metal substrates .
we expect it to be applicable with some modifications to take into account exchange moves which are believed to be favorable in al , pt , pd and au .
we believe that the approach presented here will provide useful models for adatom self diffusion on fcc(111 ) surfaces , for which reliable calculations of energy barriers are becoming available @xcite , and also to a variety of heteroepitaxial systems .
ob would like to thank oded millo and nira shimoni for helpful discussions .
we would like to acknowledge partial support from the nsf under grants dmr-9217284 and dmr-9119735 at syracuse university during the initial stages of this work and partial support from intel - israel college relations committee during the final stages .
w. li , g. vidali and o. biham , _ phys .
b _ * 48 * ( 1993 ) 8336 ; g. vidali , o. biham , h. zeng , j .- s .
lin and w. li , in : the structure of surfaces iv , eds .
s.y . tong and x. xide ( world scientific , syngapore , 1995 ) .
m. schroeder and d.e .
wolf , _ phys .
lett . _ * 74 * ( 1995 ) 2062 ; d. e. wolf in _
scale invariance , interfaces , and non - equilibrium dynamics _ , edited by m. droz , a. j. mckane , j. vannimenus and d. e. wolf , vol .
344 of nato advanced study institute , series b : physics , ( plenum , new york , 1994 ) . in experiments , no significant nucleation was observed on top of islands up to nearly full coverage @xcite .
therefore , in our simulations an atom deposited on top of an island hops randomly until it jumps down to the lower layer . to demonstrate that our models are applicable to ni(001 ) and ag(001 ) we present here the results of eam calculations for the six configurations in the first row of table i. for ni(001 ) these barriers are @xmath162 , @xmath163 , @xmath164 , @xmath165 , @xmath166 and @xmath167 ev for @xmath168 and @xmath33 respectively . for ag(001 )
these barriers are @xmath169 , @xmath163 , @xmath170 , @xmath171 , @xmath172 and @xmath173 ev for @xmath168 and @xmath33 respectively .
we observe that in both cases barriers which belong to different peaks are typically separated by @xmath174ev or more . | a model that describes self diffusion , island nucleation and film growth on fcc(001 ) metal substrates is presented .
the parameters of the model are optimized to describe cu diffusion on cu(001 ) , by comparing activation energy barriers to a full set of barriers obtained from semi - empirical potentials via the embedded atom method .
it is found that this model ( model i ) , with only three parameters , provides a very good description of the full landscape of hopping energy barriers .
these energy barriers are grouped in four main peaks . a reduced model ( model ii ) with only two parameters , is also presented , in which each peak is collapsed into a single energy value . from the results of our simulations , we find that this model still maintains the essential features of diffusion and growth on this model surface .
we find that hopping rates along island edges are much higher than for isolated atoms ( giving rise to compact island shapes ) and that vacancy mobility is higher than adatom mobility .
we observe substantial dimer mobility ( comparable to the single atom mobility ) as well as some mobility of trimers .
mobility of small islands affects the scaling of island density @xmath0 vs. deposition rate @xmath1 , @xmath2 , as well as the island size distribution . in the asymptotic limit of slow deposition ,
scaling arguments and rate equations show that @xmath3 where @xmath4 is the size of the largest mobile island .
our monte carlo results , obtained for a range of experimentally relevant conditions , show @xmath5 for the eam , @xmath6 for model i and @xmath7 for model ii barriers .
these results are lower than the anticipated @xmath8 due to dimer ( and trimer ) mobility . |
the following assumptions and conditions are necessary for a light curve to yield a unique solution of planet and star parameters : + @xmath1 the planet orbit is circular ( valid for tidally - circularized extrasolar planets ) ; + @xmath1 @xmath2 @xmath3 @xmath4 and the companion is dark compared to the central star ; + @xmath1 the stellar mass - radius relation is known ; + @xmath1 the light comes from a single star , rather than from 2 or more blended stars ; + @xmath1 the eclipses have flat bottoms .
this implies that the companion is fully superimposed on the central star s disk and requires that the data is in a band pass where limb darkening is negligible ; + @xmath1 the period can be derived from the light curve ( e.g. , the two observed eclipses are consecutive ) .
+ in this article @xmath5 is mass , @xmath6 is radius , @xmath7 is density , @xmath8 is period , @xmath9 is orbital semi - major axis , @xmath10 is the orbital inclination , and @xmath11 is the gravitational constant . where required
the subscript @xmath12 is for planet , @xmath13 for stellar , and @xmath14 for solar .
five equations are used to uniquely solve for @xmath4 , @xmath15 , @xmath9 , @xmath10 , and @xmath16 .
the simplified equations presented below require the additional assumption that @xmath17 .
transit depth @xmath18 total transit duration @xmath19 transit shape ( @xmath20 flat part of transit and @xmath21 total transit duration ) @xmath22 kepler s third law @xmath23 stellar mass - radius relation @xmath24 here @xmath25 is a constant coefficient for each stellar sequence ( main sequence , giants , etc . ) and @xmath26 describes the power law of the sequence ( e.g. , @xmath27 and @xmath28 for f k main sequence stars ( cox 2000 ) ) . note that kepler s third law and the stellar mass - radius relation set a physical scale to two disks passing in front of each other .
this breaks the geometrical degeneracy and allows a unique solution .
the five parameters @xmath4 , @xmath15 , @xmath9 , @xmath10 , and @xmath16 can be solved for uniquely from the above five equations . moreover
, the impact parameter @xmath29 and stellar density @xmath30 can be solved for uniquely without the stellar mass - radius relation .
@xmath31^{1/2}.\ ] ] @xmath32 @xmath33^{\frac{1}{1 - 3x}}.\ ] ] @xmath34^{\frac{x}{(1 - 3x)}}.\ ] ] @xmath35^{1/3}.\ ] ] @xmath36 @xmath37^{\frac{x}{(1 - 3x ) } } \sqdepth.\ ] ]
the above analytical solution has many applications , all related to selecting the best transit candidates for radial velocity mass follow - up . here
we only have room to describe one application ; for others see seager & malln - ornelas ( 2002 ) .
the stellar density @xmath30 can be uniquely determined from the light curve alone without using the stellar mass - radius relation , as seen from equation ( 7 ) .
a measured @xmath30 can be used in three ways .
( 1 ) from the light curve alone a main sequence star and a giant star can be distinguished because main sequence stars occupy a unique position in a @xmath30 vs. spectral type diagram ( figure 1 ) .
hence a giant star with an eclipsing stellar companion can be ruled out .
( 2 ) from the light curve and the stellar mass - radius relation @xmath16 can be estimated ( equation ( 12 ) ) .
even for slightly evolved stars an upper limit on @xmath15 and hence @xmath16 can be derived .
( 3 ) a common false positive planet transit can be ruled out by comparing @xmath30 derived from the light curve with @xmath30 derived from a spectral type .
if the two @xmath30 differ then something is amiss with the assumptions in 1 .
the common case is the situation where a binary star system has its eclipse depth reduced to a planet - size eclipse due to the light from a third , contaminating , star ( figure 2 ) .
the contaminating star may be a chance alignment of a foreground or background star , or a third star as part of a triple star system . for a real example of this `` blended star '' situation , see malln - ornelas et al . | a unique analytical solution of planet and star parameters can be derived from an extrasolar planet transit light curve under a number of assumptions .
this analytical solution can be used to choose the best planet transit candidates for radial velocity follow - up measurements , with or without a known spectral type . in practice , high photometric precision ( @xmath0 0.005 mag ) and high time sampling ( @xmath0 5 minutes ) are needed for this method .
see seager & malln - ornelas ( 2002 ) for full details .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in # 1to 0pt#1 |
let @xmath0 be a connected semisimple complex algebraic group , and let @xmath2 be the fixed point subgroup of an algebraic involution @xmath3 of @xmath0 .
then @xmath2 is a reductive group , which is connected if @xmath0 is simply - connected .
the lie algebra @xmath4 of @xmath0 splits into the sum of eigenspaces of @xmath3 , @xmath5 where the lie algebra @xmath6 of @xmath2 is the eigenspace of eigenvalue 1 , and @xmath7 is the eigenspace of eigenvalue @xmath8 .
the adjoint action of @xmath0 on @xmath4 , once restricted to @xmath2 , leaves @xmath6 and @xmath7 stable .
therefore @xmath7 provides an interesting representation of @xmath2 , called the isotropy representation , where one may want to study the geometry of the @xmath2-orbits . with this aim
, one looks at the so - called nilpotent cone @xmath9 , which consists of the elements whose @xmath2-orbit closure contains the origin . in this case
, @xmath10 actually consists of the nilpotent elements of @xmath4 which belong to @xmath7 . by a fundamental result of kostant and rallis @xcite , as in the case of the adjoint action of @xmath0 on @xmath4 , there are finitely many nilpotent @xmath2-orbits in @xmath7 .
provided @xmath2 is connected , we restrict our attention to the spherical nilpotent @xmath2-orbits in @xmath7 . here
spherical means with an open orbit for a borel subgroup of @xmath2 , or equivalently with a ring of regular functions which affords a multiplicity - free representation of @xmath2 .
the classification of these orbits is known and due to king @xcite . in the present paper , we begin a systematic study of the closures of the spherical nilpotent @xmath2-orbits in @xmath7 . in particular , we analyze their normality , and describe the @xmath2-module structure of the coordinate rings of their normalizations .
this is done by making use of the technical machinery of spherical varieties , which is recalled in section 1 . here
we will deal with the case where @xmath1 is a classical symmetric pair with @xmath2 semisimple , the other cases will be treated in forthcoming papers .
the semisimplicity of @xmath2 is equivalent to the fact that @xmath7 is a simple @xmath2-module , in which case @xmath11 is also called a symmetric space of non - hermitian type .
let @xmath12 be a real form of @xmath0 with lie algebra @xmath13 and cartan decomposition @xmath14 , so that @xmath3 is induced by the corresponding cartan involution of @xmath12 .
then @xmath2 is the complexification of a maximal compact subgroup @xmath15 , and the kostant - sekiguchi - dokovi correspondence @xcite establishes a bijection between the set of the nilpotent @xmath12-orbits in @xmath16 and the set of the nilpotent @xmath2-orbits in @xmath7 .
let us briefly recall how it works , more details and references can be found in @xcite .
every non - zero nilpotent element @xmath17 lies in an @xmath18-triple @xmath19 .
every @xmath18-triple @xmath19 is conjugate to a _ cayley triple _ @xmath20 , that is , an @xmath18-triple with @xmath21 , @xmath22 and @xmath23 . to a cayley triple in @xmath13 one can associate its _ cayley transform _
@xmath24 this is a normal triple in @xmath4 , that is , an @xmath18-triple @xmath25 with @xmath26 and @xmath27 . by @xcite , any non - zero nilpotent element @xmath28 lies in a normal triple @xmath29 , and any two normal triples with the same nilpositive element @xmath30 are conjugated under @xmath2 .
then the desired bijective correspondence is constructed as follows : let @xmath31 be an adjoint nilpotent orbit , choose an element @xmath32 belonging to a cayley triple @xmath33 , consider its cayley transform @xmath25 and let @xmath34 : then @xmath35 is the nilpotent @xmath2-orbit corresponding to @xmath36 . among the nice geometrical properties of the kostant - sekiguchi - dokovi correspondence , we just recall here one result concerning sphericality : the spherical nilpotent @xmath2-orbits in @xmath7 correspond to the adjoint nilpotent @xmath37-orbits in @xmath13 which are multiplicity free as hamiltonian @xmath38-spaces @xcite . in accordance with the philosophy of the orbit method ( see e.g. @xcite ) , the unitary representations of @xmath12 should be parametrized by the ( co-)adjoint orbits of @xmath12 . in particular
one is interested in the so - called unipotent representations of @xmath12 , namely those which should be attached to nilpotent orbits .
the @xmath2-module structure of the ring of regular functions on a nilpotent @xmath2-orbit in @xmath7 ( which we compute in our spherical cases ) should give information on the corresponding unitary representation of @xmath37 .
unitary representations that should be attached to the spherical nilpotent @xmath2-orbits are studied in @xcite ( when @xmath0 is a classical group ) and @xcite ( when @xmath0 is the special linear group ) .
when @xmath0 is the symplectic group , for particular spherical nilpotent @xmath2-orbits , such representations are constructed in @xcite and @xcite .
the normality and the @xmath2-module structure of the coordinate ring of the closure of a spherical nilpotent @xmath2-orbit in @xmath7 have been studied in several particular cases , with different methods , by nishiyama @xcite , @xcite , by nishiyama , ochiai and zhu @xcite , and by binegar @xcite . in appendix
[ a ] we report the list of the spherical nilpotent @xmath2-orbits in @xmath7 for all symmetric pairs @xmath39 of classical non - hermitian type . in the classical cases , the adjoint nilpotent orbits in real simple
algebras are classified in terms of signed partitions , as explained in ( * ? ? ?
* chapter 9 ) . in the list ,
every orbit is labelled with its corresponding signed partition . for every orbit we provide an explicit description of a representative @xmath28 , as element of a normal triple @xmath33 , and the centralizer of @xmath30 , which we denote by @xmath40 .
all these data can be directly computed using king s paper on the classification of the spherical nilpotent @xmath2-orbits @xcite ( but we point out a missing case therein , see remark [ remark : missingcase ] ) .
the first datum which is somewhat new in this work is the luna spherical system associated with @xmath41 , the normalizer of @xmath40 in @xmath2 , which is a wonderful subgroup of @xmath2 .
it is equal to @xmath42}$ ] , the stabilizer of the line through @xmath30 , and notice that @xmath42}/k_e\cong\mathbb c^\times$ ] .
the luna spherical systems are used to deduce the normality or non - normality of the @xmath2-orbits , and to compute the corresponding @xmath2-modules of regular functions .
appendix [ b ] consists of two sets of tables , where we summarize our results on the spherical nilpotent @xmath2-orbits in @xmath7 . given such an orbit @xmath43 , in the first set ( tables 211 ) we describe the normality of its closure @xmath44 , and
if @xmath45 denotes the normalization , we describe the @xmath2-module structure of @xmath46 $ ] by giving a set of generators of its weight semigroup @xmath47 ( that is , the set of the highest weights occurring in @xmath46 $ ] ) .
the second set ( tables 1220 ) contains the luna spherical systems of @xmath41 . in section 1
we compute the luna spherical systems . in section 2
we study the multiplication of sections of globally generated line bundles on the corresponding wonderful varieties , which turns out to be always surjective in all cases except one . in section 3 we deduce our results on normality and semigroups .
simple roots of irreducible root systems are denoted by @xmath48 and enumerated as in bourbaki , when belonging to different irreducible components they are denoted by @xmath48 , @xmath49 , @xmath50 , and so on . for the fundamental weights we adopt the same convention ,
they are denoted by @xmath51 , @xmath52 , @xmath53 , and so on . in the tables for the orthogonal cases at the end of the paper
we use a variation of the fundamental weights @xmath54 which is explained in appendix [ b ] . by @xmath55
we denote the simple module of highest weight @xmath56 , the acting group will be clear from the context .
in this section we compute the luna spherical systems given in the tables at the end of the paper , in appendix [ b ] .
first , let us briefly explain what a luna spherical system is , see e.g. @xcite for a plain introduction .
recall that a subgroup @xmath57 of @xmath2 is called wonderful if the homogeneous space @xmath58 admits an open equivariant embedding in a wonderful @xmath2-variety .
a @xmath2-variety is called wonderful if it is smooth , complete , with an open @xmath2-orbit whose complement is union of @xmath59 smooth prime @xmath2-stable divisors with non - empty transversal crossings such that two points @xmath60 lie in the same @xmath2-orbit if and only if @xmath61 the wonderful embedding of @xmath58 is unique up to equivariant isomorphism , and is a projective spherical @xmath2-variety .
the number @xmath62 of the prime @xmath2-stable divisors is called the rank of @xmath63 .
let us fix , inside @xmath2 , a maximal torus @xmath64 and a borel subgroup @xmath65 containing @xmath64 .
this choice yields a root system @xmath66 and a set of simple roots @xmath67 in @xmath66 .
let us also denote by @xmath68 the scalar product in the euclidean space spanned by @xmath66 , by @xmath69 the coroot associated with @xmath70 , and by @xmath71 the usual cartan pairing @xmath72 for any spherical @xmath2-variety @xmath63 , the set of colors , which is denoted by @xmath73 , is the set of prime @xmath65-stable non-@xmath2-stable divisors of @xmath63 .
it is a finite set . in our case ,
if @xmath63 is the wonderful embedding of @xmath58 , the colors of @xmath58 are just the irreducible components of the complement of the open @xmath65-orbit , and the colors of @xmath63 are just the closures of the colors of @xmath58 , so that the two sets @xmath73 and @xmath74 are naturally identified . for
any spherical @xmath2-variety @xmath63 one can also define another finite set , the set of spherical roots , usually denoted by @xmath75 . here
we recall its definition only in the wonderful case .
suppose @xmath63 is the wonderful embedding of @xmath58 . by definition @xmath63
contains a unique closed @xmath2-orbit , therefore every borel subgroup of @xmath2 fixes in @xmath63 a unique point .
let us call @xmath76 the point fixed by @xmath77 , the opposite of the borel subgroup @xmath65 .
for all @xmath2-stable prime divisors @xmath78 , let @xmath79 be @xmath64-eigenvalue occurring in the normal space of @xmath78 at @xmath76 @xmath80 then the set of spherical roots is the set @xmath81 , also denoted by @xmath82 .
the spherical roots are linearly independent and the corresponding reflections @xmath83 generate a finite group of orthogonal transformations which is called the little weyl group of @xmath63 .
in our case , in which the center of @xmath2 acts trivially , the spherical roots are elements of @xmath84 , that is , linear combinations with non - negative integer coefficients of simple roots .
the picard group of a wonderful variety @xmath63 is freely generated by the equivalence classes of the colors of @xmath63 . expressing the classes of the @xmath2-stable divisors in terms of the basis given by the classes of colors @xmath85=\sum_{d\in\delta_{k / h}}c_{k / h}(d,\sigma_i)[d]\ ] ]
we get a @xmath86-bilinear pairing , which is also called cartan pairing , @xmath87 it is known to satisfy quite strong restrictions , as follows . for any simple root @xmath88 , the set of colors moved by @xmath70 , which is denoted by @xmath89 , is the set of colors that are not stable under the action of the minimal parabolic subgroup @xmath90 .
any simple root @xmath70 moves at most two colors , and more precisely there are exactly four cases : * @xmath70 moves no colors ; * @xmath70 moves two colors , this happens if and only if @xmath91 , and in this case we have 1 .
@xmath92 , 2 .
@xmath93 for all @xmath94 and @xmath95 , 3 .
@xmath96 for all @xmath95 ; * @xmath70 moves one color and @xmath97 , in this case if @xmath94 we have @xmath98 for all @xmath95 ; * @xmath70 moves one color and @xmath99 , in this case if @xmath94 we have @xmath100 for all @xmath95 . the set of simple roots moving no colors is denoted by @xmath101 .
the set of colors @xmath74 is a disjoint union of subsets @xmath102 , @xmath103 , @xmath104 which consist of colors moved by simple roots of type ( a ) , ( 2a ) , ( b ) , respectively .
the set @xmath102 is also denoted by @xmath105 . * a color in @xmath105 may be moved by several simple roots . * a color in @xmath103
is moved by a unique simple root . * a color in @xmath104
may be moved by at most two simple roots , in this case two simple roots @xmath70 and @xmath106 move the same color if and only if @xmath70 and @xmath106 are orthogonal and @xmath107 .
notice that the full cartan pairing @xmath108 is determined by its restriction to @xmath109 .
if @xmath57 is a wonderful subgroup of @xmath2 , the triple @xmath110 , endowed with the map @xmath111 , is called the spherical system of @xmath57 . in appendix [ b ] , we present the spherical systems of the wonderful subgroups @xmath112 of @xmath2 by providing the sets of spherical roots @xmath82 and the luna diagrams .
the luna diagram of a spherical system consists of the dynkin diagram of @xmath2 decorated with some extra symbols from which one can read off all the data of the spherical system .
let us briefly explain how it works , here we only explain how to read off the missing data ( the set @xmath101 and the map @xmath113 ) , see e.g. @xcite for a complete description .
every circle ( shadowed or not ) represents a color .
circles corresponding to the same color are joined by a line .
the colors moved by a simple root are close to the corresponding vertex of the dynkin diagram : * no circle is placed in correspondence of the vertex , * two circles are placed one above and one below the vertex , * one circle is placed below the vertex , * one circle is placed around the vertex .
therefore , the set @xmath114 is given by the vertices with no circles .
it is worth saying that in general @xmath114 is included in @xmath115 . to read off the map @xmath116 one has to know that an arrow ( it looks more like a pointer but it has a source and a target ) starting from a circle @xmath117 above a vertex @xmath70 and pointing towards a spherical root @xmath118 non - orthogonal to @xmath70 means that @xmath119 .
vice versa , the luna diagram is organized in order that the colors @xmath117 corresponding to circles that lie above the vertices have @xmath120 for all @xmath121 , so if the there is no arrow starting from a circle @xmath117 above a vertex @xmath70 and pointing towards a spherical root @xmath118 non - orthogonal to @xmath70 ( with @xmath122 ) this means that @xmath123 .
these together with the properties of the cartan pairing for colors of type ( a ) , explained above , allows to recover the map @xmath116 .
the two colors moved by @xmath124 will be denoted by @xmath125 and @xmath126 , the former refers to the circle placed above the vertex while the latter refers to the circle placed below .
the color moved by a simple root @xmath127 will be denoted by @xmath128 . as an example
we show in detail how to recover the map @xmath116 for the first case of the list where a non - empty set @xmath105 occurs , the case [ sss : ccacy ] with @xmath129 .
the group @xmath2 is of type @xmath130 , with @xmath131 and @xmath132 greater than 2 , the set of spherical roots is @xmath133 and the luna diagram is as follows .
@xmath134 here the set @xmath114 is given by the simple roots @xmath135 for all @xmath136 and @xmath137 for all @xmath138 .
the elements of @xmath139 , i.e. the colors of type ( a ) , are five : @xmath140 we know that for all colors @xmath117 of type ( a ) @xmath141 if @xmath142 and @xmath143 , and @xmath144 otherwise .
therefore , let us show how to determine @xmath145 for all @xmath121 .
first @xmath146 then @xmath147 .
since there is an arrow from @xmath148 to @xmath149 @xmath150 , furthermore @xmath151 , thus we have @xmath152 .
the other spherical roots @xmath118 are orthogonal to @xmath153 , so @xmath154 , if @xmath145 is @xmath155 then @xmath156 must be @xmath157 but this happens only if @xmath158 . therefore , @xmath159 while it is zero on the other two spherical roots @xmath160 .
the entire map @xmath116 is as follows .
@xmath161 here we briefly recall the definition and the essential properties of some combinatorial operations on spherical systems which correspond to geometric operations on wonderful varieties , see e.g. @xcite for some more details and references .
all ( irreducible ) @xmath2-subvarieties of a wonderful @xmath2-variety @xmath63 are wonderful , they are exactly the @xmath2-orbit closures of @xmath63 , and are in correspondence with the subsets of @xmath75 . if @xmath59 are the @xmath2-stable prime divisors of @xmath63 , recall that the spherical roots @xmath162 are @xmath64-eigenvalues occurring respectively in the normal spaces of @xmath78 at @xmath76 , @xmath163 . therefore , every @xmath2-subvariety @xmath164 of @xmath63 is the intersection of some @xmath2-stable prime divisors @xmath165 for some @xmath166 .
its spherical system is thus given by * @xmath167 , * @xmath168 , * @xmath169 with the map @xmath170 restricted to @xmath171 .
let @xmath172 and @xmath173 be the wonderful embeddings of @xmath174 and @xmath175 , respectively . if @xmath176 is included in @xmath177 with connected quotient @xmath178 , there exists a surjective equivariant morphism from @xmath172 to @xmath173 with connected fibers . in terms of sperical systems
this is equivalent to an operation called quotient , as follows .
a subset @xmath179 of @xmath180 is called distinguished if there exists a linear combination with positive coefficients @xmath181 such that @xmath182 for all @xmath183 .
if @xmath179 is distinguished , the monoid @xmath184 is known to be free @xcite .
therefore , we can consider the following triple , which is called the quotient of the spherical system of @xmath172 by @xmath179 : * @xmath185 , * @xmath186 , the basis of @xmath187 , * @xmath188 endowed with the map @xmath189 restricted to @xmath190 .
if @xmath172 and @xmath173 are wonderful @xmath2-varieties with a surjective equivariant morphism with connected fibers @xmath191 , then @xmath192 is distinguished and the spherical system of @xmath173 is equal to the quotient of the spherical system of @xmath172 by @xmath193 . if @xmath172 is a wonderful @xmath2-variety , every distinguished subset @xmath179 of @xmath180 corresponds in this way to a surjective equivariant morphism with connected fibers onto a wonderful variety whose spherical system is equal to the quotient of the spherical system of @xmath172 by @xmath179 .
let @xmath194 be a parabolic subgroup of @xmath2 , with levi decomposition @xmath195 .
a wonderful @xmath2-variety @xmath63 is said to be obtained by parabolic induction from the wonderful @xmath196-variety @xmath197 if @xmath198 where @xmath199 acts trivially on @xmath197 .
further , since @xmath197 is a wonderful @xmath196-variety , the radical of @xmath196 acts trivially on @xmath197 , as well . clearly ,
if the wonderful @xmath2-variety @xmath63 is obtained by parabolic induction from the wonderful embedding of @xmath200 , then @xmath63 is the wonderful embedding of @xmath201 . in terms of spherical systems
this corresponds to the following situation .
assume that @xmath194 contains @xmath77 and @xmath196 contains @xmath64 , denote by @xmath202 the subset of @xmath67 generating the root subsystem of @xmath196 .
the wonderful @xmath2-variety @xmath63 is obtained by parabolic induction from a wonderful @xmath196-variety @xmath197 if and only if @xmath203 in this case , the spherical system of @xmath197 , after the above inclusion , is equal to the triple @xmath204 . in plain words ,
the spherical system of @xmath63 is obtained from the spherical system of @xmath197 by letting the extra simple roots in @xmath205 move one extra color each so that they are all of type ( b ) .
let @xmath194 be a parabolic subgroup of @xmath2 , containing @xmath77 , and let @xmath195 be its levi decomposition , with @xmath196 containing @xmath64 . denote by @xmath206 the radical of @xmath196 , and by @xmath202 the subset of @xmath67 generating the root subsystem of @xmath196 .
let @xmath63 be a wonderful @xmath2-variety .
consider the subset of @xmath63 of points fixed by @xmath206 and take its connected component which contains @xmath76 , the unique point fixed by @xmath77 .
it is a wonderful @xmath196-variety @xmath197 called @xmath196-localization of @xmath63 .
the spherical system of @xmath197 is obtained from the spherical system of @xmath63 as follows : * @xmath207 , * @xmath208 , * @xmath209 with the map @xmath210 restricted to @xmath211 . in this case
the spherical system of @xmath197 is said to be obtained from the spherical system of @xmath63 by localization in @xmath202 . here
we recall the statement of luna s theorem of the classification of wonderful varieties , @xcite . in our case
the center of @xmath2 always acts trivially , so here we assume for convenience that @xmath2 is a semisimple complex algebraic group of adjoint type .
let @xmath64 , @xmath65 and @xmath67 as above . every spherical root of any wonderful @xmath2-variety is the spherical root of a wonderful @xmath2-variety of rank 1 , and the wonderful varieties of rank 1 are well - known . in particular ,
the set @xmath212 of the spherical roots of all the wonderful @xmath2-varieties is finite and is described by the following .
every spherical root @xmath118 of any wonderful @xmath2-variety , for any semisimple complex algebraic group @xmath2 of adjoint type belongs to table [ tab : sphroots ] .
@xmath213 there is an abstract notion of luna spherical system , the following .
[ def : sphsystem ] a triple @xmath214 , where @xmath114 is a subset of @xmath67 , @xmath215 is a subset of @xmath212 without proportional elements and @xmath139 is a finite set endowed with a map @xmath216 , is called a _
spherical @xmath2-system _ if the following axioms hold .
* for all @xmath217 , @xmath218 for all @xmath121 , and @xmath141 only if @xmath142 .
* for all @xmath124 , @xmath219 has cardinality 2 and for all @xmath121 @xmath220 * for all @xmath217 there exists @xmath124 with @xmath221 . * for all @xmath88
such that @xmath222 , @xmath223 for all @xmath224 .
* for all @xmath70 and @xmath106 in @xmath67 such that @xmath70 and @xmath106 are orthogonal and @xmath225 , @xmath226 for all @xmath121 .
* for all @xmath121 , * * if @xmath227 with @xmath228 of type @xmath229 , @xmath230 * * if @xmath231 with @xmath228 of type @xmath232 , @xmath233 * * otherwise @xmath234 the following is known as luna s theorem of classification of wonderful varieties .
[ thm : luna ] the map which associates to a wonderful @xmath2-variety @xmath63 its spherical system @xmath204 is a bijection between the set of wonderful @xmath2-varieties up to equivariant isomorphism and the set of spherical @xmath2-systems . here we show that the spherical systems given in the tables of appendix [ b ] are indeed the spherical sytems associated with @xmath41 , the normalizers of the centralizers of the representatives @xmath30 given in appendix [ a ] . for all @xmath2 , every spherical system given in the tables
satisfies the axioms of definition [ def : sphsystem ] , so by theorem [ thm : luna ] it is equal to the spherical system associated with a ( uniquely determined up to conjugation ) wonderful subgroup of @xmath2 .
here we compute this wonderful subgroup for any spherical system of appendix [ b ] .
in all the spherical systems of appendix [ b ] the set @xmath235 is properly contained in @xmath67 , therefore the corresponding wonderful @xmath2-varieties @xmath63 can be obtained by parabolic induction from wonderful @xmath196-varieties @xmath197 , where @xmath196 is properly contained in @xmath2 .
we set @xmath236 .
furthermore , in general @xmath237 and @xmath238 are orthogonal , so that @xmath196 is a direct product @xmath239 , where @xmath240 and @xmath241 , with @xmath242 acting trivially on @xmath197 . in many cases
@xmath243 is non - empty .
notice that the above decomposition @xmath244 is not uniquely determined , but here the center of @xmath196 acts trivially on @xmath197 , so we do not care of which part of the center of @xmath196 is contained in the two factors @xmath245 and @xmath242 . in the following
we will compute , in all our cases , the wonderful subgroups associated with the spherical systems obtained by localization in @xmath240 . in the cases [ sss : symm1 ] ( @xmath246 ) ,
[ sss : symm2 ] ( @xmath246 ) , [ sss : dao ] ( @xmath246 ) , [ sss : symm4.1 ] ( @xmath246 ) , [ sss : trvial1 ] , [ sss : trvial2 ] , [ sss : bdaa ] ( @xmath246 ) , [ sss : symm8 ] ( @xmath246 ) and [ sss : ddaa ] ( @xmath246 ) the set @xmath215 is empty , so the spherical system obtained by localization in @xmath237 is trivial
. more explicitly , the parabolic subgroups @xmath194 of @xmath2 given in appendix [ a ] are the wonderful subgroups associated with the given spherical @xmath2-systems . in the cases [ sss : symm1 ] , [ sss : symm2 ] ,
[ sss : dao ] , [ sss : symm4.1 ] , [ sss : ccaac ] ( @xmath247 ) , [ sss : ccaac2 ] ( @xmath248 ) , [ sss : bdaa ] , [ sss : bdady ] ( @xmath249 ) , [ sss : bdbay ] ( @xmath249 ) , [ sss : symm8 ] , [ sss : bbaby ] ( @xmath249 ) , [ sss : bbbay ] ( @xmath249 ) , [ sss : ddaa ] , [ sss : ddady ] ( @xmath249 ) and [ sss : ddday ] ( @xmath249 ) the spherical system obtained by localization in @xmath237 is the spherical system of a symmetric subgroup @xmath250 of @xmath245 , where @xmath251 is the fixed point subgroup of an involution @xmath3 of @xmath245 .
the wonderful symmetric subgroups and their spherical systems are well - known , see e.g. @xcite . more precisely , * in the case [ sss : symm1 ] we get the case 6 of @xcite ; * in the cases [ sss : symm2 ] , [ sss : dao ] , [ sss : bdbay ] ( @xmath252 ) , [ sss : bbaby ] ( @xmath253 ) and [ sss : bbbay ] ( @xmath252 ) we get the case 5 of @xcite ; * in the cases [ sss : symm4.1 ] , [ sss : ccaac ] ( @xmath247 ) , [ sss : ccaac2 ] ( @xmath248 ) , [ sss : bdaa ] , [ sss : bdady ] ( @xmath254 ) , [ sss : symm8 ] , [ sss : ddaa ] , [ sss : ddady ] ( @xmath254 ) and [ sss : ddday ] ( @xmath255 ) we get the case 2 of @xcite ; * in the cases [ sss : bdbay ] ( @xmath256 ) , [ sss : bbaby ] ( @xmath257 ) and [ sss : bbbay ] ( @xmath256 ) we get the case 9 of @xcite ; * in the cases [ sss : bdady ] ( @xmath258 ) , [ sss : ddady ] ( @xmath258 ) and [ sss : ddday ] ( @xmath259 ) we get the case 15 of @xcite . in the cases [ sss : ccaac ] ( @xmath260 ) ,
[ sss : ccaac2 ] ( @xmath261 ) , [ sss : ccabx ] and [ sss : ccbax ] the spherical system obtained by localization in @xmath237 is the spherical system of a wonderful reductive ( but not symmetric ) subgroup of @xmath245 .
more precisely , * in the cases [ sss : ccaac ] ( @xmath260 ) and [ sss : ccaac2 ] ( @xmath261 ) we get the case 42 of @xcite ; * in the cases [ sss : ccabx ] and [ sss : ccbax ] we get the case 46 ( @xmath262 ) of @xcite .
notice that in all the above cases the levi subgroup @xmath196 such that @xmath236 is equal to @xmath264 , the centralizer of @xmath265 given in the list of appendix [ a ] . in the remaining cases this is no longer true , but we have the following situation . in the remaining cases , [ sss : ccacy ] , [ sss : cccay ] , [ sss : bdady ] ( @xmath266 ) , [ sss : bdbay ] ( @xmath266 ) , [ sss : bbaby ] ( @xmath266 ) , [ sss : bbbay ] ( @xmath266 ) , [ sss : ddady ] ( @xmath266 ) and [ sss : ddday ] ( @xmath266 ) , the given spherical @xmath2-system @xmath267 admits a distinguished set of colors @xmath179 such that the corresponding quotient @xmath268 is the spherical system of a wonderful @xmath2-variety which is obtained by parabolic induction from a wonderful @xmath264-variety .
indeed , @xmath269 .
such distinguished set of colors @xmath179 is minimal , that is , does not contain any proper non - empty distinguished subset . moreover , the corresponding quotient has higher defect , which means the following .
the defect of a spherical system is defined as the non - negative integer given by the difference between the number of colors and the number of spherical roots . in all our cases
, we have @xmath270 therefore , the set @xmath179 corresponds to a minimal surjective equivariant morphism with connected fibers of type @xmath263 in the sense of ( * ? ? ?
* proposition 2.3.5 ) .
in particular , the minimal quotients of higher defect have been studied in ( * ? ? ? * section 5.3 ) .
let us recall their description .
let @xmath176 be the wonderful subgroup associated with the spherical @xmath2-system @xmath267 , let @xmath179 be a distinguished subset satisfying the condition and let @xmath177 be the wonderful subgroup of @xmath2 associated with the quotient of @xmath267 by @xmath179 .
we can assume @xmath271 .
recall that the quotient @xmath178 is connected . under the condition we have that @xmath272 is properly contained in @xmath273 .
take levi decompositions @xmath274 and @xmath275 with @xmath276 , then @xmath277 is a simple @xmath278-module and @xmath278 and @xmath279 differ only by their connected center .
the defect of a spherical system is equal to the dimension of the connected center of the associated wonderful subgroup , so the codimension of @xmath278 in @xmath279 is equal to @xmath280 the quotient @xmath277 can be described as follows .
there exist @xmath281 @xmath279-submodules of @xmath282 , @xmath283 , isomorphic as @xmath278-modules but not as @xmath279-modules . denoting by @xmath284 the @xmath279-complement of @xmath285 in @xmath282 , as @xmath278-module , @xmath286 where @xmath287 is a co - simple @xmath278-submodule of @xmath285 which projects non - trivially on every summand @xmath283 . as said above , in our cases we always have @xmath288 , with @xmath289 given in the list of appendix [ a ] , @xmath290 and @xmath291 .
one can say something more about the inclusion of the @xmath283 in @xmath292 .
one has to consider the set @xmath293 , whose general definition involves the notion of external negative color ( see ( * ? ? ?
* section 2.3.5 ) and ( * ? ? ? * section 5.2 ) ) . without going into technical details , in our cases it holds @xmath294
moreover , @xmath295 , say @xmath296 .
assuming @xmath194 contains @xmath77 , we have that @xmath283 are respectively included in the simple @xmath196-submodules @xmath297 containing the root spaces of @xmath298 . in our cases
the integer @xmath281 , the cardinality of @xmath293 , is always equal to @xmath299 or @xmath300 . in the following , for all the remaining cases
, we describe the quotient of @xmath267 by @xmath179 , and @xmath279 in @xmath264 .
the knowledge of @xmath293 will be enough to uniquely determine the modules @xmath283 .
actually , the results contained in @xcite allow to reduce the computation of the wonderful subgroup associated with a spherical system to the computation of the wonderful subgroups associated with somewhat smaller spherical systems .
in particular , section 5.3 therein allows to reduce the computation of the wonderful subgroup associated with a spherical system with a quotient of higher defect to the computation of the wonderful subgroups associated with some spherical subsystems . moreover , many of the spherical systems under consideration have a tail , see section 6 therein , and these cases can also be reduced to some smaller cases .
similar general considerations could be done for the cases obtained by collapsing the tails .
we prefer to avoid as far as possible the technicalities and give a direct explicit description of our wonderful subgroups even if they are somewhat already known .
[ [ a - tail - case . ] ] * a ) tail case . * + + + + + + + + + + + + + + + localizing the spherical systems of the cases [ sss : bdbay ] ( @xmath302 ) , [ sss : bbaby ] ( @xmath303 ) and [ sss : bbbay ] ( @xmath302 ) in @xmath237 we obtain the following spherical system , which we label as @xmath304 , for a group of semisimple type @xmath305 with @xmath306 .
* @xmath307 .
* @xmath308 . * @xmath309 with @xmath310 and full cartan pairing as follows : * * @xmath311 , * * @xmath312 for @xmath313 , * * @xmath314 for @xmath315 , * * @xmath316 .
if @xmath317 the luna diagram is as follows , @xmath318 while if @xmath319 it is as follows , @xmath320 but the combinatorics is the same , so from now on we just report the diagram for @xmath319 .
consider the quotient by @xmath321 .
@xmath323 it is a spherical system obtained by parabolic induction from the direct product of case 2 and the rank one case 9 ( resp . the rank one case 4 ) if @xmath319 ( resp .
@xmath317 ) , the labels referring to @xcite .
we have @xmath324 . [
[ b - collapsed - tail . ] ] * b ) collapsed tail . * + + + + + + + + + + + + + + + + + + + + localizing the spherical systems of the cases [ sss : bdbay ] ( @xmath325 ) , [ sss : bbaby ] ( @xmath326 ) and [ sss : bbbay ] ( @xmath325 ) in @xmath237 we obtain the following spherical system , which is labeled as @xmath327 or s-6 in @xcite , for a group of semisimple type @xmath328 .
. * @xmath330 . * @xmath331 with cartan pairing as follows : * * @xmath311 , * * @xmath312 for @xmath313 , * * @xmath314 for @xmath332 , * * @xmath333 , * * @xmath334 .
the luna diagram is as follows .
@xmath335 consider the quotient by @xmath321 .
@xmath337 it is a spherical system obtained by parabolic induction from the case 2 of @xcite .
we have @xmath324 . [
[ a - tail - case.-1 ] ] * a ) tail case . * + + + + + + + + + + + + + + + localizing the spherical systems of the cases [ sss : ccacy ] ( @xmath129 ) and [ sss : cccay ] ( @xmath339 ) in @xmath237 we obtain the following spherical system , which we label as @xmath340 , for a group of semisimple type @xmath341 with @xmath342 .
. * @xmath344 . * @xmath345 with @xmath346 and full cartan pairing as follows : * * @xmath347 , * * @xmath348 , * * @xmath349 , * * @xmath350 , * * @xmath351 .
the luna diagram is as follows .
@xmath352 consider the quotient by @xmath353 .
@xmath355 it is a spherical system obtained by parabolic induction from the case 42 of @xcite , already considered in section [ sss : reductive ] .
we have @xmath356 . [
[ b - collapsed - tail.-1 ] ] * b ) collapsed tail . * + + + + + + + + + + + + + + + + + + + + localizing the spherical systems of the cases [ sss : ccacy ] ( @xmath357 ) and [ sss : cccay ] ( @xmath358 ) in @xmath237 we obtain the spherical system @xmath359 for a group of semisimple type @xmath360 , a particular case of the spherical system obtained above in section [ sss : typeb ] .
[ [ a - tail - case.-2 ] ] * a ) tail case . * + + + + + + + + + + + + + + + localizing the spherical systems of the cases [ sss : bdady ] ( @xmath362 ) , [ sss : ddady ] ( @xmath362 ) and [ sss : ddday ] ( @xmath363 ) in @xmath237 we obtain the following spherical system for a group of semisimple type @xmath364 with @xmath342 .
* @xmath307 .
* @xmath365 . * @xmath309 with @xmath310 and full cartan pairing as follows : * * @xmath311 , * * @xmath312 for @xmath313 , * * @xmath314 for @xmath315 , * * @xmath366 .
it is the case 60 of @xcite , labeled as @xmath367 .
[ [ b - collapsed - tail.-2 ] ] * b ) collapsed tail . * + + + + + + + + + + + + + + + + + + + + localizing the spherical systems of the cases [ sss : bdady ] ( @xmath368 ) , [ sss : ddady ] ( @xmath368 ) and [ sss : ddday ] ( @xmath369 ) in @xmath237 we obtain the following spherical system for a group of semisimple type @xmath370 .
* @xmath329 .
* @xmath371 . * @xmath372 with cartan pairing as follows : * * @xmath311 , * * @xmath312 for @xmath373 , * * @xmath374 , * * @xmath314 for @xmath375 , * * @xmath376 , * * @xmath377 .
it is the case 40 of @xcite , labeled as @xmath378 or s-10 in @xcite , and considered also in ( * ? ? ?
* section 5 ) as the spherical system of the comodel wonderful variety of cotype @xmath379 .
this section is devoted to prove the following result , that we need in order to study the singularities of closures of spherical nilpotent @xmath2-orbits in @xmath380 .
[ teo : projnorm ] let @xmath39 be a classical symmetric pair of non - hermitian type , let @xmath381 be a spherical nilpotent @xmath2-orbit . if @xmath382 , assume that the signed partition of @xmath36 is neither @xmath383 nor @xmath384 ( cases [ sss : ccabx ] and [ sss : ccbax ] in appendix [ a ] ) .
let @xmath63 be the wonderful @xmath2-variety associated to @xmath385 , then the multiplication of sections @xmath386 is surjective for all globally generated line bundles @xmath387 .
we point out that multiplication is not surjective if @xmath382 and @xmath385 is the spherical nilpotent orbit corresponding to the signed partitions @xmath383 or @xmath384 ) , see example [ example : counterexample ] below .
these cases will be treated separately in section [ ss : symplecticcases ] with an _ ad hoc _ argument .
let us briefly recall here some generalities about the multiplication of sections of line bundles on a wonderful variety , for more details and references see @xcite .
let @xmath63 be a wonderful @xmath2-variety with set of spherical roots @xmath388 and set of colors @xmath389 .
the classes of colors form a free basis for the picard group of @xmath63 , and for the semigroup of globally generated line bundles .
therefore the picard group of @xmath63 is identified with @xmath390 , and the semigroup of globally generated line bundles is identified with @xmath391 . given @xmath392 we will also write @xmath393 meaning @xmath394 . given @xmath395 we denote by @xmath396 the corresponding line bundle , and we fix @xmath397 a section whose associated divisor is @xmath117 . recall that every line bundle on @xmath63 has a unique @xmath2linearization .
then @xmath398 is a highest weight vector , and we denote by @xmath399 the @xmath2-submodule generated by @xmath398 . since @xmath63 is a spherical variety , @xmath400 is a multiplicity - free @xmath2-module , hence @xmath401 is uniquely determined and @xmath398 is uniquely determined up to a scalar factor . by identifying @xmath388 with the set of @xmath2-stable prime divisors of @xmath63 ,
every @xmath402 determines a line bundle @xmath403 , and the map @xmath404 is injective .
the line bundle @xmath405 is effective if and only if @xmath406 , and for all @xmath402 we fix a section @xmath407 whose associated divisor is @xmath408 . such a section is a highest weight vector of weight @xmath409 , and is uniquely determined up to a scalar factor . by identifying @xmath410 with @xmath390 , we regard @xmath411 as a sublattice of @xmath390 .
this defines a partial order @xmath412 on @xmath390 as follows : if @xmath413 , then @xmath414 if and only if @xmath415 . this allows to describe the space of global sections of @xmath416 as follows @xmath417 in particular , if @xmath418 , we have that @xmath419 is an irreducible @xmath2-module if and only if @xmath420 is minuscule in @xmath391 w.r.t .
@xmath412 or zero , that is , if @xmath421 and @xmath422 then it must be @xmath423 .
to any line bundle @xmath416 on @xmath63 , we attach two characters @xmath424 and @xmath425 as follows .
let @xmath57 be the stabilizer of a point @xmath426 in the open orbit of @xmath63 , fix a maximal torus @xmath64 and a borel subgroup @xmath65 such that @xmath427 , and let @xmath428 be the point fixed by the opposite borel of @xmath65 .
then we denote @xmath429 the character given by the action of @xmath57 over the fiber @xmath430 , and by @xmath431 the character given by the action of @xmath64 over the fiber @xmath432 .
if @xmath433 then the set of sections @xmath434 does not vanish on the closed orbit of @xmath63 , so it defines a regular map @xmath435 .
we choose a non - zero element @xmath436 in the line @xmath437 .
notice that @xmath438 is the irreducible module of highest weight @xmath439 and that @xmath440 is determined by the condition @xmath441 for all @xmath442 . for @xmath443
, the weight @xmath444 is combinatorially described as follows : if @xmath445 and @xmath446 is such that @xmath447 , then @xmath448 , otherwise @xmath449 for all @xmath446 such that @xmath447 . by making use of quotients and parabolic inductions ,
it is possible to reduce the study of the multiplication maps .
we recall such reductions from @xcite .
[ lem : projnorm quotients ] let @xmath63 be a wonderful variety with set of colors @xmath389 , let @xmath164 be a quotient of @xmath63 by a distinguished subset @xmath450 with set of colors @xmath451 and identify @xmath451 with @xmath452 .
if @xmath453 and @xmath454 and if @xmath455 and @xmath456 are the line bundles corresponding to @xmath117 regarded as an element in @xmath391 and in @xmath457 , then @xmath458 . in particular , if @xmath459 is surjective for all @xmath460 , then @xmath461 is surjective for all @xmath462 .
[ lem : projnorm parabolic induction ] let @xmath63 be a wonderful variety and suppose that @xmath63 is the parabolic induction of a wonderful variety @xmath164 .
then for all @xmath463 in @xmath410 the multiplication @xmath464 is surjective if and only if the multiplication @xmath465 is surjective .
we now explain how to reduce the study of the multiplications with respect to wonderful subvarieties .
[ lem : projnorm localizzazioni ] let @xmath63 be a wonderful variety and let @xmath466 be a wonderful subvariety . if @xmath464 is surjective for all globally generated @xmath387 , then @xmath467 is surjective for all globally generated @xmath468 .
denote by @xmath388 and @xmath389 the set of spherical roots and the set of colors of @xmath63 , and by @xmath469 and @xmath451 those of @xmath164 .
the restriction of line bundles induces a map @xmath470 , and the restriction of sections @xmath471 is surjective for all @xmath472 . given @xmath392 , the surjectivity of @xmath473 follows then from the surjectivity of @xmath393 .
set @xmath474 notice that every @xmath475 is minuscule w.r.t .
@xmath476 or zero , namely @xmath477 for all @xmath478 .
indeed if @xmath475 and @xmath479 for some @xmath406 , then it follows that @xmath480 , hence both @xmath408 and @xmath481 define effective divisors on @xmath164 . on the other hand
the cone of effective divisors of @xmath164 contains no line since @xmath164 is complete , therefore it must be @xmath482 .
let @xmath443 , reasoning as in @xcite by the combinatorial description of @xmath483 it follows that for all @xmath484 there exists @xmath485 such that @xmath486 , and conversely for all @xmath487 there exists @xmath443 with @xmath486 .
let now @xmath488 , then by the previous discussion there exist @xmath489 such that @xmath490 . on the other hand since @xmath489 we have @xmath491 and @xmath492 therefore the surjectivity of @xmath393 follows from that of @xmath493 .
a strategy to prove the surjectivity of the multiplication map was described in @xcite for wonderful symmetric varieties and in @xcite for general wonderful varieties .
such a strategy reduces the proof of the surjectivity of the multiplication maps for all pair of globally generated line bundles to a finite number of computations , which arise in correspondence to the so - called _ fundamental low tiples_. recall from @xcite that a triple @xmath494 with @xmath495 is called a _ low triple _ if , for all @xmath496 such that @xmath497 , @xmath498 and @xmath499 , it holds @xmath500 and @xmath501 .
the triple @xmath502 is called a _ fundamental triple _ if @xmath503 . to determine the low triples is useful the notion of covering difference .
let @xmath504 with @xmath505 and suppose that @xmath420 is maximal in @xmath391 with this property : then we say that @xmath506 covers @xmath420 and we call @xmath507 a _ covering difference _ in @xmath391 . for all @xmath508 , define its _ positive part _
@xmath509 , its _ negative part _ @xmath510 and its _ height _ @xmath511 .
notice that @xmath512 is a covering difference in @xmath391 if and only @xmath513 covers @xmath514 . as noticed in (
* section 2.1 , remark ) , the covering differences in @xmath391 are finitely many , therefore there is always a bound for the height of the positive part of a covering difference . in all the examples we know ( included those we will deal with in the present paper ) this bound can be taken to be 2 .
let @xmath502 be a low triple and suppose that @xmath459 is surjective , then it is a straightforward consequence of the definition that @xmath515 .
on the other hand we have the following . [
lem : riduzionetriple ] let @xmath63 be a wonderful variety and let @xmath516 be such that @xmath517 for every covering difference @xmath518 .
if @xmath515 for all low triples @xmath502 with @xmath519 , then the multiplication maps @xmath459 are surjective for all @xmath460 . to verify that @xmath515 we will make use of the following .
[ lemma : supporto moltiplicazione ] let @xmath520 be such that @xmath521
. then @xmath522 if and only if the projection of @xmath523 onto the isotypic component of highest weight @xmath524 is non - zero .
[ example : counterexample ] let @xmath525 and @xmath526 . if @xmath527 consider the spherical nilpotent @xmath2-orbit @xmath36 defined by the signed partition @xmath383 ( or similarly the one defined by @xmath384 if @xmath528 ) .
let @xmath63 be the corresponding wonderful @xmath2-variety , then there are elements @xmath460 such that @xmath459 is not surjective .
indeed , the spherical system of @xmath63 is the following : @xmath529 label the spherical roots and the colors of @xmath63 as follows : @xmath530 @xmath531 then the cartan pairing of @xmath63 is expressed as follows : @xmath532 consider the triple @xmath533 : then @xmath534 , and the triple is easily shown to be low . on the other hand if @xmath535 , then it would be @xmath536 , which is not the case .
therefore @xmath537 is not surjective .
we show in this section that in order to prove theorem [ teo : projnorm ] , we are reduced to the study of three special families of wonderful varieties .
following section [ sss : parabolic induction ] , by lemma [ lem : projnorm parabolic induction ] , the surjectivity of the multiplications on @xmath63 is reduced to that one on a wonderful @xmath245-variety @xmath197 , where @xmath245 is the levi subgroup of @xmath2 corresponding to the set of simple roots in @xmath538 .
more precisely , @xmath197 is the localization of @xmath63 at the subset @xmath539 , and the wonderful varieties arising in this way are described in sections [ sss : trivial cases ] , [ sss : symmetric cases ] , [ sss : reductive ] ( only the cases [ sss : ccaac ] and [ sss : ccaac2 ] ) , [ sss : typeb ] , [ sss : typec ] , [ sss : typed ] . analyzing all the possible cases , we now show that to prove the surjectivity of the multiplications for @xmath197 we are reduced to the following three families : * @xmath540 , @xmath342 , @xmath541 * @xmath542 , @xmath543 , @xmath544 * @xmath327 , @xmath545 . @xmath335 in the cases of section [ sss : trivial cases ] the wonderful variety @xmath63 is a flag variety , therefore the surjectivity of the multiplication of globally generated line bundles holds trivially since the space of sections of a globally generated line bundle on a flag variety is an irreducible @xmath2-module . in the cases of section [ sss : symmetric cases ] the wonderful variety @xmath197 is the wonderful compactification of an adjoint symmetric variety , and the surjectivity of the multiplication of globally generated line bundles holds thanks to @xcite . in the cases [ sss : ccaac ] and [ sss : ccaac2 ] of section
[ sss : reductive ] ( up to switching the two factors of @xmath2 ) the surjectivity of the multiplications of @xmath197 is reduced to that one of the wonderful variety @xmath546 with spherical system @xmath547 where @xmath548 .
more precisely , start with @xmath546 and consider the set of colors @xmath549 , it is distinguished and the corresponding quotient is a parabolic induction of @xmath197
. therefore the surjectivity of the multiplications of @xmath197 follows from that of @xmath546 thanks to lemma [ lem : projnorm quotients ] and lemma [ lem : projnorm parabolic induction ] . in the cases of section [ sss : typeb ]
( a ) @xmath197 is the wonderful variety with spherical system @xmath550 where @xmath551 and @xmath552 , but if @xmath553 it is just an adjoint symmetric variety . in the cases of section [ sss : typeb ] ( b ) @xmath197 is the wonderful variety with spherical system @xmath554 where @xmath545 . in the cases of section [ sss : typec ]
( a ) @xmath197 is the wonderful variety with spherical system @xmath555 where @xmath548 , whereas in the cases of section [ sss : typec ] ( b ) @xmath197 is the wonderful variety with spherical system @xmath556 . in the cases of section [ sss : typed ] ( a ) @xmath197 is the wonderful variety with spherical system @xmath557 where @xmath551 and @xmath548 .
the surjectivity of the multiplications in this case can be reduced to that of a comodel wonderful variety , which is known by ( * ? ? ?
* theorem 5.2 ) .
let indeed @xmath546 be the comodel wonderful variety of cotype @xmath558 , this is the wonderful variety with the following spherical system for a group of semisimple type @xmath559 .
@xmath560 consider the wonderful subvariety of @xmath546 associated to @xmath561 , then the set of colors @xmath562 @xmath563 is distinguished , and the corresponding quotient is a parabolic induction of @xmath197 .
therefore the surjectivity of the multiplications of @xmath197 follows from that of @xmath546 thanks to lemma [ lem : projnorm quotients ] and lemma [ lem : projnorm parabolic induction ] . finally , in the cases of section [ sss : typed ]
( b ) @xmath197 is the comodel wonderful variety of cotype @xmath564 , and the surjectivity of the multiplications for this variety follows by ( * ? ? ?
* theorem 5.2 ) . consider the wonderful variety @xmath63 for a semisimple group @xmath0 of type @xmath565 with @xmath342 defined by the following spherical system .
@xmath541 the spherical system associated to this luna diagram is described in section [ sss : typec ] . for convenience
we number the five spherical roots in the following way : @xmath566 there are six colors that we label in the following way : @xmath567 the weights of these colors are the following : @xmath568 notice that the @xmath0-stable divisor of @xmath63 corresponding to @xmath569 is a parabolic induction of a comodel wonderful variety of cotype @xmath570 ( see ( * ? ? ?
* section 5 ) ) .
therefore we can restrict our study to the covering differences and the low triples of @xmath63 which contain @xmath569 .
[ lemma : coveringxxx ] let @xmath512 be a covering difference in @xmath391 with @xmath571 , then either @xmath572 or @xmath573 . every other covering difference @xmath512 verifies @xmath574
. denote @xmath575 , then we have @xmath576 suppose that @xmath577 .
if @xmath578 then @xmath579 , and if @xmath580 then @xmath581 . therefore if @xmath582 it must be @xmath583 and @xmath584 , namely @xmath585 and @xmath586
it follows that @xmath587 and @xmath588 , suppose that @xmath589
. then @xmath590 since @xmath591 , and @xmath592 since @xmath593 .
therefore we get @xmath594 , which is absurd since @xmath595 .
as already noticed , the @xmath0-stable divisor of @xmath63 corresponding to @xmath569 is a parabolic induction of a comodel wonderful variety of cotype @xmath570 .
therefore the covering differences @xmath518 with @xmath596 coincide with those studied in ( * ? ? ?
* proposition 3.2 ) , and they all satisfy @xmath574 .
[ lem : triplexx ] let @xmath502 be a low fundamental triple , denote @xmath597 and suppose that @xmath571 .
then we have the following possibilities : * @xmath598 , @xmath599 ; * @xmath600 , @xmath601 ; * @xmath602 , @xmath603 ; * @xmath604 , @xmath605 ; * @xmath606 , @xmath607 ; * @xmath608 , @xmath609 . by lemma [ lemma : coveringxxx ] ,
@xmath610 and @xmath611 are the unique covering differences @xmath518 with @xmath612 .
therefore @xmath613 are minuscule in @xmath391 .
let @xmath502 be a fundamental triple with @xmath614 , denote @xmath615 and suppose @xmath616 .
notice that , if @xmath502 is a low triple , then @xmath584 : suppose indeed @xmath617 , then @xmath618 and @xmath619 .
therefore , if @xmath502 is a low triple , then ( [ gamma ] ) implies @xmath620 .
. then for every covering difference @xmath622 it holds @xmath623 , therefore @xmath502 is necessarily a low triple . to classify such fundamental triples ,
suppose @xmath624 .
then @xmath625 , hence @xmath626 and we get @xmath627 .
being @xmath574 , it follows then @xmath628 .
equivalently , we have the equality @xmath629 , and the inequalities @xmath625 , @xmath630 imply @xmath631 .
it follows @xmath632 and @xmath633 , and the inequality @xmath620 imply @xmath634
. therefore @xmath635 and @xmath636 .
similarly , suppose @xmath621 and @xmath637 .
then @xmath638 , hence @xmath639 , and we get @xmath640 .
being @xmath574 , it follows then @xmath641 .
equivalently , @xmath642 , and the inequalities @xmath643 , @xmath638 imply @xmath644 .
it follows @xmath645 and @xmath646 , and the inequality @xmath620 imply @xmath634
. therefore @xmath647 and @xmath648 .
suppose now @xmath621 and @xmath649 .
then the equalities @xmath650 imply @xmath651 , and it follows @xmath652 and @xmath653
. therefore the inequality @xmath620 implies @xmath634 , and the inequality @xmath643 implies @xmath654 . therefore either @xmath599 and @xmath655 , or @xmath605 and @xmath656 .
suppose finally @xmath657 .
notice that , if @xmath502 is a low triple , then @xmath624 : indeed @xmath658 , and if e.g. @xmath659 then @xmath660 and @xmath661 .
therefore @xmath625 , hence @xmath662 .
it follows @xmath663 , therefore @xmath664 .
being @xmath574 , in particular it must be @xmath665 .
. then @xmath667 , hence @xmath668 .
therefore @xmath669 and we get the equalities @xmath670 and @xmath671 .
the inequality @xmath672 implies then @xmath673 , hence @xmath632 , @xmath633 , and @xmath634 thanks to the inequality @xmath620 . therefore @xmath674 and @xmath675 , and @xmath502 is a low triple since @xmath676 are both minuscule .
suppose now @xmath677 .
then @xmath678 , and being @xmath574 it follows @xmath679 , and moreover we get @xmath632 and @xmath671 . by the inequalities @xmath672 , @xmath680
we get then @xmath626 and @xmath681 . on the other hand @xmath682 , therefore @xmath683 and it follows @xmath653 , and @xmath634 as well
thanks to the inequality @xmath620 .
therefore @xmath684 and @xmath685 , and @xmath502 is a low triple since @xmath686 is minuscule . to prove the projective normality of @xmath63 we now apply lemma [ lemma : supporto moltiplicazione ]
this requires some computations .
we first need an explicit description of the invariants .
let @xmath687 with standard basis given by @xmath688 .
let @xmath689 where @xmath690 and we choose a basis @xmath691 and fix a symplectic form such that @xmath692 for @xmath693 .
we set @xmath694 and @xmath695 .
let @xmath696 be the basis of @xmath697 dual to @xmath688 .
notice that the isomorphism from @xmath698 to @xmath697 sending @xmath699 to @xmath700 , @xmath701 to @xmath702 and @xmath703 to @xmath704 is @xmath0-equivariant .
we set @xmath705 , so that we can take @xmath57 as the stabilizer of the line spanned by the vector @xmath706 .
we denote by @xmath707 the vector @xmath708 . in
coordinates the vectors @xmath707 are given as follows * @xmath709 and @xmath710 ; * @xmath711 and @xmath712 ; * @xmath713 and @xmath714 ; * @xmath715 and @xmath716 * @xmath717 and @xmath718 .
we can now prove the multiplication @xmath459 is surjective for all @xmath460 . by lemma [ lemma : coveringxxx ]
every covering difference @xmath512 satisfies @xmath719 , therefore by lemma [ lem : riduzionetriple ] it is enough to check that @xmath720 for all low fundamental triples @xmath502 .
suppose that @xmath721 and let @xmath164 be the @xmath0-stable divisor of @xmath63 corresponding to @xmath569 .
then @xmath164 is a parabolic induction of a comodel wonderful variety of cotype @xmath570 , hence the inclusion @xmath722 follows by lemma [ lem : projnorm parabolic induction ] together with ( * ? ? ?
* theorem 5.2 ) . by lemma [ lemma : supporto moltiplicazione ]
we are reduced to prove that for all low triples @xmath723 listed in lemma [ lem : triplexx ] the projection of @xmath724 onto the isotypic component of type @xmath725 in @xmath726 is non - zero .
we have @xmath728 , the equivariant map @xmath729 given by @xmath730 and @xmath731 @xmath600 .
we have @xmath732 , the equivariant map @xmath733 given by @xmath734 and @xmath735 @xmath602 .
we have @xmath736 , the equivariant map @xmath737 given by @xmath738 and @xmath739 @xmath604 .
we have @xmath740 , the equivariant map @xmath741 given by @xmath742 and @xmath743 @xmath606 .
we have @xmath744 , the equivariant map @xmath745 given by @xmath746 and @xmath747 @xmath608 .
we have the equivariant map @xmath748 given by @xmath749 and @xmath750 consider the wonderful variety @xmath63 for a semisimple group @xmath0 of type @xmath751 with @xmath543 defined by the following spherical system .
@xmath544 the spherical data and the cartan pairing associated to this luna diagram are described in section [ sss : typeb ] . for convenience
we number the spherical roots in the following way : @xmath752 there are @xmath753 colors that we label in the following way : @xmath754 the weights of these colors are the following : @xmath755 where @xmath756 if @xmath319 and @xmath757 if @xmath317 .
notice that @xmath63 has the same cartan matrix of the spherical nilpotent orbit studied in ( * ? ? ?
* section 7.3 ) .
it follows that the covering differences and the fundamental low triples are the same as those computed therein , since they only depend on the cartan matrix .
in particular , every covering difference @xmath518 satisfies @xmath574 , and every fundamental triple is low . in order to prove the projective normality of @xmath63 , in the following lemma we summarize some properties of its fundamental triples .
[ lemma : fund - triples - yyy ] let @xmath758 be a fundamental triple , denote @xmath759 and suppose that @xmath760
. then @xmath761 are even integers and @xmath762 .
if moreover @xmath763 , then @xmath764 and @xmath765 .
take a sequence of coverings in @xmath766 @xmath767 denote @xmath768 . by proposition 3.2 and proposition 7.3 in @xcite
we have the following three possibilities : 1 .
@xmath769 , for some integers @xmath770 of different parity with @xmath771 , 2 .
@xmath772 , for some integers @xmath770 of the same parity with @xmath773 , 3 .
@xmath774 , for some even integers @xmath770 with @xmath773 .
since @xmath760 , there is at least one @xmath775 of type 3 .
let @xmath776 be minimal with @xmath777 of type 3 , because of the parity of @xmath778 and @xmath779 , the previous description implies that every @xmath780 with @xmath781 is of type 2 .
moreover , it follows that @xmath782 and @xmath783 for all @xmath784 , and that @xmath785 , @xmath786 are even ( resp .
odd ) for all @xmath787 ( resp .
@xmath788 ) . therefore @xmath789 and @xmath790 are even integers and @xmath791 , and we get the equalities @xmath792 and @xmath793 .
suppose that @xmath794 : then @xmath795 and @xmath796 are even and @xmath797 , hence @xmath798 .
suppose instead @xmath799 , then @xmath795 and @xmath796 are odd and @xmath800 , and again we get @xmath798 . to show the first claim , notice that @xmath801 if and only if @xmath802 : this is not the case if @xmath803 , and if @xmath804 it can not happen as well , since then @xmath795 and @xmath796 would be odd . similarly , @xmath763 if and only if @xmath805 if and only if @xmath806 .
this means @xmath807 , which implies @xmath808 . to prove the projective normality of @xmath63 we will apply lemma [ lemma : supporto moltiplicazione ] .
first we describe the invariants .
let @xmath809 with standard basis given by @xmath810 .
let @xmath811 where @xmath812 and we choose a basis @xmath813 and fix a bilinear symmetric form such that @xmath814 for all @xmath815 .
set @xmath816 , so that we can take @xmath57 as the stabilizer of the line spanned by the vector @xmath817 .
we have @xmath818 for @xmath819 .
if we denote by @xmath707 the vector @xmath708 then in coordinates the vectors @xmath707 are given as follows : @xmath820 the multiplication @xmath459 is surjective for all @xmath460 .
as already noticed , propositions 3.2 and 7.3 in @xcite show that every covering difference @xmath512 satisfies @xmath574 .
it follows that every @xmath443 is minimal in @xmath391 w.r.t .
@xmath412 , hence every fundamental triple is low
. therefore by lemma [ lem : riduzionetriple ] we have to check that @xmath821 for all fundamental triples @xmath502 .
let @xmath502 be such a triple and denote @xmath597 .
suppose that @xmath822 , and let @xmath164 be the @xmath0-stable divisor of @xmath63 corresponding to the spherical root @xmath823 .
then @xmath164 is a parabolic induction of a comodel wonderful variety of cotype @xmath824 ( see ( * ? ? ?
* section 5 ) ) .
hence the inclusion @xmath722 follows by lemma [ lem : projnorm parabolic induction ] together with ( * ? ? ?
* theorem 5.2 ) .
suppose that @xmath760 , and assume @xmath825 and @xmath826 , then @xmath762 by lemma [ lemma : fund - triples - yyy ] .
we show that @xmath722 proceeding by induction on @xmath827 .
suppose that @xmath828
. then @xmath829 . let @xmath830 be the @xmath0-stable subvariety of @xmath63 obtained by intersecting the @xmath0-stable divisors corresponding to @xmath831 and to @xmath832 .
if @xmath833 , then @xmath830 is a parabolic induction of the wonderful variety of type @xmath834 , therefore the multiplication of sections of globally generated line bundles on @xmath830 is always surjective by the inductive hypothesis thanks to lemma [ lem : projnorm parabolic induction ] .
if instead @xmath835 , then @xmath830 is a parabolic induction of a rank 1 wondeful symmetric variety @xmath197 , which is homogeneous under its automorphism group , therefore the multiplication of sections of globally generated line bundles on @xmath197 is always surjective , and the same holds for @xmath830 by lemma [ lem : projnorm parabolic induction ] again . in particular , since @xmath502 is a low triple , it follows the inclusion @xmath722 .
suppose now that @xmath763 .
then by lemma [ lemma : fund - triples - yyy ] it follows that @xmath836 and @xmath837 are even integers , and @xmath838 with @xmath839 .
hence by lemma [ lemma : supporto moltiplicazione ] we need to find an equivariant map @xmath840 such that @xmath841 ( the formula makes sense also when @xmath842 , by setting @xmath843 ) . notice that @xmath844 and we denote by @xmath845 and @xmath846 the projection respectively onto the first and onto the second factor . in particular the map @xmath846 , up to a scalar factor , is just the wedge product .
we will construct a map @xmath847 such that @xmath848 and @xmath849 so that the map @xmath850 will have the desired properties .
let @xmath851 be defined by @xmath852 let @xmath853 be defined by @xmath854 let @xmath855 be defined by @xmath856
. finally , set @xmath857 , so that @xmath858 .
notice that the value of @xmath859 ( resp .
@xmath860 ) on @xmath861 is the same as that on @xmath862 the first is equal to @xmath863 ( the sum being over @xmath864 ) , the second is equal to zero .
consider the wonderful variety @xmath63 for a semisimple group @xmath0 of type @xmath866 with @xmath545 defined by the following spherical system @xmath335 the spherical data and the cartan pairing associated to this luna diagram are described in section [ sss : typeb ] .
the spherical roots are simple roots , for convenience we enumerate them in the following way : @xmath867 there are @xmath868 colors that we label in the following way : @xmath869 the weights of these colors are the following : @xmath870 notice that the @xmath0-stable divisor of @xmath63 corresponding to @xmath871 is a parabolic induction of a comodel wonderful variety of cotype @xmath872 ( see ( * ? ? ?
* section 5 ) ) .
therefore we can restrict our study to the covering differences and the low triples of @xmath63 which contain @xmath871 .
[ lemma : covering - ab^y(s , s ) ] let @xmath873 be a covering difference in @xmath766 with @xmath874 .
then either @xmath875 , or @xmath876 , or @xmath877 for some @xmath878 .
every other covering difference satisfies @xmath574 .
recall that the cartan pairing is as follows ( we also set @xmath879 for all @xmath880 and all @xmath881 ) : @xmath882 set @xmath883 .
we have the following identities ( we set @xmath884 for @xmath880 and @xmath885 ) @xmath886 by hypothesis , we have @xmath887 .
let @xmath888 be minimal with @xmath889 .
then @xmath890 for all @xmath891 , and it follows @xmath892 for all @xmath893 , and @xmath894 for all @xmath895 .
if @xmath896 , set @xmath897
. then @xmath898 and @xmath899 , hence @xmath900 since @xmath518 is a covering difference .
we are left with the case @xmath901 , in particular @xmath902 .
we claim that @xmath903 is necessarily equal to @xmath904 or to @xmath905 .
assume that @xmath906 and @xmath907 .
since @xmath887 and @xmath908 , it follows @xmath909 , hence @xmath910 . since @xmath911 , it must be @xmath912 , and since @xmath913 , it must be @xmath914 . the latter implies @xmath915 hence @xmath916 . notice that @xmath917 : if indeed @xmath918 , then @xmath919 , hence @xmath920 , contradicting @xmath921 .
let @xmath922 be such that @xmath923 , and suppose that @xmath924 and @xmath925 for all @xmath926 with @xmath927 ( notice that these conditions have just been proved for @xmath928 ) .
as @xmath929 , it follows @xmath930 hence @xmath931 .
this implies @xmath932 .
if indeed @xmath933 , set @xmath934 then @xmath898 and @xmath899 , hence @xmath935 , contradicting @xmath902 in the first case and @xmath887 in the second case . applying this argument recursively for @xmath936 , it follows that @xmath937 for all @xmath938 , @xmath939 for all @xmath940 and @xmath941 .
in particular @xmath942 and @xmath941 , which are in contradiction .
as already noticed , the @xmath0-stable divisor of @xmath63 corresponding to @xmath871 is a parabolic induction of a comodel wonderful variety of cotype @xmath872 .
therefore the covering differences @xmath518 with @xmath943 coincide with those studied in ( * ? ? ?
* proposition 3.2 ) , and they all satisfy @xmath574 .
[ lemma : lowtriples - ab^y ] let @xmath502 be a low fundamental triple , denote @xmath944 and suppose that @xmath945
. then we have the following possibilities : * @xmath946 for @xmath947 , @xmath948 ; * @xmath949 , @xmath950 ; * @xmath951 , @xmath952 ; * @xmath953 , @xmath954 .
set @xmath955 , and set also @xmath956 and @xmath957 . by hypothesis , we have @xmath887 .
at least one of the two indices @xmath131 and @xmath132 must be odd . indeed , if both @xmath131 and @xmath132 were even , taking a sequence @xmath958 of coverings in @xmath766 , @xmath959 would necessarily be a covering difference of a comodel spherical system of cotype @xmath960 , see ( * ? ? ? * proposition 3.2.(2 ) ) , hence @xmath961 .
we claim that at least one of the two indices @xmath131 and @xmath132 must be equal to @xmath962 .
let us prove the claim .
assume both @xmath131 and @xmath132 are different from @xmath962 .
we can assume that @xmath132 is the minimal odd number between @xmath131 and @xmath132 .
since @xmath890 for every @xmath963 , as in the above proof , it follows that @xmath964 for every @xmath965 .
set @xmath966 and @xmath967 , then @xmath968 and @xmath969 , hence @xmath502 is not a low triple . therefore , we can assume @xmath247 .
suppose @xmath970 .
we have @xmath971 , but the latter is minuscule therefore we get only @xmath953 .
suppose @xmath248 .
we have @xmath972 and @xmath973 , hence @xmath910 .
now , we have @xmath974 , but in @xmath766 the latter covers only @xmath975 , with @xmath976 . therefore ,
if @xmath977 or @xmath978 we get only @xmath951 .
if @xmath916 and @xmath979 , since @xmath975 is minuscule , we get only @xmath949 .
suppose @xmath261 .
we have @xmath972 and @xmath980 , hence @xmath910 and , since @xmath915 also @xmath916 . for every @xmath981 , as @xmath925
, we have @xmath982 therefore , @xmath983 for all @xmath984 .
if @xmath131 is odd , set @xmath985 and @xmath986 , then @xmath987 and @xmath988 , hence @xmath502 is not a low triple .
if @xmath131 is even , we have @xmath989 but the latter is minuscule .
we get @xmath990 . to prove the projective normality of @xmath63 we will apply lemma [ lemma : supporto moltiplicazione ] .
this requires some computations , and we will need an explicit description of the invariants @xmath991 and @xmath992 .
fix @xmath993 with standard basis @xmath810 and @xmath994 with basis @xmath995 , @xmath996 and with a symmetric bilinear form @xmath997 such that @xmath998 .
set @xmath999 , so that we can take @xmath57 as the stabilizer of the line spanned by the vector @xmath817 .
we will need to use the spin module for the group @xmath1000 .
let us recall its construction .
let @xmath1001 where @xmath1002 is the span of @xmath995 and @xmath1003 is the span of @xmath1004 identified with the dual of @xmath1002 by the form @xmath997 .
let @xmath1005 and rename the basis of @xmath1003 as @xmath1006 .
define a map @xmath1007 by setting @xmath1008 equal to @xmath1009 then @xmath1010 is @xmath0-equivariant , and its alternating square @xmath1011 corresponds to the spin representation via the isomorphism @xmath1012 .
we have @xmath1013 the invariants , in coordinates , are @xmath1014 [ prop : projnorm - ab^y ] the multiplication @xmath459 is surjective for all @xmath460 . by lemma [ lemma : covering - ab^y(s , s ) ]
every covering difference @xmath512 satisfies @xmath719 , therefore by lemma [ lem : riduzionetriple ] it is enough to check that @xmath720 for all low fundamental triples @xmath502 .
suppose that @xmath1015 and let @xmath164 be the @xmath0-stable divisor of @xmath63 corresponding to @xmath871 .
then @xmath164 is a parabolic induction of a comodel wonderful variety of cotype @xmath872 , hence the inclusion @xmath722 follows by lemma [ lem : projnorm parabolic induction ] together with ( * ? ? ?
* theorem 5.2 ) .
we are left to check that @xmath1016 for all low fundamental triples @xmath502 with @xmath1017 .
consider first the triple @xmath1018 .
then we have the projection @xmath1019 given by @xmath1020 where @xmath1021 via @xmath1022 , and we get @xmath1023 .
we now proceed by induction on @xmath827 .
assume @xmath1024 .
then we are left to check the triples @xmath1025 , @xmath1026 and @xmath1027 .
we have the projection @xmath1028 given by @xmath1029 , with @xmath1030 given by the identification @xmath1031 , notice that @xmath67 is selfdual and set @xmath1032 given by @xmath1033 followed by the identification @xmath1034 .
we get @xmath1035 .
@xmath1026 and @xmath1027 .
since @xmath1036 , we can consider the intersection @xmath164 of the @xmath0-stable divisors of @xmath63 corresponding to the spherical roots @xmath1037 .
then the sections in @xmath1038 do not vanish on @xmath164 , so it is enough to prove that @xmath1039 where @xmath1040 denotes the multiplication of sections on @xmath164 .
consider in @xmath0 the levi subgroup @xmath196 associated to the roots @xmath1041 , which has semisimple factor of type @xmath1042 , and consider the comodel @xmath196-variety @xmath197 of cotype @xmath1043 .
the wonderful @xmath0-variety @xmath164 is obtained by parabolic induction from @xmath197 .
hence our claim follows by lemma [ lem : projnorm parabolic induction ] .
assume @xmath1044 .
then we are left to check the triples @xmath946 with @xmath1045 , @xmath949 , @xmath951 , @xmath953 .
let @xmath502 be such a triple , then @xmath1036 and we can consider the intersection @xmath164 of the @xmath0-stable divisors of @xmath63 corresponding to the spherical roots @xmath1037 .
take the levi subgroup @xmath196 of @xmath0 associated to the roots @xmath1046 , of semisimple type @xmath1047 , and consider the comodel @xmath196-variety @xmath197 of type @xmath1048 .
the wonderful @xmath0-variety @xmath164 is obtained by parabolic induction from @xmath197 .
hence our claim follows by lemma [ lem : projnorm parabolic induction ] and the induction hypothesis .
recall that we have fixed a maximal torus @xmath64 in @xmath2 and borel subgroup @xmath65 of @xmath2 containing @xmath64 .
we use @xmath1049 for the weight lattice of @xmath64 .
let us denote by @xmath1050 the weight semigroup of a @xmath2-spherical variety @xmath546 , @xmath1051 , v({\lambda } ) ) \neq 0 \}.\ ] ] let @xmath1052 be a spherical nilpotent orbit in @xmath380 , and let @xmath215 and @xmath1053 be the set of spherical roots and the set of colors of the wonderful compactification of @xmath1054}$ ] .
let us denote by @xmath1055 the element of @xmath1056 such that @xmath1057 .
provided that the multiplication of sections of globally generated line bundles on the wonderful compactification of @xmath1054}$ ] is surjective , we have that @xmath1058 is normal if and only if @xmath1059 is minuscule in @xmath1056 with respect to the partial order @xmath1060 .
if moreover @xmath1061 is the normalization of @xmath1062 , then @xmath1063 that is , @xmath1064 where @xmath1065 is the subsemigroup of @xmath1056 given by @xmath1066 in the present section we will study the normality of @xmath1062 and we will compute the weight semigroups @xmath1067 by computing the corresponding semigroups @xmath1065 . in particular we will prove the following theorem .
let @xmath1068 be a classical symmetric pair of non - hermitian type .
then @xmath1062 is not normal if and only if @xmath1069 and the signed partition of @xmath1052 is @xmath1070 , with @xmath1071 odd , or @xmath1072 , with @xmath1073 odd . in appendix
[ a ] these are the cases [ sss : bdbay ] , with @xmath1074 , [ sss : bbaby ] , with @xmath1075 , and [ sss : bbbay ] , with @xmath1074 . the normality or non - normality of @xmath1062 , as well as the generators of the weight semigroup @xmath1067 are given in tables 211 , in appendix [ b ] .
in the tables we also provide the codimension of @xmath1076 in @xmath1077 .
notice that , if @xmath1077 is normal and the codimension of @xmath1076 in @xmath1077 is greater than 1 , then @xmath1078 = \mathbb c[ke]$ ] , so that the weight semigroup of @xmath1052 actually coincides with @xmath1067 . notice also that in all cases where @xmath1062 is not normal , the normalization @xmath1079 is not even bijective ( see @xcite for a general procedure to compute the @xmath2-orbits in @xmath1061 and in @xmath1062 ) .
the normality of @xmath1080 is well known and may be deduced from @xcite ( see also @xcite , when @xmath1081 is spherical under the action of @xmath0 ) . in particular , if @xmath1068 is a classical symmetric pair of non - hermitian type , then @xmath1080 is normal in all but the following cases : * @xmath1082 with @xmath1083 and the partition of @xmath1081 is @xmath1084 ( cases [ sss : ccaac ] and [ sss : ccaac2 ] , with @xmath1085 ) , * @xmath1082 with @xmath1086 and the partition of @xmath1081 is @xmath1087 ( cases [ sss : ccabx ] and [ sss : ccbax ] , with @xmath1088 ) , * @xmath1089 and the partition of @xmath1081 is @xmath1090 ( case [ sss : bdbay ] , with @xmath1091 ) .
we now report the details of the computation of the semigroup @xmath1065 .
we omit the cases where @xmath1054}$ ] is a flag variety ( cases [ sss : trvial1 ] and [ sss : trvial2 ] in appendix [ a ] ) or a parabolic induction of a symmetric variety ( cases [ sss : symm1 ] , [ sss : symm2 ] , [ sss : dao ] , [ sss : symm4.1 ] , [ sss : bdaa ] , [ sss : symm8 ] , [ sss : ddaa ] , as well as cases [ sss : bdady ] , [ sss : bdbay ] , [ sss : bbaby ] , [ sss : bbbay ] , [ sss : ddady ] , [ sss : ddday ] when @xmath249 ) : in these cases the combinatorics of spherical systems is easier .
by @xcite , the normality of @xmath1077 was already known in all these cases ( see the discussion at the beginning of appendix [ b ] ) , and the corresponding weight semigroups @xmath1092 were obtained in @xcite by using different techniques . [ [ cases - sssccaac - q1-and - sssccaac2-p1 . ] ] * cases [ sss : ccaac ] @xmath1093 and [ sss : ccaac2 ] @xmath1094 . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : ccaac ] @xmath1093 , the other one is similar .
we have two spherical roots @xmath1095 and @xmath1096 , and three colors @xmath1097 , @xmath1098 and @xmath1099 .
we have @xmath1100 , which is minuscule in @xmath391 , therefore @xmath1062 is normal . furthermore @xmath1101 and @xmath1102 , therefore @xmath1103 .
let us set @xmath1104 and @xmath1105 , we have @xmath1106 and therefore @xmath1107 .
it follows that @xmath1108 [ [ cases - sssccacy - q2-and - ssscccay - p2 . ] ] * cases [ sss : ccacy ] @xmath1109 and [ sss : cccay ] @xmath1110 .
* + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : ccacy ] ( @xmath129 ) , the other one is similar .
let us keep the notation of sections [ sss : typec ] and [ ss : ay22ct ] , therefore @xmath1111 , @xmath1112 , @xmath1113 , @xmath1114 , @xmath1115 , @xmath1116 and @xmath1117 .
then we have @xmath1118 , which is minuscule in @xmath391 , therefore @xmath1062 is normal .
moreover @xmath1119 , @xmath1120 , therefore @xmath1121 .
moreover @xmath1122 , @xmath1123 , @xmath1124 , therefore @xmath1125 as well .
let us set @xmath1104 and @xmath1126 , we have @xmath1127 and therefore @xmath1128 .
it follows that @xmath1129 [ [ cases - sssccacy - q2-and - ssscccay - p2.-1 ] ] * cases [ sss : ccacy ] @xmath1130 and [ sss : cccay ] @xmath1131 . *
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : ccacy ] ( @xmath357 ) , the other one is similar .
label the spherical roots @xmath1132 , @xmath1133 , @xmath1134 , @xmath1135 , and label the colors @xmath1111 , @xmath1112 , @xmath1113 , @xmath1114 , @xmath1115 , @xmath1136 .
then we have @xmath1118 , which is minuscule in @xmath391 , therefore @xmath1062 is normal .
moreover @xmath1119 , @xmath1137 , therefore @xmath1121 . similarly @xmath1138 and @xmath1139 , therefore @xmath1140 as well .
let us set @xmath1104 and @xmath1126 , we have @xmath1141 and therefore @xmath1142 .
it follows that @xmath1143 [ [ cases - sssccabx - and - sssccbax . ] ] * cases [ sss : ccabx ] and [ sss : ccbax ] .
* + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : ccabx ] , the other one is similar . in this case
@xmath1055 is minuscule in @xmath391 .
following example [ example : counterexample ] , this does not imply that the ring @xmath1144 is generated by its degree one component @xmath1145 , indeed the multiplication of sections of globally generated line bundles on @xmath63 is not necessarily surjective .
however using our methods we are still able to compute the normality and the weight semigroups of @xmath1077 .
enumerate the spherical roots and the colors of @xmath63 as in example [ example : counterexample ] .
then @xmath1118 , and by definition @xmath1146 the following inclusions hold : * @xmath1147 ( where @xmath1148 ) , * @xmath1149 ( where @xmath1119 ) , * @xmath1150 ( where @xmath1151 ) , * @xmath1152 ( where @xmath1153 ) , * @xmath1154 ( where @xmath1155 ) .
consider the @xmath0-stable divisor @xmath466 corresponding to @xmath569 : it is a parabolic induction of a wonderful variety of type @xmath359 . by proposition [ prop : projnorm - ab^y ] together with lemma [ lem : projnorm parabolic induction ] it follows that the multiplication of sections is surjective for all pairs of globally generated line bundles on @xmath164 .
denote by @xmath1156 the restriction of line bundles and for @xmath453 set @xmath1157 . by lemma [ lemma : lowtriples - ab^y ]
@xmath1158 and @xmath1159 are low triples for @xmath164 , and since @xmath1160 and @xmath1161 are coverings in @xmath1162 it follows that @xmath1163 and @xmath1164 are low triples for @xmath164 as well . on the other hand , for all @xmath453 the @xmath0-modules @xmath401 and @xmath1165
are canonically identified , and since the restriction @xmath1166 is surjective we get the inclusions ( 1 ) , ( 2 ) , ( 3 ) , ( 4 ) . we are left with the inclusion ( 5 ) .
consider the distinguished subset of colors @xmath1167 and denote by @xmath197 the quotient of @xmath63 by @xmath1168 .
then @xmath197 is a rank 1 wonderful variety with spherical root @xmath1169 whose set of colors is identified with @xmath1170 . by lemma [ lem : projnorm quotients ]
we have that @xmath1171 for all @xmath1172 .
since @xmath1173 is a covering in @xmath1174 , the triple @xmath1175 is low in @xmath1174 . on the other hand
@xmath197 is a parabolic induction of a rank one symmetric variety , and for such a variety the multiplication of sections of globally generated line bundles is known to be always surjective . by lemma [ lem : projnorm parabolic induction ]
the same holds for @xmath197 , and since it corresponds to a low triple we get the inclusion @xmath1154 . @xmath1062 is normal , and @xmath1092 is generated by the weights @xmath1176 clearly , @xmath1177 .
on the other hand by the previous lemma we have that @xmath1178 for all @xmath443 , therefore @xmath1179 and the description of the generators follows by the description of the map @xmath1180 .
notice that the weights @xmath1181 are linearly independent .
therefore @xmath1182 is a saturated semigroup of weights ( that is , if @xmath1183 is the sublattice generated by @xmath1182 and if @xmath1184 is the cone generated by @xmath1182 , then @xmath1185 ) .
it follows that @xmath1062 is normal .
[ [ cases - without - roman - numerals . ] ] * cases without roman numerals . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : bdady ] ( @xmath1186 ) .
the cases [ sss : bdbay ] ( @xmath1187 ) , [ sss : bbaby ] @xmath1188 , [ sss : bbbay ] @xmath1189 , [ sss : ddady ] ( @xmath1190 ) and [ sss : ddday ] ( @xmath1190 ) are similar .
suppose @xmath266 , otherwise we have a symmetric case .
let us keep the notation of section [ sss : typed ] , therefore , for all @xmath1191 , @xmath1192 and @xmath1193 , furthermore @xmath1194 , @xmath1195 and @xmath1196 . to have a uniform notation
, we denote @xmath1197 we have @xmath1198 , which is minuscule in @xmath391 , therefore @xmath1062 is normal .
set @xmath1199 , @xmath1200 and , for all @xmath1201 , @xmath1202 notice that @xmath1203 the semigroup @xmath1204 is generated by @xmath1205 and @xmath1206 for all @xmath1207 .
since @xmath1208 , by induction on the even indices , it follows that @xmath1209 for all @xmath1210 . on the other hand @xmath1211 therefore , for the odd indices , we get @xmath1212 for all @xmath1213 .
let @xmath406 and suppose that @xmath1214 , let @xmath1215 and @xmath1216 .
notice that if @xmath369 then @xmath1217 is even .
therefore @xmath1218 , and write @xmath1219 expressing the coefficients @xmath1220 with respect to @xmath1221 we get that @xmath1222 , and @xmath1223
. the claim follows .
[ [ cases - with - roman - numerals . ] ] * cases with roman numerals . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : ddady ] ( @xmath1224 ) , the case ( i ) .
the case ( ii ) , as well as the cases [ sss : bdbay ] ( @xmath1225 ) and [ sss : ddday ] ( @xmath1226 ) are similar .
first , let us suppose @xmath1227 .
let us keep the notation of section [ sss : typed ] , therefore , for all @xmath1191 , @xmath1192 and @xmath1193 , furthermore @xmath1194 , @xmath1195 and @xmath1228 .
we have @xmath1198 , which is minuscule in @xmath391 , therefore @xmath1062 is normal .
set @xmath1199 and @xmath1200 , and define inductively , for all @xmath1201 , @xmath1229 notice that @xmath1230 the semigroup @xmath1204 is generated by @xmath1205 and @xmath1206 for all @xmath1207 .
since @xmath1208 , for the even indices , it follows that @xmath1209 for all @xmath1210 . on the other hand @xmath1211
therefore , for the odd indices , we get @xmath1212 for all @xmath1213 .
let @xmath406 and suppose that @xmath1214 , denote @xmath1215 and @xmath1216 . expressing the coefficients @xmath1231 with respect to @xmath1221
we get that @xmath1232 , and @xmath1233 .
the claim follows .
we are left with the case @xmath246 , the case @xmath249 being symmetric .
here we have @xmath1234 , which is minuscule in @xmath391 , therefore @xmath1062 is normal .
proceeding as above we get the same semigroup @xmath1235 [ [ cases - without - roman - numerals.-1 ] ] * cases without roman numerals .
* + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : bbaby ] ( @xmath326 ) .
the cases [ sss : bdbay ] ( @xmath1236 ) and [ sss : bbbay ] ( @xmath325 ) are similar .
let us keep the notation of sections [ sss : typeb ] and [ ss : abyss ] , therefore @xmath1192 and @xmath1193 for all @xmath1237 , @xmath1238 and @xmath1239 .
we have @xmath1198 .
notice that @xmath1240 is not minuscule , indeed @xmath1241 .
therefore @xmath1062 is not normal . to have a uniform notation set @xmath1242 set @xmath1199 and @xmath1200 , and define inductively for all @xmath1243 @xmath1244 notice that @xmath1245 the semigroup @xmath1204 is generated by @xmath1246 .
let @xmath406 and suppose that @xmath1214 , denote @xmath1215 and @xmath1216 .
notice that if @xmath369 then @xmath1217 is even .
therefore @xmath1218 , and write @xmath1219 expressing the spherical roots in terms of colors it follows that @xmath1247
. the claim follows .
[ [ case - with - roman - numerals . ] ] * case with roman numerals . * + + + + + + + + + + + + + + + + + + + + + + + + + + + we deal here with the case [ sss : bdbay ]
( @xmath1248 ) , the case ( i ) , the case ( ii ) is similar
. the notation will be slightly different than before : let us enumerate the spherical roots @xmath1249 as follows @xmath1250 accordingly , we enumerate the colors as follows : @xmath1251 and @xmath1252 for all @xmath1237 , @xmath1253 and @xmath1254 .
we have @xmath1198 .
notice that @xmath1240 is not minuscule , indeed @xmath1241 .
therefore @xmath1062 is not normal .
set @xmath1199 and @xmath1200 , and define inductively for all @xmath1243 @xmath1244 notice that @xmath1255 the semigroup @xmath1204 is generated by @xmath1246 .
let @xmath406 and suppose that @xmath1214 , let @xmath1215 and @xmath1216 . expressing the spherical roots in terms of colors it follows that @xmath1256 and @xmath1257 .
the claim follows .
[ [ cases - without - roman - numerals.-2 ] ] * cases without roman numerals . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : bdady ] ( @xmath368 ) .
the cases [ sss : ddady ] ( @xmath1226 ) and [ sss : ddday ] ( @xmath1224 ) are similar .
let us keep the notation of section [ sss : typed ] , therefore , for all @xmath1191 , @xmath1192 and @xmath1193 , furthermore @xmath1194 , @xmath1258 and @xmath1196 . to have a uniform notation
, we denote @xmath1197 we have @xmath1198 , which is minuscule in @xmath391 , therefore @xmath1062 is normal .
set @xmath1199 , @xmath1200 and , for all @xmath1201 , @xmath1202 notice that @xmath1259 furthermore , set @xmath1260 the semigroup @xmath1204 is generated by @xmath1205 , @xmath1206 for all @xmath1207 with @xmath1261 , and @xmath1262 .
since @xmath1208 , for the even indices , it follows that @xmath1209 for all @xmath1210 , and @xmath1263 as well . on the other hand @xmath1264 therefore , for the odd indices , we get @xmath1212 for all @xmath1213 .
let @xmath406 and suppose that @xmath1214 , denote @xmath1215 and @xmath1216 .
notice that if @xmath369 then @xmath1217 is even .
therefore @xmath1218 , and write @xmath1219 expressing the coefficients @xmath1220 with respect to @xmath1221 we get that @xmath1222 , and @xmath1265 .
the claim follows .
[ [ cases - with - roman - numerals.-1 ] ] * cases with roman numerals .
* + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us deal with the case [ sss : ddady ] ( @xmath1266 ) , the case ( i ) . the case ( ii ) as well as the cases [ sss : ddday ] ( @xmath1267 ) are similar .
first , let us suppose @xmath1227 .
let us keep the notation of section [ sss : typed ] , therefore , for all @xmath1191 , @xmath1192 and @xmath1193 , furthermore @xmath1268 , @xmath1269 and @xmath1228 .
we have @xmath1198 , which is minuscule in @xmath391 , therefore @xmath1062 is normal .
set @xmath1199 and @xmath1200 , and define inductively , for all @xmath1201 , @xmath1229 notice that @xmath1270 furthermore , set @xmath1271 the semigroup @xmath1204 is generated by @xmath1205 , @xmath1206 for all @xmath1207 with @xmath1261 , and @xmath1262 .
since @xmath1208 , for the even indices , it follows that @xmath1209 for all @xmath1210 , and @xmath1263 as well . on the other hand @xmath1264 therefore , for the odd indices , we get @xmath1212 for all @xmath1213 .
let @xmath406 and suppose that @xmath1214 , let @xmath1215 and @xmath1216 . expressing the coefficients @xmath1231 with respect to @xmath1221
we get that @xmath1232 , and @xmath1272 .
the claim follows .
we are left with the case @xmath246 , the case @xmath249 being symmetric .
here we have @xmath1234 , which is minuscule in @xmath391 , therefore @xmath1062 is normal .
proceeding as above we get the same semigroup , that is , @xmath1273 is generated by @xmath1274 , @xmath1275 , @xmath1276 , @xmath1277 , @xmath1278 .
here we report the list of the spherical nilpotent @xmath2-orbits in @xmath7 for all symmetric pairs @xmath1279 of classical non - hermitian type . every @xmath2-orbit in @xmath7 is labelled with the signed partition of the corresponding real nilpotent orbit , via the kostant - sekiguchi - dokovi bijection . for every orbit we provide a representative @xmath30 , as well as a normal triple containing it @xmath33 .
for all @xmath1280 , let @xmath1281 be the @xmath1282-eigenspace in @xmath6 of eigenvalue @xmath926 .
we denote by @xmath194 the parabolic subgroup of @xmath2 whose lie algebra is equal to @xmath1283 in each case we describe the centralizer of @xmath265 , which we denote by @xmath264 or by @xmath196 , which is a levi subgroup of @xmath194 .
we denote by @xmath199 the unipotent radical of @xmath194 .
then we describe the centralizer of @xmath30 , which we denote by @xmath40 . a levi subgroup of @xmath40 is always given by @xmath1284 , the centralizer of @xmath30 in @xmath196 .
the unipotent radical of @xmath40 is either equal to @xmath199 or equal to a co - simple @xmath1284-submodule of @xmath199 . in the latter case
, there always exist some simple @xmath1284-submodules in @xmath1285 , say @xmath283 , which we determine , with the following properties .
they are isomorphic as @xmath1284-modules but lie in pairwise distinct isotypical @xmath196-components . denoting by @xmath284 the @xmath1284-complement of @xmath1286 in @xmath292 , as @xmath1284-module , @xmath1287 where @xmath287 is a co - simple @xmath1284-submodule of @xmath1288 which projects non - trivially on every summand @xmath283
. actually , the integer @xmath281 , the number of the above simple @xmath1284-modules @xmath283 , will only be equal to 2 or 3 . [
remark : missingcase ] as already mentioned , the list of spherical nilpotent @xmath2-orbits in @xmath7 is in @xcite , and all the data in our list , such as a representative and its centralizer , can be directly computed using the information contained therein , with one exception .
there is one missing case in @xcite , corresponding to the signed partition @xmath1289 for the symmetric pair @xmath1290 with @xmath1291 ( cases [ sss : ccabx ] and [ sss : ccbax ] in appendix [ a ] ) .
the lack comes from a small mistake in ( * ? ? ?
* lemma 7.2 ) , we have checked that there is no further missing case arising from that lemma . the smallest case of this family , which is for @xmath1292 ,
was already present in ( * ? ? ?
* example 5.8 ) .
@xmath1294 , @xmath1295 , @xmath1296 .
let us fix a basis @xmath1297 of @xmath1298 , a skew - symmetric bilinear form @xmath1299 such that @xmath1300 for @xmath1301 and @xmath1302 .
then @xmath1299 can be seen as a linear form on @xmath1303 , @xmath1304 , and @xmath1305 @xmath1308 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1309 .
the centralizer of @xmath30 is @xmath1310 where @xmath1311 .
@xmath1313 , @xmath1295 , @xmath1314 .
let us fix a basis @xmath1315 of @xmath1316 , a symmetric bilinear form @xmath106 such that @xmath1317 for all @xmath1318 and @xmath1319 .
then @xmath106 can be seen as a linear form on @xmath1320 , @xmath1321 , and @xmath1322 @xmath1324 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1325 .
the centralizer of @xmath30 is @xmath1310 where @xmath1326 .
@xmath1328 , @xmath1329 , @xmath1314 .
if @xmath1330 , @xmath1331 .
let us fix a basis @xmath1297 of @xmath1298 , a symmetric bilinear form @xmath106 such that @xmath1317 for all @xmath1318 and @xmath1332 .
then @xmath106 can be seen as a linear form on @xmath1333 , @xmath1321 , and @xmath1334 let us denote by @xmath1335 the linear endomorphism of @xmath1298 switching @xmath1336 and @xmath1337 and fixing all the other basis vectors .
the conjugation by @xmath1335 is an involutive external automorphism of @xmath4 , living @xmath6 and @xmath7 stable , and inducing the nontrivial involution of the dynkin diagram of @xmath6 . if @xmath1340 , @xmath1324 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1341 .
the centralizer of @xmath30 is @xmath1310 where @xmath1342 .
if @xmath1343 , there exist two orbits labelled i and ii .
the case ( i ) can be described as above by specializing @xmath62 equal to @xmath516 .
the case ( ii ) can be obtained from the case ( i ) conjugating by @xmath1335 .
@xmath1345 , @xmath1346 , @xmath1347 .
let us fix a basis @xmath1348 of @xmath1349 and a skew - symmetric bilinear form @xmath1299 such that @xmath1300 for @xmath1350 .
similarly , let us fix a basis @xmath1351 of @xmath1352 and a skew - symmetric bilinear form @xmath1353 such that @xmath1354 for @xmath1355 .
then @xmath1356 and @xmath1357 @xmath1359 @xmath1360 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1361 .
the centralizer of @xmath30 is @xmath1310 where @xmath1362 , the @xmath1363 factor of @xmath1284 is embedded skew - diagonally , @xmath1364 , into the @xmath1365 factor of @xmath196 .
@xmath1367 @xmath1368 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1369 .
the centralizer of @xmath30 is @xmath1310 where @xmath1370 , the @xmath1371 factor of @xmath1284 is embedded as @xmath1372 into @xmath1373 where the @xmath1374 factor is included in the @xmath1375 factor of @xmath196 and the @xmath1376 factor is included in the @xmath1377 factor of @xmath196 .
this case can be obtained from the case [ sss : ccaac ] by switching the role of @xmath131 and @xmath132 .
@xmath1380 @xmath1381 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1382 .
the centralizer of @xmath30 is @xmath1383 where @xmath1384 , the @xmath1385 factor of @xmath1284 is embedded as @xmath1372 into @xmath1386 where the @xmath1374 factor is included in the @xmath1375 factor of @xmath196 and the @xmath1387 factor is included in the @xmath1388 factor of l , the @xmath1389 factor of @xmath1284 is embedded skew - diagonally @xmath1390 into the @xmath1391 factor of @xmath196 .
the quotient @xmath1392 is a simple @xmath1284-module of dimension 2 as follows . in @xmath1285
there are exactly two simple @xmath1284-submodules , @xmath1393 , of highest weight @xmath1394 w.r.t . the @xmath1374 factor , isomorphic as @xmath1284-modules but lying in two distinct isotypical @xmath196-components .
let @xmath284 be the @xmath1284-complement of @xmath1395 in @xmath292 .
as @xmath1284-module , @xmath1396 is the direct sum of @xmath284 and a simple @xmath1284-submodule of @xmath1395 which projects non - trivially on both summands @xmath1397 and @xmath1398 .
this case can be obtained from the case [ sss : ccacy ] by switching the role of @xmath131 and @xmath132 .
@xmath1401 @xmath1402 @xmath1403 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1404 .
the centralizer of @xmath30 is @xmath1310 where @xmath1405 , the @xmath1406 factor of @xmath1284 is embedded diagonally , @xmath1407 , into the @xmath1408 factor of @xmath196 .
this case can be obtained from the case [ sss : ccabx ] by switching the role of @xmath131 and @xmath132 .
@xmath1328 , @xmath1329 , @xmath1411 .
if @xmath1330 , @xmath1412 .
let us fix a basis @xmath1297 of @xmath1298 , a symmetric bilinear form @xmath106 such that @xmath1317 for all @xmath1318 and @xmath1332 .
then @xmath1413 @xmath1415 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1416 .
the centralizer of @xmath30 is @xmath1310 where @xmath1417 .
@xmath1313 , @xmath1295 , @xmath1411 . if @xmath1419 , @xmath1420 .
let us fix a basis @xmath1315 of @xmath1316 , a symmetric bilinear form @xmath106 such that @xmath1317 for all @xmath1318 and @xmath1319 .
then @xmath1421 @xmath1415 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1423 .
the centralizer of @xmath30 is @xmath1310 where @xmath1424 .
@xmath1426 , @xmath1427 , @xmath1428 , @xmath1429 .
if @xmath248 and @xmath1428 , @xmath1430 .
if @xmath1427 and @xmath357 , @xmath1431 .
if @xmath248 and @xmath357 , @xmath1432 .
let us fix a basis @xmath1433 of @xmath1434 and a symmetric bilinear form @xmath106 such that @xmath1317 for all @xmath1318 .
similarly , let us fix a basis @xmath1351 of @xmath1352 and a symmetric bilinear form @xmath1435 such that @xmath1436 for all @xmath1318 .
then @xmath1437 and @xmath1438 let us denote by @xmath1335 the linear endomorphism of @xmath1439 switching @xmath1440 and @xmath1441 and fixing all the other basis vectors .
the conjugation by @xmath1335 is an involutive external automorphism of @xmath4 , living @xmath6 and @xmath7 stable , and inducing the nontrivial involution of the dynkin diagram of @xmath6 .
if @xmath1443 , @xmath1359 @xmath1360 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1444 . the centralizer of @xmath30 is @xmath1310 where @xmath1445 , the @xmath1363 factor of @xmath1284 is embedded skew - diagonally , @xmath1364 , into the @xmath1365 factor of @xmath196 .
if @xmath326 , there exist two orbits labelled i and ii .
the case ( i ) can be described as above by specializing @xmath62 equal to @xmath132 .
the case ( ii ) can be obtained from the case ( i ) conjugating by @xmath1335 . if @xmath1447 , @xmath1448 @xmath1449 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1450 .
the centralizer of @xmath30 is @xmath1383 where @xmath1451 , the @xmath1452 factor of @xmath1284 is embedded as @xmath1453 into @xmath1454 where the @xmath1455 factor is included in the @xmath1456 factor of l , the @xmath1363 factor of @xmath1284 is embedded skew - diagonally @xmath1457 into the @xmath1365 factor of @xmath196 .
the quotient @xmath1392 is a simple @xmath1284-module of dimension @xmath62 as follows . in @xmath1285
there are exactly two simple @xmath1284-submodules , @xmath1393 , of highest weight @xmath1458 w.r.t . the @xmath1363 factor , isomorphic as @xmath1284-modules but lying in two distinct isotypical @xmath196-components
let @xmath284 be the @xmath1284-complement of @xmath1395 in @xmath292 .
as @xmath1284-module , @xmath1396 is the direct sum of @xmath284 and a simple @xmath1284-submodule of @xmath1395 which projects non - trivially on both summands @xmath1397 and @xmath1398 . if @xmath368 , the normal triple @xmath1459 , the parabolic subgroup @xmath195 and @xmath1284 have the same description , with @xmath1383 .
the quotient @xmath1392 remains a simple @xmath1284-module of dimension @xmath1460 but here in @xmath1285 there are exactly three simple @xmath1284-submodules , @xmath1461 , of highest weight @xmath1462 w.r.t . the @xmath1463 factor , isomorphic as @xmath1284-modules but lying in three distinct isotypical @xmath196-components .
let @xmath284 be the @xmath1284-complement of @xmath1464 in @xmath292 .
as @xmath1284-module , @xmath1396 is the direct sum of @xmath284 and a co - simple @xmath1284-submodule of @xmath1464 which projects non - trivially on every summand @xmath1397 , @xmath1398 and @xmath1465 . if @xmath1447 , @xmath1467 @xmath1468 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1469 .
the centralizer of @xmath30 is @xmath1383 where @xmath1470 , the @xmath1471 factor of @xmath1284 is embedded as @xmath1472 into @xmath1473 where the @xmath1471 factor is included in the @xmath1474 factor of l , the @xmath1363 factor of @xmath1284 is embedded skew - diagonally @xmath1457 into the @xmath1365 factor of @xmath196 .
the quotient @xmath1392 is a simple @xmath1284-module of dimension @xmath62 as follows . in @xmath1285
there are exactly two simple @xmath1284-submodules , @xmath1393 , of highest weight @xmath1394 w.r.t . the @xmath1363 factor , isomorphic as @xmath1284-modules but lying in two distinct isotypical @xmath196-components
let @xmath284 be the @xmath1284-complement of @xmath1395 in @xmath292 .
as @xmath1284-module , @xmath1396 is the direct sum of @xmath284 and a simple @xmath1284-submodule of @xmath1395 which projects non - trivially on both summands @xmath1397 and @xmath1398 . if @xmath368 , there exist two orbits labelled i and ii .
the case ( i ) can be described as above by specializing @xmath62 equal to @xmath1460 .
the case ( ii ) can be obtained from the case ( i ) conjugating by @xmath1335 .
@xmath1476 , @xmath1477 , @xmath1429 .
if @xmath248 and @xmath1478 , @xmath1430 . if @xmath1427 and @xmath247 , @xmath1479
if @xmath1480 , @xmath1331 .
let us fix a basis @xmath1433 of @xmath1434 and a symmetric bilinear form @xmath106 such that @xmath1317 for all @xmath1318 .
similarly , let us fix a basis @xmath1481 , @xmath1482 of @xmath1483 and a symmetric bilinear form @xmath1435 such that @xmath1436 for all @xmath1318 .
then @xmath1484 and @xmath1485 @xmath1359 @xmath1360 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1487 .
the centralizer of @xmath30 is @xmath1310 where @xmath1488 , the @xmath1363 factor of @xmath1284 is embedded skew - diagonally , @xmath1364 , into the @xmath1365 factor of @xmath196 .
@xmath1490 @xmath1449 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1491 .
the centralizer of @xmath30 is @xmath1383 where @xmath1492 , the @xmath1493 factor of @xmath1284 is embedded as @xmath1453 into @xmath1494 where the @xmath1493 factor is included in the @xmath1495 factor of l , the @xmath1363 factor of @xmath1284 is embedded skew - diagonally @xmath1457 into the @xmath1365 factor of @xmath196 .
the quotient @xmath1392 is a simple @xmath1284-module of dimension @xmath62 as follows . in @xmath1285
there are exactly two simple @xmath1284-submodules , @xmath1393 , of highest weight @xmath1458 w.r.t . the @xmath1363 factor , isomorphic as @xmath1284-modules but lying in two distinct isotypical @xmath196-components .
let @xmath284 be the @xmath1284-complement of @xmath1395 in @xmath292 .
as @xmath1284-module , @xmath1396 is the direct sum of @xmath284 and a simple @xmath1284-submodule of @xmath1395 which projects non - trivially on both summands @xmath1397 and @xmath1398 .
this case can be obtained from the case [ sss : bbaby ] by switching the role of @xmath131 and @xmath132 .
@xmath1498 , @xmath1499 , @xmath1429 .
if @xmath358 and @xmath1428 , @xmath1500 .
if @xmath1501 and @xmath357 , @xmath1431 .
if @xmath358 and @xmath357 , @xmath1502 .
let us fix a basis @xmath1348 of @xmath1349 and a symmetric bilinear form @xmath106 such that @xmath1317 for all @xmath1318 .
similarly , let us fix a basis @xmath1351 of @xmath1352 and a symmetric bilinear form @xmath1435 such that @xmath1436 for all @xmath1318 .
then @xmath1503 and @xmath1357 let us denote by @xmath1335 the linear endomorphism of @xmath1504 switching @xmath1505 and @xmath1506 and fixing all the other basis vectors .
similarly , let us denote by @xmath1507 the linear endomorphism of @xmath1504 switching @xmath1440 and @xmath1441 and fixing all the other basis vectors .
the conjugation by @xmath1335 ( and by @xmath1507 , respectively ) is an involutive external automorphism of @xmath4 , living @xmath6 and @xmath7 stable , and inducing the nontrivial involution of the first ( the second , respectively ) connected component of the dynkin diagram of @xmath6 . if @xmath1508 and @xmath1443 , @xmath1359 @xmath1360 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1509 .
the centralizer of @xmath30 is @xmath1310 where @xmath1510 , the @xmath1363 factor of @xmath1284 is embedded skew - diagonally , @xmath1364 , into the @xmath1365 factor of @xmath196 .
if @xmath325 and @xmath1443 , there exist two orbits labelled i and ii .
the case ( i ) can be described as above by specializing @xmath62 equal to @xmath131 .
the case ( ii ) can be obtained from the case ( i ) conjugating by @xmath1335 . if @xmath1508 and @xmath326 , there exist two orbits labelled i and ii . the case ( i ) can be described as above by specializing @xmath62 equal to @xmath132 . the case ( ii )
can be obtained from the case ( i ) conjugating by @xmath1507 .
if @xmath1511 , there exist four orbits with a double label i or ii .
the case ( i , i ) can be described as above by specializing @xmath62 equal to @xmath1512 . the case ( i , ii ) can be obtained from the case ( i , i ) conjugating by @xmath1507 . the case
( ii , i ) can be obtained from the case ( i , i ) conjugating by @xmath1335 . the case ( ii , ii )
can be obtained from the case ( i , i ) conjugating by @xmath1335 and @xmath1507 . if @xmath1514 and @xmath1447 , @xmath1448 @xmath1449 let @xmath195 be the corresponding parabolic subgroup of @xmath2 , so that @xmath1515 .
the centralizer of @xmath30 is @xmath1383 where @xmath1516 , the @xmath1452 factor of @xmath1284 is embedded as @xmath1453 into @xmath1454 where the @xmath1455 factor is included in the @xmath1456 factor of l , the @xmath1363 factor of @xmath1284 is embedded skew - diagonally @xmath1457 into the @xmath1365 factor of @xmath196 .
the quotient @xmath1392 is a simple @xmath1284-module of dimension @xmath62 as follows . in @xmath1285
there are exactly two simple @xmath1284-submodules , @xmath1393 , of highest weight @xmath1458 w.r.t . the @xmath1363 factor , isomorphic as @xmath1284-modules but lying in two distinct isotypical @xmath196-components .
let @xmath284 be the @xmath1284-complement of @xmath1395 in @xmath292 .
as @xmath1284-module , @xmath1396 is the direct sum of @xmath284 and a simple @xmath1284-submodule of @xmath1395 which projects non - trivially on both summands @xmath1397 and @xmath1398 .
if @xmath1514 and @xmath368 , the normal triple @xmath1459 , the parabolic subgroup @xmath195 and @xmath1284 have the same description , with @xmath1383 . the quotient @xmath1392 remains a simple @xmath1284-module of dimension @xmath1460 but here in @xmath1285 there are exactly three simple @xmath1284-submodules , @xmath1461 , of highest weight @xmath1462 w.r.t .
the @xmath1463 factor , isomorphic as @xmath1284-modules but lying in three distinct isotypical @xmath196-components .
let @xmath284 be the @xmath1284-complement of @xmath1464 in @xmath292 .
as @xmath1284-module , @xmath1396 is the direct sum of @xmath284 and a co - simple @xmath1284-submodule of @xmath1464 which projects non - trivially on every summand @xmath1397 , @xmath1398 and @xmath1465 . if @xmath369 , there exist two orbits labelled i and ii .
the case ( i ) can be described as above by specializing @xmath62 equal to @xmath1517 .
the case ( ii ) can be obtained from the case ( i ) conjugating by @xmath1335 .
this case can be obtained from the case [ sss : ddady ] by switching the role of @xmath131 and @xmath132 .
let @xmath1519 and let @xmath33 be a normal triple containing it .
the action of the semisimple element @xmath265 on @xmath4 induces a @xmath86-grading @xmath1520 , where we denote @xmath1521 = ix \}$ ] .
this defines the _ height _ of @xmath30 ( which actually depends only on @xmath1081 ) , defined as @xmath1522 . by (
* ? ? ? * theorem 2.6 ) , the orbit @xmath1081 is spherical if and only if @xmath1523 .
similarly , one may consider the action of @xmath265 on @xmath7 , and the corresponding @xmath86-grading @xmath1524 , where @xmath1525 .
this defines the @xmath7-_height _ of @xmath30 ( which actually depends only on @xmath1052 ) , defined as @xmath1526 . by ( * ? ? ?
* theorems 5.1 and 5.6 ) , the orbit @xmath1052 is spherical if @xmath1523 , whereas if @xmath1052 is spherical then @xmath1527 and @xmath1528 .
similarly to the adjoint case , @xmath1077 is normal if @xmath1529 , in which case hesselink s proof @xcite of the normality of the closure of a nilpotent adjoint @xmath0-orbit of height 2 essentially applies ( see ( * ? ? ?
* proposition 2.1 ) ) . in tables 211 , for every spherical orbit @xmath1530 ,
we report its signed partition ( column 2 ) , the kostant - dynkin diagram and the height of @xmath1081 ( columns 3 and 4 ) , the kostant - dynkin diagram and the @xmath7-height of @xmath1052 ( columns 5 and 6 ) , the normality of @xmath1062 ( column 7 ) , the codimension of @xmath1531 in @xmath1062 ( column 8) and the weight semigroup of @xmath1061 ( column 9 ) . in the orthogonal cases ,
the generators of the weight semigroups given in the tables are expressed in terms of the following variation of the fundamental weights of an irreducible root system @xmath66 @xmath1532 and we set @xmath1533 . in all cases with a roman numeral , ( i ) or ( ii )
, one @xmath2-orbit is obtained from the other one by applying an involutive automorphism of a factor of @xmath2 of type @xmath361 . in some of these cases ,
the generators of @xmath1067 are given just for the @xmath2-orbit labeled with ( i ) , the generators for the other one are obtained by switching @xmath1534 and @xmath1535 ( resp .
@xmath1536 and @xmath1537 ) if the first ( resp .
the second ) component is the one involved by the mentioned automorphism . which is the component that is involved by the automorphism is clear from the kostant - dynkin diagrams of the two orbits . for every family of @xmath2-orbits
, we draw the luna diagram as it looks like for values of the parameters @xmath1539 big enough with respect to @xmath62 .
when @xmath62 becomes close to @xmath516 , @xmath131 or @xmath132 the diagram may change .
let us explain how it changes .
whenever @xmath2 has a factor of type @xmath1540 , where the diagram ends with @xmath1541 ( the corresponding simple root @xmath1542 moving a color of type @xmath1543 , with @xmath1544 belonging to @xmath114 ) as in case 3.1 , the given diagram is for @xmath1545 .
if @xmath1546 , both the simple roots @xmath1547 and @xmath1548 move colors of type @xmath1543 . if @xmath1549 , with roman numeral ( i ) , @xmath1550 and @xmath1548 moves a color of type @xmath1543 . if @xmath1549 , with roman numeral ( ii ) , @xmath1551 and @xmath1547 moves a color of type @xmath1543 .
for example , the diagram of the case 3.1 becomes + if @xmath1552 @xmath1553 if @xmath1343 ( i ) @xmath1554 if @xmath1343 ( ii ) @xmath1555 whenever @xmath2 has a factor of type @xmath1540 , where the diagram ends with a tail @xmath1556 ( the corresponding simple root @xmath1542 moving a color of type @xmath1543 , with @xmath1544 belonging to @xmath114 and @xmath1557 belonging to @xmath215 ) as in case 7.2 , the given diagram is for @xmath1545 . if @xmath1546 , the simple roots @xmath1547 and @xmath1548 move the same color of type @xmath1543 ( @xmath1558 is a spherical root ) .
for example , the diagram of the case 7.2 for @xmath1559 becomes @xmath1560 whenever @xmath2 has a factor of type @xmath1561 , where the diagram ends with a tail @xmath1562 ( the corresponding simple root @xmath1542 moving a color of type @xmath1543 , with @xmath1544 belonging to @xmath114 and @xmath1563 belonging to @xmath215 ) as in case 7.3 , the given diagram is for @xmath1564 .
if @xmath1549 , the simple root @xmath1548 moves a color of type @xmath1565 ( @xmath1566 is a spherical root ) .
for example , the diagram of the case 7.3 for @xmath369 becomes @xmath1567 | given a classical semisimple complex algebraic group @xmath0 and a symmetric pair @xmath1 of non - hermitian type , we study the closures of the spherical nilpotent @xmath2-orbits in the isotropy representation of @xmath2 . for all such orbit closures ,
we study the normality and we describe the @xmath2-module structure of the ring of regular functions of the normalizations . |
in this paper we want to discuss some general and technical aspects of calculating non - perturbatively renormalization factors of bilinear quark operators by imposing renormalization conditions on off - shell quark green s functions on the lattice @xcite .
this method offers the possibility of computing non - perturbative contributions and all orders of @xmath4 perturbation theory in a simple way in contrast to the enormous efforts to get only lower orders with perturbative methods .
the disadvantages of this numerical renormalization are the need of computer power and at present the systematic uncertainties due to a quenched approximation ( that excludes fermion loops ) , discretization errors @xcite and gauge fixing .
we look here at operators @xcite @xmath5 with quark fields @xmath6 and @xmath7 constructed out of dirac operators @xmath8 and covariant lattice derivatives @xmath9 . + desy 97 - 203 + hub - ep-97/72
the renormalization method we study is a @xmath10-scheme using the lattice as regulator with the lattice spacing @xmath11 as cut - off .
lattice operators @xmath12 and quark fields @xmath13 are renormalized being then only a function of the scale parameter @xmath14 , via : @xmath15 ( @xmath16 is the bare lattice coupling . ) to determine the renormalization constants @xmath17 one imposes conditions on amputated forward quark vertex functions in momentum space with external four - momenta @xmath18 : @xmath19 with a projector method , where @xmath20 $ ] is taken , @xmath21 being a colour@xmath22spin trace , these prescriptions give : @xmath23 } \nonumber \\ \qquad & = & tr [ \big .
\gamma_{\!\!\cal o } \langle q(p ) | { \cal o}(a ) | q(p ) \rangle^{tree}_{amp } ] \big|_{p^2=\mu^2 } \end{aligned}\ ] ] as an optimal choice for the projector , we use , written in a general form : @xmath24 there are two definitions for @xmath25 : usually one projects onto the energy - momentum part @xmath26 ( continuum form : @xmath27 ) of the inverse propagator @xmath28 and gets : + + ( the normalization is chosen to give @xmath29 in the free case . ) another definition for @xmath25 can be derived from eq .
( 1 ) , setting @xmath30 equal to the conserved vector current @xmath31 and using . but
this is only compatible to the first determination of @xmath25 , if the component of @xmath31 transverse to @xmath18 is taken .
( this can be proven by a ward identity assuming , which is true in the continuum and approximately true on the lattice for not too large @xmath32 . )
in a first step monte carlo gluon field configurations @xmath33 have to be generated and numerically fixed to the landau gauge @xcite . the lattice matrix elements in eq .
( 1 ) , non - amputated , can then be calculated in those background fields : @xmath34 ( @xmath35 , ... are space - time points on the lattice . )
the propagators @xmath36 can be computed from a lattice dirac equation with a momentum source , @xmath37 being the fermion matrix : @xmath38 so here the number of matrix inversions is proportional to the number of momenta . but everything else is then there : @xmath39 and the quark propagator @xmath40 with @xmath41 for an amputation and @xmath25 .
the momentum source method automatically performs all the site sums in eq .
another possibility , instead of summing over @xmath42 ( and @xmath43 ) , is to choose a particular location for the operator , for example setting @xmath44 .
translational invariance tells us that this will give the same expectation value after averaging over all configurations . for this method
we need to solve the dirac equation with a point source at @xmath45 and ( for extended operators where @xmath46 ) for a small number of sources in the region around @xmath42 : @xmath47 ( @xmath48 has simply to be fourier transformed to @xmath49 then and we have @xmath50 . ) for local operators and operators with a small number of derivatives the point source method would need fewer inversions , but we see from fig.1 . that relying on translational invariance to carry out the @xmath51 sums leads to much larger error bars
. = 8.00 cm [ fig1 ] = 8.00 cm [ fig2 ]
our calculations are performed on @xmath0 lattices at @xmath52 for hopping parameters @xmath53 , @xmath54 , @xmath55 .
if the method works for current lattices , one expects to find a high enough @xmath56 region where the @xmath57 factors agree with , but are still not destroyed by discretisation effects .
however the available one - loop calculations probably need to be improved for @xmath58 . at low @xmath56 non - perturbative effects and
higher order perturbation theory may have a big influence as can be seen looking at the pseudo - scalar @xmath59 ( fig.2 ) : results for @xmath2 quark masses indicate a non - perturbative pion pole contribution in the chiral limit .
the @xmath57 factors for the operators of the moments of the structure functions , e.g. @xmath60 ( fig.3 ) , seem to show also big non - perturbative or higher order perturbation theory contributions at low @xmath56 .
o(a ) discretization errors are in general expected at higher @xmath56 . however , for the local vector current ( fig.3 ) such errors seem to be there at all scales : it is not constant as expected from its continuum behaviour . in the o(a ) improved theory @xcite it is . in a comparison with one - loop lattice perturbation theory ( dashed lines in figures )
@xcite : + @xmath61 , + @xmath62 and @xmath63 being the anomalous dimension , and with tadpole improved theory ( solid lines in figures ) @xcite we find that the improved theory fits better to the data . for the axial current ( fig.1 ) and the scalar operator ( fig.2 )
there is a good agreement .
we extract @xmath64 .
9 g. martinelli et al . , nucl .
b445 ( 1995 ) 81 .
p. rakow , this conference .
m. gckeler et al . , nucl .
53 ( 1997 ) 81
. m. l. paciello et al . , phys .
b276 ( 1992 ) 163 .
m. gckeler et al . ,
b ( proc . suppl . ) 53 ( 1997 ) 896 .
g. p. lepage and p. b. mackenzie , phys .
d48 ( 1993 ) 2250 . | we have technically improved the non - perturbative renormalization method , proposed by martinelli et al . , by using quark momentum sources and sinks .
composite two - fermion operators up to three derivatives have been measured for wilson fermions and sheikholeslami - wohlert improved fermions in the quenched approximation .
the calculations are performed in the landau gauge on @xmath0 lattices at @xmath1 for @xmath2 @xmath3 values in each case .
the improved sources greatly decrease the statistical noise .
we extract and discuss here renormalization factors for local operators and moments of the structure functions for wilson fermions . |
massive galaxies in the nearby universe are generally comprised of quiescent elliptical and s0 galaxies .
the formation of such massive galaxies has been an active area of study .
recent observations suggest that the properties of these quiescent galaxies ( qgs ) were much different at earlier times .
for example , the sizes of qgs have been found to be much smaller , at fixed stellar mass , at high redshift ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
this implies much higher mass densities within the effective radius for qgs at high redshift .
the size measurements are robust @xcite and the stellar mass measurements are in good agreement with dynamical mass estimates @xcite , confirming the dense nature of qgs at high redshift .
several mechanisms have been proposed to explain the growth in sizes among qgs ( see , e.g. , * ? ? ?
recent discussions have centered on the relative importance of major and minor dissipationless mergers with the latter favored to be the primary channel for size growth ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
while evidence is emerging that these compact qgs represent the cores of local ellipticals @xcite , the progenitors of these compact qgs at even higher redshifts remain a mystery . selecting galaxy samples at or above a fixed stellar mass limit
has provided important insight into the evolution in properties for such populations . however , the connection to the evolution of a typical galaxy over cosmic time is not straightforward given that galaxies grow in stellar mass due to in situ star formation and merging : the progenitors of galaxies that lie just above a stellar mass limit at low redshift would not be counted in a census of high redshift galaxies above the same mass limit since they were likely to be less massive . measuring the structural evolution of a galaxy as it grows in mass
therefore requires a method for linking its progenitors and descendants over cosmic time .
one such method involves selecting galaxies at a constant cumulative number density .
the basic principle behind this method is that the rank ordering of galaxy masses does not change drastically over time .
therefore , if one selects the 10th most massive galaxy in a comoving volume at @xmath2 , it is still likely to be approximately the 10th most massive galaxy in that comoving volume at @xmath10 , but with a higher overall mass due to star formation and merging .
this technique presents a complementary approach to mass - selected studies .
a number density selection has been used in other recent works to study the structural properties @xcite , star formation histories ( sfhs ; * ? ? ?
* ) , and mass growth @xcite of galaxies over cosmic time ( see also , * ? ? ?
the study by @xcite was carried out with ground - based imaging . in this work ,
we build on the results in @xcite and utilize high - resolution _ hubble space telescope _
( _ hst _ ) imaging , allowing us to measure structural properties more accurately and to push to @xmath2 . at these high redshifts
, we also identify the progenitors of @xmath21 galaxies , which in the local universe are bulge - dominated , quiescent systems with large effective radii .
the layout of this paper is as follows . in section [ sec_data ] ,
we discuss the data used for the study . in section [ sec_analysis ]
, we discuss the relevant derived quantities . in section [ sec_cumnum ] ,
we examine the structural assembly of a galaxy as it grows in time to become a @xmath22 galaxy by @xmath10 .
our results are further discussed in section [ sec_discussion ] and we summarize our findings in section [ sec_summary ] . for completeness ,
we briefly overview the structural properties of qgs and star - forming galaxies ( sfgs ) above a constant stellar mass limit in the appendix , as this provides an alternative view to the number density selection in the main part of the paper .
we assume a cosmology with @xmath23 km s@xmath24 mpc@xmath24 , @xmath25 , and @xmath26 .
stellar masses are based on a chabrier initial mass function ( imf ; * ? ? ?
adopting imfs of different forms below @xmath27 @xmath28 ( e.g. , * ? ? ?
* ) would lead , to first order , to an overall scaling of the stellar masses and otherwise identical results .
all magnitudes are given in the ab system .
we study the structural properties of massive galaxies at high redshift ( @xmath5 ) using a combination of data from the ukirt infrared deep sky survey ( ukidss ; * ? ? ?
* ) and the cosmic assembly near - infrared deep extragalactic legacy survey ( candels ; * ? ? ?
we employ a multi - wavelength data set ( @xmath29 and _ spitzer _ irac 3.6 and 4.5 ) in the ultra - deep survey ( uds ) field of ukidss .
this is one of the deepest wide - field near - ir data sets available , making it ideal for constructing samples of galaxies to relatively low stellar masses at high redshift .
the observations and data reduction are described in detail in @xcite and in @xcite and briefly summarized here .
the ukidss uds dr8 data include @xmath30 , which have 5@xmath31 limiting depths in @xmath32 apertures of 24.9 , 24.1 , and 24.5 ab mag , respectively .
the @xmath33 imaging were obtained as part of the subaru-@xmath34 deep survey ( sxds , * ? ? ?
the @xmath35 data were obtained with megacam on cfht ( pi : o. almaini ) .
the _ spitzer _
irac 3.6 and 4.5 data were obtained as part of the _ spitzer_-uds survey ( spuds ; pi : j. dunlop ) .
an updated catalog will be presented in detail in williams et al .
( 2013 , in preparation ) .
objects were detected in the @xmath36-band and photometry carried out in the other bands with matched apertures .
the irac photometry was measured using the point spread function ( psf ) convolution procedure of @xcite .
the irac bands in the uds are crucial for determining photometric redshifts , stellar masses , and rest - frame optical / near - ir colors of galaxies at high redshift .
an analysis of simulated number counts indicates that the catalog is @xmath37 complete at @xmath38 ab mag . to measure structural parameters , we utilize _ hst _ imaging in the uds that was acquired as part of candels @xcite .
the observations and data reduction are presented in @xcite .
we employ @xmath391.0 of the publicly available wfc3 @xmath3 and @xmath4 mosaics .
the psf fwhm in these two bands are @xmath40 and @xmath41 , respectively .
the candels _ hst _ wfc3 data cover roughly @xmath42 deg@xmath43 of the @xmath44 deg@xmath43 uds field with multi - wavelength coverage . in order to assemble a large sample of massive galaxies at low redshift ( @xmath7 ) , we use data in the cosmos field
@xcite . this @xmath45 deg@xmath43 field benefits from multi - wavelength imaging spanning the uv to ir , including wide - field _ hst _ acs coverage .
@xcite assembled a photometric catalog in cosmos incorporating cfht @xmath35 and @xmath36 , subaru @xmath46 as well as 12 intermediate optical bands ( ia427 , ia464 , ia484 , ia505 , ia527 , ia574 , ia624 , ia679 , ia709 , ia738 , ia767 , ia827 ) , ukirt @xmath47 , all four irac channels , and _
galex _ nuv and fuv .
some of these data products are further described in @xcite .
@xcite used the photometry in cosmos to derive stellar mass estimates , which could in principle be used in our analysis .
however , for consistency with the data preparation and stellar masses computed in the uds ( see section [ sec_analysis ] ) , we use a @xmath48-selected catalog , which incorporates most of the data in cosmos described above but reconstructed in a manner consistent with what was done in the uds ( e.g. , * ? ? ?
* ; * ? ? ?
a detailed account of the observations , data reduction , and catalog construction will be described in a future paper ( muzzin et al .
2013 , in preparation ) .
we briefly remark on the most relevant points of the catalog here .
objects were detected in the @xmath48-band , which reached a @xmath49 detection limit of 23.85 ab mag for a @xmath50 aperture .
an analysis of simulated number counts indicates that the 90% completeness is @xmath51 ab mag .
the primary purpose of using data in the cosmos field is the wide - field _ hst _ acs @xmath6 imaging , which we use to measure structural parameters ( see section [ sec_analysis ] ) . in this work
, we utilize the @xmath52 acs @xmath6 imaging @xcite .
the typical psf fwhm is @xmath53 .
the acs imaging employed here covers roughly @xmath54 deg@xmath43 of the cosmos field .
while the data sets used in this work are assembled from two different fields , the analysis carried out on the data is uniform .
photometric redshifts in both cosmos and uds were measured with eazy @xcite . in the uds
, a comparison to spectroscopic redshift measurements suggests that the photometric redshift uncertainties are @xmath55 at @xmath56 . at higher redshifts ,
spectroscopic samples are limited therefore making comparisons to the photometric redshifts difficult . as a result
, losses and contamination from catastrophic outliers remain unquantified .
however , at @xmath57 , where the majority of our sample in the uds lies , @xcite find that the photometric redshift uncertainties are larger based on differences in the photometric redshifts of close pairs @xcite .
sfgs in particular have larger photometric redshift uncertainties . in cosmos
, a comparison to spectroscopic redshifts at @xmath7 indicates that the photometric redshift uncertainties are @xmath58 .
rest - frame @xmath18 and @xmath19 colors were also computed with eazy .
stellar masses were computed with the spectral energy distribution ( sed ) fitting code fast @xcite using exponentially declining sfhs .
* hereafter bc03 ) stellar population synthesis ( sps ) models with a chabrier imf were used in the sed fitting . in the uds ,
@xcite parameterize the stellar mass limit as a function of redshift as @xmath59 . using the same technique to compute the mass limit as @xcite , but focusing solely on galaxies near @xmath60 , we find that a mass limit of @xmath61 @xmath28 encompasses @xmath62 of galaxies .
we adopt this value as our stellar mass limit in the uds at @xmath60 .
as will be made clear in section [ sec_cumnum ] , galaxies in our cosmos sample at @xmath7 are well above the limiting stellar mass at @xmath63 of @xmath64 @xmath28 .
structural parameters were obtained with galfit @xcite , which provided measurements of srsic indices ( @xmath65 ) , effective radii ( @xmath66 ) , and axis ratios ( @xmath67 ) .
the effective radii reported here are circularized , @xmath68 . in order to carry out these measurements at the same rest - frame wavelengths
, we use the _ hst _ imaging that is closest in rest wavelength to @xmath69 .
this choice of @xmath70 represents a tradeoff between ( @xmath27 ) the desire to push to high redshifts , and ( @xmath71 ) the need to probe rest - frame light as redward as possible from the 4000 break , beyond which the seds of galaxies are generally smooth and serve as a closer tracer of the stellar mass . as a result , in cosmos
, we use acs @xmath6 imaging at @xmath7 , while in the uds we use wfc3 @xmath3 imaging at @xmath72 and wfc3 @xmath4 imaging at @xmath73 . for a given object ,
the nearest star is chosen to serve as the psf model when running galfit .
nearby objects were masked .
no constraints were placed on the range of srsic indices .
the axis ratios were constrained to have @xmath74 .
the semi - major axis , @xmath75 , was constrained to be smaller than the box size .
our results and conclusions are dependent on the quality of the structural parameters measured from galfit .
we ran simulations to test the reliability of these measurements at the highest redshifts , where objects become faint in the wfc3 @xmath4 imaging .
we created @xmath76 mock galaxies from a range of srsic models , varying @xmath66 , @xmath65 , @xmath67 , and the magnitude .
these models were added to different regions of the wfc3 @xmath4 imaging that contained blank sky .
we then processed the mock images in the same manner as used to derive our measurements above .
the sample in our highest redshift bin ( @xmath77 ) , defined in section [ sec_cumnum ] , reaches magnitudes of @xmath78 ab mag and @xmath79 ab mag at the 50th and 90th percentile , respectively .
our simulations show that we recover @xmath66 , @xmath65 , and @xmath67 to precisions of @xmath80 , @xmath81 , and @xmath82 for @xmath78 ab mag and @xmath83 , @xmath84 , and @xmath81 for @xmath85 ab mag .
systematic offsets are @xmath86 for all three parameters .
effective radii as small as @xmath87 ( i.e. , less than the fwhm/2 ) , or @xmath88 kpc at @xmath89 , are recovered to similar precisions as noted above , consistent with other works ( e.g. , * ? ? ?
our results and conclusions based on structural parameters measured from galfit are therefore not strongly biased at high redshift in any way .
as one might expect at higher redshifts , the higher resolution _ hst _ imaging reveals that a small portion of the objects detected in the ground - based @xmath36-band imaging are comprised of two or more objects blended together .
we remove galaxies in the uds from our analysis ( @xmath90 ) if nearby objects contribute more than @xmath91 of the total flux within the @xmath32 color aperture , as the photometric redshifts , rest - frame colors , and stellar masses are not as reliable .
we note that this procedure may introduce a small bias against galaxies in close pairs .
we define our sample in this section by selecting galaxies at a fixed cumulative number density , @xmath8 .
we then study the structural properties of galaxies in narrow mass bins at this value of @xmath8 at different redshifts .
the analysis that follows is overall similar to that of @xcite but differs in a few aspects .
first , we study galaxies at a constant _ cumulative _ number density .
second , our analysis extends to @xmath2 .
finally , our work employs _ hst _ imaging as opposed to ground based , allowing us to more accurately measure structural properties , especially at high redshift .
the deep _ hst _ imaging also allows us to characterize the properties of individual galaxies as opposed to stacks , which were employed in @xcite .
figure [ fig_cumnum_selection](a ) shows the cumulative number density of galaxies at different redshifts .
these curves were derived using the mass functions of @xcite , which were computed with data from several fields and as a result minimized uncertainties from cosmic variance .
in addition , @xcite carefully accounted for completeness limits in stellar mass in deriving their mass functions .
we used best - fit schechter parameters from set 7 in that work , which used solar metallicity bc03 models with a chabrier imf and a @xcite dust law to determine stellar masses , similar to the sfhs employed in this work .
we derived mass functions for our redshift bins by interpolating between the best - fit schechter functions computed by @xcite , which cover @xmath92 . in carrying out the interpolation
, we also included the @xcite mass function at @xmath93 , as reported in @xcite but scaled to a chabrier imf .
we then integrated these mass functions , @xmath94 , to determine the cumulative number density of galaxies as a function of stellar mass at different redshifts : @xmath95 we chose a cumulative number density for our study of @xmath0 mpc@xmath1 ( dashed line ) as this value represents the number density of galaxies with stellar masses slightly above the stellar mass limit at @xmath60 ( @xmath61 @xmath28 ) .
we also note that the @xcite mass functions are complete well below this value at these redshifts .
as seen in figure [ fig_cumnum_selection](a ) , the selected value of @xmath8 intersects the cumulative number density curves at lower redshifts at higher stellar masses , tracing out the mass growth at that particular number density .
the corresponding stellar masses at @xmath0 mpc@xmath1 are shown as a function of redshift in figure [ fig_cumnum_selection](b ) . in order to quantify the redshift dependence
we fit a second - order polynomial to these data points , resulting in the following relation between stellar mass and redshift : @xmath96 the scatter about this relation is only @xmath97 dex , suggesting that this parameterization is adequate for the redshift range studied here .
note that this scatter does not reflect the systematic uncertainty in measuring stellar masses of galaxies at high redshift , which can be substantial ( e.g. , * ? ? ?
poisson uncertainties in the schechter parameters computed by @xcite propagate into an uncertainty in the derived stellar mass in equation ( [ eq_mass_redshift ] ) at a given redshift of @xmath98 dex .
for the high - mass end at lower redshifts ( @xmath99 ) , where cosmic variance is likely an important factor , an additional uncertainty of up to @xmath98 dex may be warranted . for a given redshift
, we study the properties of galaxies within a bin of size @xmath100 dex in stellar mass centered on the predicted mass from equation ( [ eq_mass_redshift ] )
. the actual boundaries of the bin are adjusted such that the median mass is close to the value given by equation ( [ eq_mass_redshift ] ) . given the steepness of the mass function , in practice this results in selecting galaxies at @xmath101 .
the bin size is broad enough to allow for robust measurements of median structural parameters . given the narrow redshift and mass bins employed , scattering of galaxies into and out of the sample due to photometric uncertainties is unavoidable and is a larger effect at higher redshifts ( e.g. , @xmath102 ) where monte carlo simulations of the photometry suggest uncertainties in the redshifts of @xmath103 and in the stellar masses of @xmath104 dex . as a consequence , within a given redshift and mass bin at @xmath102 , just under half of the original sample is recovered in our simulations while the remainder is made up of galaxies near the bin boundaries and therefore display similar properties to that of the original sample . in order to avoid confusion
, we emphasize that at a given redshift we are _ not _ selecting _ all _ galaxies with masses above the mass limit implied by the given value of @xmath8 , but instead , we are selecting galaxies in a narrow mass bin _ at _ the mass determined by @xmath8 .
equation ( [ eq_mass_redshift ] ) indicates that galaxies at @xmath0 mpc@xmath1 grow by a factor of @xmath9 from @xmath105 to @xmath10 , resulting in a galaxy at low redshift with a stellar mass of @xmath106 @xmath28 ( i.e. , @xmath21 ) . from @xmath107 to @xmath108 , equation ( [ eq_mass_redshift ] ) predicts that the stellar mass grows by a factor of @xmath45 , which is similar to what is found in @xcite .
we note that the stellar mass growth inferred from our purely observational motivated method ( equation ( [ eq_mass_redshift ] ) ) is less than what is predicted from abundance matching techniques ( e.g. , * ? ? ?
* ) , though the latter analysis is quite uncertain at @xmath109 .
more recent efforts that combine dark matter merger trees with observational constraints indicate similar mass growth to what is found here at @xmath110 @xcite . at @xmath10
, the latter work indicates that a @xmath21 galaxy occupies a dark matter halo of mass @xmath111 @xmath28 , which is typical of galaxy groups .
both @xcite and @xcite show with simulations that in selecting galaxies at a fixed number density , the completeness fraction declines with cosmic time , meaning that some of the objects selected in a given number density bin at high redshift are no longer found in that bin at lower redshift .
contaminants from other number density bins also enter the sample .
however , most of the contaminants scatter in from neighboring bins and likely display properties that are very similar to those of galaxies in the number density bin of interest .
finally , we note that although we use mass functions from @xcite , which are based on different data from what is employed here , we arrive at qualitative and quantitative conclusions that are quite similar to what was found in @xcite .
this suggests the systematic uncertainties as a result of this choice are minimal .
lcccccc @xmath112 & 120 & 11.16 & @xmath113 & @xmath114 & @xmath115 & @xmath116 + @xmath117 & 448 & 11.12 & @xmath118 & @xmath119 & @xmath120 & @xmath121 + @xmath56 & 20 & 11.05 & @xmath122 & @xmath123 & @xmath123 & @xmath124 + @xmath125 & 41 & 10.95 & @xmath126 & @xmath127 & @xmath128 & @xmath129 + @xmath130 & 37 & 10.84 & @xmath131 & @xmath132 & @xmath133 & @xmath134 + @xmath77 & 35 & 10.70 & @xmath135 & @xmath132 & @xmath136 & @xmath137 we first examine how the star formation properties of galaxies have evolved since @xmath2 .
as is well known , galaxies can be classified in two distinct categories : star forming and quiescent , at least out to @xmath2 @xcite .
figure [ fig_uvj_ncum ] shows the rest - frame @xmath18 versus @xmath19 colors of galaxies in different redshift bins selected at @xmath0 mpc@xmath1 .
this @xmath138 diagram is commonly used to separate qgs from sfgs ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is preferred over a color - magnitude or a color - mass selection because of its ability to separate red galaxies that are quiescent from reddened sfgs . shown in each redshift panel
are galaxies within a narrow mass bin around the predicted stellar mass from equation ( [ eq_mass_redshift ] ) for galaxies at @xmath8 .
the typical mass is therefore increasing toward low redshift .
the subset of galaxies with measured structural parameters from the _ hst _ imaging are color - coded according to their srsic index .
the @xcite boundary distinguishing qgs ( top left ) from sfgs ( bottom right ) is shown for each redshift bin . at @xmath102 ,
we use the boundary condition defined at @xmath139 since @xcite provide selection criteria up to those redshifts .
we also slightly modify the diagonal boundary at @xmath117 to better accommodate the cosmos data using the following relation : @xmath140 .
small modifications in the @xmath138 selection such as this are not uncommon ( e.g. , * ? ? ?
* ) and possibly reflect variations in the observed filter set and sed templates used to derive the rest - frame colors . based on the @xmath138 selection , figure [ fig_cumnum_qg_frac ] shows the fraction of qgs and sfgs as a function of redshift .
the error bars in figure [ fig_cumnum_qg_frac ] at @xmath109 are larger due to the smaller sample size of the uds data .
below @xmath141 , the wide area cosmos data allow us to better constrain the properties of the most massive galaxies .
the qg fraction increases from @xmath83 at @xmath142 to @xmath143 at @xmath144 .
we note that our quiescent and star - forming fractions are in good agreement with those of @xcite who reported on galaxies to @xmath11 ( open symbols in figure [ fig_cumnum_qg_frac ] ) .
together , figures [ fig_uvj_ncum ] and [ fig_cumnum_qg_frac ] show that the progenitors of nearby massive galaxies at @xmath2 were likely to be star forming given that their colors coincide with those of sfgs in @xmath138 color space . from @xmath2 to @xmath20 there appears to have been a substantial buildup in the population of qgs , which become the dominant population at @xmath99 for galaxies selected at @xmath8 . at @xmath145 ,
the population of qgs continues to grow .
we examine how the structural properties have evolved in the next section .
we now turn to the structural properties of massive galaxies as they grow in time .
the general properties of galaxies selected at @xmath8 are summarized in table [ table_ncum ] .
figure [ fig_cumnum_structure ] shows the evolution of the median effective radius , srsic index , and axis ratio for galaxies at @xmath0 mpc@xmath1 .
we include in these figures a sample of sdss galaxies at @xmath146 from @xcite selected at the appropriate mass from equation ( [ eq_mass_redshift ] ) . the sdss data is not used in any of the fits that follow .
below @xmath13 , where qgs become the dominant population in the sample , figure [ fig_cumnum_structure](a ) shows that the effective radius increases substantially from @xmath15 kpc at @xmath11 to @xmath147 kpc at @xmath10 .
the size evolution at @xmath148 follows @xmath149 with the exponent being consistent with the value of @xmath150 found in @xcite over roughly the same redshift range .
a striking feature in figure [ fig_cumnum_structure](a ) is the lack of evolution at @xmath151 in the median effective radius .
we investigate this further by showing the evolution of qgs and sfgs separately in figure [ fig_cumnum_structure ] .
the constant median @xmath66 arises because sfgs are larger than qgs and their relative abundance changes as a function of redshift . above @xmath152 ,
the size evolution is likely determined almost solely by sfgs since they become an overwhelming majority of the population .
we therefore expect the sizes of galaxies to decrease above @xmath153 ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) for samples selected at @xmath8 .
we can test whether the apparent constant value of @xmath66 at @xmath151 is a generic feature or a consequence of the particular value of @xmath8 selected for our study . at lower values of @xmath8 ( i.e. , higher masses at a given redshift )
, we find that @xmath66 can increase gradually from @xmath2 to @xmath11 .
the srsic index determines the distribution of light and hints at the presence of a bulge or disk . the median srsic index in figure [ fig_cumnum_structure](b ) increases from @xmath16 at @xmath105 to @xmath154 at @xmath10 .
the srsic index evolution at @xmath155 can be characterized by @xmath156 which is consistent with the exponent of @xmath157 found in @xcite .
the srsic index evolution indicates that while most of the stars in @xmath21 galaxies in the nearby universe are distributed in a bulge , the stars in their progenitor galaxies at @xmath102 were distributed in structures resembling exponential disks .
we note that @xcite also find that sfgs at high redshift , which represent the majority at @xmath8 , generally have low srsic indices around @xmath16 .
the fact that qgs at @xmath89 in figure [ fig_uvj_ncum ] generally have higher srsic indices than sfgs ( see also , * ? ? ?
* ) , as is also the case at lower redshifts , further suggests that our srsic profile fitting measurements are not significantly biased by the limiting depth of the _ hst _ imaging for higher redshift galaxies . in the appendix
, we show this to also be the case with a much larger , stellar mass limited sample ( figure [ fig_uvj ] ) .
while the srsic indices can be suggestive of a bulge or disk component , the axis ratio distribution provides a better constraint on the shapes of galaxies . owing to the high resolution of the _ hst _ imaging
, we can examine the axis ratios of galaxies selected at @xmath8 to @xmath2 .
the median axis ratio of galaxies at @xmath0 mpc@xmath1 has increased significantly since @xmath2 . at @xmath105 ,
the typical axis ratio is @xmath17 , a low value that is indicative of a distribution of randomly oriented thin disks .
meanwhile , at @xmath146 , the axis ratio is @xmath158 , closer to what is expected for elliptical galaxies .
this value is in good agreement with sdss studies of massive qgs at @xmath146 @xcite . at the highest redshifts ( @xmath77 ) , the residuals to the single component srsic profile fits are smooth and visual inspection of these residuals suggests that the lower axis ratios at high redshift are not generally driven by multiple components .
in addition , varying the number density selection to lower values , such as @xmath159 mpc@xmath1 , does not impact the general decreasing trend of the median axis ratios towards higher redshifts .
finally , as seen in figure [ fig_cumnum_structure](c ) the axis ratios of sfgs are generally lower than that of qgs at a given redshift and sfgs increasingly become the dominant population at @xmath57 .
the axis ratios therefore also indicate , in addition to the srsic indices , that the stars in the progenitors of @xmath21 galaxies were distributed in disks at @xmath2 .
figure [ fig_psncum ] shows example postage stamps of galaxies selected at @xmath8 at different redshifts . for illustrative purposes , we selected galaxies with structural parameters that follow the general trends seen in figure [ fig_cumnum_structure ] but with more consideration for the trend in axis ratios .
each postage stamp is @xmath160 kpc on a side and note that the size of the galaxy in each redshift bin is larger than the psf .
the relative sizes between redshift bins are therefore easy to compare in this figure when paired with the indicated half - light ellipses . toward lower redshifts , especially at @xmath13 , figure [ fig_psncum ] shows how light is added to the outer parts , leading to the increasing size and srsic index of galaxies selected at @xmath8 . at higher redshifts , the axis ratios decrease , suggesting a larger contribution from disks . synthesizing the structural information above with the star formation activity discussed in section [ sec_cumnum_uvj ] , we see that the progenitors of @xmath21 galaxies at @xmath2 were star - forming disks with @xmath15 kpc .
by @xmath20 , many of these star - forming disks disappear and give rise to a population of compact qgs . by @xmath10 ,
these compact qgs evolved into the large - sized , bulge dominated , quiescent @xmath21 galaxies . a more intuitive view of how the change in structural properties has impacted the evolution of galaxies at @xmath0 mpc@xmath1 is given by the evolution of the mass surface density profiles in figure [ fig_cumnum_mass_profile](a ) .
in order to compute these profiles , we first used the best fitting srsic index and effective radius for each galaxy to determine the light profile using the standard formula for a srsic profile : @xmath161 ) , \label{eq_sersic}\end{aligned}\ ] ] where @xmath162 is the surface brightness at radius @xmath163 , @xmath66 the half - light radius , @xmath164 the surface brightness at @xmath66 , @xmath65 the srsic index , and @xmath165 a constant that depends on @xmath65 .
@xcite find that the surface brightness profiles of galaxies at high redshift generally follow srsic profiles quite well . for simplicity
, we therefore use the analytic representation of the profile for each galaxy .
these light profiles were converted into mass profiles by normalizing the integrated light in the srsic profile to the stellar mass of each galaxy .
note that this conversion neglects radial gradients in the mass - to - light ratio . for a given redshift bin ,
the median of the mass profiles was computed at each radius resulting in the profiles shown in figure [ fig_cumnum_mass_profile](a ) . the uncertainty in the median of the mass profile at a given radius
was computed by bootstrapping the sample .
we note that one - dimensional srsic fits to these median profiles recover the srsic index and half - light radius to within @xmath166 of the median values reported in table [ table_ncum ] .
a naive interpretation of figure [ fig_cumnum_structure](b ) would be that the bulges of galaxies grow in time given the increase in the srsic index , a crude proxy for the bulge - to - disk ratio ( e.g. , * ? ? ?
* ) . however , the mass profiles in figure [ fig_cumnum_mass_profile](a ) generally overlap at small radii and diverge at large radii , suggesting a buildup of mass in the outer parts of the galaxy with time .
figure [ fig_cumnum_mass_profile](b ) shows the cumulative proportion of mass assembled at different radii relative to the total mass within @xmath167 kpc of the median galaxy at @xmath146 .
roughly @xmath168 of the total mass of the galaxy is assembled within @xmath169 kpc at @xmath146 , as expected given that @xmath147 kpc at that redshift . at @xmath170 , the assembled mass within @xmath169
kpc is @xmath171 of the total mass at @xmath146 indicating that much of the mass within @xmath169 kpc was already in place @xmath172 gyr ago .
note that the small sample in the @xmath56 bin likely leads to this curve falling slightly above the @xmath173 data at @xmath174 kpc . in figure [ fig_cumnum_mass_rinout ]
, we compare the mass growth between the central and outer regions of galaxies selected at @xmath0 mpc@xmath1 .
the total stellar mass as a function of redshift is shown by the black line , while the projected mass inside and outside of @xmath175 kpc is given by the red and blue lines , respectively .
these values are determined by integrating equation ( [ eq_sersic ] ) as follows : @xmath176 where @xmath177 and @xmath178 are the inner and outer radii enclosing the mass , @xmath179 . for the central regions ( @xmath180 kpc , @xmath181 kpc )
, the stellar mass appears to grow from @xmath2 to @xmath11 but then levels off around @xmath182 @xmath28 .
in contrast , in the outer regions ( @xmath183 kpc , @xmath184 kpc ) mass continues to build up over the entire redshift range studied , growing by a factor of @xmath9 .
the stellar mass that has been added to the outer parts of galaxies over time is therefore the dominant source of assembled mass , as the central parts appear to have been assembled by @xmath11 .
increasing the value of @xmath177 so as to avoid the central regions of the srsic fit ( e.g. , @xmath185 kpc ) , which can be less secure , does not qualitatively impact the latter result .
the results above are in qualitative agreement with those of @xcite .
an alternative projection of figure [ fig_cumnum_mass_rinout ] is shown in figure [ fig_cumnum_radii_fixedmass ] .
this figure shows the radius enclosing a given stellar mass as a function of redshift ( analogous to figure 1 in * ? ? ?
* ) for galaxies at @xmath0 mpc@xmath1 .
it is the equivalent of taking horizontal cuts in figure [ fig_cumnum_mass_profile](b ) , and scaling by a stellar mass .
the figure therefore depicts the evolution in the `` onion layering '' of stellar mass .
the lines of constant mass terminate in high - redshift bins where the given mass had not yet been assembled .
below @xmath13 , the radius enclosing a given mass remains roughly constant .
this again highlights that new stellar mass growth below @xmath13 occurs at larger radii and that the mass within the inner parts remains roughly unchanged . between @xmath2 and @xmath11 , the radius enclosing a given stellar mass appears to decrease by roughly a factor of @xmath45 .
in section [ sec_cumnum ] , we have shown that as a galaxy grows in stellar mass by a factor of @xmath9 from @xmath186 @xmath28 at @xmath105 to @xmath106 @xmath28 at @xmath10 , most of the new stellar mass that is added contributes toward mass growth at larger radii as one moves toward lower redshifts .
the mass profiles presented in figure [ fig_cumnum_mass_profile ] highlight this point as they show that most of the mass in the core is in place by @xmath11 . below @xmath11 ,
figure [ fig_cumnum_mass_rinout ] shows how mass growth continues in the outer parts to the present epoch . as a consequence ,
the effective radius of the galaxy grows from @xmath187 kpc at @xmath188 to @xmath189 kpc at @xmath10 .
fitting the mass and effective radius evolution at @xmath13 , we find that @xmath190 , which is almost exactly the relation found in @xcite at @xmath191 . the high resolution of the _ hst _ imaging allowed us to probe the mass distribution of galaxies at @xmath102 at very small radii .
this was not possible in the @xcite study as it relied on ground - based imaging . with the higher resolution _ hst _ imaging , we find that the mass within @xmath175 kpc , which is roughly the median effective radius for a galaxy selected at @xmath8 at @xmath102 , remains roughly constant at @xmath13 .
thus , there does not appear to be any substantial growth at @xmath13 in the central part of the bulge that will characterize this galaxy at @xmath10 . from @xmath2 to @xmath11 , there is evidence for mass growth in the inner part of the galaxy , as seen in figures [ fig_cumnum_mass_rinout ] and [ fig_cumnum_radii_fixedmass ] .
we note , however , that selecting galaxies at different values of @xmath8 can result in a more constant enclosing radius across redshift for a given mass .
larger samples at @xmath102 may help in clarifying the potential mass buildup in the inner parts of galaxies at those early times .
dissipationless minor mergers are thought to play a significant role in the buildup of mass for qgs .
in addition , such processes are predicted to assemble mass at large radii , thereby contributing toward the size growth of galaxies ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
for example , using a cosmological hydrodynamical simulation , @xcite suggest that a @xmath10 galaxy with @xmath106 @xmath28 grew by a factor of @xmath9 after @xmath2 , primarily through dry minor mergers .
this mass growth is almost exactly what is found in this work for the same mass descendant at @xmath10 . over this 11 gyr timespan ,
the galaxy grew in size by a factor of @xmath192 , also in agreement with our results .
one caveat in this comparison is that the overall sizes of the simulated galaxy are a factor of @xmath193 lower than what is found in this work .
with more recent simulations , @xcite show that the relation @xmath194 ( see also , * ? ? ?
* ) , found in this work and in @xcite , is most easily reproduced by @xmath195 mergers with mass ratios of 1:5 .
in contrast , major mergers lead to a close to linear dependence of @xmath66 on @xmath179 , which is not supported by the observations .
it remains to be seen whether these mergers are actually observed ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) . in the local universe ,
the most massive systems are generally bulge dominated elliptical galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) with large effective radii ( e.g. , * ? ? ?
* ) . in this work and in @xcite
, we showed that the progenitors of these local ellipticals at @xmath196 , are primarily compact qgs . as discussed in the previous section , these compact qgs
are generally considered the cores of ellipticals at @xmath10 ( see also , * ? ? ?
* ) , growing in mass and size potentially through dissipationless minor mergers to match the properties , namely sizes , of local ellipticals . with the _ hst _ imaging and the deep - ir uds data , our number density selection allows us to the trace the properties of progenitors of local ellipticals to @xmath2 .
at these redshifts , figures [ fig_uvj_ncum ] and [ fig_cumnum_qg_frac ] show that most of the progenitors are sfgs .
the srsic indices in figure [ fig_cumnum_structure](b ) suggest that these sfgs have exponential profiles , which is typically associated with the surface brightness profiles of disks .
the srsic index alone , however , is not definitive in defining the shapes of these progenitors at @xmath2 .
instead , the most compelling evidence comes from the axis ratios shown in figure [ fig_cumnum_structure](c ) . at @xmath10
, galaxies selected at @xmath0 mpc@xmath1 have median observed axis ratios of @xmath197 .
the implied intrinsic axis ratio is roughly 2:3 @xcite , indicative of spheroidal systems .
meanwhile , at @xmath2 , figure [ fig_cumnum_structure](c ) shows that galaxies at @xmath8 have a median axis ratio of @xmath17 .
this value for the axis ratio is very low considering that for a population of randomly oriented , infinitely thin disks , the median axis ratio would be @xmath198 .
it is therefore likely that the progenitors of massive , quiescent , local elliptical galaxies at @xmath2 are star - forming disks . many of the star - forming disks at @xmath2 disappear over the redshift range @xmath151 and give way to compact qgs , as seen in figure [ fig_uvj_ncum ] .
while the details of this transition require further investigation , we note that for a much broader mass range than what is considered here , @xcite identify a subset of the star forming population at @xmath89 with similar structure to that of compact qgs ( see also , * ? ? ?
we used _ hst _ imaging to study the structural properties of galaxies selected at a constant cumulative number density of @xmath0 mpc@xmath1 at redshifts of @xmath155 .
this selection allowed us to trace the evolution of galaxies with stellar mass @xmath199 @xmath28 at @xmath105 , as they grew by a factor of @xmath9 to become @xmath21 galaxies in the local universe .
this work builds on the previous analysis by @xcite , who also selected galaxies at a constant number density . here ,
we employ high - resolution _ hst _ imaging and extend the analysis to @xmath2 .
in contrast to mass - selected studies , our selection at a constant cumulative number density allows for a more straightforward evolutionary link between progenitors and descendants at different redshifts . at @xmath5
, we employed candels wfc3 @xmath3 and @xmath4 imaging in the uds , while at @xmath7 we used wide - field acs @xmath6 imaging in cosmos to fit single - component srsic profiles at a common rest - frame wavelength .
the resulting srsic indices , effective radii , and axis ratios were used to aid in our analysis . the uniform data sets and analysis methods carried out in this work serve to minimize systematics .
our main conclusions are the following : 1 .
the typical galaxy at the selected value of @xmath8 has grown in effective radius by a factor of @xmath192 , mostly since @xmath11 ( figure [ fig_cumnum_structure ] ) .
the evolution in the stellar mass surface density profiles of galaxies selected at @xmath8 indicates that most of the mass in the central regions was in place by @xmath11 , while almost all of the new mass growth took place in the outer parts ( figures [ fig_cumnum_mass_profile]-[fig_cumnum_radii_fixedmass ] ) .
this inside - out mass growth is responsible for the increase in size and srsic index toward low redshift .
3 . at @xmath13
, we find that as the stellar mass builds up , the effective radius grows as @xmath14 , in excellent agreement with @xcite .
recent simulations show that such a dependence is consistent with mergers with mass ratios of 1:5 being responsible for much of the size growth @xcite .
4 . at @xmath2 ,
the rest - frame @xmath138 colors , srsic indices , and axis ratios indicate that the progenitors of present day massive galaxies were star - forming disks with @xmath200 kpc and a third of the @xmath10 stellar mass . at @xmath201 , these galaxies doubled in stellar mass and were mostly compact qgs .
these galaxies evolved further into the @xmath21 galaxies in the local universe that are known to be quiescent , bulge dominated , elliptical galaxies with large effective radii .
we thank the anonymous referee for helpful comments and suggestions .
we also thank daniel szomoru for helpful discussions .
we acknowledge funding from erc grant highz 227749 .
this research was supported by an nwo - spinoza grant .
67 natexlab#1#1 , g. , faber , s. m. , perez - gonzalez , p. g. , et al .
2012 , arxiv : 1206.5000 behroozi , p. s. , wechsler , r. h. , & conroy , c. 2012 , arxiv:1207.6105 , e. f. , van der wel , a. , papovich , c. , et al .
2012 , , 753 , 167 , r. , van dokkum , p. g. , tal , t. , et al .
2009 , , 697 , 1290 , f. , jog , c. j. , & combes , f. 2007 , , 476 , 1179 , g. b. , van dokkum , p. g. , & coppi , p. 2008
, , 686 , 1503 , g. b. , whitaker , k. e. , van dokkum , p. g. , et al .
2011 , , 739 , 24 , g. & charlot , s. 2003 , , 344 , 1000 , d. , armus , l. , bohlin , r. c. , et al .
2000 , , 533 , 682 , p. , aussel , h. , ajiki , m. , et al .
2007 , , 172 , 99 , m. , di serego alighieri , s. , cimatti , a. , et al .
2009 , , 704 , l34 , a. j. & trujillo , i. 2009 , , 696 , l43 , g. 2003 , , 115 , 763 , a. , cassata , p. , pozzetti , l. , et al .
2008 , , 482 , 21 , s. , norberg , p. , baugh , c. m. , et al .
2001 , , 326 , 255 , c. & wechsler , r. h. 2009 , , 696 , 620 , e. , renzini , a. , pirzkal , n. , et al . 2005 , , 626 , 680 , j. , kuhlen , m. , & madau , p. 2007
, , 667 , 859 , m. , van dokkum , p. g. , schreiber , n. m. f. , et al .
2008 , , 688 , 770 , n. a. , kocevski , d. d. , faber , s. m. , et al .
2011 , , 197 , 35 hilz , m. , naab , t. , & ostriker , j. p. 2013
, , 429 , 2924 , b. p. , van der wel , a. , rix , h .- w . , & franx , m. 2012 , , 749 , 96 , p. f. , bundy , k. , hernquist , l. , wuyts , s. , & cox , t. j. 2010 , , 401 , 1099 , p. f. , bundy , k. , murray , n. , et al .
2009 , , 398 , 898 , o. , capak , p. , salvato , m. , et al .
2009 , , 690 , 1236 , o. , salvato , m. , le floch , e. , et al .
2010 , , 709 , 644 , a. m. , aussel , h. , calzetti , d. , et al .
2007 , , 172 , 196 , a. m. , faber , s. m. , ferguson , h. c. , et al .
2011 , , 197 , 36 , m. , van dokkum , p. g. , labb , i. , et al .
2009 , , 700 , 221 , i. , bouwens , r. , illingworth , g. d. , & franx , m. 2006 , , 649 , l67 , c. n. & gunn , j. e. 2012 , , 421 , 2277 laporte , c. f. p. , white , s. d. m. , naab , t. , ruszkowski , m. , & springel , v. 2012 , , 424 , 747 , a. , warren , s. j. , almaini , o. , et al .
2007 , , 379 , 1599 , a. & peebles , p. j. e. 2003 , , 589 , 29 , d. , van dokkum , p. g. , frster schreiber , n. m. , et al .
2009 , , 701 , 1765 , r. , stoughton , c. , leauthaud , a. , et al .
2010 , , 401 , 371 , m. , williams , r. j. , franx , m. , et al .
2012 , ( arxiv : 1207.6634 ) , t. , johansson , p. h. , & ostriker , j. p. 2009
, , 699 , l178 , a. b. , ellis , r. s. , bundy , k. , & treu , t. 2012 , , 746 , 162 , p. a. , bouwens , r. j. , carollo , c. m. , et al .
2010 , , 709 , l21 , c. , finkelstein , s. l. , ferguson , h. c. , lotz , j. m. , & giavalisco , m. 2011 , , 412 , 1123 , s. g. , holden , b. p. , kelson , d. d. , et al .
2012 , , 748 , l27 , s. g. , kelson , d. d. , holden , b. p. , franx , m. , & illingworth , g. d. 2011 , , 735 , 53 , c. y. , ho , l. c. , impey , c. d. , & rix , h .- w .
2002 , , 124 , 266 , r. f. & williams , r. j. 2010 , , 725 , 794 , r. f. , williams , r. j. , franx , m. , & hildebrandt , h. 2012 , , 744 , 88 , n. , abraham , r. g. , aussel , h. , et al .
2007 , , 172 , 38 , n. , aussel , h. , brusa , m. , et al .
2007 , , 172 , 1 , k. & sxds team .
2004 , in bulletin of the american astronomical society , vol .
36 , american astronomical society meeting abstracts , 1478 , s. , mo , h. j. , white , s. d. m. , et al .
2003 , , 343 , 978 stefanon , m. , marchesini , d. , rudnick , g. h. , brammer , g. b. , & whitaker , k. e. 2013 , arxiv:1301.7063 , d. , franx , m. , & van dokkum , p. g. 2012 , , 749 , 121 , d. , franx , m. , van dokkum , p. g. , et al .
2010 , , 714 , l244 szomoru , d. , franx , m. , van dokkum , p. g. , et al .
2013 , , 763 , 73 , s. , van dokkum , p. , franx , m. , et al .
2007 , , 671 , 285 , j. , kriek , m. , franx , m. , et al .
2011 , , 736 , l9 , a. , holden , b. p. , zirm , a. w. , et al .
2008 , , 688 , 48 , a. , rix , h .- w .
, holden , b. p. , bell , e. f. , & robaina , a. r. 2009 , , 706 , l120 , a. , rix , h .- w . , wuyts , s. , et al .
2011 , , 730 , 38 , p. g. & conroy , c. 2010 , , 468 , 940 , p. g. , franx , m. , kriek , m. , et al .
2008 , , 677 , l5 , p. g. , kriek , m. , & franx , m. 2009 , , 460 , 717 , p. g. , whitaker , k. e. , brammer , g. , et al .
2010 , , 709 , 1018 , k. e. , labb , i. , van dokkum , p. g. , et al .
2011 , , 735 , 86 , r. j. , quadri , r. f. , & franx , m. 2011 , , 738 , l25 , r. j. , quadri , r. f. , franx , m. , van dokkum , p. , & labb , i. 2009 , , 691 , 1879 , r. j. , quadri , r. f. , franx , m. , et al . 2010 , , 713 , 738 , s. , frster schreiber , n. m. , van der wel , a. , et al .
2011 , , 742 , 96 , s. , labb , i. , franx , m. , et al .
2007 , , 655 , 51 , a. w. , van der wel , a. , franx , m. , et al .
2007 , , 656 , 66
the primary purpose of the paper was to examine the structural properties of galaxies selected at a constant cumulative number density .
in doing so , we were able to trace the evolution of @xmath21 galaxies from their progenitors at @xmath2 .
here we show results for galaxies selected above a constant mass limit , a more commonly used selection that provides an alternative view to the number density selection .
we distinguish qgs and sfgs with a @xmath138-selection , as these two subpopulations are known to have differing properties . as the mass selection allows for a larger sample
, we can confirm some of the general findings in the main part of the paper where the sample was much smaller due to the number density selection .
figure [ fig_uvj ] shows @xmath138 diagrams for galaxies in different redshift bins at @xmath155 above a stellar mass of @xmath202 . above this mass limit
, the sample is complete for qgs and sfgs to @xmath203 @xcite , thus the highest redshift bin at @xmath77 likely exhibits some incompleteness for qgs .
the subset of objects with measured structural parameters from the _ hst _ imaging is color - coded according to their srsic index .
for the cosmos sample at @xmath7 , each color - color bin is color - coded according to the median srsic index for galaxies in that bin and the symbol size reflects the size of the sample within the bin ( on a logarithmic scale ) .
the boundary distinguishing qgs from sfgs is shown for each redshift bin .
for the redshift range studied here , figure [ fig_uvj ] shows the buildup in the proportion of qgs , down to a fixed mass limit , with time .
the figure also shows that qgs generally have higher srsic indices relative to sfgs at a given redshift .
this is also the case for optically red galaxies ( e.g. , @xmath204 ) , where @xmath138-selected qgs have higher srsic indices relative to reddened sfgs .
the @xmath138 selection therefore works efficiently to distinguish galaxies to @xmath2 based not only on their sfhs , but also their structural properties .
this was suggested in the analysis at constant number density , but is confirmed here with a much larger sample .
the fact that qgs and sfgs have different structural properties at the highest redshifts provides additional confidence in the structural parameters obtained from galfit : high and low srsic index galaxies are found where we think they should lie , in the quiescent and star - forming regions of @xmath138 color space , respectively .
figure [ fig_sersic ] shows the median srsic index , @xmath65 , of qgs and sfgs at different redshifts .
error bars ( @xmath205 ) were computed by bootstrapping the sample in different redshift bins .
note that the random errors at @xmath141 are generally smaller than the size of the data point given the large sample size in cosmos .
qgs ( red circles ) show mild evolution in their srsic index over @xmath155 , increasing slightly from @xmath206 to @xmath207 at low redshift
. a fit of the form @xmath208 indicates that @xmath209 .
we ignore the @xmath142 qg data point for this fit as incompleteness likely plays a role for qgs at the highest redshifts . in carrying out these fits
, we also add a 10% systematic uncertainty in quadrature to the random errors to account for potential differences between the cosmos and uds samples . the median srsic index for sfgs ( blue circles ) increases from @xmath210 at @xmath89 to @xmath211 at low redshift .
a fit of the form @xmath208 indicates that @xmath212 .
the srsic indices of both qgs and sfgs with mass @xmath202 are therefore decreasing toward higher redshifts .
a potential technical explanation for the declining srsic indices of qgs with redshift is the presence of sfgs which scatter into the qg bin .
the rest - frame colors are more uncertain at higher redshifts and there are relatively more sfgs , which generally have lower srsic indices .
we test this explanation by comparing the ssfrs of qgs with high srsic indices ( @xmath213 ) with those that have low srsic indices ( @xmath214 ) . using the ssfrs from the sed fit we find that for the different redshift bins above @xmath57 , the qgs with low and high srsic indices generally have median ssfrs within the uncertainties .
we note that for the highest redshift bin , @xmath77 , the difference in ssfr is somewhat larger .
however , incompleteness for qgs likely plays a role for this redshift bin .
it is worth noting that for a similar mass limit as in this work , @xcite also find that the srsic indices of qgs decline to @xmath215 with qgs selected based on an ssfr limit and a lack of detection in mips 24 imaging . likewise for sfgs at low redshift ,
the more numerous qg population could scatter into the sfg selection , leading to an elevated median srsic index for sfgs .
the smaller uncertainties for the rest - frame colors make this scenario less likely .
we confirm that the typical ssfrs of sfgs at @xmath7 with low and high srsic indices are the same within the uncertainties .
we therefore conclude that the decline with redshift in the srsic index for qgs and sfgs with mass @xmath202 is not a consequence of galaxies of either type scattering into the other bin .
again , we note that at the highest redshifts , the srsic indices obtained from galfit are generally different for qgs and sfgs , as is the case at lower redshifts where the signal - to - noise of the measurements is more robust .
in addition to our simulations discussed earlier , this point affirms that our structural parameters are reasonable at high redshift .
figure [ fig_size](a ) shows the median effective radius ( @xmath66 ) of qgs and sfgs at different redshifts .
qgs above @xmath202 show substantial evolution in their sizes over @xmath155 , increasing by roughly a factor of @xmath9 from @xmath216 kpc to @xmath217 kpc .
a fit of the form @xmath218 indicates that @xmath219 for qgs .
the qg data point at @xmath77 was again not included in the fit due to incompleteness .
meanwhile , sfgs increase in size by a factor of @xmath45 over the same redshift range from @xmath220 kpc to @xmath221 kpc .
we find @xmath222 for sfgs .
we note that at high redshifts , the median sizes for qgs and sfgs are always larger than the measurement limits imposed by the psf of the wfc3 @xmath3 ( fwhm/2 @xmath223 ) and @xmath4 ( fwhm/2 @xmath224 ) imaging .
figure [ fig_size](b ) shows the median mass - normalized effective radius versus redshift for qgs ( red ) and sfgs ( blue ) .
the effective radii were normalized to a stellar mass of @xmath225 @xmath28 assuming a size - mass relation @xmath226 , where @xmath227 for qgs and @xmath228 for sfgs @xcite .
the mass - normalized radii therefore represent the size that galaxies would have if they lie on the given size - mass relation with a stellar mass of @xmath225 @xmath28 .
a fit of the form @xmath229 , where @xmath230 @xmath28 , indicates that @xmath231 for qgs and @xmath232 for sfgs . for our mass - limited sample ,
the median size of qgs is generally smaller than that of sfgs at a given redshift and the sizes of both qgs and sfgs decrease toward higher redshift as also found by several other authors ( e.g. , * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the distribution of axis ratios provides additional insight into the shapes of qgs and sfgs .
figure [ fig_boa ] shows the median axis ratio of qgs and sfgs with mass @xmath202 at different redshifts .
the dashed lines are fits of the form @xmath233 . for qgs ,
a fit to the data at @xmath234 indicates @xmath235 , consistent with no change over this redshift range in the median @xmath67 .
the median axis ratio for all qgs at @xmath234 is @xmath236 , which is very close to the value computed by @xcite for a similar mass limit ( after accounting for imf differences ) and redshift range .
incompleteness impacts the qg data point at @xmath77 and is therefore ignored in the fit above .
also shown in figure [ fig_boa ] is the median @xmath67 value at @xmath10 from sdss determined by @xcite for galaxies with stellar mass @xmath237 .
@xcite find no change in the distribution of axis ratios of qgs from @xmath10 to @xmath238 , consistent with our results at those redshifts .
for a sample of 14 qgs at @xmath239 with stellar masses @xmath240 @xmath28 , @xcite computed a median axis ratio of @xmath241 .
for the same redshift range and mass cut , we measure a median axis ratio of @xmath242 for 46 qgs , which is consistent with that of @xcite .
they suggest that qgs may be more disk dominated at high redshift based on the proportion of qgs with low axis ratios .
a more detailed analysis of the distribution of axis ratios with larger samples may lead to a more definitive answer .
the median axis ratio of sfgs also does not change significantly over @xmath155 .
we find @xmath243 , consistent with no evolution .
the median value for the axis ratio of sfgs over @xmath155 is @xmath244 .
the deviation at @xmath125 is within @xmath245 of this value .
interestingly , for a population of infinitely thin disks , the median observed axis ratio would be @xmath198 , not far from the value found here .
the mass selection affords a larger sample at @xmath89 than was possible with the number density selection .
we find that the general trends found with the latter selection still hold with this much larger mass - limited sample .
in particular , the axis ratios of sfgs , which dominate both the mass - limited sample and the number density selected sample at @xmath102 , display median axis ratios consistent with randomly oriented disks .
the implemented stellar mass limit , @xmath202 , to first order limits variations in the distribution of stellar masses across different redshift bins .
this works to minimize residual correlations between various parameters and stellar mass . for qgs ,
the range of median stellar masses for the different redshift bins is only @xmath246 dex .
for sfgs , the range in median stellar masses is somewhat higher at @xmath247 dex , with the higher redshift bins ( @xmath57 ) having more massive sfgs above the mass limit .
we do not expect this difference for sfgs to have a significant impact on our conclusions .
in fact , it strengthens many of our key points .
for example , if higher mass sfgs have higher srsic indices , then the srsic indices for sfgs at @xmath151 should be even lower than those at low redshift after correcting to the same median mass . | we study the structural evolution of massive galaxies by linking progenitors and descendants at a constant cumulative number density of @xmath0 mpc@xmath1 to @xmath2 .
structural parameters were measured by fitting srsic profiles to high - resolution candels _ hst _
wfc3 @xmath3 and @xmath4 imaging in the ukidss - uds at @xmath5 and acs @xmath6 imaging in cosmos at @xmath7 . at a given redshift
, we selected the _
hst _ band that most closely samples a common rest - frame wavelength so as to minimize systematics from color gradients in galaxies . at fixed @xmath8 ,
galaxies grow in stellar mass by a factor of @xmath9 from @xmath2 to @xmath10 .
the size evolution is complex : galaxies appear roughly constant in size from @xmath2 to @xmath11 and then grow rapidly to lower redshifts .
the evolution in the surface mass density profiles indicates that most of the mass at @xmath12 kpc was in place by @xmath11 , and that most of the new mass growth occurred at larger radii .
this inside - out mass growth is therefore responsible for the larger sizes and higher srsic indices of the descendants toward low redshift . at @xmath13 ,
the effective radius evolves with the stellar mass as @xmath14 , consistent with scenarios that find dissipationless minor mergers to be a key driver of size evolution .
the progenitors at @xmath2 were likely star forming disks with @xmath15 kpc , based on their low srsic index of @xmath16 , low median axis ratio of @xmath17 , and typical location in the star - forming region of the @xmath18 versus @xmath19 diagram .
by @xmath20 , many of these star - forming disks disappeared , giving rise to compact quiescent galaxies . toward lower redshifts
, these galaxies continued to assemble mass at larger radii and became the local ellipticals that dominate the high - mass end of the mass function at the present epoch . |
in the past decade , many near - threshold charmonium - like states have been reported by belle , babar , besiii , lhcb and other collaborations , which triggers lots of studies on the molecule - like hadrons containing heavy quarks .
for example , the triplet of excited @xmath14 baryons , @xmath15 , was observed by belle @xcite , and they tentatively identified the quantum numbers of these states as @xmath16 .
the same neutral state @xmath17 was also observed in @xmath18 decays by the babar collaboration with the mean value of mass higher by about @xmath19 from that obtained by belle @xcite , although the widths from these two measurements are consistent .
moreover , a new charmed hadron @xmath11 with mass @xmath20 and width @xmath21 was reported by babar collaboration by analyzing the @xmath22 invariant mass spectrum @xcite , and it is confirmed as a resonant structure in the final state of @xmath23 by belle @xcite .
the experimental observations have stimulated extensive interest in understanding the structures of the states @xmath15 and @xmath11 .
since the @xmath15 and @xmath11 are near the threshold of @xmath0 and @xmath7 , respectively , many work treat them as candidates of molecular states . for the @xmath15 , m. lutz and e. kolomeitsev interpreted it as a chiral molecule @xcite , while c. jimnez - tejero _ et al .
_ found it was a dynamically generated resonance with a dominant @xmath0 configuration @xcite .
y. b. dong _
_ estimated the strong @xmath24 decays of the @xmath15 state for different spin - parity assignments by assuming a dominant molecular @xmath0 structure of the state and showed that the decay widths of @xmath25 were consistent with current data for the @xmath26 and @xmath27 assignments @xcite .
moreover , j. r. zhang investigated @xmath15 and @xmath11 as the @xmath28wave @xmath0 state with @xmath29 and @xmath7 state with @xmath27 , respectively in the framework of qcd sum rules .
the results showed that the masses of these two states were bigger than the experimental data , but the compact structure of the states could be ruled out @xcite . for the @xmath11 , x. g. he _
et al . _
proposed that it may be a @xmath30 molecular state with @xmath29 @xcite , while y. b. dong _
_ showed the @xmath11 was a molecular state composed of a nucleon and @xmath31 mesons with @xmath26 by studying the decay widths of the strong two - body decay channels @xmath32 , @xmath33 and @xmath34 @xcite , and this conclusion was confirmed by investigating the width of the radiative decay process @xmath35 @xcite and the strong three - body decay process @xmath36 and @xmath37 @xcite .
moreover , j. he and x. liu explained the @xmath11 as an isoscalar @xmath28wave or @xmath38wave @xmath39 system with @xmath40 or @xmath26 in the framework of the one - boson - exchange model @xcite . and in ref .
@xcite , they found a possible molecular candidate with @xmath27 for the @xmath11 .
in addition , a bound state @xmath39 with @xmath27 , which can be explained as the @xmath11 , was also obtained in a constituent quark model @xcite . in the work of ref .
@xcite , the total cross section of the @xmath41 reaction was calculated within an effective lagrangian approach , which indicated that the spin - parity assignment of @xmath42 for @xmath11 gave a sizable enhancement for the total cross section in comparison with a choice of @xmath43 . another way to describe the states @xmath15 and
@xmath11 is based on the assumption that they are conventional charmed baryons .
a relativized potential model predicted that the masses of @xmath44 with @xmath27 or @xmath45 and @xmath46 with @xmath47 or @xmath45 are close to the value of @xmath15 and @xmath11 , respectively @xcite .
the strong decays of @xmath15 and @xmath11 as charmed baryons have been studied by using @xmath48 model @xcite , heavy hadron chiral perturbation theory @xcite , and chiral quark model @xcite .
moreover , ebert _
_ suggested @xmath15 as one of the orbital ( @xmath49 ) excitations of the @xmath14 with @xmath29 , @xmath50 or @xmath45 , and proposed @xmath11 as the first radial excitation of @xmath14 with @xmath47 @xcite . j. he _
et al . _
estimated the production rate of @xmath11 as a charmed baryon at panda @xcite .
h. garcilazo _
et al . _ solved exactly the three - quark problem by means of the faddeev method in momentum space , and showed that @xmath15 would correspond to an orbital excitation with @xmath29 or @xmath50 , and the @xmath11 may constitute the second orbital excitation of the @xmath51 baryon @xcite .
although many theoretical explanations to @xmath15 and @xmath11 were proposed , the properties of these two states are still in ambiguous . therefore
, more efforts are needed to reveal the underlying structure of these two states .
quantum chromodynamics ( qcd ) is widely accepted as the fundamental theory of the strong interaction . however , the direct use of qcd for low - energy hadron physics , for example , the properties of hadrons , the nucleon - nucleon ( @xmath52 ) interaction , is still difficult because of the nonperturbative complications of qcd .
qcd - inspired quark models are still the main approach to study the hadron - hadron interaction . it is well known that the forces between nucleons ( hadronic clusters of quarks ) are qualitative similar to the forces between atoms ( molecular force ) .
this similarity is naturally explained in the quark delocalization color screening model ( qdcsm ) @xcite , which has been developed and extensively studied . in this model , quarks confined in one nucleon are allowed to delocalize to a nearby nucleon and the confinement interaction between quarks in different baryon orbits is modified to include a color screening factor .
the latter is a model description of the hidden color channel coupling effect @xcite .
the delocalization parameter is determined by the dynamics of the interacting quark system , thus allows the quark system to choose the most favorable configuration through its own dynamics in a larger hilbert space .
the model gives a good description of nucleon - nucleon and hyperon - nucleon interactions and the properties of deuteron @xcite .
it is also employed to calculated the baryon - baryon scattering phase shifts and the dibaryon candidates in the framework of the resonating group method ( rgm ) @xcite .
recently , it has been used to investigate the pentaquarks with heavy quarks , and the @xmath53 can be explained as the molecular pentaquark of @xmath54 with quantum numbers @xmath55 in qdcsm @xcite . in present work , qdcsm is employed to study the properties of @xmath0 systems , and the channel - coupling effect of @xmath7 , @xmath56 , and @xmath12 channels are included .
our purpose is to investigate whether @xmath15 and @xmath11 could be explained as a molecular state composed of a nucleon and @xmath57 or @xmath31 mesons . on the other hand ,
we also want to see if any other bound or resonance state exist or not .
extension of the study to the bottom case is also interesting and is performed here .
the structure of this paper is as follows .
after the introduction , we present a brief introduction of the quark model used in section 2 .
section 3 devotes to the numerical results and discussions .
the summary is shown in the last section .
.[mass]the masses of the baryons and charmed and bottom mesons ( in mev ) . [ cols="^,^,^,^,^,^ " , ] [ bound_c ] for the @xmath8 system , the single channel calculation shows that the energy of @xmath0 is above its threshold although there is a weak attraction between @xmath82 and @xmath57 .
it is unbound ( labeled as `` @xmath83 '' in table [ bound_c ] ) , and the @xmath7 is also unbound , because the interaction between @xmath82 and @xmath31 is repulsive as mentioned above . however , by taking into account the channel - coupling effect , we obtain a stable state , the mass of which is lower than the threshold of @xmath0 .
the binding energy and the mass of this bound state is shown in table [ bound_c ] , which is labeled as `` @xmath84 '' .
first , we should mention how we obtain the mass of a bound molecular pentaquark .
generally , the mass of a molecular pentaquark can be written as @xmath85 , where @xmath86 and @xmath87 stand for the theoretical masses of a baryon and a meson respectively , and @xmath18 is the binding energy of this molecular state . in order to minimize the theoretical errors and to compare calculated results to the experimental data , we shift the mass of molecular pentaquark to @xmath88 , where the experimental masses of baryons and mesons are used . taking this bound state as an example ,
the calculated mass of pentaquark is @xmath89 mev , then the binding energy @xmath18 is obtained by subtracting the theoretical masses of @xmath82 and @xmath57 , @xmath90 ( mev ) . adding the experimental masses of the hadrons , the mass of the pentaquark @xmath91 ( mev ) is arrived .
secondly , we find that the mass of this bound state is close to the mass of the observed @xmath9 , which was reported by belle collaboration .
therefore , in our quark model calculation @xmath9 can be explained as a @xmath0 molecular state with the quantum number @xmath92 .
finally , the coupling between the @xmath28wave @xmath0 and @xmath7 channels , which is through the central force , is of crucial importance for obtaining a bound state here . in order to see the strength of these channel - coupling
, we calculate the transition potential of these two channels , which is shown in fig .
obviously , it is a strong coupling among these channels that makes @xmath0 the bound state .
this mechanism to form a bound state has been proposed before .
eric s. swanson proposed that the admixtures of @xmath93 and @xmath94 states were important for forming @xmath95 state @xcite , which was also demonstrated by t. fernndez - carams and collaborators @xcite .
the mechanism also applied to the study of @xmath96-dibaryon @xcite , in which the single channel @xmath97 is unbound , but when coupled to the channels @xmath98 and @xmath99 , it becomes a bound state .
lccccccc & & @xmath1 & @xmath100 & @xmath101 & @xmath102 & @xmath84 + @xmath103 & @xmath104 & @xmath105 & @xmath106 & @xmath107 & @xmath107 & @xmath108 + @xmath109 & @xmath104 & @xmath107 & @xmath110 & @xmath107 & @xmath107 & @xmath110 + @xmath111 & @xmath104 & @xmath105 & @xmath106 & @xmath107 & @xmath112 & @xmath105 + @xmath113 & @xmath104 & @xmath107 & @xmath106 & @xmath114 & @xmath115 & @xmath106 + @xmath116 & @xmath104 & @xmath107 & @xmath107 & @xmath107 & @xmath117 & @xmath117 + [ bound_b ] for the @xmath10 system , it includes only one channel @xmath7 , and it is a bound state with the mass of @xmath118 mev , which is close to the mass of @xmath11 .
therefore , in our quark model calculation , @xmath11 can be explained as a @xmath7 molecular state with the quantum number @xmath16 .
this result is consistent with the conclusion of ref.@xcite , in which they proposed that the @xmath11 could be explained as isoscalar @xmath28wave or @xmath38wave @xmath39 systems with @xmath27 or @xmath26 in the framework of the one boson exchange model .
meanwhile , a constituent quark model calculation also supported the existence of @xmath11 as a molecular state composed by nucleon and @xmath31 mesons with @xmath27 @xcite . for the @xmath119 system ,
the @xmath0 is unbound because of the repulsive interaction between @xmath82 and @xmath57 as shown in fig .
and for @xmath7 channel , the attraction is too weak to tie the two particles together , so it is also unbound . due to the stronger attraction
, the energy of @xmath12 is below its threshold , so the standalone @xmath12 is a bound state here .
then , we do a channel - coupling calculation .
the results show that no stable state can be obtained , i.e. , all the energies obtained are higher than the threshold of @xmath0 , which indicates that the channel - coupling effect is not strong enough to make @xmath0 bound here .
the transition potential of these three channels are shown in fig .
3(b ) , and we find they are smaller than the one of @xmath8 @xmath0 and @xmath7 channels , which is shown in fig .
moreover , coupling to the @xmath0 and @xmath7 channels , the energy of state @xmath12 is pushed above its threshold , thus preventing a resonance from materializing . for the @xmath120 system ,
the results are similar with those of the @xmath119 system .
the @xmath7 is unbound due to the repulsive potential between @xmath82 and @xmath31 as shown in fig .
both the standalone @xmath56 and @xmath12 states are bound because of the strong attractions between the corresponding two hadrons .
however , these two states disappear by coupling to the @xmath7 channel . for the @xmath13 system , there is only one channel @xmath12
, its energy , @xmath121 mev , is below the corresponding threshold .
it is a good resonance state after coupling to @xmath0 by means of tensor interaction .
this result is consistent with the one of ref @xcite , in which they showed that the @xmath12 with @xmath13 was an attractive state , presenting a resonance close to threshold by means of a chiral constituent quark model .
because of the heavy flavor symmetry , we also extend the study to the bottom case of @xmath1 system , the numerical results for which are listed in table [ bound_b ] .
the results are similar to the @xmath0 system . from table
[ bound_b ] , we find there are several interesting states : a @xmath1 bound state with @xmath8 by two channels ( @xmath1 and @xmath100 ) coupling ; a @xmath100 resonance state with @xmath10 ; and a @xmath102 resonance state with @xmath13 .
in summary , we perform a dynamical calculation of the @xmath0 systems with @xmath2 and @xmath3 , @xmath4 , @xmath5 , and @xmath6 in the framework of qdcsm .
our results show : ( 1 ) all the positive parity states are unbound in our calculation .
( 2 ) the pure @xmath0 with @xmath8 is unbound , but a bound state with mass of @xmath122 mev can be obtained by coupling the @xmath7 channel .
the mass of this bound state is close to the observed @xmath9 , which shows that @xmath9 can be explained as a @xmath0 molecular state with the quantum number @xmath92 in our quark model calculation .
( 3 ) the @xmath7 with @xmath10 is also a resonance state with the mass of @xmath118 mev , closing to the mass of @xmath11 , which indicates that @xmath11 can be explained as a @xmath7 molecular state with the quantum number @xmath16 in qdcsm .
( 4 ) the @xmath13 @xmath12 is also a resonance state with mass of @xmath121 mev . besides
, the calculation is also extended to the bottom case of @xmath1 system .
the results are similar to the case of the @xmath0 system . on the experimental side , confirming the existence of the charmed hadrons @xmath9 and @xmath11 is an interesting subject .
besides , searching for other molecular states with heavy quarks , such as @xmath12 , @xmath1 , @xmath100 and @xmath102 will be challenging topics in future .
this work is supported partly by the national science foundation of china under contract nos .
11205091 , 11035006 and 11535005 , the natural science foundation of the jiangsu higher education institutions of china ( grant no .
16kjb140006 ) , and jiangsu government scholarship for overseas studies .
99 r. mizuk _ et al _ ( belle collaboration ) , _ phys .
lett . _ * 94 * , 122002 ( 2005 ) .
b. aubert _ et al _ ( babar collaboration ) , _ phys .
d _ * 78 * , 112003 ( 2008 ) .
b. aubert _
et al _ ( babar collaboration ) , _ phys .
lett . _ * 98 * , 012001 ( 2007 ) .
r. mizuk _ et al _ ( belle collaboration ) , _ phys .
lett . _ * 98 * , 262001 ( 2007 ) .
m. f. m. lutz and e. e. kolomeitsev , _ nucl .
a _ * 755 * , 29 ( 2005 ) . c. e. jimnez - tejero , a. ramos and i. vidaa , _ phys .
c _ * 80 * , 055206 ( 2009 ) . y. b. dong , a. faessler , t. gutsche and v. e. lyubovitskij , _ phys .
d _ * 81 * , 074011 ( 2010 ) .
j. r. zhang , _ phys .
* 89 * , 096006 ( 2014 ) .
x. g. he , x. q. li , x. liu and x. q. zeng , _ eur .
j. c _ * 51 * , 883 ( 2007 ) . y. b. dong , a. faessler , t. gutsche and v. e. lyubovitskij , _ phys .
d _ * 81 * , 014006 ( 2010 ) . y. b. dong , a. faessler , t. gutsche , s. kumano and v. e. lyubovitskij , _ phys .
d _ * 82 * , 034035 ( 2010 ) . y. b. dong , a. faessler , t. gutsche , s. kumano and v. e. lyubovitskij , _ phys .
d _ * 83 * , 094005 ( 2011 ) .
j. he and x. liu , _ phys .
d _ * 82 * , 114029 ( 2010 ) . c. garca - recio , v. k. magas , t. mizutani , j. nieves , a. ramos , l. l. salcedo and l. tolos , _ phys .
d _ * 79 * , 054004 ( 2009 ) .
g. ortega , d. r. entem and f. fern@xmath123ndez , _ phys .
b _ * 718 * , 1381 ( 2013 )
. j. j. xie , y. b. dong and x. cao , _ phys .
d _ * 92 * , 034029 ( 2015 ) .
s. capstick and n. isgur , _ phys .
d _ * 34 * , 2809 ( 1986 ) .
c.chen , x. l.chen , x. liu , w. z. deng and s. l. zhu , _ phys .
d _ * 75 * , 094017 ( 2007 ) .
h. y. cheng and c. k. chua , _ phys .
d _ * 75 * , 014006 ( 2007 ) .
x. h. zhong and q. zhao , _ phys .
d _ * 77 * , 074008 ( 2008 ) .
d. ebert , r. n. faustov and v. o. galkin , _ phys .
b _ * 659 * , 612 ( 2008 ) .
j. he , z. ouyang , x. liu and x. q. li , _ phys .
d _ * 84 * , 114010 ( 2011 ) .
h. garcilazo , j. vijande and a. valcarce _ j. phys .
g _ * 34 * , 961 ( 2007 ) .
f. wang , g. h. wu , l. j. teng and t. goldman , phys .
* 69 * , 2901 ( 1992 ) ; g. h. wu , l. j. teng , j. l. ping , f. wang and t. goldman , phys .
c * 53 * , 1161 ( 1996 ) .
h. x. huang , p. xu , j. l. ping and f. wang , phys .
c * 84 * , 064001 ( 2011 ) .
j. l. ping , f. wang and t. goldman , nucl .
a * 657 * , 95 ( 1999 ) ; g. h. wu , j. l. ping , l. j. teng _
et al . _ ,
a * 673 * , 279 ( 2000 ) ; h. r. pang , j. l. ping , f. wang and t. goldman , phys .
c * 65 * , 014003 ( 2001 ) ; j. l. ping , f. wang and t. goldman , nucl .
a * 688 * , 871 ( 2001 ) ; j. l. ping , h. r. pang , f. wang and t. goldman , phys .
c * 65 * , 044003 ( 2002 )
. l. z. chen , h. r. pang , h. x. huang , j. l. ping and f. wang , phys .
c * 76 * , 014001 ( 2007 ) ; j. l. ping , h. x. huang , h. r. pang , f. wang and c. w. wong , phys . rev .
c * 79 * , 024001 ( 2009 ) .
m. chen , h. x. huang , j. l. ping and f. wang , phys .
c * 83 * , 015202 ( 2011 ) .
h. x. huang .
j. l. ping , and f. wang , _ eur .
j. c _ * 76 * , 624 ( 2016 ) .
a. valcarce , h. garcilazo , f. fernndez and p. gonzalez , rep .
. phys . * 68 * , 965 ( 2005 ) and reference there in .
j. vijande , f. fernandez and a. valcarce , j. phys .
g * 31 * , 481 ( 2005 ) .
m. m. xu , m. yu and l. s. liu , phys . rev .
* 100 * , 092301 ( 2008 ) .
eric s. swanson , phys .
b. * 588 * , 189 ( 2004 ) .
t. fernndez - carams , a. valcarce , and j. vijande , phys .
* 103 * , 222001 ( 2009 ) .
h. r. pang , j. l. ping , f. wang , t. goldman and e. g. zhao , phys .
c * 69 * , 065207 ( 2004 ) . t. f. carams and a. valcarce , phys .
d * 85 * , 094017 ( 2012 ) . | the @xmath0 and @xmath1 systems with @xmath2 and @xmath3 , @xmath4 , @xmath5 , and @xmath6 are investigated within the framework of quark delocalization color screening model .
the results show that all the positive parity states are unbound . by coupling to the @xmath7 channel
, the state @xmath0 with @xmath8 can form a bound state , which can be invoked to explain the observed @xmath9 state .
the mass of the @xmath7 with @xmath10 is close to that of the reported @xmath11 , which indicates that @xmath11 can be explained as a @xmath7 molecular state in qdcsm . besides , the @xmath12 with @xmath13 is also a possible resonance state .
the results of the bottom case of @xmath1 system are similar to those of the @xmath0 system .
searching for these states will be a challenging subject of experiments . |
infinite dimensional morse theory ( see @xcite for a a systematic exploration ) is very useful in obtaining multiple solutions for nonlinear variational problems .
the central concept in this theory is the critical group @xmath0 for a @xmath1-functional @xmath2 at an isolated critical point @xmath3 .
the critical group describes the local property of @xmath4 near the critical point @xmath3 .
on the other hand , bartsch and li @xcite introduced the critical group @xmath5 of @xmath4 at infinity , which describes the global property of the functional @xmath4 . with these concepts we have the morse inequalities@xmath6 where @xmath7 is a formal series with nonnegative integer coefficients,@xmath8 in most applications
, we may distinguish critical points using critical group , and we may find new critical points using the morse inequality .
therefore , the study of the critical group is very important . in 1979 , amann @xcite established the theory of saddle point reduction ( also called lyapunov - schmidt reduction in some literature ) .
since then , saddle point reduction becomes an important method in critical point theory , and has been widely applied to various nonlinear boundary value problems @xcite .
let @xmath9 be a separable hilbert space with norm @xmath10 , and @xmath11 .
the basic assumption in saddle point reduction is the following @xmath12 : : @xmath13 are closed subspaces of @xmath14 such that @xmath15 , and there exists a real number @xmath16 such that@xmath17 for all @xmath18 and @xmath19 .
then by saddle point reduction , there exist @xmath20 and @xmath21 such that @xmath22 is a critical point of @xmath23 if and only if @xmath24 is a critial point of @xmath4 ; moreover we have@xmath25 for case @xmath26 and with ` @xmath27 ' replaced by ` @xmath28 ' for case @xmath29 , see @xcite for a good proof of these results .
thus , to find critical points of @xmath4 we may consider the reduced functional @xmath23 . since @xmath23 is defined on a subspace , it should be easier to study .
as mentioned before , morse theory is a powerful tool in the study of variational problems .
if we want to apply morse theory , naturally we need to study the relation between the critical group of @xmath23 and that of @xmath4 . in our previous work @xcite
, we proved the results described in the following theorem .
let @xmath14 be a separable hilbert space and @xmath30 .
if @xmath29 holds , @xmath4 satisfies the palais - smale @xmath31 condition with critical values bounded from below , then @xmath32 for @xmath33 .
if @xmath26 holds and @xmath34 , @xmath4 satisfies @xmath31 condition with critical values bounded from below , then@xmath35 if @xmath29 holds and @xmath36 is an isolated critical point of @xmath23 , then@xmath37 in view of theorem a , we naturally expect that in case @xmath26 there should have a relation similar to for the critical groups at isolated critical points . in this paper
, we will establish such a relation in the following theorem .
[ t1]let @xmath14 be a separable hilbert space and @xmath38 .
assume @xmath39 holds and @xmath40 ; @xmath36 is an isolated critical point of @xmath23 such that @xmath41 is an isolated critical value .
if moreover @xmath23 satisfies the @xmath31 condition , then@xmath42 this theorem and theorem a completely describe the relation between the critical groups of the original functional @xmath4 and the reduced functional @xmath23 . under the assumption of theorem [ t1 ]
, @xmath43 is an isolated critical point of @xmath4 .
it is also well known that if @xmath4 satisfies the @xmath31 condition , so does @xmath23 , see ( * ? ? ? * lemma 1 ) .
however , we emphasizes that in theorem [ t1 ] it is not necessary to require that @xmath4 satisfies @xmath31 .
this is very important in applications .
let @xmath44 denotes the leray
schauder index for @xmath45 ( a compact perturbation of the identity map ) at its isolated zero point @xmath3 and assume that @xmath46 is a compact perturbation of the identity .
similar to ( * ? ?
* corollary 2.4 ) , as a corollary of theorem [ t1 ] and the poincar - hopf formula for @xmath1-functional ( * ? ? ?
* theorem 3.2 ) @xmath47 if @xmath26 holds and @xmath34 , we have @xmath48 the corresponding result for case @xmath29 , namely ( * ? ? ? * corollary 2.4 ) , is originally due to lazer and mckenna @xcite . as far as we know
, the identity does not appear elsewhere .
our investigation of the relation between the critical groups of the original functional and the reduced functional is motivated by the study of multiple solutions for nonlinear boundary value problems . in the second part of this paper ,
as applications of our abstract results we consider asymptotically linear elliptic systems . such problems have attracted some attentions in recent years , see @xcite .
our assumptions on the nonlinearity are so weak that the corresponding euler - lagrange functional does not satisfy the @xmath49 condition .
nevertheless , using some idea from @xcite , by taking advantage of saddle point reduction we can overcome this difficulty . as we will see in remark [ rek2 ] , because the asymptotic limits may be different variable matrices , the local linking argument used in @xcite does not apply . to prove our multiplicity results ( theorems [ st1 ] and [ st2 ] ) , theorem a ( iii ) and theorem [ t1 ] are crucial .
let @xmath14 be a banach space and @xmath50 .
let @xmath3 be an isolated critical point of @xmath4 with critical value @xmath51 , @xmath52 be an arbitrary neighborhood of @xmath3 .
then the group@xmath53 is called the @xmath54 critical group of @xmath4 at @xmath3 .
here @xmath55 $ ] , @xmath56 stands for the @xmath54 singular relative homology group of the topological pair @xmath57 with coefficients in a field @xmath58 . by the excision property of homology ,
the critical groups of @xmath4 at @xmath3 described the local property of @xmath4 near @xmath3 .
if @xmath4 satisfies the @xmath31 condition and the critical values of @xmath4 are bounded from below by @xmath59 , then according to ( * ? ? ?
* definition 3.4 ) , the group@xmath60 is called the @xmath54 critical group of @xmath4 at infinity . since @xmath4 satisfies @xmath31 , by the deformation lemma , the right hand side of does not depend on the choose of @xmath61 . since all critical points of @xmath4 are contained in @xmath62 , we can say that the critical groups of @xmath4 at infinity describe the global property of @xmath4 . from the definitions of critical groups , we see that analytically , @xmath0 is simpler than @xmath63 , because the former does not require the @xmath64 condition ; while topologically
, @xmath0 is more complicated than @xmath5 , because the topological pair on the right hand side of is simpler .
this explains why the results in theorem a ( i ) and ( ii ) were proved first .
[ proof of theorem [ t1 ] ] assume @xmath65 .
since @xmath66 is an isolated critical value of @xmath23 , there is an @xmath67 such that @xmath23 has no critical value in @xmath68 $ ] . since @xmath23 satisfies @xmath64 , by the second deformation lemma @xcite
, there is a continuous @xmath69 \times\varphi_{a+\varepsilon}\rightarrow\varphi_{a+\varepsilon}$ ] such that@xmath70{ll}\eta(0,u)=u\text { , } & u\in\varphi_{a+\varepsilon}\text{,}\\ \eta(1,\varphi_{a+\varepsilon})\subset\varphi_{a}\text { , } & \\
\eta(t , u)=u\text { , } & ( t , u)\in\left [ 0,1\right ] \times\varphi_{a}\text{.}\end{array } \right\ } \label{defo}\ ] ]
let @xmath71 be a neighborhood of @xmath22 such that @xmath23 has no critical point in @xmath72 , and set@xmath73 } \eta(t , o)\bigg ) \cup\varphi _ { a}\text{.}\ ] ] then @xmath74 is an @xmath75-invariant neighborhood of @xmath22 , and @xmath76 is a neighborhood of @xmath77 . by the property of @xmath23 and @xmath78 described in , if @xmath79 , then for any @xmath80 we have @xmath81 .
thus , setting@xmath82 we have @xmath83 . under the assumption @xmath39
, it has been shown in the proof of ( * ? ? ?
* theorem 1.2 ) that @xmath84 is homotopically equivalent to@xmath85 via a homotopy @xmath86 \times a\rightarrow a$ ] constructed in that proof ( using condition @xmath39 and the implicit function theorem ) .
moreover , denoting@xmath87 a homeomorphism @xmath88 between @xmath89 and@xmath90 has also been given there .
the deformations @xmath91 and @xmath88 have been illustrated in figure [ fig1 ] , where the thick segments with endpoint @xmath22 represent the level set @xmath92 ; while the shadowed rigions in the three subfigures represent the sets @xmath84 , @xmath89 and @xmath93 respectively .
let @xmath94 be the homotopy inverse of @xmath91 , then @xmath95 is a homotopic equivalance between @xmath84 and @xmath93 . by the definitions of @xmath91 and @xmath88 ( see the proof of ( * ? ? ?
* theorem 1.2 ) ) , we see that @xmath96 does not change the @xmath97-variable .
therefore , restricting @xmath96 to @xmath98 , we obtain@xmath99 since @xmath74 is invariant under the flow @xmath75 , we can define @xmath100 \times ( b\cap\omega ) \rightarrow b\cap\omega$],@xmath101{ll}(v , w)\text { , } & \text{if } ( v , w)\in\varphi_{a}\times x^{+}\text{,}\\ ( \eta ( t , v ) , w ) \text { , } & \text{if } ( v , w)\in\big ( ( x^{-}\backslash\varphi _ { a})\cap u\big ) \times s\text{.}\end{array } \right.\ ] ] using , it is easy to see that @xmath102 is continuous , and @xmath103 is a homotopic equivalance between @xmath104 and @xmath105 . combining the above homotopies , we can deform @xmath98 to @xmath105 continuously .
the deformation maps @xmath77 to @xmath106 , therefore it is also a homotopic equivalance between @xmath107 and @xmath108 . noting that @xmath109
, we have@xmath110 passing to homology and applying the knneth formula , we deduce@xmath111 where we have used the fact that @xmath112 , since @xmath113 . in critical point theory
, it will be very convenient if the gradient of the functional under consideration is a compact perturbation of the identity operator .
hence , if the original functional @xmath4 has this property , we hope that the reduced functional @xmath23 also has such property .
this is true if @xmath114 maps bounded sets to bounded sets .
[ ( * ? ? ?
* corollary 2.2)][p2]let @xmath14 be a separable hilbert space and @xmath50 .
assume @xmath29 or @xmath26 holds . if @xmath114 is bounded and there is a compact operator @xmath115 such that @xmath116 , then there is a compact operator @xmath117 such that @xmath118 .
finally , for the convenience of our later application , we recall a homological version of the famous three critical points theorem . [
* theorem 2.1)][ls]let @xmath14 be a banach space and @xmath119 satisfy the palais - smale @xmath64 condition .
assume that @xmath4 is bounded from below .
if @xmath120 for some @xmath121 , then @xmath4 has at least three critical points .
we note that according to ( * ? ? ?
* page 33 ) , @xmath122 for some @xmath121 implies that @xmath123 is not a local minimizer of @xmath4 , as required in the original statement of ( * ? ? ?
* theorem 2.1 ) .
in this section , as application of our abstract results on critical groups , we consider elliptic gradient systems of the form @xmath124{ll}-\delta u = f_{u } ( x , u , v)\text{,}\qquad & \text{in } \omega\text{,}\\ -\delta v = f_{v } ( x , u , v)\text { , } & \text{in } \omega\text{,}\\ \phantom{-\delta } u = v=0\text { , } & \text{on } \partial\omega\text{,}\end{array } \right .
\label{se1}\ ] ] where @xmath125 is a bounded smooth domain , @xmath126 satisfies the linear growth condition@xmath127 for some constant @xmath128 . here , to simplify the notations we denote @xmath129 .
the gradient is taken with respect to @xmath130 . without lost of generality
we may assume @xmath131 .
we also assume @xmath132 , so that @xmath133 is a trivial solution of .
therefore , we will focus on nontrivial solutions . to apply variational methods ,
let @xmath14 be the hilbert space @xmath134 endowed with the inner product@xmath135 and corresponding norm @xmath136 , here @xmath137 , @xmath138 , the dot ` @xmath139 ' represents the standard inner product in @xmath140 . under the growth condition ,
the functional @xmath141,@xmath142 is well defined and of class @xmath1 .
the critical points of @xmath143 are solutions of the system . before state our assumptions on @xmath91 and our main results ,
let us denote by @xmath144 the set of those positive definite symmetric matrix functions @xmath145 whose entries are continuous real functions on @xmath146 . for given @xmath147 , there is an associated weighted eigenvalue problem@xmath124{ll}-\delta{\pmb{z}}=\lambda a(x){\pmb{z } } & \text{in } \omega\text{,}\\ { \pmb{z}}={\pmb{0 } } & \text{on } \partial\omega\text{.}\end{array } \right .
\label{se3}\ ] ] since @xmath148 is positive definite , using the spectral theory of compact self - adjoint operator , it is well known that there is a complete list of distinct eigenvalues@xmath149 such that @xmath150 as @xmath151 .
our multiplicity results depend on the interaction between the nonlinearity @xmath91 and the eigenvalues of .
we also need the following concepts introduced by da silva ( * ? ? ?
* definition 1.4 ) .
[ sd1]let @xmath152 .
we define @xmath153 if @xmath154 for all @xmath155 .
we define @xmath156 if @xmath153 and @xmath157 is positive definite on @xmath158 with @xmath159 , where @xmath160 stands for the lebesgue measure . to state our assumptions on the nonlinearity @xmath161
, we assume that there exists @xmath162 with @xmath163 for some @xmath164 , such that@xmath165 uniformly for @xmath166 .
we then assume the following conditions on @xmath91 .
@xmath167 : : there exists some @xmath168 such that @xmath169 for @xmath170 . @xmath171
: : there exists @xmath172 with @xmath173 for some @xmath174 , such that@xmath175 if @xmath176 verifies @xmath177 , @xmath178 , then by the taylor formular we see that holds with @xmath179 being the hessian of @xmath180 at @xmath133 .
if @xmath181 for some @xmath164 , namely the problem is non resonant at the origin , then the condition @xmath182 can be removed .
the same remark applies to @xmath171 if holds .
for the sake of simplicity , we denote @xmath183 by @xmath184 , and @xmath185 by @xmath186 .
set@xmath187 our main results are the following theorems [ st1]suppose that @xmath188 satisfies and @xmath189 .
suppose moreover that there exists @xmath190 , @xmath191 such that@xmath192 then the system has at least two nontrivial solutions in each of the following cases : @xmath193 holds with @xmath194 .
@xmath195 holds with @xmath196 .
[ st2]suppose that @xmath188 satisfies and @xmath197 .
suppose moreover that there exists @xmath190 , @xmath198 such that@xmath199 then the system has at least two nontrivial solutions in each of the following cases : @xmath193 holds with @xmath200 .
@xmath195 holds with @xmath201 .
obviously , if in addition to we have @xmath202 then holds . in this case
we say that the problem is asymptotically linear at infinity . since the pioneer work of amann and zehnder @xcite , asymptotically linear problems for a single equation have captured great interest .
we referr to @xcite and references therein for some interesting results .
the asymptotically linear elliptic systems have captured some attentions in recent years . in @xcite ,
the authors considered the case that @xmath203 are constant matrices . in @xcite ,
furtado and paiva studied the case that @xmath204 and @xmath205 are variable matrices . under some conditions that ensure the euler - lagrange functional @xmath143 satisfying the ceremi type compactness condition , they obtained a nontrivial solution . in @xcite , also for the case that @xmath204 and @xmath205 are variable matrices , silva obtained two nontrivial solutions for the problem by applying morse index type argument to @xmath143 .
thus it is essential to require @xmath206 so that @xmath143 is also of class @xmath207 .
similar to @xcite , the reduction conditions and are used to control the morse index .
finally , he also required some conditions to guarantee ceremi type compactness for @xmath143 .
recently , furtado and paiva obtained a multiplicity result ( * ? ? ?
* theorem 1.1 ) under and @xmath208 .
but they only considered the case that @xmath181 for some @xmath164 , and they also required @xmath206 .
therefore , our theorem [ st1 ] is an improvement . under our assumptions
the functional @xmath143 may not satisfy the @xmath209 condition . to overcome this difficulty , as in @xcite
, we will perform the saddle point reduction and turn to consider the reduced functional @xmath23 .
it turns out that the case of theorem [ st2 ] is more difficult , because the reduced functional @xmath23 is defined on an infinite dimentional subspace .
although our proof of theorem [ st2 ] is based on some idea from @xcite , there are some real differences .
the most significant one is that if @xmath204 and @xmath205 are different matrix functions , the local linking approach used in @xcite does not work any more , see remark [ rek2 ] for details .
hence we must use critical groups and our abstract result ( theorem [ t1 ] ) is crucial .
let @xmath143 be the functional introduced in . to prove our theorems it suffices to show that @xmath143 has two nonzero critical points .
recall that for our hilbert space @xmath14 and @xmath210 $ ] , by sobolev inequality there is a constant @xmath211 such that@xmath212 that is , the embedding @xmath213 is continuous .
moreover , using the rellich - kondrachov theorem we see that the embedding is also compact if @xmath214 .
we only prove case ( ii ) .
set @xmath217,@xmath218 then @xmath219 .
we will show that @xmath143 has a local linking with respect to the decomposition @xmath220 .
namely , there exists @xmath221 such that@xmath124{l}\phi ( { \pmb{z } } ) \leq0\text{\qquad for } { \pmb{z}}\in v_{-}\text { , } \left\vert { \pmb{z}}\right\vert \leq\rho\text{,}\\ \phi ( { \pmb{z } } ) > 0\text{\qquad for } { \pmb{z}}\in v_{0}\oplus v_{+}\text { , } 0<\left\vert { \pmb{z}}\right\vert \leq\rho\text{.}\end{array } \right . \label{slo}\ ] ] the desired result will then follow from ( * ? ? ?
* theorem 2.1 ) . to prove
, we argue as in ( * ? ? ?
* page 24 ) .
since @xmath222 is an isolated eigenvalue , it is well know that there exists positive number @xmath16 such that@xmath223 using , , there exists @xmath224 such that@xmath225 for @xmath226 , using and , we obtain@xmath227 provided @xmath228 . on the other hand , since @xmath229 , there exists @xmath230 such that@xmath231 for @xmath232 with @xmath233 , we may write @xmath234 , where @xmath235 , @xmath236 . set@xmath237 for @xmath238 , we have@xmath239 by , we see that for @xmath238,@xmath240 this is also true for @xmath241 , because in this case@xmath242 hence @xmath243 by our assumption @xmath244 . therefore using we deduce @xmath245 now ,
let @xmath232 be such that @xmath246 if @xmath247 , since @xmath248 , by we may deduce @xmath249 .
if @xmath250 , then @xmath251 and @xmath252 , using @xmath253 again , we also have @xmath254 combining the above argument , we see that is true with @xmath255 .
if we replace @xmath256 by the weaker condition : @xmath257 for @xmath258 , then we can only obtain @xmath259 in the second line of .
namely , we only have a weak local linking in the sense of brezis and nirenberg @xcite , which is not sufficient to obtain lemma [ sl2 ] via the result of j.q .
liu @xcite .
however , if is replaced by the stronger condition@xmath260 then the functional @xmath143 is of class @xmath261 . according to perera ( * ? ? ?
* theorem 2.6 ) , we can still obtain the conclusion of lemma [ sl2 ] . as mentioned before , the proof of theorem [ st2 ] is more difficult .
therefore , in what follows we will only prove theorem [ st2 ] .
let@xmath262 to verify the condition @xmath26 and perform saddle point reduction , we need the following result .
now , for @xmath267 and @xmath268 , using proposition [ sp3 ] and our assumption we obtain@xmath269 therefore , @xmath143 satisfies the condition @xmath39 and we obtain a reduced functional @xmath270 , which is of class @xmath1 .
it suffices to find two non - zero critical points of @xmath23 .
we want to show that @xmath23 is coercive . for this , it is quite natural to pick a sequence @xmath271 in @xmath272 such that @xmath273 .
then consider the normalization sequence @xmath274 .
however , since @xmath275 , the weak limit of the normalization sequence may be the zero element in @xmath272 .
this makes it difficult to prove that @xmath276 .
the proof is quite similar to that of ( * ? ? ?
* lemma 3.2 ) , where instead of @xmath279 , it is assumed that @xmath284 .
since @xmath285 is bounded , up to a subsequence , we may assume that @xmath286 in @xmath272 .
the compactness of the embedding@xmath287 implies that @xmath288 in @xmath289 . by
we have@xmath290 therefore @xmath291 multiplying by @xmath292 on both sides , we deduce@xmath293 since @xmath294 and @xmath295 , the above inequality implies that @xmath296 and hence @xmath283 .
this is the only place where we need . in the case of theorem [ st1 ] ,
the reduced functional @xmath23 is defined on a finite dimensional subspace .
hence it is easy to obtain the coerciveness of @xmath23 using the assumption @xmath208 .
therefore , in theorem [ st1 ] , we may replace with a subcritical growth condition .
assume for a contradiction that for some @xmath298 and @xmath299 we have @xmath300 let @xmath301 , by lemma [ sl4 ] , up to a subsequence we have @xmath286 for some @xmath283 .
let@xmath302 then @xmath303 .
for @xmath304 we have@xmath305 by @xmath197 and the fatou lemma,@xmath306 on the other hand , @xmath197 also implies the existence of @xmath307 such that@xmath308 therefore,@xmath309 this contradicts with .
thus @xmath277 is coercive .
it follows that @xmath277 is bounded from below .
[ rek1 ] in ( * ? ? ? * lemma 3.2 ) , another version of ` non vanishing lemma ' ( as mentioned in the proof of lemma [ sl4 ] ) is proved and used in ( * ? ? ?
* lemma 3.4 ) to show that @xmath277 satisfies the @xmath31 condition .
then the coerciveness of @xmath277 is obtained via a well - known result of li @xcite , see also @xcite . our argument in lemmas [ sl4 ] and [ sll3 ] does not involve the derivative information of @xmath277 , hence is considerably simpler .
let @xmath115 be defined as@xmath310 then @xmath311 is compact and @xmath312 .
obviously @xmath313 maps bounded sets to bounded sets . by proposition [ p2 ] , @xmath314 is also a compact perturbation of @xmath315 . by the definition of the reduced functional @xmath23 , we have@xmath316 using lemma [ sll3 ]
we see that @xmath23 is also coercive and bounded from below . in particular , any @xmath31 sequence of @xmath23 is bounded . applying ( * ? ? ?
* proposition 2.2 ) , we deduce that @xmath23 satisfies @xmath31 .
[ proof of theorem [ st2]]we prove the case ( i ) . by lemma [ sl3 ]
, @xmath23 satisfies the @xmath31 condition , and bounded from below . note that@xmath317 by theorem [ t1 ] and lemma [ sl2 ] we obtain@xmath318 now , if @xmath200 , the deseired result follows from proposition [ ls ] .
[ rek2 ] if @xmath203 , then the decompositions of @xmath14 in lemma [ sl2 ] and in are related to the same eigenvalue problem . for the case proved in this section , namely theorem [ st2 ] ( ii ) , we have @xmath319 and @xmath320 . as in @xcite , it is then easy to show that @xmath23 has a local linking with respect to the decomposition @xmath321 .
then the local linking version of the three critical points theorem @xcite yields the desired result , we do nt need theorem [ t1 ] . on the other hand ,
if @xmath322 , then the above inclusion of the decompositions is false , the local linking property of @xmath143 does not descend to @xmath23 .
hence the local linking version of the three critical points theorem is not applicable , and our theorem [ t1 ] is crucial . this work was completed while s.b .
liu was visiting the institute of mathematics , chinese academy of sciences ; and the institute of mathematics , peking university .
liu would like to thank both institutes for invitation and hospitality .
amann , h. , zehnder , e. : http://www.numdam.org/item?id=asnsp_1980_4_7_4_539_0[nontrivial solutions for a class of nonresonance problems and applications to nonlinear differential equations ] .
ann . scuola norm .
( 4 ) * 7 * , 539603 ( 1980 ) bartsch , t. , li , s. : http://dx.doi.org/10.1016/0362-546x(95)00167-t[critical point theory for asymptotically quadratic functionals and applications to problems with resonance ] .
nonlinear anal . *
28 * , 419441 ( 1997 ) lazer , a.c . , mckenna , p.j .
: http://dx.doi.org/10.1016/0022-247x(85)90320-8[multiplicity results for a class of semilinear elliptic and parabolic boundary value problems ] . j. math . anal .
appl . * 107 * , 371395 ( 1985 ) li , s. , willem , m. : http://dx.doi.org/10.1007/s000300050058[multiple solutions for asymptotically linear boundary value problems in which the nonlinearity crosses at least one eigenvalue ] .
nodea nonlinear differential equations appl . * 5 * , 479490 ( 1998 ) struwe , m. : variational methods , _ ergebnisse der mathematik und ihrer grenzgebiete ( 3 ) [ results in mathematics and related areas ( 3 ) ] _ , volume 34 .
springer - verlag , berlin , 1996 , 2nd edition , applications to nonlinear partial differential equations and hamiltonian systems zou , w. , li , s. , liu , j.q . : http://dx.doi.org/10.1016/s0362-546x(98)00191-6[nontrivial solutions for resonant cooperative elliptic systems via computations of critical groups ] .
nonlinear anal .
* 38 * , 229247 ( 1999 ) | in the setting of saddle point reduction , we prove that the critical groups of the original functional and the reduced functional are isomorphic . as application ,
we obtain two nontrivial solutions for elliptic gradient systems which may be resonant both at the origin and at infinity .
the difficulty that the variational functional does not satisfy the palais - smale condition is overcame by taking advantage of saddle point reduction .
our abstract results on critical groups are crucial .
_ keywords : _ critical groups ; saddle point reduction ; knneth formula ; resonant elliptic systems _ msc20000 : _ 58e05 ; 35j60 |
rogue waves " , sometimes knows as freak waves " or monster waves " , are waves appearing as extremely large , localized waves in the ocean which have been of considerable interest recently , cf . @xcite .
the average height of rogue waves is at least twice the height of the surrounding waves , are very unpredictable and so they can be quite unexpected and mysterious .
a feature of rogue waves is that they
come from nowhere and disappear with no trace " @xcite .
in recent years , the concept of rogue waves has been extended beyond oceanic waves : to pulses emerging from optical fibres @xcite ; waves in bose - einstein condensates @xcite ; in superfluids @xcite ; in optical cavities @xcite , in the atmosphere @xcite ; and in finance @xcite .
the most commonly used mathematical model for rogue waves involves rational solutions of the focusing nonlinear ( nls ) equation @xmath0 where subscripts denote partial derivatives , with @xmath1 the wave envelope , @xmath2 the temporal variable and @xmath3 the spatial variable in the frame moving with the wave , see [ sec2 ] . in this paper
we are concerned with rational solutions of the boussinesq equation @xmath4 which are algebraically decaying have a similar appearance to rogue - wave solutions of the nls equation .
equation was introduced by boussinesq in 1871 to describe the propagation of long waves in shallow water @xcite ; see , also @xcite .
the boussinesq equation is also a soliton equation solvable by inverse scattering @xcite which arises in several other physical applications including one - dimensional nonlinear lattice - waves @xcite ; vibrations in a nonlinear string @xcite ; and ion sound waves in a plasma @xcite .
there has been considerable interest in partial differential equations solvable by inverse scattering , the _ soliton equations _ , since the discovery in 1967 by gardner , greene , kruskal and miura @xcite of the method for solving the initial value problem for the korteweg - de vries ( kdv ) equation @xmath5 during the past forty years or so there has been much interest in rational solutions of the soliton equations . for some soliton equations ,
solitons are given by rational solutions , e.g. for the benjamin - ono equation @xcite further applications of rational solutions to soliton equations include : in the description of vortex dynamics @xcite ; vortex solutions of the complex sine - gordon equation @xcite ; and in the transition behaviour for the semi - classical sine - gordon equation @xcite . in [ sec2 ]
, we discuss rational solutions of the focusing nls equation , including some generalised rational solutions which involve two arbitrary parameters . in
[ sec3 ] , we discuss rational solutions of the boussinesq equation , also including some generalised rational solutions which involve two arbitrary parameters . in [ sec4 ]
, we discuss rational solutions of the i ( kpi ) equation @xmath6 which are derived in two ways , first from rational solutions of the focusing nls equation and second from rational solutions of the boussinesq equation . in the simplest nontrivial case
, it is shown that these two types of rational solutions are different .
further we derive a more general rational solution which has those related to the focusing nls and boussinesq equations as special cases . in [ sec5 ]
we discuss our results .
the nonlinear ( nls ) equation @xmath7 is one of the most important nonlinear partial differential equations . in 1972 ,
zakharov and shabat @xcite developed the inverse scattering method of solution for it .
prior to the discovery that the nls equation was solvable by inverse scattering , it had been considered by researchers in water waves @xcite ( see also @xcite ) . in 1973 ,
hasegawa and tappert @xcite discussed the relevance of the nls equation in optical fibres and their associated solitary wave solutions .
hasegawa and tappert showed that optical fibres could sustain envelope solitons both bright and dark solitons .
bright solitons , which decay as @xmath8 , arise with anomalous ( positive ) dispersion for with @xmath9 , the _ focusing nls equation_. dark solitons , which do not decay as @xmath8 , arise with normal ( negative ) dispersion for with @xmath10 , the _ de - focusing nls equation_. rational solutions of the focusing nls equation have the general form @xmath11 where @xmath12 and @xmath13 are polynomials of degree @xmath14 in both @xmath15 and @xmath16 , with total degree @xmath14 , and @xmath17 is a polynomial of degree @xmath18 in both @xmath15 and @xmath16 , with total degree @xmath18 and has no real zeros .
the polynomials @xmath19 , @xmath12 and @xmath13 satisfy the hirota equations @xmath20 with @xmath21 and @xmath22 the hirota operators @xmath23_{x'=x , t'=t}.\end{aligned}\ ] ] the first two rational solutions of the focusing nls equation have the form @xcite @xmath24 where @xmath25 the solution @xmath26 given by is known as the peregrine solution " @xcite .
further @xmath27 dubard @xcite show that the rational solutions of the focusing nls equation can be generalised to include some arbitrary parameters .
the first of these generalized solutions has the form @xmath28 where @xmath29 with @xmath30 and @xmath31 arbitrary constants , see also @xcite .
these generalized solutions have now been expressed in terms of wronskians , see gaillard @xcite , guo , ling and liu @xcite , ohta and yang @xcite .
we note that the polynomial @xmath32 has the form @xmath33 where @xmath34 and @xmath35 are linear functions of @xmath15 and @xmath16 . in figure
[ fig : nlsu2gen ] , plots of the generalised rational solution @xmath36 given by of the focusing nls equation for various values of the parameters @xmath30 and @xmath31 .
the solution has a single peak when @xmath37 , which splits into three peaks as @xmath38 and @xmath39 increase ; this solution is called a rogue wave triplet " in @xcite and the `` three sisters '' in @xcite .
@xmath40{nls_u2_0.jpeg } & \includegraphics[height=2in , width=2in]{nls_u2_5.jpeg } & \includegraphics[height=2in , width=2in]{nls_u2_10.jpeg } \\
\a=\b=0 & \a=\b=5 & \a=\b=10\\ \includegraphics[height=2in , width=2in]{nls_u2_20.jpeg } & \includegraphics[height=2in , width=2in]{nls_u2_50.jpeg } & \includegraphics[height=2in , width=2in]{nls_u2_100.jpeg}\\ \a=\b=20 & \a=\b=50 & \a=\b=100 \end{array}\ ] ]
clarkson and kruskal @xcite showed that boussinesq equation has symmetry reductions to the first , second and fourth equations ( , , ) @xmath41 with @xmath42 , and @xmath30 , @xmath31 arbitrary constants .
vorobev @xcite and yablonskii @xcite expressed the rational solutions of in terms of polynomials , now known as the _ yablonskii
vorobev polynomials _ ( see also @xcite ) .
okamoto @xcite derived analogous polynomials , now known as the _ okamoto polynomials _ , related to some of the rational solutions of .
subsequently okamoto s results were generalized by noumi and yamada @xcite who showed that all rational solutions of can be expressed in terms of logarithmic derivatives of two sets of special polynomials , called the _
generalized hermite polynomials _ and the _ generalized okamoto polynomials _ ( see also @xcite ) .
consequently rational solutions of can be obtained in terms the yablonskii
vorobev , generalized hermite and generalized okamoto polynomials , cf . @xcite .
some of the rational solutions that are expressed in terms of the generalized okamoto polynomials are generalized to give the rational solutions of the boussinesq equation obtained in @xcite , which are analogs of the rational solutions of the kdv equation @xcite .
however none of these rational solutions of the boussinesq equation are bounded for all real @xmath3 and @xmath2 , so are unlikely to have any physical significance .
it is known that there are additional rational solutions of the boussinesq equation which do nt arise from the above construction .
for example , ablowitz and satsuma @xcite derived the rational solution @xmath43 by taking a long - wave limit of the two - soliton solution , see also @xcite .
this solution is bounded for real @xmath3 and @xmath2 , and tends to zero algebraically as @xmath44 .
if in the boussinesq equation , we make the transformation @xmath45 then we obtain the bilinear equation @xmath46 which can be written in the form @xmath47 where @xmath21 and @xmath22 are hirota operators .
since the boussinesq equation has the rational solution then we seek solutions in the form @xmath48 where @xmath49 is a polynomial of degree @xmath18 in @xmath15 and @xmath16 , with total degree @xmath18 , of the form @xmath50 with @xmath51 constants which are determined by equating powers of @xmath3 and @xmath2 . using this procedure
we obtain the following polynomials [ bq : fn]@xmath52 and the polynomials @xmath53 and @xmath54 are given in the appendix .
we note that these polynomials have the following form @xmath55 where @xmath12 is a polynomial of degree @xmath14 in both @xmath15 and @xmath16 .
we remark that the polynomials @xmath56 which arise in the rational solutions of the focusing nls equation have a similar structure , see for example , though the coefficients in the polynomials @xmath12 are different .
the polynomials @xmath57 , for @xmath58 , in scaled variables , are given by pelinovsky and stepanyants @xcite see their equations ( 6)(8 ) . however whilst they state that the polynomials are associated with solutions of their equation ( 2 ) , which is a scaled variant of the boussinesq equation , pelinovsky and stepanyants do nt mention , or reference , the boussinesq equation .
@xmath59{bqu1.jpeg } & \includegraphics[width=2in]{bqu2.jpeg } & \includegraphics[width=2in]{bqu3.jpeg } \\
u_1(x , t ) & u_2(x , t ) & u_3(x , t)\\ \includegraphics[width=2in]{bqu4.jpeg } & \includegraphics[width=2in]{bqu5.jpeg } & \includegraphics[width=2in]{bqu6.jpeg } \\ u_4(x , t ) & u_5(x , t ) & u_6(x , t ) \end{array}\ ] ] in figure [ fig31 ] , plots of the rational solutions @xmath60 , for @xmath61 , of the boussinesq equation .
these show that the maxima of the solutions all lie on the line @xmath62 , with @xmath63 local maxima for the rational solution @xmath60 .
@xmath59{bq_f3root2.jpeg } & \includegraphics[width=2in]{bq_f4root2.jpeg } & \includegraphics[width=2in]{bq_f5root2.jpeg}\\ f_3(x , t ) & f_4(x , t ) & f_5(x , t ) \end{array}\ ] ] @xmath59{bq_f3rootloci2.jpeg } & \includegraphics[width=2in]{bq_f4rootloci2.jpeg } & \includegraphics[width=2in]{bq_f5rootloci2.jpeg}\\ f_3(x , t ) & f_4(x , t ) & f_5(x , t ) \end{array}\ ] ] @xmath59{bqgrats_u60 } & \includegraphics[width=2in]{bqgrats_u61 } & \includegraphics[width=2in]{bqgrats_u6250 } \\
t=0 & t=1 & t=2.5 \\ \includegraphics[width=2in]{bqgrats_u65 } & \includegraphics[width=2in]{bqgrats_u68 } & \includegraphics[width=2in]{bqgrats_u615 } \\
t=5 & t=8 & t=15 \end{array}\ ] ] in figure [ fig32a ] , plots of the complex roots of @xmath56 , for @xmath64 , for @xmath62 and @xmath65 , i.e. @xmath66 for @xmath67 , @xmath68 for @xmath69 and @xmath70 for @xmath71 , are given .
these show a triangular " structure for both @xmath62 and @xmath65 , though with a different orientation . in figure [ fig32b ] ,
plots of the loci of the complex roots of @xmath56 , for @xmath64 , as @xmath2 varies , between between the triangular " structures for @xmath62 and @xmath65 are given .
these show that as @xmath2 increases the roots move away from the real axes . in figure [ fig34 ] , plots of the loci of the complex roots of @xmath56 with the solution
@xmath60 superimposed , for @xmath64 , as @xmath2 varies are given .
these show that as the roots move away from the real axes , the solution decays to zero .
since the focusing nls equation has generalised rational solutions , see , then a natural question is whether the boussinesq equation also has generalised rational solutions . to investigate this , we are concerned with the following theorem .
[ thm : fntilde ] the boussinesq equation has generalised rational solutions in the form @xmath72 for @xmath73 , where @xmath74 where @xmath56 is given by , @xmath75 and @xmath76 are polynomials of degree @xmath18 in @xmath15 and @xmath16 , and @xmath30 and @xmath31 are arbitrary constants . since the generalised polynomial @xmath32 for the focusing nls equation has the structure given by , we suppose that the boussinesq equation has a solution in the form , with @xmath56 given by and the polynomials @xmath75 and @xmath76 , which are of degree @xmath18 in @xmath15 and @xmath16 , have the form @xmath77 where the coefficients @xmath78 and @xmath79 are to be determined . substituting into the bilinear equation with @xmath80 , @xmath81 , @xmath82 and @xmath53 given by , @xmath75 and @xmath76 in the form , then by equating powers of @xmath3 and @xmath2
we find that [ bq : pq]@xmath83 with @xmath30 and @xmath31 arbitrary constants ; the polynomials @xmath84 , @xmath85 , @xmath86 and @xmath87 are given in the appendix .
@xmath59{bq_u2_0.jpeg } & \includegraphics[width=2in]{bq_u2_5.jpeg } & \includegraphics[width=2in]{bq_u2_10.jpeg}\\ \a=\b=0 & \a=\b=5 & \a=\b=10 \\
\includegraphics[width=2in]{bq_u2_20.jpeg } & \includegraphics[width=2in]{bq_u2_50.jpeg } & \includegraphics[width=2in]{bq_u2_100.jpeg}\\ \a=\b=20 & \a=\b=50 & \a=\b=100 \end{array}\ ] ] @xmath59{bq_u3_0.jpeg } & \includegraphics[width=2in]{bq_u3_100.jpeg } & \includegraphics[width=2in]{bq_u3_500.jpeg } \\
\a=\b=0 & \a=\b=100 & \a=\b=500 \\ \includegraphics[width=2in]{bq_u3_1000.jpeg } & \includegraphics[width=2in]{bq_u3_5000.jpeg } & \includegraphics[width=2in]{bq_u3_10000.jpeg}\\ \a=\b=1000 & \a=\b=5000 & \a=\b=10000 \end{array}\ ] ] @xmath59{bqgrat_u40.jpeg } & \includegraphics[width=2in]{bqgrat_u43.jpeg } & \includegraphics[width=2in]{bqgrat_u45.jpeg } \\
\a=\b=0 & \a=\b=10 ^ 3 & \a=\b=10 ^ 5 \\
\includegraphics[width=2in]{bqgrat_u46.jpeg } & \includegraphics[width=2in]{bqgrat_u47.jpeg } & \includegraphics[width=2in]{bqgrat_u48.jpeg}\\ \a=\b=10 ^ 6 & \a=\b=10 ^ 7 & \a=\b=10 ^ 8 \end{array}\ ] ] the first two generalised rational solutions are @xmath88 where @xmath89 and @xmath90 with @xmath30 and @xmath31 arbitrary constants . plots of the solutions @xmath91 , @xmath92 and @xmath93 of the boussinesq equation for various values of the parameters @xmath30 and @xmath31 are given in figures [ fig : u2gen ] , [ fig : u3gen ] and [ fig : u4gen ] , respectively .
figure [ fig : u2gen ] shows that the solution @xmath91 has two peaks when @xmath37 , then as @xmath38 and @xmath39 increase a third peak appears , creating a triangular structure .
numerical evidence suggests that as @xmath38 and @xmath39 increase then the three peaks all tend to the same height @xmath94 .
figure [ fig : u3gen ] shows that the solution @xmath92 has three peaks when @xmath37 , then as @xmath38 and @xmath39 increase three more peaks appear , for @xmath30 and @xmath31 sufficiently large with one central peak and five in a ring around it .
again , numerical evidence suggests that as @xmath38 and @xmath39 increase then the three peaks all tend to the same height @xmath95 .
figure [ fig : u4gen ] shows that the solution @xmath93 has four peaks when @xmath37 , then as @xmath38 and @xmath39 increase five more peaks appear , with for @xmath30 and @xmath31 sufficiently large with two central peaks and seven in a ring around it . as for @xmath91 and @xmath92
, numerical evidence suggests that as @xmath38 and @xmath39 increase then the three peaks all tend to the same height @xmath96 .
ohta and yang ( * ? ? ?
* figure 1 ) show that for focusing nls equation , the generalised rational solution @xmath97 has a single peak when @xmath37 , and three peaks otherwise .
ohta and yang ( * ? ? ?
* figure 2 ) also show that the generalised rational solution @xmath98 has a single peak when unperturbed , and six peaks otherwise .
define the polynomials @xmath99 , for @xmath100 , by @xmath101 with @xmath102 and @xmath103 the polynomials in the generalised rational solution .
then for @xmath102 and @xmath103 given by , it is easily verified that @xmath99 , for @xmath104 , satisfy the bilinear equation . hence in the general case
we have the following conjecture .
[ conj : theta ] the polynomials @xmath99 given by satisfy the bilinear equation .
consequently , from this and theorem [ bq : untilde ] we have the following result .
let @xmath99 be given by , then the polynomial @xmath105 given by can be written as @xmath106 which is a linear combination of four solutions @xmath107 , @xmath108 and @xmath109 of the bilinear equation .
the ( kp ) equation @xmath110 which is known as kpi if @xmath111 , i.e. , and kpii if @xmath112 , was derived by kadomtsev and petviashvili @xcite to model ion - acoustic waves of small amplitude propagating in plasmas and is a two - dimensional generalisation of the kdv equation .
the kp equation arises in many physical applications including weakly two - dimensional long waves in shallow water @xcite , where the sign of @xmath113 depends upon the relevant magnitudes of gravity and surface tension , in nonlinear optics @xcite , ion - acoustic waves in plasmas @xcite , two - dimensional matter - wave pulses in bose - einstein condensates @xcite , and as a model for sound waves in ferromagnetic media @xcite .
the kp equation is also a completely integrable soliton equation solvable by inverse scattering and again the sign of @xmath113 is critical since if @xmath111 , then the inverse scattering problem is formulated in terms of a riemann - hilbert problem @xcite , whereas for @xmath112 , it is formulated in terms of a @xmath114 ( `` dbar '' ) problem @xcite .
the first rational solution of the kpi equation , is the so - called lump solution " @xmath115 = -4\frac{(\xi-3\tau)^2-\eta^2 - 1}{[(\xi-3\tau)^2+\eta^2 + 1]^2},\ ] ] which was found by manakov @xcite .
subsequent studies of rational solutions of the kpi equation include ablowitz @xcite , ablowitz and villarroel @xcite , dubard and matveev @xcite , gaillard @xcite , johnson and thompson @xcite , ma @xcite , pelinovsky @xcite , pelinovsky and stepanyants @xcite , satsuma and ablowitz @xcite , and singh and stepanyants @xcite .
we remark that the kp equation is invariant under the _ galilean transformation _
@xmath116 with @xmath117 an arbitrary constant .
in fact the rational solutions of the kpi equation derived by dubard and matveev @xcite and gaillard @xcite are equivalent under the galilean transformation .
dubard and matveev @xcite derive rational solutions of the kpi equation from the generalised rational solution @xmath97 of the focusing nls equation ; see also @xcite .
specifically dubard and matveev @xcite show that @xmath118 is a solution of the kpi equation .
if we define @xmath119 , then @xmath120 the polynomial @xmath121 satisfies @xmath122 which is the bilinear form of the kpi equation , and so @xmath123 is a rational solution of the kpi equation .
the boussinesq equation is a symmetry reduction of the kpi equation and so the generalised rational solutions @xmath124 given by of the boussinesq equation can be used to generate rational solutions of the kpi equation .
if in the kpi equation we make the travelling wave reduction @xmath125 then @xmath126 satisfies the boussinesq equation . consequently given a solution of the boussinesq equation , then we can derive a solution of the kpi equation .
in particular , if @xmath45 for some known @xmath127 , is a solution of the boussinesq equation , then @xmath128 is a solution of the kpi equation .
for example the choice @xmath129 gives the lump solution of kpi .
using the generalised rational solution @xmath91 of the boussinesq equation we obtain the rational solution of the kpi equation given by @xmath130 where @xmath131 , i.e. @xmath132 we remark that this polynomial , in scaled coordinates , is given by gorshkov , pelinovsky and stepanyants @xcite , see their equation ( 4.2 ) , though the authors do nt mention the boussinesq equation .
if we compare the polynomials @xmath133 and @xmath134 , respectively given by and , then we see that they are fundamentally different .
as we shall now demonstrate , they are special cases of a more general polynomial . consider the polynomial @xmath135 , with parameters @xmath136 , @xmath30 and @xmath31 , given by @xmath137 \eta\right\}\nonumber\\ & + 2\b\left\{{\xi}^{3}-9\tau { \xi}^{2}-6({\eta}^{2}-9{\tau}^{2}+{\mu}^{2 } ) \xi+9\tau { \eta}^{2}-27{\tau}^{3 } + \ 3 ( 3{\mu}^{2}+12\mu+4 ) \tau\right\ } \nonumber\\ & + \a^2 + \b^2.\label{kp1:ff2}\end{aligned}\ ] ] this polynomial has both the polynomials @xmath133 and @xmath134 as special cases , specifically @xmath138 furthermore @xmath139 with @xmath135 given by , is a solution of the kpi equation , which includes as special cases the solutions , when @xmath140 and @xmath141 , and , when @xmath142 , as is easily shown .
@xmath59{kpiu2gen_mum1.jpeg } & \includegraphics[width=2in]{kpiu2gen_mum066.jpeg } & \includegraphics[width=2in]{kpiu2gen_mubq.jpeg } \\[2pt ] \mu=-1 & \mu=-\tfrac23 & \mu=-\tfrac13\\ \includegraphics[width=2in]{kpiu2gen_mu0.jpeg } & \includegraphics[width=2in]{kpiu2gen_mu0511.jpeg } & \includegraphics[width=2in]{kpiu2gen_mu075.jpeg } \\
\mu=0 & \mu=0.5115960325 & \mu=0.75 \\ \includegraphics[width=2in]{kpiu2gen_mu1.jpeg } & \includegraphics[width=2in]{kpiu2gen_mu1618.jpeg } & \includegraphics[width=2in]{kpiu2gen_mu2.jpeg } \\
\mu=1 & \mu=\tfrac12(1+\sqrt{5 } ) & \mu=2\\ \includegraphics[width=2in]{kpiu2gen_mu25.jpeg } & \includegraphics[width=2in]{kpiu2gen_mu3.jpeg } & \includegraphics[width=2in]{kpiu2gen_mu4.jpeg } \\ \mu=2.5 & \mu=3 & \mu=4 \end{array}\ ] ] in figure [ fig : vgen ] , the initial solution @xmath143 given by is plotted for various choices of the parameter @xmath136 .
when @xmath140 , then this arises from the solution derived from the focusing nls equation whilst when @xmath144 , then this arises from the solution derived from the boussinesq equation . from figure
[ fig : vgen ] we can see that for @xmath145 , the solution @xmath143 has two peaks on the line @xmath146 , which coalesce when @xmath147 to form one peak at @xmath148 . by considering when @xmath149 then @xmath150 is the real positive root of @xmath151\left[\mu^2 + 2(1+\tfrac13\sqrt{6})\mu+2+\sqrt{6}\right]=0,\ ] ] i.e. @xmath152 . for @xmath153
, it can be shown that @xmath154 increases until it reaches a maximum height of @xmath155 when @xmath156 , which is the golden mean !
in this paper we have derived a sequence of algebraically decaying rational solutions of the boussinesq equation which depend on two arbitrary parameters , have an interesting structure and have a similar appearance to rogue - wave solutions in the sense that they have isolated lumps " .
the associated special polynomial , which has equal weight in @xmath3 and @xmath2 , satisfies a bilinear equation of hirota type and comprises of a linear combination of four independent of the bilinear equation , something remarkable for a solutions of a bilinear equation .
the derivation of a representation of these special polynomials as determinants is currently under investigation and we do not pursue this further here . using our rational solutions of the boussinesq equation , we derived rational solutions of the the kpi equation and compared them to those obtained from rational solutions of the focusing nls equation by dubard and matveev @xcite .
it was shown that the two sets of solutions are fundamentally different and both are special cases of a more general rational solution .
we remark that ablowitz @xcite derived a hierarchy of algebraically decaying rational solutions of the kpi equation which have the form @xmath157 where @xmath158 is a polynomial of degree @xmath159 in @xmath160 , @xmath161 and @xmath162 .
this is a fundamentally different hierarchy of solutions of the kpi equation compared to hose discussed in [ sec4 ] ,
not least because it involves polynomials of all even degrees , not just of degree @xmath163 , with @xmath100 .
pac thanks andrew bassom for helpful comments and illuminating discussions and the school of mathematics & statistics at the university of western australia , perth , australia , for their hospitality during his visits when some of this research was done . | we study rational solutions of the boussinesq equation , which is a soliton equation solvable by the inverse scattering method .
these rational solutions , which are algebraically decaying and depend on two arbitrary parameters , are expressed in terms of special polynomials that are derived through a bilinear equation , have a similar appearance to rogue - wave solutions of the focusing nonlinear ( nls ) equation and have an interesting structure .
further rational solutions of the i ( kpi ) equation are derived in two ways , from rational solutions of the nls equation and from rational solutions of the boussinesq equation .
it is shown that the two families of rational solutions of the kpi equation are fundamentally different . |
@xmath3-meson decays are an excellent source of information about the ckm mechanism and allow us to test our understanding of the cp violation . in nonleptonic
@xmath3 decays we must deal with final states interactions ( fsi ) as well , since they may modify the values of the extracted parameters .
it is hard to take fsi into consideration properly since there are a lot of possible decay channels .
+ during the recent years several authors have investigated various possible corrections due to fsi .
most of the analyses take into account the elastic and inelastic effects arising from intermediate states containing light quarks ( @xmath4 , @xmath5 , @xmath6 ) [ 1 - 6 ] and apply symmetries of strong interactions ( isospin , su(3 ) ) to reduce the number of parameters .
some authors argued that intermediate states containing charmed quarks ( @xmath7 ) may also play an important role [ 7 - 13 ] .
+ in the present paper we analyse @xmath3 decays into two light noncharmed pseudoscalar mesons .
we consider fsi originating only from the intermediate states containing c quarks . in sections 2 and 3
we introduce our parametrisation and relations between the amplitudes .
section 4 contains the description of the fit procedure and our results together with the cp asymmetry predictions .
finally , a short summary is given in section 5 .
the decays of @xmath3 meson into two noncharmed pseudoscalar mesons are characterised by 10 @xmath8 invariant amplitudes corresponding to the specific quark - line diagrams . as in @xcite
we use four dominant amplitudes : tree @xmath9 , colour - suppressed @xmath10 , penguin @xmath11 and singlet penguin @xmath12 .
unprimed ( primed ) amplitudes denote strangeness conserving ( violating ) processes and are related to each other .
topological decompositions of decay amplitudes can be found in @xcite .
+ we use the wolfenstein parameters : @xmath13 , @xmath14 , @xmath15 and @xmath16 @xcite . all relations bellow
are calculated up to @xmath17 unless explicitly written otherwise .
terms proportional to @xmath18 are kept on account of complex factor in @xmath19 , which may interfere with fsi correction .
we assume that all short - distance ( sd ) strong phases are negligible . for the tree amplitude
we have @xmath20 where @xmath21 is the su(3 ) breaking factor .
both @xmath22 and @xmath23 amplitudes have a weak phase equal @xmath1 .
we assume that sd penguin amplitudes are dominated by @xmath24 quark contribution .
when terms of order @xmath25 are included , the strangeness violating penguin amplitude @xmath19 acquires a small weak phase .
thus , @xmath19 can be represented as a sum of two terms , the second one due to the @xmath17 correction : @xmath26 penguin amplitude @xmath27 has weak phase -@xmath28 .
we used in our fits the value of @xmath29 consistent with the world average .
the singlet penguin has the same phase as penguin @xmath30 : @xmath31 + finally , we accept relations between the tree and colour - suppressed amplitudes : @xmath32 @xmath33 where @xmath34 and @xmath35 .
the last equation includes electroweak penguin @xmath36 .
the ew penguin contribution @xmath37 was calculated ( see e.g. @xcite ) without @xmath25 corrections .
this fact should not affect the fits much since @xmath38 @xcite and the small correction in @xmath12 is practically invisible in the fits ( the only changes we observed were in the asymmetry for the @xmath39 decay channel ) .
it was argued [ 7 - 13 ] that the intermediate states composed of charmed mesons ( @xmath40 , etc . ) , generated from the @xmath41 tree amplitudes @xmath42 , may lead via rescattering to amplitudes of penguin topology with an internal @xmath7 quark ( the `` charming penguin '' ) .
our calculations are similar as in the case of long - distance @xmath4-type penguins @xcite . assuming su(3 ) symmetry
, we can redefine penguins : @xmath43 where @xmath44 is related to the size of the ld charming penguin and is a complex number in general . because we do not have information about @xmath44 ( or @xmath42 ) , it is convenient to introduce the following parametrisation : @xmath45 strong phase @xmath46 and size @xmath47 of the charming penguin are additional free parameters in our fits .
the weak phases are determined by the tree amplitudes @xmath48 and are either @xmath49 or 0 .
we can eliminate @xmath50 using the relation @xmath51 + short - distance charming penguin @xmath52 has the same weak phase as @xmath53 .
it can be included in a new redefined charming penguin @xmath54 with new effective size and strong phase .
we minimise function f defined as : @xmath55 where @xmath56 denote theoretical ( experimental ) cp - averaged branching fractions and @xmath57 is an experimental error for i - th decay channel .
the sum is over all 16 decay channels as in @xcite .
experimental branching ratios and their errors are listed in tables 1 and 2 .
the connection between the amplitudes and branching ratios was corrected in our calculations for the lifetime difference between @xmath58 and @xmath59 : @xmath60 + we considered two sets of data .
the first one was the same as in @xcite .
the second one was used in @xcite .
data in table 2 are more recent and differ from the previous ones in a couple of entries .
we performed fits in three general cases : 1 . without long - distance charming penguin contributions and with @xmath61 , @xmath62 , @xmath63 , @xmath1 treated as free parameters 2 . with long - distance charming penguins described by real @xmath64 as an additional parameter and @xmath65 , which is consistent with calculations done in @xcite but without any assumed connection between @xmath66 and @xmath67 3 . with long - distance charming penguins described by two additional free parameters : @xmath68 , @xmath64 .
.fits to the first set of data ( in units of @xmath69 ) [ cols="<,<,^,^,^,^ " , ] charming penguins with a nonvanishing strong phase may be a source of direct cp asymmetries .
the predicted values were calculated for the same points as in tables 1,2 .
the results are given in table 3 together with the averages from belle , babar and cleo experiments @xcite .
the main features are large asymmetries in the @xmath70 sector with relatively small asymmetries for the @xmath71 decays channels .
we are not able to predict the absolute signs of the asymmetries since we have two allowed signs of @xmath68 .
the asymmetry for @xmath72 is a pure @xmath25 effect and shows a potential influence of this correction .
our results permit to draw the following conclusions : 1 . even without the charming penguins the value of angle @xmath1 extracted from the fit depends on the details of data .
more recent data prefer the value of @xmath1 more in accordance with the expectations of the standard model .
2 . if we admit the non - zero value of the charming penguin ( with strong phase equal zero ) , the fitted values of @xmath1 may move toward the sm value by @xmath73 .
admitting strong phase of the charming penguin as a free parameter leads to a relatively flat function @xmath74 i.e. it allows a wide range of @xmath1 .
this means that there is probably too much freedom in the fits .
however , the fitted strong phase @xmath68 is relatively stable and close to @xmath2 .
gerard , j. weyers , _ eur .
phys . j. _ * c7 * , 1 ( 1999 ) .
p. enczykowski , _ acta phys . pol .
_ * b28 * , 1605 ( 1997 ) .
j. f. donoghue , e. golowich , a. petrov and j. m. soares , _ phys .
lett . _ * 77 * , 2178 ( 1996 ) .
p. enczykowski , _ phys .
* d63 * , 014016 ( 2001 ) .
p. enczykowski , _ acta phys .
_ * b33 * , 1833 ( 2002 ) .
p. ach and p. enczykowski , _ phys .
rev . _ * d66 * , 054011 ( 2002 ) .
a. j. buras , r. fleischer,_phys .
lett . _ * b341 * , 379 ( 1995 ) .
m. ciuchini , e. franco , g. martinelli and l. silvestrini , _ nucl .
_ * b501 * , 271 ( 1997 ) .
m. ciuchini , r. contino , e. franco , g. martinelli and l. silvestrini , _ nucl .
phys . _ * b512 * , 3 ( 1998 ) .
m. ciuchini , e. franco , g. martinelli , m. pierini and l. silvestrini , _ phys .
lett . _ * b515 * , 33 ( 2001 ) . c. isola , m.ladisa , g. nardulli ,
t. n. pham and p.santorelli , _ phys .
_ * d64 * , 014029 ( 2001 ) .
j. l. rosner , _ phys .
rev . _ * d64 * , 094002 ( 2001 ) .
gerard and c. smith , _ eur .
phys . j. _ * c30 * , 69 ( 2003 ) . c. w. chiang , j. l. posner , _ phys .
_ * d65 * , 074035 ( 2002 ) .
p. enczykowski,_phys .
lett . _ * b590 * , 63 ( 2004 ) . c. w. chiang , m. gronau , j. l. posner and d. a. suprun , _ phys .
rev . _ * d70 * , 034020 ( 2004 ) . | charmless @xmath0 decays are studied using flavor su(3 ) symmetry .
amplitude with charming penguin topology is considered for two cases : with zero and with arbitrary strong phase .
two sets of data ( an older and the most recent one ) are used in the fits , so that the stability of the fits is tested .
it is shown that within the present uncertainties in the data the parameters of the fit may be significantly modified , especially the @xmath1 angle .
the fits indicate the strong phase of the charming penguin amplitude to be around @xmath2 .
pacs numbers : 13.25.hw , 12.15.hh |
in @xcite , garoufalidis and kricker introduced a powerful invariant @xmath0 which is a lift of the kontsevich integral to a space spanned by trivalent graphs whose edges are colored by elements of @xmath4 $ ] , at least when the alexander polynomial vanishes . in this note
we calculate how @xmath0 behaves when evaluated on the boundary of a grope . a grope ( of class @xmath5 )
is a specific @xmath6-complex that topologically models an element of the @xmath5th term of the lower central series of a group .
it is formed by gluing together surfaces called _ stages . _
some examples are pictured in figure [ fig1 ] .
see @xcite and @xcite for details .
[ rational ] let @xmath7 be an embedded grope of class @xmath1 @xmath8 in @xmath9 which has one boundary component , is of class @xmath1 , and has surface stages of genus one .
let @xmath10 denote the unknot .
then * @xmath11 vanishes in euler degree below @xmath2 .
* @xmath12 is a sum of connected graphs where every edge but one is colored by @xmath13 $ ] , and the other edge is colored by @xmath3 .
, @xmath14 and @xmath15.,title="fig:",height=163 ] , @xmath14 and @xmath15.,title="fig:",height=163 ] , @xmath14 and @xmath15.,title="fig:",height=163 ] note that the knots in this theorem automatically have trivial alexander polynomial since they bound gropes of class exceeding three .
this will imply that the kontsevich integral of @xmath16 vanishes in degree @xmath17 and that @xmath18 where @xmath19 denotes the subspace of @xmath20 spanned by connected graphs with two hairs on a single edge .
moreover we will show every element of @xmath19 is realized for some @xmath7 . since one can easily show that @xmath21 using an @xmath22 weight system , we achieve the following results , at least when the ambient manifold is @xmath9 .
( these were both proven in @xcite with a lot more effort . )
[ vassiliev ] let @xmath7 be an embedded genus one grope of class @xmath1 in a homology @xmath23-sphere ( @xmath24 ) .
* rational - valued goussarov - vassiliev invariants of degree @xmath17 do not distinguish @xmath25 from the unknot @xmath10 .
* there exists a goussarov - vassiliev invariant , @xmath26 , of degree @xmath27 and a grope @xmath7 of class @xmath1 such that @xmath28 .
the case of arbitrary homology spheres follows easily from the @xmath9 case , as we show at the end of the paper .
we conclude by emphasizing the fact that this note deals with knots that bound gropes , which is a more restricted class than those cobounding a grope with the unknot , as studied in @xcite .
we briefly review some properties of @xmath0 here .
this is distilled from @xcite , which is further distilled from @xcite .
we restrict to the case of knots with trivial alexander polynomial , since knots which bound gropes of class @xmath29 have that property .
let @xmath30 $ ] .
then a @xmath31-colored graph is a trivalent graph together with a map @xmath32 .
( a coloring of an edge is sometimes called a bead . )
then @xmath33 is a vector space generated by @xmath31-colored graphs , modulo certain relations .
this space is graded by the euler degree which is the number of vertices of the graph .
letting @xmath41 be the degree @xmath1 part of @xmath0 , we will use the fact that @xmath41 is a universal @xmath42-valued finite type invariant of null degree @xmath1 .
that is , it is universal with respect to null clasper surgeries , which are clasper surgeries whose leaves link the knot trivially .
so @xmath0 vanishes on alternating sums of surgeries of null claspers of total degree @xmath43 , and its value on alternating sums of null - claspers of total degree @xmath1 is given by the so - called _ complete contraction _ , defined by gluing the leaves together using the equivariant linking number , and coloring an edge by a @xmath35 or @xmath44 every time it passes through a fixed seifert surface .
the equivariant linking number of two leaves which are meridians to dual bands of a seifert surface is @xmath3 .
more generally , suppose that one leaf is a meridian to a band of a seifert surface and another leaf links only the dual band with linking number @xmath45
. then the equivariant linking number of these two leaves is @xmath46 . on the other hand ,
the equivariant linking number of leaves which are null homotopic in the complement of a seifert surface is just the standard linking number .
suppose a knot bounds an embedded grope of class @xmath1 in @xmath9 , where all the surface stages are of genus one
. then the knot can be represented as a rooted tree clasper surgery , @xmath47 , in the complement of the unknot which forms a meridian to the root leaf @xcite .
note that the other leaves can be embedded arbitrarily in the complement of the unknot s spanning disk .
break @xmath47 into a union of a @xmath48 at its root and a clasper @xmath49 which sits on the unknot @xmath50 .
then @xmath49 is a null clasper of euler degree @xmath2 .
the complete contraction @xmath51 is defined by thinking of @xmath49 as a unitrivalent tree and gluing up its leaves using the equivariant linking form .
one of the two dual bands of @xmath50 has exactly one leaf of @xmath49 as a meridian .
( because the grope is of class exceeding @xmath6 . )
the equivariant linking number of this leaf with any leaf that links the dual band is an integral multiple of @xmath3 . for every other pair of leaves
the equivariant linking numbers are integers .
thus @xmath52 is a sum of trivalent graphs of euler degree @xmath2 which have a bead @xmath3 on a single edge . by the universality of @xmath0
, we have that @xmath53 , and that @xmath54 note that any terms of @xmath0 of euler degree exceeding @xmath2 must have at least two hairs and so be of vassiliev degree exceeding @xmath27 .
thus there are no terms with vassiliev degree less than or equal to @xmath59 , and @xmath60 .
observe that @xmath61 as claimed . to see that every element of @xmath19 is realized ,
take a generator @xmath62 .
remove one of the two hairs and embed @xmath63 as a rooted clasper on the unknot . breaking some edges of @xmath63 into hopf pairs of leaves
we get a rooted clasper with @xmath1 leaves as desired .
this can be detected using an @xmath22 weight system ( * ? ? ?
* appendix ) .
this weight system satisfies two relations .
any closed loop gives a multiplicative factor of @xmath23 , whereas an edge can be expanded into a difference of two terms , where the edge is removed and replaced either by two parallel edges , or by two edges which cross . which consists of an ellipse with @xmath64 parallel vertical edges and two hairs .
we show it is nonzero by showing that the diagram formed by joining the two univalent vertices is nonzero .
the resulting diagram is an ellipse with @xmath59 parallel edges added .
applying the @xmath22 relations to these @xmath59 edges , starting from the right , we get down to @xmath65 , where @xmath66 is a single circle . hence our diagram evaluates to @xmath67 .
we have already established theorem 1 for knots in @xmath9 .
any homology sphere , @xmath68 , can be obtained from @xmath9 by surgery on a disjoint union of @xmath48 claspers . by general position
, we may assume that the clasper @xmath47 corresponding to our grope is disjoint from these claspers .
letting @xmath49 denote the clasper formed by surgering the unknot along the @xmath48 containing the root , we see that modulo euler degree @xmath43 we can pull @xmath49 into a ball
. then @xmath0 decomposes as a product of @xmath69 and @xmath70 , which easily implies theorem 1 .
999 , _ a knot bounding a grope of class @xmath59 is @xmath71-trivial _ , ph.d .
thesis , uc san diego 2000 , _ grope cobordism of classical knots _ , topology , vol .
43 , issue 1 ( 2004 ) ; 119 - 156 , _ a rational noncommutative invariant of boundary links _ ,
( 2004 ) 115 - 204 , _ on knots with trivial alexander polynomial _ , to appear in j. diff .
geom . , _ what is ... a grope ? _ , notices of the amer .
51 , no . 8 , ( 2004 ) , _ wheeling : a diagrammatic analogue of the duflo isomorphism _ ,
d thesis 2000 , uc berkeley | in this note , we calculate the leading term of the rational lift of the kontsevich integral , @xmath0 , introduced by garoufalidis and kricker , on the boundary of an embedded grope of class @xmath1 .
we observe that it lies in the subspace spanned by connected diagrams of euler degree @xmath2 and with a bead @xmath3 on a single edge .
this places severe algebraic restrictions on the sort of knots that can bound gropes , and in particular implies the two main results of the author s thesis @xcite , at least over the rationals . - |
the spin-@xmath0 chain with nearest - neighbor exchange coupling @xmath1 and next - nearest - neighbor coupling @xmath2 is one of the representatives of frustrated spin systems @xcite . if @xmath2 is antiferromagnetic ( af ) i.e. , @xmath7 , the system has magnetic frustration regardless of the sign of @xmath1 .
recently , several quasi one - dimensional ( 1d ) edge - sharing cuprates @xcite such as @xmath8 , @xmath9 , @xmath5 , @xmath4 and @xmath6 , have been investigated as quantum multiferroic materials or frustrated magnets , and many of them are believed to be approximately described by a @xmath1-@xmath2 spin chain with ferromagnetic ( fm ) @xmath1 and af @xmath2 @xcite at least in the high temperature regime above their ordering temperatures .
recent theoretical studies have elucidated effects of external magnetic field @xmath10 on the @xmath1-@xmath2 chain @xcite .
it was found that the field induces three multipolar tomonaga - luttinger ( tl ) liquid phases , i.e. , quadrupolar , octupolar , and hexadecapolar tl - liquid phases , in the wide region of @xmath11 . in these three phases , multipolar operators
@xmath12 with @xmath13 , @xmath14 , and @xmath15 are respectively shown to exhibit a quasi long - range order .
in addition , in these phases , the longitudinal spin correlator also decays in an algebraic form , while the transverse spin correlator obeys an exponential - decay law .
it is generally difficult to experimentally distinguish magnetic multipolar phases in spin systems .
there is no established method of measuring multipolar correlations , and static quantities such as specific heat and susceptibility exhibit no large difference between the multipolar and standard tl liquids .
recently , we have shown @xcite for these multipolar tl liquids that ( i ) in contrast to usual tl - liquid phase , nuclear - magnetic - resonance ( nmr ) relaxation rate @xmath3 of the multipolar liquid phases decreases with lowering temperature near the saturation , and ( ii ) gapless points and peak positions of dynamical structure factors are clearly different from those of usual af spin chains .
furthermore , in ref .
@xcite we have numerically evaluated @xmath3 of the quadrupolar phase as a function of temperature and magnetic field , combining field theoretical approach with numerical data .
the obtained result demonstrated that field dependence of @xmath3 is also characteristic : @xmath3 decreases with increasing the field except for the vicinity of saturation .
these features of dynamical quantities would be helpful to detect multipolar tl liquids in experiments for the above quasi-1d magnets . in this paper , using the same method as ref .
@xcite , we determine the field and temperature dependence of @xmath3 in the octupolar phase of spin-@xmath0 @xmath1-@xmath2 chain .
the result is expected to be relevant for quasi-1d @xmath1-@xmath2 magnets with large fm @xmath16 .
for instance , @xmath4 might be a possible candidate for the realization of octupolar tl liquid @xcite .
here we shortly review the low - energy properties of multipolar tl - liquid phases of spin-@xmath0 @xmath1-@xmath2 chain @xcite .
the hamiltonian of @xmath1-@xmath2 chain under magnetic field @xmath10 is defined as @xmath17 where we focus on the case with @xmath18 and @xmath7 . for the saturated state in high - field regime ,
two- , three- and four - magnon bound states with momentum @xmath19 , respectively , become the lowest - energy excitation in the regions @xmath20 ( @xmath21 ) , @xmath22 ( @xmath23 ) , and @xmath24 ( @xmath25 ) . with decreasing field @xmath10 below the saturation field ,
these bound magnons are condensed , and the system can be regarded as a hard - core bose gas of bound magnons .
we call these gases of two- , three- , and four - magnon bound states , respectively , the second , third , and fourth order multipolar ( quadrupolar , octupolar , and hexadecapolar ) tl liquids @xcite .
these multipolar phases expand from the saturation down to the low - field regime and turn into the low - field vector - chiral phase with @xmath26 . in the @xmath27-th order
multipolar phase , correlation functions of both the longitudinal spin operator @xmath28 and the @xmath27-th multipolar operators @xmath29 ( i.e. , creation / annihilation operator of bound magnons ) can be calculated from the effective tl - liquid theory for the hard - core bose gas of multi - magnon bound states @xcite .
they are shown to decay algebraically as follows : @xmath30 \label{eq : correlation_1}\\ \langle \prod_{n=1}^p s^+_{n}\prod_{n=1}^p s^-_{j+n}\rangle & = & ( -1)^j c_1 |j|^{-1/(2\kappa)}+\cdots , \label{eq : correlation_2}\end{aligned}\ ] ] where @xmath31 is the uniform magnetization per site , @xmath32 are nonuniversal constants , and @xmath33 is the tl - liquid parameter . on the other hand , correlators of @xmath34-th multipolar operators including
transverse spin operator @xmath35 obey an exponential - decay form @xmath36 @xcite since any @xmath37-magnon excitation requires a finite excitation energy to break @xmath27-magnon bound states .
the correlation length @xmath38 must be proportional to the inverse of the binding energy @xmath39 of magnons .
the absence of the quasi - long - range order in the transverse - spin correlation plays a key role in the following discussion of nmr relaxation rate .
for spin systems in solids , nmr relaxation time @xmath40 is usually determined through weak hyperfine interactions between electron and nuclear spins .
therefore , we can obtain information about electron - spin correlations from measurement of @xmath40 . the formula for @xmath40
@xcite is given by @xmath41 + a^2_{\parallel}{\cal s}^{zz}(\omega ) , \label{eq : t_1}\end{aligned}\ ] ] where @xmath42 are the hyperfine coupling constants , and @xmath43 is the local dynamical structure factor @xmath44 the strength of @xmath42 depends on kinds of nuclei , crystal structures , and direction of field @xmath10 .
the value of @xmath45 is fixed to a nearly resonating frequency of nuclear - spin precession motion , and it is generally much smaller than the energy scale of electron spin systems .
therefore , we take the limit @xmath46 in the following discussions . utilizing these formulas , we will discuss the temperature and field dependence of @xmath3 in the octupolar liquid phase .
from eqs . ( [ eq : correlation_1 ] ) , ( [ eq : t_1 ] ) and ( [ eq : dynamical_st ] ) and field - theoretical method @xcite , @xmath3 in the multipolar liquid phases is obtained as @xmath48 here , @xmath49 , @xmath50 is the excitation velocity , @xmath51 is lattice spacing , and @xmath52 is the beta function .
we note that eq .
( [ eq : nmr_boso ] ) is valid for the temperature regime @xmath53 , where the low - energy effective theory of the bound - magnon bose gas is reliable . if spatial dependence of hyperfine couplings is taken account of beyond the formula ( [ eq : t_1 ] ) , the prefactors @xmath54 in the first and second term of eq .
( [ eq : nmr_boso ] ) are generally modified and become different .
the form of @xmath3 in eq .
( [ eq : nmr_boso ] ) consists of contributions from the longitudinal spin correlation ; the first and second terms in eq .
( [ eq : nmr_boso ] ) originate from the second and third terms in eq .
( [ eq : correlation_1 ] ) , respectively . remarkably , both terms are power - law type functions of temperature . on the other hand ,
since the transverse spin correlation exhibits a gapped behavior , its contribution to @xmath3 is a thermal - activation form @xmath55 .
therefore , if we consider the temperature region lower than the binding energy @xmath56 , the contribution from transverse spin correlation is negligible .
the disappearance of transverse - spin contribution is a characteristic feature of multipolar tl liquids and makes the multipolar liquids distinct from the usual tl liquid in various quasi-1d af magnets as discussed below .
parameters @xmath33 , @xmath50 , and @xmath57 can be evaluated as a function of field @xmath10 ( or magnetization @xmath58 ) and @xmath59 at zero temperature by using density - matrix renormalization group method @xcite . substituting them into eq .
( [ eq : nmr_boso ] ) , we obtain the temperature and field dependence of @xmath3 .
in particular , the temperature dependence of each term is determined by the tl - liquid parameter @xmath33 . for the multipolar phases ,
@xmath33 is known to increase almost monotonically with increasing @xmath10 ( or @xmath58 ) and approach unity at the saturation @xcite .
we can thereby find that @xmath3 decreases with lowering temperature in the high - field region with @xmath60 , while it diverges in the low - field region with @xmath61 @xcite .
this behavior highly contrasts to the usual tl liquid such as that of the spin-@xmath0 af chain , in which one of the longitudinal- or transverse - spin contributions diverges at @xmath62 and @xmath3 is a monotonically increasing function of temperature in sufficiently low - temperature regime .
furthermore , for the quadrupolar phase ( @xmath13 ) , the field dependence of @xmath3 has also been determined rather accurately @xcite .
it has then been shown that with increasing @xmath10 , @xmath3 decreases for low field regime and increases in the vicinity of the saturation field ( see fig . 1 of ref
. @xcite ) .
this peculiar non - monotonic field dependence of @xmath3 is expected to be a common property of all the multipolar tl - liquid phases in the @xmath1-@xmath2 chain and would be also a hallmark of them .
let us now examine the temperature and field dependence of @xmath3 in the octupolar tl - liquid phase ( @xmath63 ) .
we evaluate @xmath3 combining the formula ( [ eq : nmr_boso ] ) with numerical estimates of @xmath33 , @xmath50 , and @xmath57 which are obtained in the same manner as refs .
the coupling constant @xmath64 is simply set to unity .
hereafter , we concentrate on the point of @xmath65 ( @xmath66 ) , at which the octupolar phase widely exists for @xmath67 ( @xmath68 ) @xcite .
we note that in the octupolar phase , the excitation velocity @xmath50 , i.e. , the width of low - energy excitation band , is rather small compared to those of quadrupolar phase @xcite .
for instance , we evaluate @xmath69 for @xmath70 and @xmath71 for @xmath72 at the point of @xmath65 .
this is a natural result because the parameter region of the octupolar liquid phase is closer to the ferromagnetic instability point @xmath73 at zero field .
the magnon binding energy is also evaluated as @xmath74 for @xmath70 and @xmath75 for @xmath72 at @xmath65 @xcite . therefore , when we apply the formula ( [ eq : nmr_boso ] ) at this point , temperature must be low enough , say , @xmath76 .
we also note that the numerical estimates of @xmath57 in the present calculation are subject to a rather large finite - size effect and the evaluated @xmath3 is less accurate than that of the quadrupolar liquid in ref .
however , we have checked that the result below is stable and semi - quantitatively unchanged against such an error of @xmath57 . in the octupolar tl - liquid phase for @xmath65 ( @xmath66 ) .
we simply set the coupling constant @xmath77 . ]
( [ eq : nmr_boso ] ) ] in the octupolar tl - liquid phase for @xmath65 ( @xmath66 ) and @xmath78 . ]
( [ eq : nmr_boso ] ) ] in the octupolar tl - liquid phase for @xmath65 ( @xmath66 ) and @xmath78 . ]
we show in fig . [
fig : temperature ] the temperature dependence of @xmath3 .
the result confirms the previous prediction that , as temperature decreases , @xmath3 increases divergently for the low - field regime ( @xmath79 ) and decreases for the high - field regime ( @xmath80 ) .
figure [ fig : field ] shows the field dependence of @xmath3 for several fixed temperatures .
it is remarkable that , for the sufficiently low temperature @xmath81 , @xmath3 decreases with increasing @xmath10 ( or @xmath58 ) except for the vicinity of the saturation field .
the contribution from each term in eq .
( [ eq : nmr_boso ] ) is plotted in fig .
[ fig : each - comp ] separately .
it is clear that the first term in eq .
( [ eq : nmr_boso ] ) proportional to @xmath82 contributes dominantly to a peculiar behavior of @xmath3 , decrease with increasing @xmath10 , for a wide field range .
this feature becomes less significant as temperature increases and the field dependence of @xmath3 becomes nearly flat for a rather high temperature , @xmath83 .
the @xmath84 and @xmath10 dependence of @xmath3 found above is qualitatively the same as that of the quadrupolar tl liquid @xcite , as expected . even in the neighboring hexadecapolar liquid phase , similar properties
must be realized ( although the formula ( [ eq : nmr_boso ] ) would be valid only in extremely low temperature regime ) .
quasi-1d @xmath1-@xmath2 magnets with high crystallographic symmetry often have principal axes for hyperfine coupling tensor .
for such a case , we can eliminate all the contributions from the longitudinal spin correlations , i.e. , @xmath64 terms , in @xmath3 , setting the field @xmath10 parallel to a principal axis . as a result ,
thermal activated contribution @xmath55 proportional to @xmath85 , which is the hallmark of the multipolar liquids , becomes detectable in @xmath3 . in this case , therefore , we can easily distinguish the multipolar tl liquid and usual tl liquid phases from temperature dependence of @xmath3 .
we have studied the nmr relaxation rate @xmath3 in field - induced octupolar tl liquid of the spin-@xmath0 @xmath1-@xmath2 chain . substituting numerical estimates of parameters @xmath33 , @xmath50 , and @xmath57 into the field - theoretical result ,
we have evaluated @xmath3 as a function of temperature and magnetic field .
we have then found that , similarly in the quadrupolar liquid , @xmath3 in the octupolar tl liquid also exhibits characteristic features which are qualitatively different from those in the usual tl liquid ; in the octupolar liquid , ( i ) @xmath3 decreases ( increases ) with increasing temperature @xmath84 for the low- ( high-)field regime , and ( ii ) for sufficiently low temperatures , @xmath3 decreases with increasing magnetic field @xmath10 for a wide field range except the vicinity of the saturation field .
these features can be understood from the absence of the quasi - long - range order in the transverse - spin correlation , which is a natural consequence of the formation of magnon bound states , and therefore be generally expected in all the multipolar tl liquids in @xmath1-@xmath2 chain .
we note that the above features of @xmath3 in the octupolar phase is observed only in sufficiently low temperature regime ( @xmath86 at @xmath65 ) .
this is a quantitative difference from the case of the quadrupolar phase , which has relatively small @xmath16 .
the result in this paper indicates that nmr relaxation rate @xmath3 could be useful to detect a signature of the multipolar tl - liquid phases .
however , only the temperature dependence of @xmath3 can not be a sufficient evidence for them .
for instance , @xmath3 also decreases with lowering temperature in high - field spin - gapped states such as magnetization plateaus .
therefore , we propose that observations of both characteristic behavior in @xmath3 and gapless behavior of bulk quantities ( specific heat , susceptibilities , etc . ) present an indirect but strong evidence for multipolar tl liquids .
it has been reported @xcite that three quasi-1d cuprates @xmath4 , @xmath5 , and @xmath6 have relatively large value of @xmath16 ( @xmath87 ) .
we expect that our result will be helpful for searching for the appearance of octupolar phase in these compounds under high magnetic field .
this work was supported by grants - in - aid for scientific research from mext , japan ( grants no . 21740277 , no . 21740295 , and no .
22014016 ) .
9 enderle m , mukherjee c , fk b , kremer r k , broto j -m , rosner h , drechsler s -l , richter j , malek j , prokofiev a , assmus w , pujol s , raggazzoni j -l , rakoto h , rheinstdter m and rnnow h m 2005 _ europhys .
lett . _ * 70 * 237 in the narrow range @xmath88 , the soft two - magnon bound state has an incommensurate momentum @xmath89 to realize a two - component tl liquid . in this paper
, we do not consider this incommensurate case and focus on the one - component multipolar tl liquids of the multimagnon bound states with momentum @xmath90 .
however , the qualitative behavior of @xmath3 in the incommensurate multipolar phase is expected to be similar to those of the one - component multipolar phase since it mainly results from the formation of multimagnon bound states .
we determine the magnon binding energy @xmath39 in the octupolar tl liquid as follows .
first we prepare @xmath91 , where @xmath92 is the lowest energy eigenvalue in the subspace with fixed @xmath93 ( @xmath94 is total number of sites ) .
then , using @xmath95 we define four kinds of energy gaps @xmath96 , @xmath97 , @xmath98 , and @xmath99 .
finally the true binding energy is defined as the minimum value of them , @xmath100 . | in the spin-@xmath0 frustrated chain with nearest - neighbor ferromagnetic exchange @xmath1 and next - nearest - neighbor antiferromagnetic exchange @xmath2 under magnetic field , magnetic multipolar - liquid ( quadrupolar , octupolar , and hexadecapolar ) phases are widely expanded from the saturation down to a low - field regime .
recently , we have clarified characteristic temperature and field dependence of the nmr relaxation rate @xmath3 in the quadrupolar phase . in this paper
, we examine those of @xmath3 in the octupolar phase combining field theoretical method with numerical data .
the relevance of the results to quasi one - dimensional @xmath1-@xmath2 magnets such as @xmath4 , @xmath5 and @xmath6 is shortly discussed . |
doubly heavy baryons take the special place among baryons which contain heavy quarks .
the existence of two heavy quarks causes the brightly expressed quark - diquark structure of the @xmath1 baryon , in a wave function which one s the configuration with compact heavy @xmath6-diquark dominates .
regularity in a spectrum of mass of doubly heavy baryons appear in many respects to a similar case of mesons containing one heavy quark @xcite .
production mechanisms for @xmath7-baryons and @xmath8-mesons also have common features . at the first stage compact heavy @xmath6-diquark
is formed , than it fragments in a final @xmath7-baryon , picking up a light quark .
the calculations of production cross sections for doubly heavy baryons in @xmath9- and @xmath10-interactions were made recently as in the model of a hard fragmentation of a heavy quark in doubly heavy diquark @xcite as within the framework of the model of precise calculation of cross section of a gluon
gluon fusion into doubly heavy diquarks and two heavy antiquarks in the leading order of the perturbation theory of qcd @xcite .
mechanism of production of hadrons containing charmed quarks , based on consideration of hard parton subprocesses with one @xmath3-quark in an initial state , was discussed earlier in papers @xcite .
it was shown , that in the region of a large transferred momentum @xmath11 , where @xmath12 is charmed quark mass ) the concept of a charm excitation in a hadron does not contradict parton model and allows to effectively take into account the contribution of the high orders of the perturbative qcd theory to the born approximation . however , there is open problem of the `` double score '' , which is determined by the fact that the part of the born diagrams of birth of two heavy quarks in a gluon - gluon fusion can be interpreted , as the diagrams with charm excitation in one of initial protons .
these diagrams give leading in @xmath13 contribution in the @xmath3-quark perturbative , so - called point - like structure function ( sf ) of a proton .
as to the non - perturbative contribution in @xmath14quark sf of a proton @xcite it does not depend from @xmath15 and becomes very small at @xmath16 .
for example , fig .
1 shows one of 36 born diagrams , which have order @xmath17 , describing production of the @xmath2-diquark in the gluon - gluon fusion subprocess .
the experience in calculation of heavy quark production cross sections in a gluon - gluon fusion demonstrates that the contribution of the next order of the perturbative qcd in @xmath18 can be comparable with the contribution of the born diagrams . in the case of gluon gluon production of the two pairs of heavy quarks there will be more than three hundred diagrams with additional gluon in the final state , which have order @xmath19 , and their direct calculation is now considered difficultly feasible .
in this paper the model of @xmath2-diquark production in proton - proton interactions , based on the mechanism of the charm excitation in a proton is considered .
it is supposed , that the @xmath2-diquark is formed during scattering of @xmath3-quarks from colliding protons with radiation of a hard gluon , i.e. in the parton subprocess : @xmath21 the feynman diagrams of the parton subprocess ( 1 ) are shown in fig .
2 , where @xmath22 and @xmath23 are 4-momenta of the initial @xmath3-quarks , @xmath24 is 4-momentum of the final gluon , @xmath25 is 4-momentum of the diquark , which one is divided equally between the final @xmath3-quarks .
the doubly heavy diquark is considered as bound state of two @xmath3-quarks in the antitriplet colour state and in the vector spin state .
if @xmath26 and @xmath27 are colour indexes of initial quarks , and @xmath28 is colour index of a final diquark , the amplitude of production of the @xmath2-diquark @xmath29 is connected with the amplitude of production of two @xmath3-quarks with 4-momenta @xmath30 as follows : @xmath31 where @xmath32 , @xmath33 is the diquark mass , @xmath34 is the diquark wave function in zero point , @xmath35 is the colour part of a diquark wave function . considering spin degrees of freedom of @xmath14quarks and @xmath2-diquark ,
we have following conformity between amplitudes of birth of free quarks and a diquark with fixed spin projections ( without colour indexes and common factor @xmath36 ) : @xmath37.\end{aligned}\ ] ] because the wave function of the @xmath2-diquark is antisymmetric on colour index and symmetric on remaining indexes , the production of the scalar @xmath2-diquark is forbidden , i.e. @xmath38 amplitudes adequate to the diagrams in fig .
2 , where the final c - quarks are in the arbitrary spin states , are written out below , without the colour factors and the common factor @xmath36 : @xmath39 where @xmath40 , @xmath13 is strong coupling constant , @xmath41 .
let s remark , that the amplitudes @xmath42 are received by replacement of the initial quarks momenta @xmath43 in the amplitudes @xmath44 ans are taken with a minus sign , that allows for the antisymmetrization of the initial state of two identical @xmath3-quarks .
the corresponding colour factors are presented by the following expressions : @xmath45 using known property of a completely antisymmetric tensor of the third rank @xmath46 it is easy to find products of the colour factors @xmath47 , which ones are presented in the appendix a. the method of the calculation of a production amplitude of a bound nonrelativistic state of quarks in the fixed spin state is based on a formalism of the projection operator @xcite .
using properties of the charge conjugation matrix @xmath48 , we can link a scattering amplitude of a quark on a quark with a scattering amplitude of an antiquark on a quark , for example : @xmath49 as it may be shown , at @xmath50 one has : @xmath51 \sim \hat\varepsilon ( p , s_z=0)(\hat p+m_{cc}),\nonumber\end{aligned}\ ] ] where @xmath52 is polarization 4-vector of a spin-1 particle . after following effective replacements @xmath53 where @xmath54 and @xmath55 , amplitudes @xmath56 , with corresponding colour factors @xmath57 , describe production of the @xmath2-diquark with fixed polarization .
the square of the module of amplitude of @xmath2-diquark production after average on spin and colour degrees of freedom is given by the following expression : @xmath58 where in the amplitudes @xmath56 we have put @xmath59 . the summation on vector diquark polarizations in the square of the amplitude of the process ( 1 )
was done using the standard formula : @xmath60 the calculation of the value @xmath61 have been executed using the package of an analytical calculations feyncalc @xcite .
the answer is shown in the appendix b , as a function of standard mandelstam variables @xmath62 and @xmath63 .
in the parton model the cross section of a @xmath2-diquark production in @xmath10-interactions is represented as follows : @xmath64 where @xmath65 @xmath66 @xmath67 is the @xmath3-quark distribution function in a proton at @xmath68 , @xmath69 is the diquark transverse momentum , @xmath70 is the rapidity of the diquark in c.m.f . of colliding protons ,
@xmath71 it is supposed that spin-@xmath72 and spin-@xmath73 @xmath74baryons relative yield is @xmath75 as it is predicted by the simple counting rule for the spin states .
the production cross section of @xmath1-baryons plus @xmath76baryons in our approach is connected with the production cross section of @xmath2-diquark within the framework of a model of a non - perturbative fragmentation as follows : @xmath77 where @xmath78 is the phenomenological function of a fragmentation , normalized approximately on unity , as a total probability of transition @xmath2-diquark in final doubly charmed baryon . at @xmath79 the fragmentation function
is selected in the standard form @xcite : @xmath80 where @xmath81 is the @xmath1-baryon mass , @xmath82 is the light quark mass , @xmath83 the is rate - fixing constant .
the fragmentation function for @xmath84 can be determined by the solving the dglap evolution equation @xcite .
following to paper @xcite , at numerical calculations we have used following values of parameters : @xmath85 gev , @xmath86 , @xmath87 gev@xmath88 , @xmath89 gev . for a @xmath14quark distribution function in a proton @xmath90
the parametrization cteq5 @xcite was used . in figures 3 and 4 at @xmath91
tev and @xmath92 tev , accordingly , the curves show results of our calculations of @xmath69-spectra latexmath:[$(|y the calculations from paper @xcite , adequate to the contribution of the gluon - gluon fusion production of @xmath1-baryons in a born approximation .
thus , our calculations demonstrate , that the observed production cross section of @xmath1-baryons on colliders tevatron and lhc can be approximately 2 times more at the expense of the contribution of the parton subprocess @xmath94 , than it was predicted earlier in the papers @xcite .
99 , prog.theor.phys.*82 * , 760 ( 1989 ) . , z.phys .
* c64 * , 57 ( 1994 ) .
, phys.lett . *
b332 * , 411 ( 1994 ) . , z.phys.*c76 * , 111 ( 1997 ) .
, phys.rev . *
d49 * , 555 ( 1994 ) .
, phys.lett . *
b385 * , 297 ( 1996 ) .
, phys.lett . , * b426 * , 384 ( 1998 ) . , yad.fiz .
* 59 * , 909 ( 1996 ) .
, phys.rev . *
d54 * , 3228 ( 1996 ) .
, phys.rev . *
d57 * , 4385 ( 1998 ) .
, mod.phys.lett . *
a9 * , 1083 ( 1994 ) .
, phys.lett .
* b343 * , 381 ( 1995 ) . , phys.rev . *
d56 * , 3046 ( 1997 ) .
, nucl.phys . * 478 * , 311 ( 1996 ) .
phys . , * b174 * , 317 ( 1980 ) . , the feyncalc book , mertig research and consulting , 1999 .
, phys.rev . *
d27 * , 105 ( 1983 ) .
, sov.j.nucl.phys . * 15 * , 438 ( 1972 ) ; _ yu.a .
dokshitser _ , sov.phys.jetp , * 46 * , 641 ( 1977 ) ; _ g. altarelli and g.parisi_ , nucl.phys .
, * b126 * , 298 ( 1977 ) . , ( cteq coll . ) , preprint hep - ph 9903282 . cross section of @xmath1-baryon production at @xmath99 tev and @xmath100
. stars ( * ) show the results of calculation from paper [ 10 ] , curve is our result obtained in the model of a charm excitation in colliding protons . cross section of @xmath1-baryon production at @xmath101 tev and @xmath100 . stars ( * ) show the results of calculation from paper [ 10 ] , curve is our result obtained in the model of a charm excitation in colliding protons . | the production of baryons containing two charmed quarks @xmath0 in hadronic interactions at high energies and large transverse momenta is considered .
it is supposed , that @xmath1-baryon is formed during a non - perturbative fragmentation of the @xmath2-diquark , which was produced in the hard process of @xmath3-quark scattering from the colliding protons : @xmath4 . it is shown that such mechanism enhances the expected doubly charmed baryon production cross section on tevatron and lhc colliders approximately 2 times in contrast to predictions , obtained in the model of gluon - gluon production of @xmath2-diquarks in the leading order of perturbative qcd .
* hadronic production of doubly charmed + baryons via charm excitation in proton * d.a .
g@xmath5nter and v.a.saleev _ samara state university , samara , russia _ |
this paper discusses the infinite class of higher - derivative hamiltonians @xmath9 where @xmath10 and @xmath11 are arbitrary positive real parameters .
these hamiltonians are all @xmath0 symmetric ; that is , they are symmetric under combined space reflection @xmath12 and time reversal @xmath13 .
the literature on @xmath0 symmetry is extensive and growing rapidly ; some early references @xcite and recent review articles @xcite provide background for this work . the hamiltonians in ( [ eq1 ] )
are special because they are spectrally identical to another class of hamiltonians @xmath2 , which have entirely real spectra .
for example , for the case @xmath14 the hamiltonian @xmath15 has a real positive discrete spectrum and this spectrum is identical to the spectrum of the conventionally dirac hermitian hamiltonian @xmath16 the spectral equivalence of these two hamiltonians has long been known and has been examined in many papers @xcite .
spectral equivalences between pairs of non - hermitian and hermitian hamiltonians have been discussed in the past @xcite and a number of approximately equivalent pairs of hamiltonians have been constructed ( see , for example , refs .
a few other exactly equivalent pairs of hamiltonians have been found @xcite .
a brief demonstration that the hamiltonians in ( [ eq2 ] ) and ( [ eq3 ] ) are spectrally identical is given in ref .
@xcite . the prescription introduced in ref .
@xcite makes use of several elementary transformations of differential equations .
we review this analysis here . to begin
, we construct the coordinate - space eigenvalue problem for @xmath17 in ( [ eq2 ] ) by substituting @xmath18 . the formal eigenvalue problem
@xmath19 then becomes the differential - equation eigenvalue problem @xmath20 the eigenfunction @xmath21 is required to satisfy homogeneous boundary conditions inside a pair of stokes wedges in the complex-@xmath22 plane .
these wedges lie below and adjacent to the real-@xmath22 axis and have angular opening @xmath23 , as shown in fig .
[ f1 ] . to establish that the eigenvalues of ( [ eq4 ] ) are also eigenvalues of @xmath24 in ( [ eq3 ] )
, we follow the simple four - step recipe introduced in ref .
@xcite : * step 1 : * map the integration path onto the real axis by the change of variable @xmath25 as @xmath26 runs along the real axis from @xmath27 to @xmath28 , @xmath22 runs along the complex contour that asymptotes in the stokes wedges shown in fig .
[ f1 ] . under the change of independent variable in ( [ eq5 ] ) ,
the derivatives transform as follows : @xmath29 thus , with the substitution @xmath30 , ( [ eq4 ] ) becomes @xmath31 - 16\gamma(1+it)^2\phi(t)=e\phi(t ) .
\label{eq8}\ ] ] * step 2 : * take the fourier transform of ( [ eq8 ] ) .
the fourier transform of @xmath32 is @xmath33 so the fourier transform of ( [ eq8 ] ) is achieved by making the replacements @xmath34 and we obtain @xmath35 * step 3 : * introduce the new dependent variable @xmath36 by @xmath37 and choose @xmath38 to remove the one - derivative term and thereby change ( [ eq11 ] ) to schrdinger form .
the condition on @xmath38 that eliminates terms containing @xmath39 is @xmath40 differentiating ( [ eq13 ] ) , we get @xmath41 and substituting ( [ eq12 ] [ eq14 ] ) into ( [ eq11 ] ) , we obtain the schrdinger equation @xmath42 * step 4 : * finally , rescale the independent variable @xmath43 : @xmath44 this rescaling makes ( [ eq15 ] ) resemble the original schrdinger equation in ( [ eq4 ] ) : @xmath45 the schrdinger equation ( [ eq17 ] ) is associated with the hamiltonian in ( [ eq3 ] ) .
since the differential equation ( [ eq17 ] ) is real and the boundary conditions on the eigenfunctions @xmath36 are imposed on the real axis , it follows that the eigenvalues are real .
this constitutes a rigorous proof that the non - hermitian hamiltonian @xmath17 in ( [ eq2 ] ) has a real spectrum .
the term in ( [ eq3 ] ) that is proportional to @xmath6 is a quantum anomaly that arises as a remnant of the complex boundary conditions on the eigenfunctions associated with ( [ eq4 ] ) .
( these boundary conditions are illustrated in fig .
[ f1 ] . ) in the classical limit @xmath46 , we obtain the classical hamiltonian @xmath47 it is an interesting but previously unnoticed fact that at the classical level the two hamiltonians ( [ eq2 ] ) and ( [ eq18 ] ) are equivalent . to demonstrate this equivalence we show that the classical trajectories determined by these hamiltonians in complex - coordinate space have identical periods .
we plot in fig .
[ f2 ] two classical orbits of @xmath48 for @xmath49 , @xmath50 , and energy @xmath51 .
note that there are four turning points , which are located at @xmath52 , @xmath53 , @xmath54 , and @xmath55 .
a trajectory starting at @xmath56 gives a closed orbit in the upper - half plane , and a trajectory starting at @xmath57 gives a closed orbit in the lower - half plane .
both orbits have the same period @xmath58 , where @xmath59 we calculate @xmath58 by solving hamilton s equations @xmath60 and @xmath61 . eliminating @xmath43 , we express the period @xmath58 as the contour integral @xmath62 we then use cauchy s theorem to distort the contour to one that goes along a ray from the origin to a turning point , encircles the turning point , and follows the ray back to the origin .
the contour then continues along another ray out to the other turning point , encircles this turning point , and follows that ray back to the origin .
the resulting integral becomes a standard representation for a beta function . in fig .
[ f3 ] we plot the classical orbit for a particle of energy @xmath51 described by the hamiltonian @xmath63 in ( [ eq18 ] ) .
note that there are now two and not four turning points located at @xmath64 .
the figure shows a closed periodic classical orbit that begins at @xmath56 .
the period @xmath58 of this orbit is exactly that given in ( [ eq19 ] ) .
figure [ f3 ] exhibits a remarkable new topological feature that is not found in the complex classical trajectories of hamiltonians of the form @xmath65 , namely , that the classical orbit makes a @xmath66 loop around the turning points .
there have been many studies of complex classical systems @xcite and in previous numerical studies of complex trajectories the hamiltonians that were examined were quadratic in the momentum @xmath43 .
when the momentum term in the hamiltonian is quadratic , the trajectory always makes a @xmath67 u - turn about the turning points .
in contrast , with hamiltonians of the form @xmath68 , the classical trajectory makes a @xmath69 turn and for hamiltonians of the form @xmath70 , the classical trajectory makes a @xmath66 turn about the turning points .
these behaviors are illustrated in fig .
[ f4 ] . to understand the angular rotation of a classical trajectory about a turning point , let us consider a hamiltonian of the form @xmath71 . from the first of hamilton s equations , @xmath72
, we can eliminate @xmath43 from the hamiltonian and we find that a particle of energy @xmath73 obeys the equation @xmath74^{(n-1)/n}$ ] .
we can now see that the turning point is also a branch point of an @xmath75-sheeted riemann surface .
note that as @xmath76 , a classical trajectory that approaches a turning point will make a full @xmath77 turn and continue going in the same direction !
( we encounter many examples of such higher - order turning points throughout this paper .
however , we emphasize that , in general , the properties of higher - order classical turning points are not easy to predict analytically , and require the kind of sophisticated analysis used in catastrophe theory . )
the main objective of this paper is to show how to construct isospectral pairs of hamiltonians , one of which is associated with a complex eigenvalue problem while the other is associated with a real eigenvalue problem .
there have been many attempts to find isospectral pairs of hamiltonians of the form @xmath78 by using the four - step differential - equation procedure outlined above .
however , all such attempts have proved fruitless . in this paper
we consider a wider class of hamiltonians in which @xmath79 is replaced by @xmath80 , and we suggest by means of illustrative examples ( rather than by presenting a proof ) that for each integer value of @xmath75 it is now possible to construct such an isospectral pair of @xmath0-symmetric hamiltonians .
( to construct a proof , one would have to follow the procedure detailed in the illustrative examples in secs .
[ s3 ] and [ s4 ] . ) the eigenfunctions for the first member of the pair @xmath81 satisfy boundary conditions in stokes wedges in the complex plane .
however , the eigenfunctions for the second member of the pair @xmath82 satisfy a real differential equation with homogeneous boundary conditions given on the real axis .
therefore , our construction constitutes a rigorous proof that entire eigenspectrum of the complex hamiltonian @xmath81 is real .
the hamiltonian @xmath2 that is spectrally equivalent to @xmath1 has quantum anomaly terms containing powers of @xmath6 up to @xmath83 .
if we discard these anomaly terms , we obtain a pair of _ classical _ hamiltonians that are equivalent in the following sense : for each closed periodic classical trajectory of the first hamiltonian , it appears that there exists a closed trajectory of the second hamiltonian that has exactly the same period .
this paper is organized as follows : in sec . [ s2 ] we present some general introductory calculations .
then , in secs . [ s3 ] and [ s4 ] we treat the cases @xmath84 and @xmath85 . finally , in sec . [ s5 ]
we make some concluding remarks .
in order to carry out step 1 on @xmath1 in ( [ eq3 ] ) we must generalize the mapping in ( [ eq5 ] ) : @xmath86 where @xmath87 . note that this change of variable reduces to ( [ eq5 ] ) when @xmath14 ( @xmath88 ) .
now , as @xmath26 runs from @xmath27 to @xmath28 along the real axis , @xmath22 runs along a complex contour that is appropriate for the hamiltonian @xmath1 . the transformations from @xmath22 derivatives to @xmath26 derivatives
can then be written as @xmath89 this set of equations generalizes ( [ eq6 ] ) and ( [ eq7 ] ) ; ( [ eq22 ] ) and ( [ eq23 ] ) reduce to ( [ eq6 ] ) and ( [ eq7 ] ) when @xmath88 .
in the next two sections we show how use these results to transform a general class of hamiltonians of the form ( [ eq1 ] ) .
note that the hamiltonian in ( [ eq2 ] ) corresponds to @xmath14 .
we begin with the simplest generalization , namely , @xmath84 .
in this section we consider the hamiltonian ( [ eq1 ] ) for the case @xmath84 : @xmath90 the corresponding differential - equation eigenvalue problem @xmath91 is @xmath92 to solve this boundary - value problem the integration path must lie in stokes wedges . for the above equation these wedges do not include the real axis .
the wedges are determined by the asymptotic solutions to ( [ eq27 ] ) , which for large @xmath93 have the form @xmath94 , where @xmath95 is a positive constant and @xmath96 is a cube root of unity : @xmath97 . from this asymptotic behavior
we see that each of the stokes wedges has angular opening @xmath98 .
we impose boundary conditions in the wedges @xmath99 and @xmath100 .
( these wedges are the shaded regions in the lower - half plane shown in fig .
[ f5 ] . ) requiring that @xmath21 vanish as @xmath101 with @xmath102 in the above two wedges eliminates two of the three possible solutions to ( [ eq27 ] ) and keeps only the solution whose asymptotic behavior is given by @xmath103 .
we now follow step 1 of the procedure explained in sec .
[ s1 ] and make the change of variable in ( [ eq21 ] ) with @xmath104 . from ( [ eq24 ] )
we then obtain the differential equation @xmath105\phi(t)\nonumber\\ & & \qquad-3 ^ 9\gamma(1+it)^3\phi(t)=e\phi(t ) .
\label{eq28}\end{aligned}\ ] ] this differential equation is to be solved on the _ real _ axis in the @xmath26 variable and the solution satisfies homogeneous boundary conditions on the _ real _ axis . note that as @xmath26 runs from @xmath27 to @xmath106 , the variable @xmath22 runs from an angle of @xmath107 to an angle of @xmath108 in the complex-@xmath22 plane .
this integration in the complex-@xmath22 plane is shown in fig .
[ f5 ] .
following step 2 , we perform the fourier transform in ( [ eq9 ] ) and use ( [ eq10 ] ) to obtain @xmath109f(p)\nonumber\\ & & \qquad-3 ^ 9\gamma\left(1+\hbar{\textstyle{\frac{\rm d}{{\rm d}p}}}\right)^3f(p)=ef(p ) .
\label{eq29}\end{aligned}\ ] ] we then simplify ( [ eq29 ] ) by transforming it to a new differential equation that does not have a second - derivative term . to do
so we let @xmath110 , as in step 3 of sec .
the condition on @xmath38 that eliminates the second - derivative term is @xmath111 and upon differentiation we get @xmath112 the resulting equation for @xmath36 is @xmath113 this completes step 3 .
finally , we perform the scaling @xmath114 and obtain @xmath115 this completes step 4 and from this result we make the _ ansatz _ @xmath116 to identify the hamiltonian that is equivalent to that in ( [ eq26 ] ) : @xmath117 observe that there is an @xmath118 anomaly term as well as two @xmath6 anomaly terms .
therefore , in the classical limit @xmath46 , @xmath119 in ( [ eq35 ] ) becomes @xmath120 we now argue that the two classical hamiltonians , @xmath121 and @xmath122 , are equivalent by computing numerically the classical trajectories for each hamiltonian and verifying that trajectories having the same energy have the same periods . in fig .
[ f6 ] we plot a classical trajectory for @xmath121 with @xmath123 and @xmath50 for a particle of energy @xmath51 starting at @xmath124 . the trajectory in fig .
[ f6 ] has an elaborate structure .
it winds around six of the nine turning points in the negative ( clockwise ) direction . because each of the nine turning points is also a branch point of cubic type , with each wind the direction of the path changes by @xmath69 , as shown in fig .
thus , after the path leaves its starting point at @xmath124 , it winds in the negative direction and crosses from sheet zero to sheet two of the riemann surface ( which is equivalent to sheet @xmath125 ) .
the number of the sheet is indicated in a circle on fig .
the path then heads upward and to the left , but it does not cross itself because it is on the second sheet and no longer on the zeroth sheet .
the path then proceeds to visit each sheet of the riemann surface twice before returning to its starting point at @xmath124 on the zeroth sheet .
we can calculate the period of this orbit exactly by distorting the path shown in fig .
[ f6 ] into a path that begins at the origin , travels outward to a turning point along a ray , encircles the turning point , and then returns to the origin , and then repeats this trip five more times .
however , in evaluating the complex line integrals we must be careful to remember that the _ phase _ of the integrand changes as the path encircles the turning point .
upon adding together the contributions from all six turning points we obtain the exact result @xmath126\gamma(10/9)\gamma(1/3)/\gamma(4/9)= 3.377\,981\,999\,015\ldots~.$ ] in fig .
[ f7 ] we plot the classical trajectory of a particle of energy @xmath51 governed by the hamiltonian @xmath122 with @xmath123 and @xmath127 .
this trajectory begins at @xmath128 .
the time for the particle to follow the path shown in this figure is @xmath129 , which is precisely @xmath130 of the period of the closed orbit shown in fig .
we emphasize that the trajectory of this orbit is _ not yet closed_. figure [ f7 ] shows that the particle must trace this path three times before the orbit closes , and the time required to do this is exactly the period of the orbit shown in fig .
. one might think ( wrongly ! ) that there should be a total of _ six _ rather than three turning points on fig .
we determine the positions of the turning points from the first of hamilton s equations @xmath131 the turning points are the points where @xmath132 , and this seems to occur when @xmath133 and when @xmath134 .
however , if we then substitute each of these conditions into the equation @xmath135 , we find that only the first of these conditions is consistent , so we learn that there are three turning points situated at the three roots of @xmath136 .
the second condition is inconsistent .
in this section we consider the hamiltonian @xmath137 the substitution @xmath18 converts the formal eigenvalue problem @xmath138 into the differential - equation eigenvalue problem @xmath139 the asymptotic behavior of the eigenfunctions @xmath21 has the form @xmath140 , where @xmath141 is a positive constant .
thus , the eigenfunctions vanish exponentially fast in a pair of @xmath0-symmetric wedges in the lower - half complex-@xmath22 plane centered about @xmath142 and @xmath143 .
* step 1 : * we map this differential - equation eigenvalue problem onto the real axis by taking @xmath144 in ( [ eq21 ] ) so that @xmath145 .
the mapped eigenvalue problem is @xmath146\phi(t)\nonumber\\ & & \qquad -4^{16}\gamma(1+it)^4\phi(t)=e\phi(t ) .
\label{eq40}\end{aligned}\ ] ] * step 2 : * taking the fourier transform of ( [ eq40 ] ) , we get @xmath147f(p ) \nonumber\\ & & \qquad -4^{16}\gamma\left(1+\hbar{\textstyle{\frac{\rm d}{{\rm d}p}}}\right)^4f(p)=ef(p )
. \label{eq41}\end{aligned}\ ] ] * step 3 : * we transform the dependent variable using @xmath110 and choose @xmath38 to eliminate the third - derivative term @xmath148 .
the function @xmath38 that does this is @xmath149 the result in ( [ eq42 ] ) simplifies ( [ eq41 ] ) to @xmath150\nonumber\\ & & \quad+\textstyle{\frac{\hbar^2\eta^2}{4^{19}\gamma}}\left[24p^8g''(p)+240p^7 g'(p)+483p^6g(p)\right]\nonumber\\ & & \quad-\textstyle{\frac{\hbar^3\eta}{32}}\left[-48p^3g''(p)-6p^2g'(p ) + 87pg(p)\right]=eg(p ) .
\label{eq43}\end{aligned}\ ] ] * step 4 : * next , we perform the scaling transformation @xmath151 and obtain the _ real _ eigenvalue problem @xmath152\nonumber\\ & & \quad+\hbar^2 4 ^ 5\sqrt{\gamma\eta}\left[24p^8g''(p)+240p^7g'(p)+483p^6g(p ) \right]\nonumber\\ & & \quad-8\hbar^3\gamma^{1/4}\eta^{3/4}\left[-48p^3g''(p)-6p^2g'(p)+87pg(p ) \right]=eg(p ) , \label{eq45}\end{aligned}\ ] ] whose boundary conditions are posed on the real axis .
finally , we identify the hamiltonian that gives rise to this eigenvalue problem by substituting @xmath116 : @xmath153 we have thus shown that the eigenvalues of @xmath154 are real and are identical to the eigenvalues of @xmath155 .
notice that @xmath154 has first- , second- , and third - order anomaly terms . in the classical limit @xmath46
we obtain @xmath156 we will now demonstrate that the two classical hamiltonians , @xmath157 and @xmath158 , are equivalent by showing that the classical periods of the motion are identical .
let us first examine the complex particle motion due to @xmath157 for the case @xmath123 and @xmath50 . in fig .
[ f8 ] we plot a closed orbit of a particle of energy @xmath51 that begins at @xmath159 .
the orbit takes the approximate shape of a rectangle and makes @xmath66 turns at the four corners .
the period of this orbit is @xmath160 .
we can calculate this period exactly by distorting it into rays that go from the origin out to each of the turning points and back to the origin , taking great care to keep track of the phase of the integrand .
the exact formula for the period is @xmath161\gamma(1/4)\gamma(17/16)/ \gamma(5/16)$ ] .
next , we examine the classical trajectories of @xmath158 for the case @xmath123 and @xmath50 . in fig .
[ f9 ] we plot the classical trajectory for a particle of @xmath51 that begins at @xmath162 .
there are four turning points of fourth - root type .
however , in this figure it is not possible to see the detailed structure of this trajectory near the origin .
therefore , in fig .
[ f10 ] we plot a blow - up of the square region in fig .
[ f9 ] near the origin . in fig .
[ f10 ] we can now see the four turning points and a rather complicated trajectory .
the regions surrounding the upper pair of turning points have a highly detailed structure , and thus the square region surrounding the upper - right turning point is blown up again in fig .
[ f11 ] ( left side ) .
the vicinity of the turning point contains additional structure , and this region must be blown up still more in fig . [ f11 ] ( right side ) .
we have shown by using a number of examples that it is possible to construct an infinite tower of pairs of isospectral hamiltonians @xmath1 and @xmath2 , for which the first member of the pair is a complex @xmath0-symmetric hamiltonian .
the differential - equation eigenvalue problem for the second hamiltonian is entirely real , and therefore the eigenvalues of both hamiltonians are real .
the second member of the pair has quantum anomalies of order @xmath163 .
we have also shown that at the classical level , where the anomaly terms in @xmath2 are discarded , the hamiltonians are equivalent by demonstrating that they have closed orbits of the same period .
many unsolved problems remain .
for example , we do not know if for _ every _ classical orbit of the hamiltonians @xmath7 there is a corresponding classical orbit of the hamiltonians @xmath8 of exactly the same period . because complex coordinate space is a riemann surface having an elaborate sheet structure , it is possible to find various classical orbits having many different periods .
for example , for the classical hamiltonian @xmath121 with @xmath123 and @xmath50 , we find that a particle of energy @xmath51 that starts at @xmath164 has a closed trajectory of period @xmath165\gamma(10/9)\gamma(4/3)/\gamma(4/9)= 4.143\,708\,353\ldots$ ] .
this orbit is shown in fig .
[ f12 ] .
we calculate the period of this orbit exactly by distorting the path shown in fig .
[ f12 ] into a path that begins at the origin , travels outward to a turning point along a ray , encircles the turning point , and then returns to the origin , and then repeats this trip five more times . in the complex line
integrals the _ phase _ for each path advances as the path encircles the turning point . upon adding together the contributions from all six turning points we obtain an analytic expression for the period of the motion .
although we are convinced that one exists , we have not yet been able to find a corresponding orbit of the hamiltonian @xmath122 .
h. f. jones , j. mateo , r. j. rivers , phys .
d * 74 * , 125022 ( 2006 ) .
a. a. andrianov , phys .
d * 76 * , 025003 ( 2007 ) .
h. f. jones and r. j. rivers , phys .
d * 75 * , 025023 ( 2007 ) .
f. scholtz , h. geyer , and f. hahne , ann .
phys . * 213 * , 74 ( 1992 ) . c. m. bender , s. f. brandt , j .- h .
chen , and q. wang , phys .
d * 71 * , 025014 ( 2005 ) . c. m. bender and p. d. mannheim , arxiv : hep - th/0706.0207 .
h. f. jones , arxiv : hep - th/0711.4967 .
d. krejcirik , h. bila , and m. znojil , j. phys .
a : math . gen . * 39 * , 10143 ( 2006 ) . c. m. bender , d. d. holm , and d. w. hook , j. phys . a : math .
theor . * 40 * , f81-f89 ( 2007 ) . c. m. bender , d. d. holm , and d. w. hook , j. phys .
a : math . theor . * 40 * , f793-f804 ( 2007 ) . c. m. bender ,
chen , d. w. darg , and k. a. milton , j. phys . a : math . gen .
* 39 * , 4219 - 4238 ( 2006 ) .
a. fring , j. phys .
a : math . theor . * 40 * , 4215 ( 2007 ) . c. m. bender and d. w. darg , j. math .
phys . * 48 * , 042703 ( 2007 ) .
a. nanayakkara , czech .
j. phys . * 54 * , 101 ( 2004 ) and j. phys .
a : math . gen . * 37 * , 4321 ( 2004 ) . | a technique for constructing an infinite tower of pairs of @xmath0-symmetric hamiltonians , @xmath1 and @xmath2 ( @xmath3 ) , that have exactly the same eigenvalues is described and illustrated by means of three examples ( @xmath4 ) .
the eigenvalue problem for the first hamiltonian @xmath5 of the pair must be posed in the complex domain , so its eigenfunctions satisfy a complex differential equation and fulfill homogeneous boundary conditions in stokes wedges in the complex plane .
the eigenfunctions of the second hamiltonian @xmath2 of the pair obey a real differential equation and satisfy boundary conditions on the real axis .
this equivalence constitutes a proof that the eigenvalues of both hamiltonians are real .
although the eigenvalue differential equation associated with @xmath2 is real , the hamiltonian @xmath2 exhibits quantum anomalies ( terms proportional to powers of @xmath6 ) .
these anomalies are remnants of the complex nature of the equivalent hamiltonian @xmath1 . for the cases
@xmath4 in the classical limit in which the anomaly terms in @xmath2 are discarded , the pair of hamiltonians @xmath7 and @xmath8 have closed classical orbits whose periods are identical . |
consider a distributed system composed of a team of mobile computational entities , called _ robots _ , moving and operating in the euclidean plane @xmath10 , initially each at a distinct point .
each robot can move freely in the plane , and operates in _ look_-_compute_-_move_cycles . during a cycle
, a robot determines the position ( in its own coordinate system ) of the other robots ( _ look _ ) ; it executes a protocol ( which is deterministic and it is the same for all robots ) to determine a destination point ( _ compute _ ) ; and moves towards the computed destination ( _ move _ ) . after each cycle
, a robot may be inactive for an arbitrary but finite amount of time .
the robots are anonymous , without a central control , and oblivious ( i.e. , at the beginning of a cycle , a robot has no memory of any observation or computation performed in its previous cycles ) .
what is computable by such entities has been the object of extensive research within distributed computing ; e.g. , see @xcite ; for a recent review see @xcite .
vision and mobility provide the robots with _ stigmergy _ , enabling the robots to communicate and coordinate their actions by moving and sensing their relative positions ; they are otherwise assumed to lack any means of explicit direct communication .
this restriction could enable deployment in extremely harsh environments where communication is impossible or can be jammed .
nevertheless , in many other situations it is possible to assume the availability of some sort of direct communication .
the theoretical interest is obviously for weak communication capabilities .
a model employing a weak explicit communication mechanism is that of _ robots with lights _ , or _ luminous robots _ , initially suggested by peleg @xcite . in this model , each robot is provided with a local externally - visible _ light _ , which can assume colors from a fixed set .
the robots explicitly communicate with each other using these lights .
the lights are persistent ( i.e. , the color is not erased at the end of a cycle ) , but otherwise the robots are oblivious @xcite .
notice that a light with only one possible color is the same as no light ; hence the luminous robots model generalizes the classical one . both in the classical model and in that with lights , depending on the assumptions on the activation schedule and the duration of the cycles , different settings are identified . in the synchronous
setting , the robots operate in rounds , and all the robots that are activated in a round perform their cycle in perfect synchrony . in this case ,
the system is _ fully synchronous _ ( or fsynch ) if all robots are activated at all rounds , and it is _ semi - synchronous _ ( or ssynch ) otherwise . in the _ asynchronous _ setting ( or asynch )
, there is no common notion of time , and no assumption is made on the timing and duration of each computation and movement , other than that it is finite .
( for the ssynchand asynchmodels there are bland fairness assumptions that prevent robots from remaining inactive forever , which are discussed in section [ s : robotmodels ] ) .
the choice of when a robot is activated ( in ssynch ) and the duration of an activity within a cycle ( in asynch ) is made under the control of an _ adversary _ , or _
scheduler_. similarly , the choices of the initial location of each robot and of its private coordinate system are made under adversarial conditions .
a crucial distinction is whether or not the adversary has also the power to stop a moving robot before it reaches its destination .
if so , the moves are said to be _ non - rigid_. the only constraint is that , if interrupted before reaching its destination , the robot moves at least a minimum distance @xmath11 ( otherwise , the adversary would be able to prevent robots from reaching any destination , in any amount of cycles ) .
if the adversary does not have such a power , the moves are said to be _
rigid_. the model with rigid moves is referred to as rigid , and the other one is called non - rigid . in the rest of the paper , with abuse of terminology , we will often refer to fsynch , ssynch , or asynchrobots or schedulers ( as opposed to systems ) , and to rigidor non - rigidrobots or schedulers ( as opposed to models ) .
the classical model and the more recent model of robots with lights share a common assumption : that three or more collinear robots are mutually visible .
it can be easily argued against such an assumption , and for the importance of investigating computability when visibility is _ obstructed _ by the presence of other robots : that is , if two robots @xmath12 and @xmath13 are located at @xmath14 and @xmath15 at time @xmath2 , they can see each other if and only if no other robot lies on the segment @xmath16 at that time .
very little is known on computing with obstructed visibility .
in fact , the few studies on obstructed visibility have been carried out in other models : the model of robots in the _ one - dimensional _ space @xmath17 @xcite ; and the so - called _ fat robots _
model , where robots are not geometric points but occupy unit disks , and collisions are allowed and can be used as an explicit computational tool ( e.g. , @xcite ) . in our model , collisions can create unbreakable symmetries : since robots are oblivious and anonymous and execute the same protocol , if @xmath18 ( a collision ) , then the activation adversary can force @xmath19 for all @xmath20 if the two robots do not have lights or their lights have the same color . thus , unless this is the intended outcome , collision avoidance is always a requirement for all algorithms in the model considered here .
in this paper we focus on luminous robots in the presence of obstructed visibility , and investigate computing in such a setting .
clearly , obstructed visibility increases the difficulty of solving problems without the use of additional assumptions .
for example , with unobstructed visibility , every active robot can determine the total number @xmath21 of robots at each activity cycle .
with obstructed visibility , unless a robot has a - priori knowledge of @xmath21 and this knowledge is persistently stored , the robot might be unable to decide if it sees all the robots ; hence it might be unable to determine the value @xmath21 .
the main problem we investigate , called mutual visibility , is perhaps the most basic in a situation of obstructed visibility : starting from arbitrary distinct positions in the plane , within finite time the robots must reach a configuration in which they are in distinct locations , they can all see each other , and they no longer move . this problem is clearly at the basis of any subsequent task requiring complete visibility .
notice that this problem does not exist under unobstructed visibility , and has never been investigated before . among the configurations that achieve mutual visibility ,
a special class is that where all robots are in a strictly convex position ; within that class , of particular interest are those where the robots are on the perimeter of a circle , possibly equally spaced .
the problems of forming such configurations ( respectively called convex formationand circle formation ) have been extensively studied both directly ( e.g. , @xcite ) and as part of the more general pattern formationproblem ( e.g. , @xcite ) .
unfortunately , none of these investigations consider obstructed visibility , and those algorithms do not work in the setting considered here .
note that a requirement of the mutual visibilityproblem is that robots stop moving after they have reached a configuration in which they all see each other . to this end , we will grant robots the ability to perform a special operation called _ termination _ , after which they can no longer be activated by the scheduler .
the termination operation is especially useful in practice when the robots have to perform several tasks in succession .
of course , even if this operation is not directly available , it can still be simulated via the addition of an extra color , which can be used by a robot to indicate ( to the other robots , as well as to itself ) that it has terminated .
moreover , if the termination operation is removed from the algorithms presented in this paper ( and some straightforward adjustments are made , which do not require extra colors ) , then a weaker form of the mutual visibilityproblem is solved , in which the robots get to permanently see each other , but they never stop moving .
that is , all obstructions are permanently removed , but the termination condition is not met . in some cases , removing the termination operation will even allow us to successfully apply our algorithms to different problems , such as circle formationand near gathering , as discussed in section [ s : extra ] . in this paper
we investigate under what conditions luminous robots can solve mutual visibilityand at what cost ( i.e. , with how many colors ) .
we establish a spectrum of results , depending on the power of the adversary , on the number @xmath5 of colors , and on the a - priori knowledge the robots have about the systems .
we first consider the case when the adversary can choose the activation schedule ( in ssynch ) and the duration of each robot s operations ( in asynch ) , but can not interrupt the movements of the robots ; that is , movements are rigid . in this case , we show the following . [
t : sum1 ] mutual visibilityis solvable without collisions by rigidrobots * with _ no _ colors in ssynch , if the robots know their number , @xmath21 ; * with @xmath22 colors in ssynch , always ; * with @xmath23 colors in asynch , always .
we then consider the case when the adversary has also the power to interrupt the movements of the robots ; that is , movements are non - rigid .
the only restriction is that there exists a constant absolute length @xmath11 such that , even if a robot s move is interrupted before it reaches the destination , it travels at least a length @xmath24 towards it ( otherwise it many never be able to reach any destination ) . in the case of non - rigid movements , we prove the following . [
t : sum2 ] mutual visibilityis solvable without collisions by non - rigidrobots * with _ no _ colors in ssynch , if the robots know @xmath24 and their number , @xmath21 ; * with @xmath22 colors in ssynch , if the robots know @xmath24 ; * with @xmath23 colors in ssynch , always ; * with @xmath23 colors in asynch ,
if the robots agree on the direction of one coordinate axis .
all these results are established constructively .
we present and analyze two protocols , algorithm [ alg1 ] ( _ shrink _ ) and algorithm [ alg2 ] ( _ contain _ ) , whose goal is to allow the robots to position themselves at the vertices of a convex polygon , solving convex formation , and thus mutual visibility .
these two algorithms are based on different strategies , and are tailored for different situations . protocol _
shrink _ uses two colors and requires rigid movements , while protocol _
contain _ uses more colors but operates also with non - rigid movements .
we prove their correctness for ssynchrobots ( sections [ s : ssynch ] and [ s : ssynch2 ] ) .
we then show how , directly or with simple expansions and modifications of these two algorithms , all the claimed results follow ( sections [ s : asynch ] and [ s : extra ] ) . finally , we propose some open problems ( section [ s : conclusions ] ) .
let us point out that , to prove the correctness of _ shrink _ , we solve a seemingly unrelated problem , communicating vessels , which is interesting in its own right . as a byproduct of our solutions ,
we provide the first obstructed - visibility solution to a classical problem for oblivious robots : _ collision - less convergence to a point _ ( near - gathering ) ( see @xcite ) , indeed , if the robots continue to follow algorithm _ shrink _ once they reach full visibility , the convex hull of their positions converges to a point , and the robots approach it without colliding , thus solving near - gathering(section [ s : nearg ] ) .
this algorithm has an interesting fault - tolerance property : if a single robot is faulty and becomes unable to move , the robots will still solve near - gathering , converging to the faulty robot s location ( section [ s : fault ] ) . additionally , both protocols can be modified so that the robots can position themselves on the perimeter of a circle , thus providing an obstructed - visibility solution to the classical problem of circle formation . the problem can be solved with @xmath22 colors in rigid ssynch , with @xmath23 colors in non - rigid ssynch , and with @xmath25 colors in rigid asynch , and non - rigid asynchwith agreement on one axis ( section [ s : circle ] ) .
we mostly follow the terminology and definitions of the standard model of oblivious mobile robots ( e.g. , see @xcite ) . by @xmath26
we denote a set of oblivious mobile computational entities , called _ robots _ , operating in the euclidean plane , and initially placed at distinct points .
each robot is provided with its own local coordinate system centered in itself , and its own notion of unit distance and handedness .
we denote by @xmath27 the position occupied by robot @xmath28 at time @xmath2 ; these positions are expressed here in a global coordinate system , which is used for description purposes , but is unknown to the robots .
two robots @xmath12 and @xmath13 are said to _ collide _ at time @xmath2 if @xmath18 .
a robot @xmath12 can _ see _ another robot @xmath13 ( equivalently , @xmath13 is _ visible _ to @xmath12 ) at time @xmath2 if and only if no other robot lies in the segment @xmath16 at that time .
the robots are luminous : each robot @xmath12 has a persistent state variable , called _ light _ , which may assume any value in a finite set @xmath29 of _ colors_. the color of @xmath12 at time @xmath2 can be seen by all robots that can see @xmath12 at that time .
the robots are autonomous ( i.e. , without any external control ) , anonymous ( i.e. , without internal identifiers ) , indistinguishable ( i.e. , without external markings ) , without any direct means of communication , other than their lights . at any time
, robots can be performing a variety of operations , but initially ( i.e. , at time @xmath30 ) they are all still and idle . when activated , a robot performs a _
look - compute - move _ sequence of operations : it first obtains a snapshot of the positions , expressed in its local coordinate system , of all visible robots , along with their respective colors ( _ look _ phase ) ; using the last obtained snapshot as an input , the robot executes a deterministic algorithm , which is the same for all robots , to compute a destination point @xmath31 ( expressed in its local coordinate system ) and a color @xmath32 , and it sets its light to @xmath5 ( _ compute _ phase ) ; finally , it moves towards @xmath33 ( _ move _ phase ) .
it then starts a new cycle , whenever the scheduler ( which is an abstract entity controlling to some extent the behavior of the robots ) decides to activate it again . in the _
compute _ phase , a robot may also decide to terminate its execution .
when a robot has terminated , it remains still forever , and its light remains the same color that it was at the moment of termination .
the robots are _ oblivious _ in the sense that , when a robot transitions from one cycle to the next , all its local memory , except for the light , is reset . in other words ,
a robot has no memory of past computations and snapshots , except for the light . with regards to the activation and timing of the robots , there are two basic settings : _ semi - synchronous _ ( ssynch ) and _ asynchronous _ ( asynch ) . in ssynch ,
the time is discrete ; at each time instant @xmath34 ( called a _ round _ or a _ turn _ ) a subset of the robots is activated by the scheduler and performs a whole _
look - compute - move _ cycle atomically and instantly . at
any given round , any subset of robots may be activated , from the empty set to all of @xmath35 . in particular ,
if all robots are activated at every round , the setting is called _ fully synchronous _ ( fsynch ) .
there is a bland _ fairness _ constraint on the choices that the scheduler can make : every robot must be activated infinitely many times ( unless it terminates ) . in asynch , there is no common notion of time : each robot executes its cycles independently , the _ look _ operation is instantaneous , but the _ compute _ and _ move _ operation can take an unpredictable ( but finite ) amount of time , unknown to the robot . in a _ move _ phase
there are no constraints on the speed of a robot , as long as it always moves directly towards its destination point at non - negative speed .
the scheduler that controls the activations ( in ssynch ) and the durations of the operations ( in asynch ) can be thought of as an _ adversary _ , whose purpose is to prevent the robots from doing their task .
other than acting as a scheduler , the adversary also determines the initial position of the robots and their local coordinate systems ; in particular , the coordinate system of a robot might not be preserved over time and might be modified by the adversary between one cycle and the next . in the simplest model , the robots do not necessarily agree on the orientation of the coordinate axes , on the unit distance , and on the clockwise direction ( i.e. , the handedness of the system ) .
however , in section [ sec : axis ] , we will discuss the special model in which all the robots agree on the direction of one axis , and the adversary is unable to change it .
the adversary might or might not have the power to interrupt the movement of a robot before it reaches its destination in the _ move _ operation .
if it does , the system is said to be non - rigid .
the only constraint on the adversary is that there exists a constant @xmath11 such that , if interrupted before reaching its destination , a robot moves at least a distance @xmath24 .
the value of @xmath24 is decided by the scheduler once and for all , and typically it is not known by the robots , which therefore can not use it in their computations ( we will discuss the scenario in which the robots know the value of @xmath24 in section [ s : delta ] ) .
notice that , without this constraint , the adversary would be able to prevent a robot from reaching any given destination in a finite number of turns .
if movements are not under the control of the adversary , and every robot reaches its destination at every turn , the system is said to be rigid .
the mutual visibilityproblem requires @xmath21 robots to form a configuration in which they occupy @xmath21 distinct locations , and no three of them are collinear .
subsequently , the robots have to terminate .
a protocol @xmath36 is a solution of mutual visibilityif it allows the robots to solve mutual visibilitystarting from any initial configuration in which their positions are all distinct , and regardless of the decisions of the adversary ( including the activation schedule , the local coordinate systems of the robots , and the value of @xmath24 ) .
let us stress that , since robots are oblivious and anonymous and execute the same protocol , if @xmath18 ( a collision ) , then the adversary can force @xmath19 for all @xmath20 if the two robots do not have lights or their lights have the same color . hence the two robots will never again occupy distinct locations , and will no longer be able to solve mutual visibility . thus , collision avoidance of robots with the same color is a requirement for any solution protocol . among the configurations that solve the mutual visibilityproblem
, a special class is that in which all robots are in a strictly convex position . within this class , of particular interest
are the configurations in which the robots lie on the perimeter of a circle . among these , there are the notable configurations in which the robots occupy the vertices of a regular @xmath21-gon .
the problems of forming such configurations are called convex formation , circle formation , and uniform circle formation , respectively .
a finite set of points @xmath37 is said to be _ convex _ if all the points of @xmath38 lie on the perimeter of the convex hull of @xmath38 .
if a point @xmath1 of a convex set @xmath38 lies in the relative interior of an edge of the convex hull of @xmath38 , then @xmath1 is said to be a _ degenerate _ vertex of the convex hull .
if none of the points of a convex set @xmath38 is a degenerate vertex of the convex hull , then @xmath38 is said to be a _
strictly convex _ set . on the other hand
, we will say that a polygon is _ degenerate _ if its area is zero .
these two notions of degeneracy are used in different contexts ( one refers to vertices , the other refers to whole polygons ) , hence they can hardly be confused .
let @xmath39 denote the convex hull of @xmath40 at time @xmath2 .
the robots lying on its boundary are called _ external robots _ at time @xmath2 , while the ones lying in its interior are the _ internal robots _ at time @xmath2 .
observe that a robot may not know where the convex hull s vertices are located , because its view may be obstructed by other robots .
however , it can easily determine whether it is an external or an internal robot .
in fact , a robot @xmath41 is external at time @xmath2 if and only if there is a half - plane bounded by a straight line through @xmath14 whose interior contains no robots at time @xmath2 . in other words , @xmath12 is external if and only if it lies on the boundary of the convex hull of the robots that it can currently see .
note also that the neighbors of an external robot on its visible convex hull are indeed its neighbors on the actual convex hull .
if , in addition , @xmath12 lies at a non - degenerate vertex of the ( visible ) convex hull , it is said to be a _
vertex _ robot .
moreover , a robot is able to tell if @xmath42 is a line segment , i.e. , if all the robots are collinear . in particular
, if a robot can see only one other robot , it understands that it is an _
endpoint robot_. conversely , non - endpoint robots can always see more than one other robot .
the points of @xmath10 are treated like vectors , and as such they can be added , subtracted , multiplied by scalars , etc .
the dot product between vectors @xmath43 and @xmath44 will be indicated by the expression @xmath45 .
in this section we consider the mutual visibilityproblem in the rigid ssynchsetting .
we present and analyze a protocol , algorithm [ alg1 ] ( _ shrink _ ) , and we prove it solves mutual visibilityin such a setting using only two colors . the main idea of algorithm [ alg1 ] is to make only the external robots move , so as to shrink the convex hull .
when a former internal robot becomes external , it starts moving as well .
eventually , all the robots reach a strictly convex configuration , and at this point they all see each other and they can terminate .
[ alg1 ] @xmath46 myself @xmath47 @xmath48 convex hull of @xmath49 if an active robot @xmath50 , located at @xmath1 , realizes that it is not a vertex robot , it does not move .
otherwise , it locates its clockwise and counterclockwise neighbors ( in its own coordinate system ) on the convex hull s boundary , say located at @xmath43 and @xmath44 , which are necessarily visible .
then , @xmath50 attempts to move somewhere in the triangle @xmath51 , in such a way to shrink the convex hull , and possibly make one more robot become a vertex robot . to avoid collisions with other robots that may be moving at the same time , @xmath50 s movements are restricted to a smaller triangle , shaded in gray in figure [ fig1 ] . moreover , to avoid becoming a non - vertex robot , @xmath50 does not cross any line parallel to @xmath52 that passes through another robot , and it carefully positions itself on the closest of such lines , as shown in figure [ fig1:a ] . in particular ,
if no such line intersects the gray area , @xmath50 makes a _ default move _ , and it moves halfway toward the midpoint of the segment @xmath52 , as indicated in figure [ fig1:b ] . in order to recognize that the mutual visibilityproblem has been solved , and to correctly terminate , the robots carry visible lights of two possible colors : namely , @xmath53 .
all robots lights are initially set to _
if an active robot realizes that it is a vertex of the convex hull , it sets its light to the other value , _
vertex_. hence , when a robot sees only robots whose lights are set to _
vertex _ , it knows it can see all the robots in the swarm , and hence it terminates .
the above rules are sufficient to solve the mutual visibilityproblem in most cases , but there are some exceptions .
it is easy to see that there are configurations in which mutual visibilityis never solved until an internal robot moves , regardless of the algorithm employed .
for instance , suppose that the configuration is centrally symmetric , with one robot lying at the center .
let the local coordinate systems of any two symmetric robots be oriented symmetrically and have the same unit distance , and assume that the scheduler chooses to activate all robots at every turn .
then , every two symmetric robots have symmetric views , and therefore they move symmetrically . if the central robot which is an internal robot never moves , then the configuration remains centrally symmetric , and the central robot always obstructs all pairs of symmetric robots .
hence mutual visibilityis never solved , no matter what algorithm is executed .
it turns out that our rules can be fixed in a simple way to resolve also this special case : whenever an internal robot sees only robots whose lights are set to _
( except its own light ) , it moves to the midpoint of any edge of the convex hull .
finally , the configurations in which all the robots are collinear need special handling . in this case
it is impossible to solve mutual visibilityunless some robots leave the current convex hull .
suppose that a robot @xmath12 realizes that all robots lie on a line , and that it is not an endpoint ( i.e. , @xmath12 can see only two other robots , which are collinear with it ) .
then , @xmath12 moves by any positive amount , orthogonally to the line formed by the other two visible robots .
when this is done , the previous rules apply . in the following we discuss some basic invariants , which will serve to prove the correctness of algorithm [ alg1 ] .
suppose that , for some @xmath54 , @xmath55 is not a line segment : the situation is illustrated in figure [ fig2 ] .
if a vertex robot is activated , it is bound to remain in the corresponding gray triangle , called _ movement region _ of the robot .
more precisely , the movement region consists of the interior of the gray triangle , plus the vertex where the robot currently is , plus the interior of the edge that is opposite to the robot .
hence all movement regions are disjoint .
moreover , if there is only one internal robot and it sees only robots whose light is set to _
, it moves to the midpoint of an edge of @xmath39 , which does not lie in any movement region .
it follows that , no matter which robots are activated at time @xmath2 , they will not collide at time @xmath56 .
also , @xmath57 .
recall that a robot @xmath58 is a vertex robot if an only if it lies at the vertex of a reflex angle whose interior does not contain any robots .
now , referring to figure [ fig1 ] , it is clear that a vertex robot will remain a vertex robot after a move
. additionally , if no new vertex robots are acquired between time @xmath2 and @xmath56 , then the ordering of the vertex robots around the convex hull is preserved from time @xmath2 to time @xmath56 . this easily follows from the fact that every robot remains in its own movement region ( cf . figure [ fig2 ] ) .
we seek to prove that algorithm [ alg1 ] makes every robot eventually become a vertex robot . as it will be apparent in the proof of theorem [ t : correctness ] ,
the crux of the problem is the situation in which only default moves are made ( cf . figure [ fig1:b ] ) .
we first prove that , if all robots perform only default moves , then they all converge to the same point ( see lemma [ l : default ] below ) .
since we are assuming that only the vertex robots move , and that their movements depend only on the positions of other visible vertex robots , we may as well assume that all robots are vertex robots , and that their indices follow their order around the convex hull .
indeed , by the invariants observed in section [ s : invariants ] , all robots will remain vertex robots throughout the execution , and their ordering around the convex hull will remain the same .
so , let @xmath59 , @xmath50 , @xmath60 be three vertex robots , which appear on the boundary of @xmath55 consecutively in this order .
let @xmath50 perform a default move at time @xmath2 .
then , the new position of @xmath50 is a convex combination of the current positions of these three robots , and precisely @xmath61 in general , as different sets of vertex robots are activated in several rounds , and nothing but default moves are made , the new location of each robot is always a convex combination of the _ original _ positions of all the robots , obtained by applying to the set of active robots , at every round . in formulas ,
@xmath62 with @xmath63 and @xmath64 , assuming that the robots start making only default moves at time @xmath65 .
let @xmath66 .
we fix @xmath67 , and we let @xmath68 , where indices are taken modulo @xmath21 . we claim that @xmath69 if such a claim is true ( for all @xmath67 ) , it implies that the robots get arbitrarily close to each other , as @xmath2 grows .
this , paired with the fact that @xmath70 for every @xmath2 , as observed in section [ s : invariants ] , allows us to conclude that the robots converge to the same limit point .
a proof of this statement can be obtained using the theory of convergence of asynchronous algorithms in the book @xcite .
indeed , the update rule corresponds to performing time stepping on a markov chain with circulant transition matrix @xmath71 it is proven in a statement on @xcite that the time - stepping iteration converges even when performed asynchronously , under a model that generalizes our ssynch .
nevertheless , we give here an alternative self - contained proof .
first we reformulate the problem in the following terms .
[ [ section ] ] suppose that @xmath21 vessels containing water are arranged in a circle , and there is a pipe between each pair of adjacent vessels , regulated by a valve . at every second , some of the valves are opened and others are closed , in such a way that each of the @xmath21 valves stays open for infinitely many seconds , in total . if a valve between two adjacent vessels stays open between seconds @xmath2 and @xmath56 , then @xmath72 of the surplus of water , measured at second @xmath2 , flows from the fuller vessel to the emptier one .
our claim is that the amount of water converges to the same limit in all vessels , no matter how the valves are opened and closed .
we call this problem communicating vessels
. in this formulation , the amount of water in the @xmath73-th vessel at time @xmath54 would be our previous @xmath74 .
however , here we somewhat abstract from the mutual visibilityproblem , and we consider a slightly more general initial configuration , in which the @xmath75 s are arbitrary real numbers . this problem is a special case of a diffusion model on a simple circular graph . to solve it , we shall introduce a quadratic energy functional @xmath76 , and prove that it is decreasing .
the use of such an energy functional in this class of problems is well known in the literature ( see for instance @xcite ) , but the fact that the iteration is performed semi - synchronously on each node separately is less standard , so we need to do a little more work .
we set @xmath77 if the valve between the @xmath73-th and the @xmath78-th vessel is open between time @xmath2 and @xmath56 ( indices are taken modulo @xmath21 ) , and @xmath79 otherwise .
it is easy to verify that activating robot @xmath50 at time @xmath2 in our previous discussion corresponds to setting @xmath77 in the communicating vesselsformulation .
let us denote by @xmath80 the vector whose @xmath73-th entry is @xmath74 , and let @xmath81 .
we first prove an inequality on the euclidean norms of the vectors @xmath80 .
note that the inequality holds regardless of what assumptions are made on the opening pattern of the valves .
[ l : main ] for every @xmath54 , @xmath82 for brevity , let @xmath83 , @xmath84 , @xmath85 ; hence , @xmath86 and @xmath87 .
suppose first that @xmath88 , i.e. , both valves connecting the @xmath73-th vessel with its neighbors are open .
then , @xmath89 . we have @xmath90 which can be obtained by dropping the term @xmath91 from the algebraic identity @xmath92 now , suppose instead that @xmath93 and @xmath79
. then we have @xmath94 , and @xmath95 where the first equality comes from the identity @xmath96 if @xmath97 and @xmath77 , an analogous argument gives @xmath98 finally , if @xmath99 , @xmath100 , and trivially @xmath101 we sum for each @xmath102 the relevant inequality among , , , , depending on the value of @xmath103 and @xmath104 . each of the terms @xmath105 appears twice if and only if @xmath77 , and the coefficients of the terms in @xmath106 sum to 1 for every @xmath73 , hence we get . from the previous lemma
, it immediately follows that the sequence @xmath107 is non - increasing .
since it is also bounded below by @xmath108 , it converges to a limit , which we call @xmath109 .
let @xmath110 and @xmath111 .
observe that each entry of @xmath112 is a convex combination of entries of @xmath80 , hence @xmath113 is non - increasing and @xmath114 is non - decreasing .
therefore they both converge , and we let @xmath115 and @xmath116 .
[ c : limits ] @xmath117 for every @xmath54 , we have @xmath118 which proves the second inequality . as for the first inequality , for every @xmath119 and large - enough @xmath2
, we have @xmath120 . for the next lemma ,
we let @xmath121 .
[ l : vessels ] suppose that @xmath122 for at least @xmath123 distinct values of @xmath102 .
then , @xmath124 due to corollary [ c : limits ] , it is enough to prove that @xmath125 . by contradiction ,
assume @xmath126 , and let @xmath127 .
we have @xmath128 hence there exists @xmath129 such that @xmath130 for every @xmath131 . by lemma
[ l : main ] , @xmath132 for every @xmath131 and every @xmath73 such that @xmath77 .
this implies @xmath133 , that is , a necessary condition for the valve between the @xmath73-th and the @xmath78-th vessel to be open at time @xmath131 is that @xmath134 .
consider now the @xmath135 open intervals @xmath136 each of width @xmath24 .
since @xmath137 and @xmath138 , there are @xmath139 s above and below all these intervals . moreover , by the pigeonhole principle , at least one of the intervals contains no @xmath139 s , for any @xmath102 . in other words
, we can find a partition @xmath140 , with @xmath141 and @xmath142 both non - empty , and a threshold value @xmath143 such that @xmath144 for every @xmath145 , and @xmath146 for every @xmath147 .
hence , at time @xmath148 , only valves between entries of @xmath80 whose indices belong to the same @xmath149 can be open .
it is now easy to prove by induction on @xmath131 the following facts : * @xmath150 , * @xmath151 , * @xmath79 whenever @xmath73 and @xmath152 belong to two different classes of the partition .
since @xmath141 and @xmath142 are non - empty , there must be at least two distinct indices @xmath153 and @xmath154 such that @xmath155 and @xmath156 ( where indices are taken modulo @xmath21 ) .
it follows that the @xmath157-th and @xmath158-th valve are never open for @xmath131 , and this contradicts the hypothesis that @xmath159 for at most one choice of @xmath102 .
this solves the communicating vesselsproblem .
[ c : vessels ] under the hypotheses of lemma [ l : vessels ] , for every @xmath102 , @xmath160 by lemma [ l : vessels ] , since @xmath161 , all the limits coincide . moreover
, the sum of the @xmath74 s does not depend on @xmath2 ; hence their average , taken at any time , must be equal to the joint limit .
let us return to the mutual visibilityproblem , to prove our final lemma .
[ l : default ] if , at every round , each robot makes a default move ( cf .
figure [ fig1:b ] ) or stays still , all external robots have their lights set to _
vertex _ , and
no new robots become vertex robots or terminate , then all robots locations converge to the same limit point . as discussed at the beginning of section [ s : convergence ] , this is implied by . recall that @xmath162 , and hence @xmath163 .
then , follows immediately from corollary [ c : vessels ] .
we are now ready to prove our main theorem .
[ t : correctness ] algorithm [ alg1 ] solves mutual visibilityfor rigid ssynchrobots with 2-colored lights .
if the initial convex hull is a line segment , it becomes a non - degenerate polygon as soon as one or more of the non - vertex robots are activated .
it is also easy to observe ( cf .
figure [ fig2 ] ) that , from this configuration , the convex hull may never become a line segment .
so the invariants discussed in section [ s : invariants ] apply , possibly after a few initial rounds : no two robots will ever collide , and a vertex robot will never become a non - vertex robot .
assume by contradiction that the execution never terminates .
note that a robot terminates if and only if all robots terminate .
indeed , if there are any non - vertex robots ( whose lights are still set to _ off _ ) , then each vertex robot can see at least one of them .
hence we are assuming that all robots execute the algorithm forever . at some point ,
the set of vertex robots reaches a maximum @xmath164 , and as soon as all of these robots have been activated , they permanently set their lights to _ vertex_. let @xmath165 be a time at which all the robots in @xmath166 have their lights set to _
suppose that there are external robots that are not vertex robots after time @xmath148 , and let @xmath12 be one such robot that is adjacent to a vertex robot @xmath167 .
then , after @xmath167 is activated and moves , @xmath12 becomes a vertex robot as well , contradicting the maximality of @xmath166 .
hence the external robots are exactly the robots in @xmath166 , and no other robot may become external after time @xmath148 .
if there is only one internal robot at time @xmath131 , it becomes external as soon as it is activated , due to line 23 of the algorithm , which is impossible , as argued in the previous paragraph .
therefore there are at least two internal robots at every time @xmath131 . on the other hand ,
if a vertex robot makes a non - default move at any time @xmath131 , a new robot becomes external at time @xmath56 .
indeed , referring to figure [ fig1:a ] , the line @xmath168 passes through @xmath169 and @xmath170 , and no robot lies above this line at time @xmath56 .
hence @xmath5 becomes a new external robot , which again is impossible . as a consequence ,
only default moves are made after time @xmath148 .
moreover , no robot becomes external or becomes a vertex robot after time @xmath148 , and no robot ever terminates . therefore lemma [ l : default ] applies , and the robots converge to the same limit point .
but since there are at least two internal robots , this means that at least one of them has to move , implying that it becomes a vertex robot at some point ( by the above assumptions , only vertex robots can move ) , a contradiction
. hence the execution terminates , meaning that at some point one of the robots sees only vertex robots .
this implies that there are no non - vertex robots , hence the configuration is strictly convex , all robots can see each other , and they all terminate without moving as soon as they are activated , thus solving the mutual visibilityproblem .
here we give a protocol , algorithm [ alg2 ] ( _ contain _ ) , for the mutual visibilityproblem that works for non - rigidrobots and the ssynchscheduler .
recall that , in the non - rigidmodel , the robots make unreliable moves , that is , the scheduler can stop them before they reach their destination point , but not before they have moved by at least a constant @xmath11 .
since these robots are weaker than the ones considered in section [ s : ssynch ] , they will require lights of three possible colors , as opposed to two .
our goal is also to design an algorithm that can be applied to robots in the rigid asynchmodel , as well as the non - rigid ssynchone .
this model will be discussed in section [ s : rasynch ] . in order to do this , we introduce a couple of extra technical subtleties into algorithm [ alg2 ] , which are irrelevant here , but will turn out to be necessary in section [ s : rasynch ] .
algorithm [ alg2 ] consists of three phases , to be executed in succession : a _ segment breaking _ phase , an _ interior depletion _ phase , and a _ vertex adjustments _ phase .
the first phase deals with the special configuration in which the robots are all collinear , and makes them not collinear .
if the robots are not initially collinear , this phase is skipped . in the second phase , the internal robots move toward the boundary of the convex hull , thus forming a convex configuration , perhaps with some degenerate vertices . in the third phase the robots ( which are now all external ) make small movements to finally reach a strictly convex configuration .
three colors are used by the robots : @xmath171 .
initially , all robots lights are set to _
off_. for added clarity , in the algorithm the line numbers of instructions belonging logically to different phases are typeset in different colors , according to the following table .
[ cols= " > , < " , ] [ alg2 ] @xmath46 myself @xmath47 @xmath48 convex hull of @xmath49 @xmath172 boundary of @xmath42 recall that we denote by @xmath39 the convex hull of the positions of all the robots at time @xmath173 . in this section
we also denote by @xmath174 the convex hull of the positions of the internal robots at time @xmath173 .
note that the `` global '' notions of @xmath42 and @xmath175 may differ from the ones computed by the robots executing algorithm [ alg2 ] , because a robot may be unable to see the positions of all the other robots in the swarm . in the following discussion ,
when referring to @xmath42 and @xmath175 , we will typically mean the `` global '' ones , unless we explicitly state otherwise .
we first describe the interior depletion phase , starting from a non - collinear initial configuration . to begin with
, all the robots lights are set to _ off_. as soon as an external robot is activated , it sets its own light to _ external _ ( lines 34 , 35 ) and does not move as long as it can still see robots whose light is _ off _ ( lines 26 , 27 ) .
note that a robot @xmath12 that occupies a vertex of @xmath175 eventually becomes aware of it , by looking at the convex hull of the visible robots whose lights are _
off_. these may not all be internal robots , because perhaps not all external robots have been activated yet , but eventually @xmath12 gets to see a good - enough approximation of a `` neighborhood '' of @xmath175 , and it realizes it occupies one of its vertices .
so , when a robot understands that it lies on a vertex of @xmath175 , it moves toward the boundary of @xmath42 , part of which is also identifiable by @xmath12 .
we distinguish three cases . 1 .
if @xmath12 realizes it is the only internal robot , it moves toward the midpoint of an edge of the convex hull ( line 41 ) .
to avoid bouncing back and forth at different turns , it always chooses the closest of such midpoints .
2 . if @xmath12 realizes that @xmath175 is a line segment and it occupies one endpoint of it , it moves like in figure [ fig3:a ] .
that is , it moves to the boundary of @xmath42 , while remaining within a right angle oriented away from @xmath175 ( lines 4345 ) .
finally , if @xmath12 `` believes '' that @xmath175 is a non - degenerate polygon and that it lies on one of its vertices , it moves as in figure [ fig3:b ] ( lines 4754 ) . remember that @xmath12 may believe so even if @xmath175 is actually degenerate , because some external robots may still be _
off_. however , @xmath12 gets an approximation of @xmath175 , which we call @xmath176 , and it knows it lies on a vertex of @xmath176 , implying that it also lies on a vertex of the `` real '' @xmath175 .
now , if the internal angle of @xmath176 at @xmath14 is acute , @xmath12 moves as the robot in @xmath1 in figure [ fig3:b ] : it moves to the boundary of @xmath42 while remaining between the extensions of its two incident edges of @xmath176 .
otherwise , if the angle is not acute , @xmath12 moves as the robot in @xmath3 in figure [ fig3:b ] : it moves to the boundary of @xmath42 while staying between the two perpendiculars to its incident edges of @xmath176 .
moreover , @xmath12 actually performs the move only if it is sure that its destination point lies on the boundary of the `` real '' @xmath42 . for this reason
, it has to check if the destination point computed as described above lies on a completely - visible edge of the observed convex hull whose endpoints are both set to _ external _ ( line 53 ) .
for instance , in figure [ fig13 ] , the robot in @xmath1 can not move to the gray area even if the robots in @xmath43 and @xmath44 are set to _
external _ , because the robot in @xmath3 prevents the one in @xmath1 from seeing the whole edge @xmath52 . on the other hand
, the robot in @xmath3 can move to its own gray area , provided that @xmath43 and @xmath44 are set to _
external_. indeed , the robot in @xmath3 can see all of @xmath52 , and it is therefore sure that it is an edge of the `` real '' convex hull
. can not move even if @xmath43 and @xmath44 are set to _
external _ , because @xmath3 may be hiding some other external robots , and @xmath52 may not be an edge of the convex hull ] now to the vertex adjustments phase .
when a robot lies at a vertex of @xmath42 and it sees only robots whose light is set to _ external _ , it makes the `` default move '' of figure [ fig1:b ] , where @xmath43 and @xmath44 are the locations of its two neighbors on @xmath42 ( line 29 ) .
moreover , while doing so it also sets its light to the third value , _ adjusting _ , as a `` self - reminder '' ( line 28 ) .
so , when it is activated again , it knows it has already adjusted its position , and it terminates , after reverting its light to _ external _ ( lines 2025 ) . this way we make sure that each vertex robot adjusts its position exactly once , and we ensure termination .
when the adjustment is done , the robots at @xmath43 and @xmath44 are guaranteed to occupy vertices of @xmath42 , instead of lying in the middle of an edge .
so , each external robot becomes a vertex robot at some point , then it adjusts its position while remaining a vertex , possibly making its adjacent robots become vertices as well , and it terminates .
when all robots have terminated , the configuration is strictly convex , and therefore mutual visibilityis solved .
finally , the segment breaking phase deals with the special case in which all robots are initially collinear .
let robots @xmath12 and @xmath13 be the two endpoints of @xmath42 : as soon as one of them is activated ( possibly both ) , it sets its light to _ adjusting _ , moves orthogonally to @xmath42 , and then waits ( lines 11 , 12 ) .
meanwhile , the other robots do not do anything until some conditions are met ( lines 3134 ) .
if only @xmath12 moves , @xmath13 realizes it ( line 32 ) and sets its own light to _ external _ ( and vice versa ) .
if both @xmath12 and @xmath13 move together , some other robot realizes that it is a non - degenerate vertex of the convex hull and that it can see both @xmath12 and @xmath13 set to _ adjusting _ ( line 33 ) : in this case , it sets itself to _
external_. when @xmath12 or @xmath13 sees some robots set to _
external _ , it finally sets itself to _ external _ , as well ( lines 22 , 23 ) .
additionally , it may terminate , provided that neither of its neighboring robots on the convex hull s boundary has still its light set to _ off _
( line 24 ) and that it recognizes no robots as internal ( line 25 ) .
this is to force @xmath12 and @xmath13 to make at least one default move in the unfortunate case that a third external robot is found between them after their initial move , or gets there during the interior depletion phase ( refer to the complete discussion in section [ s : segbreak ] ) .
after this is done , the execution transitions seamlessly into one of the general cases . if @xmath177 this is not sufficient .
suppose first that @xmath178 .
then , @xmath12 and @xmath13 may move in such a way that the configuration remains centrally symmetric , with the middle robot @xmath3 obstructing @xmath12 and @xmath13 .
however , after moving once , @xmath12 and @xmath13 become _ external _ and terminate ( lines 8 , 9 )
. meanwhile @xmath3 waits until it sees both @xmath12 and @xmath13 set to _ external _ , and finally it moves orthogonally to @xmath42 ( lines 1316 ) , thus solving mutual visibilityalso in this special case .
if @xmath179 , each robot moves once ( lines 11 , 12 ) , and then it detects a situation in which it can safely terminate ( lines 69 ) .
we first prove that no collisions occur during the interior depletion phase , and then that the phase itself eventually terminates , with all the robots becoming external . in this section
we will assume that the robots are not initially collinear .
the collinear case will be discussed in section [ s : segbreak ] , and it will be shown that is seamlessly transitions into one of the other cases .
it is easy to observe that , during the interior depletion phase , all external robots keep seeing ( internal ) robots whose lights are set to _ off _ , and therefore none of them moves . on the other hand , no internal robot moves outside of the convex hull .
[ obs : depletion ] if there are internal robots at time @xmath2 , no external robot moves , and @xmath180 .
[ l : coll1 ] if @xmath12 and @xmath13 are two internal robots at time @xmath2 , then @xmath181 if @xmath12 is not activated at time @xmath2 , or it is activated but it does not move , then the left - hand side is zero , and therefore the inequality holds .
suppose now that @xmath12 moves by a positive amount , so @xmath182 is not the null vector . let @xmath109 be the line through @xmath14 that is orthogonal to the segment @xmath183 . by construction ,
@xmath12 moves in such a way that @xmath184 lies in the open half - plane bounded by @xmath109 that does not contain @xmath174 ( note that this holds _ a fortiori _ also if some external robots have not set their lights to _ external _ yet , and therefore the @xmath175 computed by @xmath12 is larger than the real one ) .
since @xmath185 , @xmath15 lies on @xmath109 or in the half - plane bounded by @xmath109 that does not contain @xmath184 .
this is equivalent to saying that the dot product between @xmath182 and @xmath186 is not positive .
[ l : coll2 ] as long as there are internal robots , no collisions occur .
if there are internal robots , every external robot sees robots whose light is set to _ off _ , and hence it does not move . by construction ,
the internal robots avoid moving on top of external robots , and therefore there can be no collision involving external robots .
suppose by contradiction that two robots @xmath12 and @xmath13 that are internal at time @xmath2 collide for the first time at @xmath56 , and therefore @xmath187 .
by lemma [ l : coll1 ] applied to @xmath12 and @xmath13 , we have @xmath188 applying lemma [ l : coll1 ] again with @xmath12 and @xmath13 inverted , we also have @xmath189 adding [ eq : coll1 ] and [ eq : coll2 ] together and doing some algebraic manipulations , we obtain @xmath190 @xmath191 @xmath192 the latter is equivalent to @xmath193 , implying that @xmath18 .
this contradicts the fact that @xmath12 and @xmath13 collide for the first time at @xmath56 .
we still have to prove that the interior depletion phase terminates , that is , eventually all robots become external .
due to observation [ obs : depletion ] , when a robot becomes external , it stops moving and remains external , at least as long as there are other internal robots .
thus , if by contradiction this phase does not terminate , the set of internal robots reaches a non - empty minimum , and from that time on no new robot becomes external .
after possibly some more turns , say at time @xmath194 , all external robots have been activated and have set their lights to _ external _ , and hence no robot changes its light any more . in the following lemmas , we will show that these assumptions on @xmath148 yield a contradiction .
we will prove that , if @xmath195 is a non - degenerate polygon , then either its area or its diameter will grow unboundedly . therefore , at some point in time , @xmath175 will not be a subset of @xmath42 any more .
( the analysis when @xmath195 is a degenerate polygon is easy , and it will be carried out in the proof of lemma [ l : intfinal ] . )
recall that , due to line 50 of algorithm [ alg2 ] , when a robot computes its destination , it remains within the extensions of its incident edges of @xmath174 .
hence , referring to figure [ fig4 ] , it is easy to observe the following .
[ o : edges ] let robots @xmath12 and @xmath13 lie at adjacent vertices of @xmath174 at time @xmath131 , and let the area of @xmath174 be positive . then , @xmath184 and @xmath196 lie on the same side of the line through @xmath14 and @xmath15 ( or possibly on the line itself ) .
[ l : intgrow ] if @xmath131 and the area of @xmath174 is positive , then @xmath197 .
let @xmath198 be the set of robots that lie at vertices of @xmath174 , at time @xmath131 .
let @xmath49 be the polygon ( illustrated in figure [ fig4 ] as a thick dashed polygon ) whose vertices are the locations at time @xmath56 of the robots of @xmath199 , taken in the same order as they appear around the boundary of @xmath175 .
note that , since the robots are non - rigidand not all of them are necessarily activated at time @xmath2 , @xmath49 is not necessarily a convex polygon .
because the property stated in observation [ o : edges ] holds for all the edges of @xmath174 and @xmath49 , we have that @xmath200 .
but , by definition of @xmath148 , none of the robots of @xmath199 ever becomes external , and hence @xmath201 is the convex hull of @xmath49 .
we conclude that @xmath202 .
[ c : areadiameter ] for @xmath131 , the area of @xmath174 and the diameter of @xmath174 do not decrease as @xmath2 increases . by lemma
[ l : intgrow ] , if @xmath203 , then @xmath204 .
hence the area of @xmath205 can not be greater than the area of @xmath206 , and the diameter of @xmath205 can not be greater than the diameter of @xmath206 .
[ c : intint ] if @xmath12 is an internal robot at time @xmath131 , and @xmath14 is not a non - degenerate vertex of @xmath174 , then @xmath12 is internal at any time @xmath207 , and @xmath208 is not a non - degenerate vertex of @xmath209 . by lemma [ l : intgrow ] , if @xmath207 , then @xmath210 .
moreover , according to line 46 of algorithm [ alg2 ] , an internal robot that does not lie at a degenerate vertex of @xmath175 does not move .
it follows that , after time @xmath2 , robot @xmath12 will not move , hence it will remain internal , and it will never lie at a non - degenerate vertex of @xmath175 .
recall that , due to line 53 of algorithm [ alg2 ] , a robot on the perimeter of @xmath175 is able to move only if it completely sees an entire edge of @xmath42 .
next we prove that , after time @xmath148 , there exists at least one robot that is able to move .
[ l : canmove ] at any time after @xmath148 , there is a robot that , if activated , makes a non - null movement .
suppose for a contradiction that no robot is able to move , and let @xmath211 be a non - degenerate vertex of @xmath175 .
if @xmath212 is the internal angle of @xmath175 at @xmath211 , we let @xmath213 be the point at which the bisector of @xmath214 intersects the perimeter of @xmath42 , as figure [ fig16 ] shows . , @xmath215 , @xmath216 , @xmath217 are unable to move ; the one in @xmath218 is able to move ] let @xmath52 be an edge of @xmath42 such that @xmath219 ( note that @xmath213 may coincide with either @xmath43 or @xmath44 ) .
since @xmath42 and @xmath175 are convex , either the segment @xmath220 is completely visible to @xmath211 , or the segment @xmath221 is . without loss of generality ,
let @xmath220 be completely visible to @xmath211 , i.e. , @xmath222 . by definition of @xmath148 , both robots in @xmath43 and
@xmath44 have their lights let to _ external_. also , note that @xmath213 lies within @xmath223 , as computed by the robot in @xmath211 executing line 52 of algorithm [ alg2 ] .
hence , by line 53 , the robot in @xmath211 can move , provided that @xmath224 .
but by assumption no robot can move , and therefore there must be one robot occupying a non - degenerate vertex of @xmath175 neighboring @xmath211 , say @xmath215 , such that the ray from @xmath211 through @xmath215 intersects the segment @xmath52 , say in @xmath225 .
observe that @xmath215 is strictly closer to the line @xmath52 than @xmath211 . by the convexity of @xmath175 , the bisector of the explementary of the internal angle at @xmath215 intersects the segment @xmath52 ,
say in @xmath226 .
in fact , @xmath226 lies between @xmath213 and @xmath225 ( refer to figure [ fig16 ] ) .
also , @xmath215 can see all the points in the segment @xmath227 . as we argued for @xmath211 in the previous paragraph
, there must be another vertex @xmath228 of @xmath175 , closer to the line @xmath52 , that prevents @xmath215 from seeing the entire segment @xmath52 .
proceeding in this fashion , we obtain a sequence @xmath211 , @xmath215 , @xmath228 , @xmath229 , @xmath216 of vertices of @xmath175 , which get closer and closer to the line @xmath52 .
since these vertices must be all distinct , and there are only finitely many robots in the swarm , there must be one last element of the sequence , @xmath218 .
it follows that @xmath218 can see all of @xmath52 , and the corresponding angle bisector intersects @xmath52 as well , say in @xmath230 .
this implies that the robot in @xmath218 can actually move to a neighborhood of @xmath230 , contradicting our assumption . as a consequence of corollary [ c : intint ] ,
no new robot becomes a non - degenerate vertex of @xmath175 after time @xmath148 , but some robots may indeed cease to be non - degenerate vertices of @xmath175 , and stop moving forever .
hence , at some time @xmath231 , the set of robots that lie at non - degenerate vertices of @xmath175 reaches a minimum @xmath232 . by lemma [ l : intgrow ] ,
the area of @xmath175 is positive at every time @xmath233 , and hence @xmath234 . in the next lemma
we prove two fundamental properties of @xmath175 that hold after time @xmath235 .
[ l : intmisc ] after time @xmath235 , the following statements hold . *
the cyclic order of the robots of @xmath166 around @xmath175 is preserved . *
the length of the shortest edge of @xmath174 does not decrease as @xmath2 increases . similarly to section [ s : ssynch ] , we call the gray regions in figures [ fig3 ] and [ fig4 ] the _ movement regions _ of the respective robots
. let @xmath233 , and consider the polygon @xmath49 as defined in the proof of lemma [ l : intgrow ] . by our assumptions on @xmath166 , @xmath49 is a convex polygon , or else some robot would cease to occupy a vertex of @xmath175 .
hence @xmath236 and , by lemma [ l : intgrow ] , @xmath237 .
let @xmath238 occupy two adjacent edges of @xmath175 at time @xmath2 , and let @xmath109 be the axis of the segment @xmath16 .
it follows from algorithm [ alg2 ] that @xmath109 separates the movement regions of @xmath12 and @xmath13 at time @xmath2 ( cf .
figure [ fig4 ] ) .
this , paired with the fact that the movement region of a robot of @xmath166 at time @xmath2 does not intersect the interior of @xmath175 , implies ( a ) .
now , to prove ( b ) , it is sufficient to note that , with the previous paragraph s notation , the distance between the movement regions of @xmath12 and @xmath13 at time @xmath2 is precisely the distance between @xmath14 and @xmath15 ( see figure [ fig4 ] ) .
therefore , the segment @xmath239 is not shorter than @xmath16 , implying that the length of the shortest edge of @xmath175 can not decrease .
we need one last geometric observation .
is not @xmath52 ] [ l : diameter ] if the internal angle at vertex @xmath241 of a convex polygon has measure less than @xmath242 , then the diameter of the polygon is the distance between @xmath241 and another vertex .
the diameter of a polygon is the longest distance between two of its vertices .
suppose for a contradiction that vertices @xmath43 and @xmath44 have the maximum distance , with @xmath243 , as in figure [ fig14 ] .
then , since the polygon is convex , the angle @xmath244 is containted in the internal angle at @xmath241 , and therefore its measure is less than @xmath242 .
since the sum of the internal angles of a triangle is @xmath245 , it follows that either @xmath246 or @xmath247 . by the law of sines , in the first case @xmath248
is longer than @xmath52 , and in the second case @xmath249 is longer than @xmath52 . in both cases , @xmath52 is not the longest segment joining two vertices of the polygon , which is a contradiction .
we are finally ready to prove the termination , and therefore the correctness , of the interior depletion phase .
[ l : intfinal ] if algorithm [ alg2 ] is executed from a non - collinear configuration , after finitely many turns all robots become external , no robot s light is set to _ adjusting _ , and no two robots collide . the non - collision part has already been proven in lemma [ l : coll2 ] , so we need to prove that eventually all robots become external . by assumption @xmath42
has positive area , and we have to show that all robots become external in finitely many turns . if there is just one internal robot , it keeps moving somewhere within @xmath42 , until it either becomes external , or all external robots have been activated . when all external robots have their lights set to _
, if there is still a single internal robot , it keeps moving toward the same edge of @xmath42 , until it finally reaches it . if there there are exactly two internal robots , they move as shown in figure [ fig3:a ] .
it is easy to see that , each time at least one of the two internal robots moves ( by at least @xmath11 ) , their distance increases by more than @xmath250 .
therefore , after finitely many turns , at least one of the two internal robots becomes external , and at most one internal robot remains .
suppose now that there are at least three internal robots , but @xmath175 has null area , that is , all the internal robots are collinear .
then , according to algorithm [ alg2 ] , only the two endpoint robots of @xmath175 are allowed to move , as figure [ fig3:a ] shows .
if they move in such a way that the internal robots keep remaining collinear , eventually one of them reaches the boundary of @xmath42 , and there is one less internal robot .
otherwise , they reach a situation in which @xmath175 has strictly positive area .
therefore we may assume that @xmath175 has positive area , and we suppose for a contradiction that some internal robots never become external .
by the previous lemmas and observations , we know that at some time @xmath235 the situation becomes `` stable '' .
specifically , @xmath42 never changes , the set @xmath166 of robots that occupy non - degenerate vertices of @xmath175 keeps remaining the same , and these robots positions preserve their order around @xmath175 , by lemma [ l : intmisc].a .
also , the area and diameter of @xmath175 do not decrease , by corollary [ c : areadiameter ] .
let @xmath251 be the length of the shortest edge of @xmath174 .
by lemma [ l : intmisc].b , we know that @xmath251 is a weakly increasing function of @xmath233 .
suppose that , at some time @xmath252 , some robot @xmath253 that is able to move is activated . by lines 53 , 54 of algorithm [ alg2 ] ,
the destination point of @xmath12 lies on the boundary of @xmath42 .
hence , the adversary must stop @xmath12 before it reaches its destination , or it would become external . but this can not happen before @xmath12 has moved by at least @xmath24 , implying that @xmath254 .
we distinguish two cases , based on the measure @xmath255 of the internal angle of @xmath256 corresponding to vertex @xmath257 .
we will prove that , if @xmath258 , then the square of the diameter of @xmath256 increases by at least a constant ; if @xmath259 , then the area of @xmath256 increases by at least a constant . increases by at least @xmath260 suppose that @xmath258 . let @xmath261 the diameter of @xmath174 , and let @xmath262 be such that @xmath263 .
due to line 50 of algorithm [ alg2 ] , and referring to figure [ fig15 ] , it is easy to prove that @xmath264 hence @xmath265 because @xmath266 , it follows that @xmath267 , and @xmath268 .
therefore , by the law of cosines applied to triangle @xmath269 , @xmath270 hence , in this case , the square of the diameter of @xmath175 increases by at least @xmath271 .
increases by at least @xmath272 let @xmath259 , and let @xmath273 occupy the two vertices of @xmath256 adjacent to vertex @xmath257 , in such a way that @xmath274 as figure [ fig5 ] shows .
it follows that @xmath275 then , recalling that @xmath266 , we have that the area of @xmath175 increases at least by the area of the triangle @xmath276 , which in turn is at least @xmath277 @xmath278 concluding , every time a robot of @xmath166 moves , either the square of the diameter of @xmath175 increases by at least @xmath271 , or the area of @xmath175 increases by at least @xmath279 . but lemma [ l : canmove ] states that there is always at least one robot of @xmath166 that is able to move , and therefore the robots will move infinitely many times , due to the fairness of the ssynchscheduler . from corollary [ c :
areadiameter ] , and from the fact @xmath271 and @xmath279 are positive constants , it follows that either the diameter or the area of @xmath175 grows unboundedly .
this contradicts the fact that @xmath280 , and that @xmath39 is independent of @xmath2 .
due to lemma [ l : intfinal ] , all robots become external at some point , and it remains to show that they finally reach a _ strictly _ convex configuration and correctly terminate .
this turns out to be a significantly easier task .
[ l : extfinal ] if algorithm [ alg2 ] is executed from a configuration in which all robots are external and no robot s light is set to _ adjusting _ , then after finitely many turns all robots have terminated , no two of them have collided , and the configuration is strictly convex . in the vertex adjustments phase , each robot eventually sets its own light to _ external _ ( if it has not already done so in the interior depletion phase ) .
meanwhile , as soon as a vertex robot @xmath12 sees only robots whose lights are set to _
external _ , it sets its own light to _ adjusting _ and makes a default move , as in figure [ fig1:b ] .
recall that robots are non - rigid , hence a vertex robot may be stopped before reaching its destination , but not before having moved by at least @xmath11 .
when @xmath12 is activated again , it sees itself in the _ adjusting _ state and , if it sees a robot set to _ off _ , it necessarily also sees a robot set to _
external_. indeed , after the default move , @xmath12 can see all the robots in the swarm .
note that , if a robot is set to _ adjusting _ , neither of its two neighbors on the perimeter of the convex hull can be _
off_. hence , if @xmath12 sees a robot set to _ off _ , and since it sees itself set to _ adjusting _ , it must also see a robot set to _
external_. therefore @xmath12 sets its light back to _ external _ ( lines 2123 ) , thus allowing other robots to move .
@xmath12 also terminates because , as already noted , its two neighbors can not be _ off _
( line 24 ) and , since @xmath12 sees every robot in the swarm and the configuration is convex , @xmath12 does not see any internal robots ( line 25 ) .
as observed in section [ s : ssynch ] , where a similar procedure was used to reduce the size of the convex hull while increasing the set of vertex robots , when a robot occupying a vertex of the convex hull moves , it becomes a vertex of the new convex hull . on the other hand ,
if a robot @xmath12 lies in the interior of an edge of the convex hull and one of its two neighbors @xmath13 lies on a vertex , then , as soon as @xmath13 moves , @xmath12 becomes a vertex of the new convex hull .
hence , eventually all robots become vertices of the convex hull . after that
, whenever a robot is activated , it permanently sets its light to _ external _ and terminates .
it follows that eventually all robots terminate in a strictly convex configuration .
moreover , by the observations already made in section [ s : invariants ] , and referring to figure [ fig2 ] , it is clear that no collisions can occur in this phase . from lemmas [ l : intfinal ] and [ l : extfinal ]
, the correctness of algorithm [ alg2 ] immediately follows , provided that the robots are not initially collinear .
this case is considered in the following .
[ t : correctness2 ] algorithm [ alg2 ] solves mutual visibilityfor non - rigid ssynchrobots with 3-colored lights . due to lemmas [ l : intfinal ] and [ l : extfinal ] , we only have to show that the case in which the robots are initially collinear correctly evolves into one of the other cases . if @xmath177 , this is easy to verify through a case analysis , following the algorithm s description of section [ s : alg2descr ] .
so , let @xmath281 , and let robots @xmath12 and @xmath13 initially occupy the vertices of the line segment @xmath282 .
nothing happens until @xmath12 or @xmath13 is activated ; then , at least one of them becomes _ adjusting _ and moves orthogonally to @xmath42 .
after @xmath12 or @xmath13 has moved , some robots eventually become _ external _ ( line 35 ) . indeed ,
if both @xmath12 and @xmath13 move , then all other robots can see both of them .
in particular , there is at least one such robot occupying a non - degenerate vertex of the convex hull , which sees two _ adjusting _ robots , and therefore becomes _ external _ ( line 33 ) .
if only @xmath12 moves ( or vice versa ) , then @xmath13 sees only three robots ( including itself ) , and its corresponding convex hull angle is acute , hence it becomes _ external _ ( line 32 ) . in the latter case , only @xmath13 is allowed to become _ external _ , while the other robots wait , because their corresponding angles are not acute .
when a robot has become _ external _ , any robot that was _ adjusting _ can see it , and hence becomes _ external _ as well .
if , in addition , neither of its two neighbors is _ off _ , it also terminates ( note that an _ adjusting _ robot must be located on the convex hull s boundary ) . at this point
the execution proceeds normally , except that there may be one of two vertex robots that have already terminated , and we have to show that this does not prevent the others from forming a strictly convex configuration . obviously the interior depletion phase causes no trouble and it is carried out correctly , but the vertex adjustments phase might `` get stuck '' .
we will prove that this is not the case .
clearly , if at most one robot has terminated , all robots except perhaps one are able to move in the vertex adjustments phase , and therefore they all become non - degenerate vertices .
if @xmath12 and @xmath13 initially move in the same direction , they become neighboring vertices , and all the other robots become consecutive external robots .
it is easy to see that in this case the external robots are still able to make adjusting movements in cascade , and become non - degenerate vertices .
finally , suppose that @xmath12 and @xmath13 initially move in opposite directions .
let @xmath167 be the robot closest to @xmath12 , and @xmath283 the robot closest to @xmath13 .
since @xmath281 , @xmath284 .
as already noted , at least one between @xmath167 and @xmath283 becomes a non - degenerate vertex of the convex hull , say @xmath167 . on the other hand , depending on how much @xmath12 and @xmath13 move
, @xmath283 may become internal , or a degenerate or non - degenerate vertex of the convex hull .
if @xmath283 becomes internal , as in figure [ fig10b : a ] , then both @xmath12 and @xmath13 see an internal robot : indeed , after the move , @xmath12 and @xmath13 can see all robots .
then , @xmath12 and @xmath13 become _ external _ but do not terminate ( line 25 ) , and everything works as intended in the later phases . if @xmath283 becomes a degenerate vertex of the convex hull , as in figure [ fig10b : b ] , then it is an _ off _ neighbor of both @xmath12 and @xmath13 , which once again become _ external _ but do not terminate ( line 24 ) .
again , the execution transitions seamlessly into another phase .
if @xmath283 becomes a non - degenerate vertex of the convex hull , as well as @xmath167 , then all the robots between them become internal .
note that , in this case , @xmath12 and @xmath13 may terminate ( indeed , they may not be able to see each other , and hence they may not realize that there are internal robots ) , but they do not lie at adjacent vertices of @xmath42 , due to the presence @xmath167 and @xmath283 . after the interior depletion phase , @xmath167 and @xmath283 are able to adjust , thus enabling all other external robots to become non - degenerate vertices , in cascade .
in this section we briefly touch on the asynchmodel . in the rigidcase , we show that algorithm [ alg2 ] solves the mutual visibilityproblem . in the non - rigidcase
, we show how to solve mutual visibilityassuming that the robots agree on the direction of one coordinate axis .
algorithm [ alg2 ] turns out to solve the mutual visibilityproblem for rigid asynchrobots , as well . for the interior depletion phase ,
the collision avoidance proof gets slightly more complex , but termination is easier to prove . on the other hand , the vertex adjustments phase and the segment breaking phase work almost in the same way .
first we state an equivalent of lemma [ l : coll1 ] .
the only difference is that , instead of a generic time @xmath173 , now we have to consider a specific time @xmath285 at which a robot @xmath12 performs a _
look_. also , instead of considering the position of @xmath12 at time @xmath56 , we consider the destination point computed after such a _ look_. after these changes , the proof of lemma [ l : coll1 ] works in the asynchcase as well , and therefore we have the following .
[ l : coll3 ] let rigid asynchrobots execute algorithm [ alg2 ] , and let @xmath12 and @xmath13 be two internal robots at time @xmath285 . if @xmath12 executes a _ look_phase at time @xmath2 , and the next destination point of @xmath12 is @xmath1 , then @xmath286 the previous lemma can be used to prove that no collisions occur during the interior depletion phase .
[ l : coll3b ] if rigid asynchrobots execute algorithm [ alg2 ] from a non - collinear configuration , no collisions occur as long as there are internal robots .
suppose for a contradiction that the internal robot @xmath12 performs a _
look_at time @xmath2 , then robot @xmath13 performs a _
look_at time @xmath207 , and they collide at time @xmath287 , in @xmath288 .
we may further assume that this is the first collision between the two robots , and therefore @xmath289 .
because the model is rigidand each internal robot s destination point is on the convex hull , it follows that each internal robot makes exactly one move and then becomes external .
therefore , @xmath14 , @xmath208 , @xmath290 , and the destination point of @xmath12 are all collinear , and the same holds for @xmath13 .
additionally , we have @xmath291 ( see figure [ fig9 ] ) . by lemma [ l : coll3 ]
applied to @xmath13 at time @xmath292 , and because @xmath293 lies between @xmath294 and the destination point of @xmath13 , we have @xmath295 on the other hand , @xmath296 , implying that @xmath297 by adding the two inequalities together , we obtain @xmath298 recall that @xmath299 and that @xmath208 lies between @xmath14 and @xmath290 , and therefore the last inequality implies @xmath300 hence @xmath301 and @xmath302 but we already know that the right - hand side is not positive , hence so is the left - hand side : @xmath303 now , by lemma [ l : coll3 ] applied to @xmath12 at time @xmath2 , and recalling that @xmath290 lies between @xmath14 and the destination point of @xmath12 , we have @xmath304 if we add together the last two inequalities and we recall that @xmath305 , we get @xmath306 because @xmath288 and @xmath305 , we finally obtain @xmath307 which is equivalent to @xmath308 , implying that @xmath18 , a contradiction .
we can now prove that algorithm [ alg2 ] works also with rigid asynchrobots .
[ t : correctness3 ] algorithm [ alg2 ] solves mutual visibilityfor rigid asynchrobots with 3-colored lights . in the interior depletion phase there can be no collisions , due to lemma [ l : coll3b ] . also , whenever an internal robot moves , its destination lies on the boundary of the `` real '' @xmath42 ( cf . line 53 of algorithm [ alg2 ] ) .
since movements are rigid , such a robot becomes external in a single move .
suppose for a contradiction that the interior depletion phase does not terminate .
then , at some point , the set of external robots reaches a maximum , all the external robots are set to _ external _ , and no robot is moving ( indeed , moving internal robots are bound to become external ) .
hence , lemma [ l : canmove ] can be applied . as a consequence
, there is some internal robot that is able to move , and which will therefore reach the convex hull s perimeter at the end of its next _ move_phase , thus becoming external .
this contradicts our assumptions .
therefore , in finite time all robots become external , and the interior depletion phase terminates . when all robots are external , none of them moves unless it sees only robots set to _ external _ ( line 27 ) .
this means that , in the vertex adjustments phase , a robot waits until it is sure that no robot is in the middle of a move ( note that this holds also for robots that it can not see , because as soon as one of them moves it becomes visible to all other robots ) .
indeed , a robot sets itself to _ adjusting _ right before starting to move and sets itself back to _
external _ when it is done moving .
hence the robots synchronize themselves , and we may pretend them to be ssynch , as opposed to asynch . then , the proof proceeds exactly as in lemma [ l : extfinal ] . in the case in which the robots are initially collinear
, the proof follows the lines of theorem [ t : correctness2 ] , with a few differences . indeed , despite being asynch , the robots manage to wait for each other and synchronize their actions .
suppose that one endpoint robot @xmath12 becomes _ adjusting _ and starts moving to its destination .
then , every robot is bound to wait for the other endpoint robot , @xmath13 , to take action .
so , @xmath13 could either become _ adjusting _ as well and start moving ( if it performed its _
look_before @xmath12 started moving ) , or it could notice @xmath12 and become _
external_. if @xmath12 and @xmath13 are both _ adjusting _ and moving asynchronously ,
some other robots eventually become _ external _ , but do not move yet .
in particular , referring to figure [ fig10 ] , at least robots @xmath167 and @xmath283 can become external in this phase .
notice that , if @xmath12 and @xmath13 move asynchronously in opposite directions ( figure [ fig10:b ] ) , @xmath167 and @xmath283 may switch between being internal and being external several times . however , as soon as they set their light to _ external _ , they do not set it back to _ off _ even if they become internal again .
but @xmath12 moves exactly by @xmath309 , and @xmath13 moves exactly by @xmath310 ( line 12 ) , because the model is rigid .
this movement length is chosen in such a way that both @xmath167 and @xmath283 eventually become vertex robots , as figure [ fig10:b ] suggests .
therefore the colors of @xmath167 and @xmath283 are eventually consistent , despite asynchrony .
so , every robot waits for both @xmath12 and @xmath13 to see some _ external _ robots and thus become _ external _ themselves . only then
do other robots start moving ( lines 2629 ) . as a consequence
, we may once again pretend that the robots in this phase are ssynch , and the proof is completed as in theorem [ t : correctness2 ] .
unfortunately , for non - rigid asynchrobots , our correctness proof of the interior depletion phase of algorithm [ alg2 ] fails .
indeed , to prove collision avoidance , we used in a crucial way the fact that , if two internal robots are moving at the same time , then at most one of them saw the other robot in the middle of a movement .
this is true under the non - rigid ssynchmodel ( obviously ) and under the rigid asynchmodel ( because each internal robot becomes external after only one move ) , but not under the non - rigid asynchmodel . in this model , an internal robot @xmath12 may perform different moves in different directions before becoming external . for instance , if @xmath12 s first movement is stopped by the adversary and , in the meantime , new robots have become _ external _ or new robots have become visible , @xmath12 may decide to move in a significantly different direction the second time .
this , paired with the ability of the asynchscheduler to hold a moving robot for an indefinitely long time and then release it and let it complete its move , does cause collisions in some ( quite pathological ) cases . on the other hand , however , the vertex adjustments phase of algorithm [ alg2 ] works under all models ; therefore we only need to replace the interior depletion phase and the segment breaking phase . with the additional assumption that all robots agree on one axis , there is an easy way to fix the interior depletion phase , which is illustrated in figure [ fig6 ] .
say that the robots agree on the @xmath311 axis , i.e. , they agree on the `` north '' direction , but they may disagree on `` east '' and `` west '' . then , if an internal robot sees that every robot that lies to the north is set to _
( i.e. , if its own @xmath311 coordinate is maximum among the internal robots ) , it is eligible to move . if there is a row of several internal robots that are all eligible to move ( as in figure [ fig6 ] ) , then only the two endpoints are allowed to move , and the others wait .
the left endpoint moves to the upper - left quadrant , and the right endpoint moves to the upper - right quadrant , and their destination points are on the convex hull , but not on locations already occupied by external robots . to guarantee termination
, we make each robot move straight to the north toward the boundary of the convex hull of the visible robots , unless there are external robots in the way . in this special case
, we make the robot move slightly sideways .
also the protocol for the segment breaking phase needs some modifications : indeed , referring to figure [ fig10:b ] , in which @xmath12 and @xmath13 move in opposite directions , it is no longer true that @xmath167 and @xmath283 will eventually be external robots when @xmath12 and @xmath13 stop ( recall that robots are non - rigidnow ) . unfortunately , @xmath167 and @xmath283 may become temporarily external while @xmath12 and @xmath13 move , and thus they may ( permanently ) set themselves to _ external _ , which could lead to inconsistent behaviors . once again , we can fix the protocol if the robots agree on the @xmath311 axis : now , in the segment breaking phase , an endpoint robot moves according to algorithm [ alg2 ] only if it has the maximum @xmath311 coordinate .
this makes only one endpoint move in most cases , which eliminates the aforementioned issue .
moreover , if both endpoints have the same @xmath311 coordinate , they will both move north , thus forming a configuration like the one in figure [ fig10:a ] , which causes no trouble .
the mutual visibilityproblem is solvable by non - rigid asynchrobots carrying 3-colored lights , provided that they agree on one axis .
we show that the above algorithm is correct . in the interior depletion phase
, there can be no collisions , and each internal robot eventually reaches the convex hull . indeed , suppose that initially there is a unique internal robot @xmath12 with largest @xmath311 coordinate .
as soon as enough external robots have set themselves to _ external _ , @xmath12 starts moving north , and no other robot moves .
eventually @xmath12 becomes external without colliding with any robot ( note that , even if @xmath12 does not initially see the boundary of the convex hull , it will eventually see it after finitely many moves ) .
if several internal robots have the largest @xmath311 coordinate , as in figure [ fig6 ] , the argument is similar . at most two robots
can move at the same time , and they can not collide because the difference of their @xmath33 coordinates can not decrease . after enough cycles , either they have reached the convex hull , or one of them has been `` left behind '' and is no longer eligible to move . either way , at least one internal robot eventually becomes external .
once an internal robot has become external , the same argument repeats for all other internal robots .
note that these `` sub - phases '' do not interfere with each other , because a new robot becomes eligible to move only after the previous eligible robots have stopped on the convex hull .
the moment the last internal robot becomes external , no robot is moving , and therefore the whole swarm correctly transitions to the vertex adjustments phase , which works exactly as described in theorem [ t : correctness3 ] and lemma [ l : extfinal ] .
if the robots are initially collinear , they correctly transition to a non - collinear configuration , as in theorem [ t : correctness3 ] .
indeed , note that the two endpoint robots can not move in opposite directions ( as in figure [ fig10:b ] ) , and hence it does not matter if they are rigidor non - rigid , since in this case it is not harmful if they move by smaller amounts than those indicated by algorithm [ alg2 ] ( cf . figure [ fig10:a ] ) . the same clearly holds if @xmath178 and the middle robot executes line 16 .
here we discuss some applications of the previous mutual visibilityalgorithms to other problems , and we also discuss different robot models .
as already observed , algorithm [ alg1 ] also solves the convex formationproblem , where the robots have to terminate in a strictly convex position .
moreover , no robot ever crosses the perimeter of the initial convex hull unless , of course , all the robots are initially collinear .
this works for rigid ssynchrobots carrying 2-colored lights . for non - rigid ssynchrobots carrying 3-colored lights ,
algorithm [ alg2 ] also solves the convex formationproblem , but it has an additional property : during the interior depletion phase , the convex hull of the robots remains unaltered ( unless all robots are collinear ) , and in the vertex adjustments phase it shrinks a little , due to the small movements of the vertices .
we can actually make these movements as small as we want , by changing line 29 of algorithm [ alg2 ] into @xmath312 where @xmath313 is an arbitrarily - chosen positive constant .
similarly , in lines 12 and 16 we can make the robot move orthogonally to @xmath42 by @xmath313 or less . as a result
, we can guarantee that the robots will terminate in a ( strictly convex ) configuration whose vertices are contained in an @xmath313-wide band around the initial convex hull s perimeter .
similar observations hold for the algorithms and models discussed in section [ s : asynch ] . as a followup to algorithms [ alg1 ] and [ alg2 ]
, the robots can even solve the circle formationproblem , in which they have to become concircular and then terminate . moreover , if they are rigid ssynch(respectively , non - rigid ssynch ) , they can do so with the same 2-colored ( respectively , 3-colored ) lights that they used to solve mutual visibility .
first , it is necessary to slightly modify the termination condition of the algorithms : in algorithm [ alg1 ] , when a robot sees only robots set to _
vertex _ , it does not terminate , but it starts executing a _ circle formation _ phase .
similarly , in algorithm [ alg2 ] , we remove lines 2325 , thus preventing vertex robots from reverting their color to _ external _ and terminating after they have adjusted their position .
instead , they wait until they only see robots set to _
adjusting_. accordingly , in lines 27 and 36 we remove the conditions that prevent robots from moving if they see other robots set to _
adjusting_. since we are assuming that robots are ssynch , it is straightforward to see that the correctness proof of section [ s : ssynch2 ] goes through even after these modifications to the protocol , and that eventually all robots are set to _
adjusting_. at this point , the circle formation phase starts . since all robots are set to _ adjusting _
, each robot knows that all of them occupy non - degenerate vertices of the convex hull , and that there are no other robots in the swarm .
hence the phase starts in a strictly convex configuration , and all the robots see each other .
in particular , the _
smallest enclosing circle _
( sec ) computed by each active robot is the same . in the circle formation phase ,
the robots move toward the perimeter of the sec in a precise order , as illustrated in figure [ fig7 ] .
if a robot lies in @xmath1 , which is not on the sec , and one of its neighbors lies in @xmath13 , which is on the sec , then the robot in @xmath1 moves along the extension of the edge of the convex hull that is incident to @xmath1 and not to @xmath13 .
if both neighbors of the robot lie on the sec ( as with the robot in @xmath3 in figure [ fig7 ] ) , it chooses one of its two incident edges , and moves along the extension of that edge .
it is clear that the combined motion of the robots does not cause collisions or obstructions , and that the sec is always preserved .
indeed , any robot that is already on the sec remains still , and those that are inside the sec are allowed to move only within the sec itself .
moreover , the direction in which each robot moves is preserved until one of them reaches the sec . hence , even if robots are non - rigid , after finitely many turns at least one of them reaches the sec , and therefore eventually they all reach the sec . at this point
, they correctly terminate .
the same circle formation phase can also be used in conjunction with the algorithms discussed in section [ s : asynch ] for asynchrobots .
the only difference is that , instead of modifying the asynchalgorithms like we did with the ssynchones , we simply add an extra state , called _ done _ , to synchronize robots and make them transition correctly from the vertex adjustments phase into the circle formation phase .
that is , instead of terminating , a robot sets itself to _ done _ , and then waits until all other robots are set to _ done _ , as well
. only then does it proceed to executing the circle formation phase described above .
of course , before the circle formation phase starts , if a robot sees another robot set to _ done _ , it treats its like an _ external _ robot .
this works with both rigidand non - rigid asynchrobots carrying 4-colored lights .
a simple modification of algorithm [ alg1 ] solves the near - gatheringproblem , which requires all the robots to converge to a point without colliding : it is sufficient to remove lines 8 , 9 , and 23 , that is , all the operations involving colors , and the termination condition .
indeed , if there is only one internal robot , either it will become external , or the other robots will converge to its location . on the other hand ,
if all robots become external , the convex hull will keep shrinking until its vertices converge to a point .
this works for rigid ssynchrobots , even without the use of colored lights .
however , if the robots carry 2-colored lights , they can also terminate when they get close enough to one another .
this is done by simply modifying the termination condition of line 9 : @xmath314 where @xmath313 is any given positive constant .
suppose that the robots are non - rigid ssynch , and as such they can be stopped by the scheduler at each turn before they reach their destination point , but not before they have moved by at least @xmath24 .
recall that in this case they have an algorithm for mutual visibilitythat uses 3-colored lights , described in section [ s : ssynch2 ] .
however , if the robots know the exact value of @xmath24 and they can use it in their computations , they can solve mutual visibilityeven with 2-colored lights , by executing a slightly modified version of algorithm [ alg1 ] . if all the robots are initially collinear , algorithm [ alg1 ] makes them reach a non - collinear configuration , even if they are non - rigid .
subsequently , the invariants discussed in section [ s : invariants ] keep holding , and in particular the convex hull of the robots never grows , and vertex robots never become non - vertex robots . we have to show that a version of lemma [ l : main ] can be obtained for this model , as well .
referring to figure [ fig1 ] , we can make the robot in @xmath1 move toward @xmath315 by a smaller amount , never passing internal robots , and never colliding with them , unless they are closer than @xmath24 .
if an internal robot @xmath12 is closer than @xmath24 and it stands between @xmath1 and @xmath315 , the robot in @xmath1 moves close enough to @xmath12 , on the line parallel to @xmath52 , and it sets its light to the correct value ( note that it knows before moving whether it will end up being a vertex robot or not ) .
this `` lateral move '' can not be stopped by the scheduler , and it is guaranteed to create a new external robot , and eventually increase by one the number of vertex robots .
on the other hand , if only `` non - lateral moves '' are made , the analysis in section [ s : convergence ] can be generalized , because equation [ e : default ] takes the form @xmath316 where @xmath317 $ ] , and @xmath318 is a constant . indeed ,
if the convex hull of the robots never grows , and its initial diameter is @xmath319 , then each moving robot is guaranteed to move by at least a fraction of @xmath320 of its computed movement vector .
therefore , all the lemmas in section [ s : convergence ] can be reproved by merely adjusting some coefficients in the formulas .
it remains to prove that , if only one internal robot is left , it eventually reaches the boundary of the convex hull without colliding with other robots .
but since @xmath24 is known , we can make this robot stay still until it either becomes external ( due to other robots movements ) , or the diameter of the convex hull becomes smaller than @xmath24 .
as soon as it is guaranteed to make a reliable move , it can reach the midpoint of an edge of the convex hull , and therefore become external .
when all robots are external , they eventually reach a strictly convex configuration and they correctly terminate , as detailed in the proof of theorem [ t : correctness ] .
suppose that the robots do not carry visible lights and have no internal memory , but they share the knowledge of the total number of robots in the swarm , @xmath21 . if the robots are rigid ssynch , it is possible to slightly modify algorithm [ alg1 ] to solve mutual visibilityin this model , as well .
note that the information given by other robots visible lights is used only when a robot has to terminate ( line 9 ) , or when it is the only internal robot and it has to move to the perimeter of the convex hull ( line 23 ) .
however , both these situations can be recognized locally by counting the vertices of the convex hull of the visible robots : if it has @xmath21 non - degenerate vertices , mutual visibilityhas been solved , and the executing robot can terminate .
if the convex hull has @xmath123 vertices and the executing robot is internal , it moves to the boundary , as in line 23 of algorithm [ alg1 ] .
the same techniques can be used to modify the algorithm of section [ s : delta ] , so that non - rigid ssynchrobots with knowledge of @xmath24 and knowledge of @xmath21 can solve mutual visibilitywithout the use of colored lights .
we are also able to optimize algorithm [ alg2 ] for robots with knowledge of @xmath21 : namely , we can achieve termination detection even if the robots do not use the _ adjusting _ color , as follows .
when all robots are external and a vertex robot makes a default move , it does not change its color , but remains _
external_. then , when a vertex robot sees @xmath21 robots , it terminates .
note that making a default move allows a robot to see all other robots at its next activation , and therefore each external robot makes at most one default move before terminating .
moreover , when all robots are collinear , we apply this simple protocol : if a robot is an endpoint of the convex hull , it moves orthogonally to it ( without changing color ) ; otherwise it stays still . this way , as soon as an endpoint is activated
, the configuration becomes non - collinear .
the only exception to this rule is the case @xmath178 , in which the central robot has to move orthogonally to the segment , while the other two robots stay still .
this technique allows non - rigid ssynchrobots with knowledge of @xmath21 to solve mutual visibilitywith 2 colors as opposed to 3 .
observe that lemma [ l : vessels ] requires only @xmath123 valves to be opened infinitely often , as opposed to @xmath21 .
this implies that , if rigid ssynchrobots execute the modification of algorithm [ alg1 ] described in section [ s : nearg ] , they converge to a point even if one robot is unable to move .
therefore , in the presence of one faulty robot , near - gatheringis still solvable , even without the use of colored lights .
( on the other hand , if two robots are faulty , near - gatheringis clearly unsolvable , because the two faulty robots can not approach each other . ) additionally , if @xmath21 is known , mutual visibilityand convex formationare solved in the presence of a faulty robot , provided that it is located on the boundary of the convex hull .
suppose that the scheduler is _ sequential _
, i.e. , it is ssynchand it activates exactly one robot at each turn . in this very special model there is a simple algorithm to solve mutual visibilitywith no termination detection , even with no colored lights and no knowledge of @xmath21 , and even if the robots are non - rigidand two of them are faulty .
( if three robots are faulty , mutual visibilityis clearly unsolvable . )
when a robot is activated , it just moves by a small amount , without crossing or landing on any line that passes through two robots that it can currently see ( including itself ) , as illustrated in figure [ fig8 ] .
clearly , when a robot moves as described , it becomes visible to all other robots .
hence , when all robots ( except possibly two of them ) have moved at least once , they can all see each other .
this protocol solves mutual visibilitywith no termination detection , in the sense that , after finitely many turns , the robots will keep seeing each other .
however they will never stop moving because they will never know if their task is terminated or not .
indeed , termination detection is not achievable under this set of assumptions and , to be able to obtain it , some other assumptions are needed ; for example , 2-colored lights or the knowledge of @xmath21 . with 2-colored lights , a robot can change its own color the first time it moves , and terminate at the next activation .
with knowledge of @xmath21 , a robot simply terminates when it sees @xmath21 robots .
in this paper we initiated the investigation of the computational capabilities of a swarm of anonymous mobile robots with obstructed visibility : in this model , which has never been considered in the literature , two robots can not see each other if a third robot lies between them .
we focused on the basic problem of mutual visibility , in which the robots , starting from an arbitrary configuration , have to reach a configuration in which they all see each other , and then terminate the execution .
this task is clearly impossible if the robots are completely oblivious , unable to communicate , and do not have any additional information . indeed , in this scenario a robot can never distinguish between an initial configuration in which it can not see some other robots , and a configuration in which all robots are visible and it is safe to terminate ( recall that the termination operation can not be undone ) .
therefore we employed the extended model of luminous robots , in which each robot is carrying a visible light that it can set to different colors .
the goal is then to minimize the number of colors required by the robots to solve the mutual visibilityproblem under different settings and restrictions .
namely , we considered ssynchand asynchrobots , and rigidand non - rigidmovements .
we also discussed how to reduce the number of used colors if some information is given to the robots , such as the size of the swarm , @xmath21 , or a minimum distance @xmath24 that a robot is guaranteed to cover in each movement .
our main results are summarized in theorems [ t : sum1 ] and [ t : sum2 ] .
we then touched on more complex problems , and proposed solutions that use our mutual visibilityprotocols as a preprocessing step .
notably , we gave the first algorithms for the near gatheringproblem ( with fault tolerance ) and the circle formationproblem that work under the obstructed visibility model .
we proposed two main algorithms , and several modifications and adaptations to various models .
algorithm [ alg1 ] ( _ shrink _ ) uses 2 colors and makes the convex hull of the robots shrink , ideally converging to a point .
algorithm [ alg2 ] ( _ contain _ ) uses 3 colors , and keeps the initial convex hull basically unaltered .
it is therefore suited for practical situations in which the robots have to surround a large - enough area , as well as solving mutual visibility . also , both algorithms keep the robots within the initial convex hull ( unless they are initially collinear ) , which is useful in practice , for instance in the presence of hazardous areas around the swarm .
some interesting research problems remain unsolved .
we would like to reduce the number of colors used by our algorithms in the various models , or to prove our algorithms optimal .
our solutions to mutual visibilityin some models use only 2 colors ( or no lights at all if @xmath21 is known ) , which is clearly optimal . for other models , such as non - rigid ssynchand rigid asynch
, we used 3 colors , and our question is whether this can be improved .
we conjecture that algorithm [ alg1 ] , which uses only 2 colors and has been designed and proven correct for rigid ssynchrobots , can be applied with no changes also to non - rigid ssynchrobots ( we could prove that 2 colors are sufficient in this model under the assumption that the robots know @xmath24 ) . in the non - rigid asynchsetting we were only able to solve mutual visibility(with 3 colors ) assuming that the robots agree on the direction of one coordinate axis .
we ask if this assumption can be dropped , perhaps if more colors are used .
another question is whether mutual visibilitycan be solved deterministically without using colored lights or extra information , and without termination detection . that is
, we allow the robots to move forever , but we require them to remain mutually visible from a certain time on .
we proposed a simple solution that works under the sequential scheduler , and we ask if this can be generalized to ssynchor even asynchschedulers .
we emphasize that obstructed visibility represents a broad line of research in the field of computation by mobile robots , and this paper explored just a few directions .
several classical problems are worth studying under this model , such as the general pattern formation problem , flocking , scattering , with or without bounded visibility , etc . c. agathangelou , c. georgiou , and m. mavronicolas . a distributed algorithm for gathering many fat mobile robots in the plane . in _ proceedings of the 32nd acm symposium on principles of distributed computing ( podc ) _ , pages 250259 , 2013 .
z. bouzid , s. dolev , m. potop - butucaru , and s. tixeuil .
robocast : asynchronous communication in robot networks . in _ proceedings of the 14th international conference on principles of distributed systems ( opodis )
_ , lncs 6490 , pages 1631 , 2010 .
k. bolla , t. kovacs , and g. fazekas .
gathering of fat robots with limited visibility and without global navigation . in _ proceedings of the international symposium on swarm and evolutionary computing _ , pages 3038 , 2012 .
i. chatzigiannakis , m. markou , and s. nikoletseas .
distributed circle formation for anonymous oblivious robots . in _ proceedings of the 3rd international workshop on experimental and efficient algorithms ( wea ) _ , pages 159174 , 2004 .
s. das , p. flocchini , g. prencipe , n. santoro , and m. yamashita .
the power of lights : synchronizing asynchronous robots using visible bits . in _ proceedings of the 32nd international conference on distributed computing systems ( icdcs ) _ , pages 506515 , 2012 .
s. datta , a. dutta , s. gan chaudhuri , and k. mukhopadhyaya .
circle formation by asynchronous fat robots . in _ proceedings of the 9th international conference on distributed computing and internet technology ( icdcit ) _ , pages 195207 , 2013 .
x. dfago and a. konagaya .
circle formation for oblivious anonymous mobile robots with no common sense of orientation . in _ proceedings of the 2nd acm international workshop on principles of mobile computing ( pomc ) _ , pages 97104 , 2002 .
a. efrima and d. peleg .
distributed models and algorithms for mobile robot systems . in _ proceedings of the 33rd international conference on current trends in theory and practice of computer science ( sofsem ) _ ,
pages 7087 , 2007 .
flocchini , g. prencipe , n. santoro , and g. viglietta .
distributed computing by mobile robots : solving the uniform circle formation problem . in _ proceedings of the 18th international conference on principles of distributed systems ( opodis ) _ , pages 217232 , 2014 .
p. flocchini , n. santoro , g. viglietta , and m. yamashita .
rendezvous of two robots with constant memory . in _ proceedings of the 20th international colloquium on structural information and communication complexity ( sirocco ) _ , pages 189200 , 2013 .
b. katreniak .
biangular circle formation by asynchronous mobile robots . in _ proceedings of the 12th international colloquium on structural information and communication complexity ( sirocco ) _
, 185199 , 2005 .
d. peleg .
distributed coordination algorithms for mobile robot swarms : new directions and challenges . in _ proceedings of the 7th international workshop on distributed computing ( iwdc ) _ ,
pages 112 , 2005 .
g. viglietta .
rendezvous of two robots with visible bits . in _ proceedings of the symposium on algorithms and experiments for sensor systems , wireless networks and distributed robotics ( algosensors ) _ , pages 291306 , 2013 . | consider a finite set of identical computational entities that can move freely in the euclidean plane operating in look - compute - move cycles .
let @xmath0 denote the location of entity @xmath1 at time @xmath2 ; entity @xmath1 can see entity @xmath3 at time @xmath2 if at that time no other entity lies on the line segment @xmath4 .
we consider the basic problem called mutual visibility : starting from arbitrary distinct locations , within finite time the entities must reach , without collisions , a configuration where they all see each other .
this problem must be solved by each entity autonomously executing the same algorithm .
we study this problem in the _ luminous robots _ model ; in this generalization of the standard model of oblivious robots , each entity , called _ robot _ , has an externally visible persistent light that can assume colors from a fixed set of size @xmath5 .
the case where the number of colors is less than 2 ( i.e. , @xmath6 ) corresponds to the classical model without lights : indeed , having lights of one possible color is equivalent to having no lights at all . the extensive literature on computability in such a model , mostly for @xmath6 and recently for @xmath7 ,
has never considered the problem of mutual visibilitybecause it has always assumed that three collinear robots are mutually visible . in this paper
we remove this assumption , and investigate under what conditions luminous robots can solve mutual visibilitywithout collisions , and at what cost , in terms of the number of colors used by the robots .
we establish a spectrum of results , depending on the power of the adversary ( i.e. , the scheduler controlling the robots actions ) , on the number @xmath5 of colors , and on the a - priori knowledge the robots have about the system . among such results
, we prove that mutual visibilitycan always be solved without collisions in ssynchwith @xmath8 colors and in asynchwith @xmath9 colors .
if an adversary can interrupt and stop a robot before it reaches its computed destination , mutual visibilityis still solvable without collisions in ssynchwith @xmath9 colors , and , if the robots agree on the direction of one axis , also in asynch .
all the results are obtained constructively by means of novel protocols . as a byproduct of our solutions ,
we provide the first obstructed - visibility solutions to two classical problems for oblivious robots : _ collision - less convergence to a point _
( also called _ near - gathering _ ) and _ circle formation_. |
there has been a considerable amount of both analytical and numerical work on the ising model living on dynamically triangulated random surfaces and the dual @xmath0 ( and also @xmath1 ) graphs in the past few years , largely motivated by the interest of the model for string theory and two - dimensional gravity .
the partition function of the model for a fixed number of points @xmath3 is @xmath4 where @xmath5 is the connectivity matrix for a given graph or triangulation and the sum over graphs @xmath6 means that we are , in effect , looking at annealed connectivity disorder in such models .
the model was solved in the planar limit in @xcite by noting that a grand canonical partition function could be obtained as the free energy of a zero dimensional @xmath7 hermitian matrix model @xcite @xmath8 in the limit @xmath9 .
the matrix propagator in the above is given by @xmath10 where @xmath11 , and the potentials for @xmath0 and @xmath1 graphs are given by @xmath12 and @xmath13 respectively .
we have suppressed the external field dependence for clarity , though unlike the case of a fixed 2d lattice an exact solution is still possible even in the presence of an external field .
the diagrams generated by the matrix model integral are `` fat '' graphs because the two indices carried by the matrix fatten the propagators into ribbons .
this gives enough structure to build surfaces with the trace round internal loops and the initial factors in the coupling and propagator producing just the right factor of @xmath14 to count the genus .
the matrix model solution has been pushed as far as toroidal topology @xcite , and both the toroidal and planar results for the ( non - onsager ) critical exponents are in agreement with continuum derivations in lightcone and conformal gauge liouville theory @xcite .
there is now a gratifyingly large body of numerical evidence supporting the dynamical lattice exponents for both ising and potts models @xcite . divorcing the ising model calculations and simulations from their origins in string theory and regarding them as an exercise in statistical mechanics , it is natural to ask if the methods employed might also be applicable to _ quenched _ disorder .
in such a case the free energy @xmath15 is calculated on each graph before summing over the random graphs @xmath16 where @xmath17 is the partition function for the ising model on a particular random graph .
the logarithm can be dealt with by using the replica trick @xcite @xmath18 at the expense of having to calculate with @xmath3 replicas of the original matter system . on a fixed lattice
this is not a problem , but the @xmath19 barrier in 2d gravity means that it is nt possible to calculate with more than 2 ising replicas .
ignoring the problem and simply taking the limit @xmath20 in the formulae of @xcite for the critical exponents , does however give apparently reasonable results that are close to the dynamical lattice exponents , which is what is seen in a direct simulation of a quenched set of @xmath0 fat graphs @xcite .
it would obviously be better to have a more trustworthy formalism for investigating quenched problems where such troubles did not arise and one could confidently calculate with an arbitrary number of replicas .
such a formalism is actually inherent in the matrix model approach in a limit that is opposite in a sense to that usually considered , namely @xmath21 @xcite . in this limit
the fat graphs that build the surfaces or their duals in the matrix model integrals degenerate to `` thin '' graphs - in other words zero dimensional feynman diagrams .
such graphs have in fact been considered before in the context of spin glass theories @xcite , without utilizing the methods of field theory and large order expansions of @xcite .
the generic problems in the matrix model approach for @xmath22 can be traced to the difficulty in evaluating angular integrals that arise from diagonalizing the matrices , and these disappear for @xmath23 . consequently the difficulties in applying the replica formalism also vanish . as the regular surface interpretation
is lost the continuum limit is taken by forcing the number of points @xmath24 rather than tuning the cosmological constant as for @xmath9 .
in the remainder of the paper we discuss the behaviour of the ising model on such thin graphs . in the next section we review the analytical calculations of @xcite for ferromagnetic couplings , which serve as a test of the method . in this context
we also discuss the saddle point solution for higher @xmath25 interactions and @xmath26 state potts models and their close correspondance with bethe lattice results .
we then describe in section 3 simulations of the ferromagnetic ising model on @xmath0 and @xmath1 graphs , carried out essentially as a test of the formalism and the simulation programs , which are in agreement with these results . in section 4
we discuss the _
anti_-ferromagnetic ising model on thin graphs and describe our simulations of both the ising and three and four state potts models with purely negative couplings on quenched ensembles of feynman diagrams .
the calculations of @xcite suggest that the antiferromagnetic model may have a spin glass phase at sufficiently negative coupling , and we investigate this numerically by histogramming the overlaps between replicas . in sections 5 and 6
we discuss briefly some exploratory simulations of the ising antiferromagnet on quenched fat graphs and of potts antiferromagnets on thin graphs respectively , largely to compare the results with those for the ising model . finally in section 7
we present our conclusions and discuss the various loose ends that have been left untied by the current work .
the ( annealed ) average of the partition function over all the thin graphs with @xmath27 vertices can be expressed as @xmath28 where @xmath29 is the number of undecorated graphs with @xmath27 points , @xmath30 for @xmath0 graphs , and the action is identical to that of the matrix model in equ.(2 ) .
if we rescale the fields @xmath31 and drop the overall factors that are not germane to the discussion then the large @xmath3 limit can be evaluated by calculating the saddle points of the action @xmath32 a non - trivial phase structure can appear if there is a crossover in the dominant saddlepoint values , and this is what does in fact occur . in the @xmath0 model
the solution to the saddle point equations are @xmath33 ( along with a trivial zero solution , and a solution exchanging @xmath34 in the second case ) .
the first set of solutions above is the high temperature unmagnetized phase and the second , which becomes dominant for @xmath35 is the low temperature magnetized phase , with the magnetization being given by @xcite @xmath36 if we denote the deviation from the critical coupling @xmath37 by @xmath38 then we see that we have mean - field like behaviour @xmath39 .
the same is true for the other exponents which can easily be calculated using the results of @xcite .
the saddle point equations can also be readily solved for the case of the ising model on @xmath1 graphs , to give @xmath40 which indicate a phase transition at @xmath41 , again with mean field exponents .
if we pause for a second , and look back over the results we can see that we have reproduced the behaviour of the ising model on bethe lattices with 3 and 4 neighbours @xcite with these results including , remarkably , the relevant critical temperatures .
the correspondence between the bethe lattice results and those on feynman diagrams is unexpected because the bethe lattice contains no loops , whereas the number of loops in the feynman diagrams which can be calculated using standard methods is extensive @xcite , being @xmath42 for @xmath0 graphs with @xmath27 points for instance .
the correspondance is not a fluke , it persists for the ising model on @xmath43 graphs and the corresponding bethe lattices .
indeed , it remains true for the 3 and 4 state potts models also , where we have @xmath44 for the critical value of @xmath45 in a @xmath46-state potts model on a bethe lattice with @xmath47 neighbours @xcite .
the action for the 3 state potts model on @xmath0 graphs is just @xmath48 where @xmath49 is now given by @xmath50 with @xmath51 .
this gives the following high and low temperature saddle point solutions @xmath52 ( along with permutations ) which again display a transition at the bethe lattice value of @xmath53 ( ie @xmath54 ) . at the risk of belabouring the point
the high and low temperature solutions for the 4 state potts model on @xmath0 graphs @xmath55 where two of the signs are to be taken positive and two negative in the second low temperature set , also fall into the pattern with @xmath56 ( @xmath57 ) .
it is thus clear that , at least for ferromagnetic bonds where there is no frustration , the feynman diagrams offer an effective way of calculating with a closed version of the bethe lattice .
in fact , they might even be considered as more respectable than the bethe lattice itself where one has to throw away the terms in the outermost shells to get a phase transition @xcite .
the correspondance between the diagrams and the bethe lattice is underlined by the following observation , due to h. hilhorst @xcite : we can construct one of our feynman diagrams in a bethe - lattice - like manner by picking a seed point and following the standard procedure for constructing a bethe lattice , but also allowing ourselves the possibility of joining branches to other branches at any step .
this then gives the correct loop counting of @xmath42 in the @xmath0 model with @xmath27 vertices , once all the branches have been closed .
as a warmup exercise for simulations of the antiferromagnetic ising model on feynman diagrams we did some smaller scale simulations of the ferromagnetic model on @xmath0 and @xmath1 graphs .
the generation of the graphs is a much easier exercise than generating the planar @xmath0 graphs used in @xmath58 gravity simulations as the absence of a constraint on the topology obviates the need to perform gymnastics with tutte s algorithm .
the simulation itself used standard finite size scaling techniques to determine the critical exponents .
the absence of frustration in the ferromagnetic model allowed the use of the wolff algorithm to give effective spin updating .
we simulated on @xmath0 and @xmath1 graphs of sizes @xmath59 and @xmath60 points for various @xmath61 values , performing 20,000 measurements ( after a suitable number of metropolis equilibration sweeps ) with the number of wolff sweeps per measurement being determined by the inverse of the mean cluster size . with ferromagnetic couplings and no frustration the quantities we use to characterize the transition , such as
the energy and magnetization are all self - averaging , so we might expect to be able to obtain the correct critical behaviour using a _ single _ graph of each size , provided our graphs are not too small and we pick a representative graph .
we shall see that this is , in fact , the case .
we commence our analysis of the results by looking at the crossing of the ( unnormalized ) binder s cumulant for the magnetization @xmath62 to locate the critical point .
this provides a good test that we have managed to generate our graphs correctly .
if we consider @xmath0 graphs first , the critical value of @xmath63 translates to @xmath64 and the crossing of the cumulant gives the estimate @xmath65 .
similarly on @xmath1 graphs @xmath41 translates to @xmath66 and the crossing of the cumulant gives the estimate @xmath67 .
the quoted errors are the differences in the crossing points for different ratios of graph sizes . as a check we have done two complete sets of simulations with a different random graph of each size , and used all the possible combinations in the cumulant analysis .
there was no appreciable variation between different graphs of the same size .
the crossing of the cumulant is shown in fig3.1 for three of the graph sizes used .
the value of the specific heat critical exponent @xmath68 for the mean field ising model is zero , signifying a discontinuity between a finite value at @xmath69 and zero at @xmath70 rather than the logarithmic divergence of the onsager solution .
if we look at fig3.2 we see clear signs of this mean field behaviour in the form of the specific heat curve and a direct fit to the data is also consistent with @xmath71 .
we can verify this by looking at the scaling of the maximum slope of the magnetization cumulant which is expected to behave as @xmath72 for a graph with @xmath3 vertices .
this gives @xmath73 which , using the hyperscaling relation @xmath74 , is again consistent with @xmath71 .
the value of @xmath75 can also be fed into the finite size scaling relation for the vanishing of the magnetization at the critical point @xmath76 to extract a value for @xmath61 .
we find that @xmath77 , so @xmath78 which is the mean field value .
similarly , the scaling of the linear susceptibility @xmath79 at the critical point @xmath80 can be fitted to extract the exponent @xmath81 , giving @xmath82 which is in respectable agreement with the mean field value of 1 .
we spare the reader the extraction of further exponents and the analysis of the @xmath1 results in a similar vein , as it is clear from the results already presented that the ising model exhibits a mean field transition ( as expected ) .
the cumulant analysis also confirms that the transition occurs at the corresponding bethe lattice transition points .
fortified by the good agreement between the analytical calculations and our simulations , we can now proceed to the case that really interests us , the antiferromagnet , to look for spin glass behaviour .
in the absence of a @xmath19 barrier , we are at liberty to calculate with as many replicas as we can manage on thin graphs . it was pointed out in @xcite that the saddle point equations for @xmath83 replicas of the ising model on @xmath0 graphs admitted a factorized solution of the form @xmath84 where the @xmath85 are individual solutions of the annealed , @xmath86 , equation .
the @xmath87 quenched free energy was equal to the annealed free energy , modulo finite size corrections .
however , the entropy per spin calculated with this solution became negative for @xmath88 and higher moments of the partition function showed that the factorized solution no longer dominated @xmath89 with the changeover point apparently converging to a finite value for @xmath83 increasing .
spin glass transitions are often signalled by the breakdown between quenched and annealed averages so , while this is not conclusive evidence for such a transition , it is certainly suggestive . despite the fact that only antiferromagnetic couplings are present in the model , there is certainly frustration due to the loops .
indeed , a variational calculation by de seze some years ago @xcite noted that an ising antiferromagnet with dilute bond disorder _ would _ have a spin glass phase if the undiluted version of the model did not allow antiferromagnetic order at @xmath90 .
the hamiltonian in this case was @xmath91 where the sum was over all @xmath92 and the @xmath93 were distributed ( in a quenched ensemble ) according to @xmath94 which has the effect of killing a fraction @xmath95 of the bonds in any given realization .
the model we consider on feynman diagrams could be cast in rather similar form , as we have some probability that two vertices are connected for a given graph , so viewed in this light spin - glass like behaviour is not so surprising .
although we would no longer expect such close similarity with the standard bethe lattice as frustration has highlighted the role of the loops , there have been previous simulations of `` closed '' versions of the bethe lattice @xcite that might be expected to be a closer approximation .
these considered mixed ferromagnetic and antiferromagnetic bonds on a bethe lattice where the branches were closed in two ways - either with fixed boundary conditions , where no spin glass behaviour was seen , or with pairing between different branches in a given shell , which did display signs of a spin glass phase .
unfortunately purely antiferromagnetic couplings were not considered , which would have allowed a more direct comparison with our results as the feynman diagrams are effectively a multi - shell generalization of the paired closure .
our tactics for simulating the antiferromagnet were rather similar to those in @xcite .
we took an explicitly quenched average by generating @xmath96 different feynman diagrams for each of the sizes that we simulated , @xmath97 and @xmath98 points .
the number of lattices simulated is justified _ a posteriori _ by checking that the results for non - self averaging observables , such as the overlap described in the next paragraph , look reasonable .
we used a massively parallel processor ( the intel paragon ) to run the different lattices concurrently and perform the requisite quenched averaging over the observables during the course of the simulation .
for the sake of simplicity we made no attempt to use either multi - canonical , simulated tempering , or replica cluster simulation schemes which are applicable , usually with some effort , to frustrated systems such as ours @xcite but stuck to the metropolis algorithm . at each @xmath61 point we employed 500,000 simulated annealing sweeps to cool the system followed by 200,000 production sweeps with a measurement every 10th sweep .
each sweep consisted of a complete metropolis update of the lattice .
in addition to all of the standard thermodynamic variables that we measured in the ferromagnetic case we paid particular attention to the overlap between two replicas of the ising model on a given graph , as non trivial behaviour in the overlap distribution is often the sign of a spin glass phase .
the overlap is defined in the standard fashion by putting two independent replicas on _ each _ graph in the ensemble @xmath99 where @xmath100 is the spin at site @xmath101 in one replica , and @xmath102 is the spin at site @xmath101 in the other .
we then measure @xmath103\ ] ] where @xmath104 $ ] denotes the quenched graph average and @xmath105 the thermal average , to get a histogram of the overlap probability .
@xmath106 is known to be strongly non - self - averaging in soluble models such as the sherrington kirkpatrick ( sk ) model @xcite
so the quenched average is vital , given the strong variation between individual graphs . a useful heursitic check in the simulations in to make sure that @xmath106 is symmetric about @xmath107 as @xmath108 analytically .
if this is not the case the data is almost certainly not equilibrated .
we have also measured the autocorrelation times in the measurements of @xmath106 in order to help judge their reliability .
it is known in the sk model that @xmath106 is related to the parisi order parameter by @xmath109 @xcite , but in principle there are an infinite number of order parameters generalizing the overlap in short range models @xcite .
it has also been observed that @xmath106 can produce misleading results in short range spin glasses , which might not exhibit the coexistence of infinitely many pure states that is seen in the mean field solution of the sk model @xcite , but rather have a two phase `` droplet '' structure .
recent numerical work @xcite on the nearest neighbour four dimensional @xmath110 ising spin glass , however , seems to support a mean field like picture so the question is far from settled . in any case ,
given that the genuine spin glass with nearest neighbour @xmath110 couplings on the bethe lattice behaves in an sk like way with the appropriate boundary conditions @xcite , it is not unreasonable to expect that @xmath106 may be an appropriate tool to use for the feynman diagrams we consider here .
in addition , observables such as the spin glass susceptibility @xmath111 \label{csg1}\ ] ] may be calculated from the moments of @xmath106 @xmath112 we also measured directly the binder s cumulant associated with @xmath113 @xmath114 whose crossing points for differently sized lattices might be expected to pinpoint any spin glass transition point , in a similar manner to the cumulant for the magnetization in ferromagnetic transitions . as in the ferromagnetic case , we commence our analysis by looking for the crossing point of the appropriate cumulant , @xmath115 .
the results for this are much less accurate than for the ferromagnetic transition , both because of the smaller lattice sizes simulated and the non - self averaging nature of quantities that concern the overlap .
nonetheless , we obtain a crossing for three of the four lattice sizes , indicating that there is a spin glass transition with @xmath116 , where the error is again from the different intercepts for different ratios of lattice sizes .
this translates to @xmath117 , which is rather smaller than the value indicated by the negative entropy in @xcite ( @xmath88 or @xmath118 ) .
however , using the high temperature solution to extract the entropy in the sk model @xcite gives an expression that becomes negative _ inside _ the spin glass phase rather than at the transition point as in the random energy model @xcite . in @xcite the entropy changes sign in what seems to be the paramagnetic phase from our simulations , but the apparent inversion here with respect to the sk model may be due to the purely antiferromagnetic couplings .
we show the crossing of the cumulant in fig4.1 , where we have dropped the results from the @xmath119 lattices that ( narrowly ) failed to cross the others in the plot .
it is known that the specific heat displays a cusp at @xmath37 in the sk model corresponding to an exponent @xmath120 , but if we look at our data in fig4.2 we see that , although there is a cusp it appears to be in the region of @xmath121 .
however , the specific heat estimations of the critical coupling need to be extrapolated to infinite volume to compare with the cumulant results . to do this
we need to extract a value for the combination @xmath75 from our data may be rather ill - defined on our diagrams , the combination is not . ] which can be done by regarding the scaling of the maximum slope of @xmath115 which , just as for @xmath122 , is expected to behave as @xmath123 we find @xmath124 , which is consistent with the cusp displayed by the specific heat as the hyperscaling relation @xmath125 still holds .
it is also compatible with the mean field sk value @xmath120 .
if we now use our value of @xmath75 to extrapolate the @xmath37 s for the specific heat to infinite volume we find @xmath126 , which is less negative than the value from the cumulant crossing .
the quoted error for the critical coupling determined from the specific heat cusp is probably too optimistic , given it is the result of a fit first to @xmath75 followed by a linear interpolation of the estimated cusp positions , so the cumulant crossing value is certainly more reliable .
the linear susceptibility @xmath127 does not diverge at a spin glass transition point , but is expected to display a cusp similar to that in the specific heat . in fig4.3
we plot @xmath79 and see clear evidence of a cusp although it is noteworthy that this is considerably displaced from @xmath128 .
the finite size scaling does not give such a convincing shift back towards this value as in the specific heat case , but the quality of the data is poorer . from the spin glass point of view
the more interesting susceptibility is @xmath129 defined in equ.([csg1 ] ) , which we measured both directly and using using equ.([csg2 ] ) .
the data from the direct measurements is plotted in fig4.4 .
@xmath129 is expected to diverge at a spin glass transition , in contrast to @xmath79 , so there will be a finite size scaling relation of the form @xmath130 analogously to equ.([cm ] ) . fitting the data from both ways of extracting @xmath129 we find @xmath131 , where the error bar does not take account of the uncertainty in determining the critical point .
this value is still marginally compatible with the sk value of 1 .
looking directly at the histograms of @xmath106 also gives indications of a transition around @xmath132 . in figs4.5 - 4.6
, we show the histograms for @xmath133 , deep in the paramagnetic phase and at @xmath134 , below the transition point .
we have plotted the histograms without the error bars coming from the quenched average ( which are not particularly large ) for clarity .
the histograms in fig4.5 are clearly gaussian in shape , as expected for a paramagnetic phase and the width of the gaussians for differently size graphs is in reasonable agreement with `` delta - function '' like behaviour , namely @xmath135 . as the region around @xmath128 is approached the histograms broaden considerably and begin to show signs of a dip at the origin .
once we are in the putative spin glass phase the histograms look generically like that in fig4.6 for @xmath134 .
comparing with histograms for the different lattice sizes shows that the peak at @xmath136 is sharpening with increasing lattice size and that the `` tail '' stretching down to @xmath107 remains stable .
these results suggest that the behaviour is strikingly similar to the sk spinglass , with a delta function at finite @xmath113 adjoined to a continuous tail down to @xmath107 in the infinite volume limit .
this also means that the distribution of @xmath106 takes on a form very similar to that seen on a bethe lattice spin - glass with closed boundary conditions in the simulations in @xcite .
once we are some distance below @xmath128 the autocorrelation times for the overlap become very large and the histograms begin to lose their symmetry and roughen , suggesting that the data has failed to equilibrate , a behaviour often seen in spin glass simulations . from these results
the ising antiferromagnet on feynman diagrams would appear to possess a spin glass phase at sufficiently negative @xmath61 with similar properties to the sk model .
it is thus behaving in a mean field like manner and all the paraphernelia of replica symmetry breaking might be expected to be applicable .
as we indicated in the introduction there is nothing to stop us simulating the ising model on a quenched ensemble of fat graphs , even though such a system is currently not accesssible to reliable analytical calculations .
the simulations of an ising ferromagnet on such an ensemble in @xcite encountered no obvious pitfalls and gave results that were very similar to the annealed ensemble of dynamical graphs .
one might expect quenched and annealed averages to give rather similar results when frustration is not playing a major role , so the antiferromagnet may well be a different story entirely .
a simulation of an ising antiferromagnet on quenched fat graphs is of interest because the graphs involved are no longer of an `` infinite - dimensional '' mean field type .
indeed , it appears that the fractal dimension of the graphs is 3 or 4 ( depending on the method of definition ) @xcite , and 3 may be the lower critical dimension of the spin - glass transition .
it might be possible to look analytically at the change , if any , from the mean field behaviour on @xmath23 thin graphs as the size of the matrices @xmath14 is increased to fatten the graphs .
such an approach has already provided surprisingly good values for critical exponents and couplings in the one matrix model by extrapolation from quite small @xmath14 @xcite .
we are less ambitious here and simply conduct a trial simulation of the ising model on @xmath96 planar @xmath0 graphs with 250 vertices in each .
the loop distribution is calculable using the methods of @xcite and is strongly peaked at small ( 3,4 .. ) loops .
this means that by comparison with the thin graphs , where the loops are in general rather large and the local structure tree - like , frustration is likely to play a stronger role .
we use the same number of graphs and simulated annealing sweeps in the simulation as in the thin graph simulations of the preceding section .
we find that the histograms of @xmath106 begin to deviate strongly from a gaussian shape in the region of @xmath137 and become very irregular and assymetric for @xmath138 .
this is a signal either that we have not equilibrated the system or failed to include enough graphs in the quenched average for these @xmath61 values .
increasing the simulated annealing steps by a factor of four makes little difference to the roughness of the @xmath106 .
the associated autocorrelation times for the measurements of @xmath113 also rapidly become astronomical below @xmath137 . in the region of @xmath137
the observed histograms are rather similar to those in the region of @xmath139 on thin graphs , being perhaps more irregular .
this would mean that one has to work harder on fat graphs than on thin graphs to obtain reliable numerical results .
one of the histograms for @xmath106 at @xmath140 for the fat graph is shown in fig5.1 , where we have included a histogram of @xmath106 at @xmath141 on a thin graph of 250 vertices for comparison . allowing for the differing `` distances '' from the apparent critical values and remembering that a size of 250 on a fat graph may not be equivalent to a size of 250 on a thin graph ,
the two distributions are very similar in form .
as we have only simulated a single graph size , we make no attempt to conduct a serious scaling analysis with our current fat graph data and content ourselves with some qualitative observations : the non - trivial distribution of @xmath106 that has been observed , and the large autocorrelation times suggest that something non - trivial is still going on for the antiferromagnet on fat graphs for sufficiently negative @xmath61 . if the form of @xmath106 persisted on larger graphs it would be suggestive of mean field like behaviour , given its similarity to the thin graph distribution and hence that of the sk model .
we have thus seen that the ising antiferromagnet on @xmath0 graphs displays a non - gaussian @xmath106 for sufficiently negative couplings , which strongly resembles those seen in @xcite for the ising spin glass on a bethe lattice with closed boundary conditions , and those for the sk model .
it would appear that the model is closer to these mean - field systems than short range spin glasses .
given this it is of interest to enquire whether higher state potts models also display non - trivial @xmath106 on feynman diagrams for purely negative couplings .
it is not immediately obvious that they will , as it is easier to `` unfrustrate '' a given loop if we have more colours .
in addition , genuine potts glasses ( @xmath110 or gaussian distributed couplings ) have a completely different order parameter distribution to the ising spin glass in both infinite range models and on bethe lattices @xcite . in the infinite range ( mean field )
case they display two transition temperatures , with a @xmath106 that is still a delta function below the first and two delta functions joined by a continuous portion below the second . on the bethe lattice the form of @xmath106 depends on the bond distribution and the number of neighbours , for @xmath110 bonds in the 3 state potts model for instance there is a delta - function when the number of neighbours @xmath142 and a continuous distribution for @xmath143 . for the 4 state potts model
there is only a delta function @xcite . for more than four states in both cases
the phase transition displays some first order properties . for p - state potts models
the overlap can be defined as @xmath144 which is arranged to be zero in the uncorrelated case @xmath145 .
this can be measured just as in the ising simulations .
a quick test simulation of the 3 state model on @xmath0 graphs reveals that the distribution of @xmath106 does not appear to change from the paramagnetic form even for large negative couplings , suggesting there is no spinglass phase at all .
the intuition that there is now enough liberty to relieve the frustration can , be substantiated by brook s theorem from graph theory @xcite : if g is a connected graph with maximal vertex valency @xmath146 , and if g is not complete , then the chromatic number @xmath147 , earlier ! ] the minimum number of colours it takes so that no adjacent vertices are the same colour , of g is @xmath47 . in this language
we have a 3-regular graph that is not complete , so @xmath148 and we can avoid frustration completely with 3 colours . by the same token , we would not expect any spinglass like behaviour on @xmath149 graphs for @xmath150 state potts models with purely negative couplings as @xmath151 for these .
chromatic numbers for more general random graphs are rather large , though some asymptotic bounds are known , that are in accordance with simulations of antiferromagnetic potts models @xcite .
there are of course , particular cases such as the solution of the celebrated four colour problem , which states that @xmath152 for any planar map @xcite .
@xmath153 trivially for a planar square lattice and a three colouring algorithm also exists as a consequence of the solution of the six vertex model @xcite .
surprisingly , another quick test simulation for a 3 state potts model with purely negative couplings on @xmath1 feynman diagrams ( so neither brooks theorem nor the four - colour theorem apply ) also produced distributions that did not change from the paramagnetic form with decreasing beta .
it would be interesting to confirm this behaviour and explore it further for different @xmath46 and graphs . a brief summary of whether or not a spin glass transition appears to occur in the cases discussed above ( @xmath154 ) or not ( @xmath155 ) is given in the table below .
we have left the @xmath51 model on thin @xmath1 graphs undecided .
@xmath154 + @xmath156 & fat & @xmath0 & @xmath154 + @xmath51 & fat , thin & @xmath0 & @xmath155 + @xmath51 & thin & @xmath1 & ?
+ @xmath157 & fat & any & @xmath155 + @xmath150 & fat , thin & @xmath149 & @xmath155 + * table 1 * : summary of whether a spin glass transition appears @xmath154 or not @xmath155 .
we have seen analytically in section 2 that both ising and higher state potts models on feynman diagrams reproduce the bethe lattice mean field behaviour , even down to the same critical couplings for bethe lattices with the same number of nearest neighbours . in section 3
we saw that the self averaging nature of the quantities that characterize the transition , such as the energy and the magnetization , meant that we could obtain numerically respectable results with a single graph .
these were in good accord with the analytical calculations .
the net conclusion is thus that , at least as far as ferromagnetic couplings are concerned , the ising and potts models do not `` see '' the loops that make the difference between the bethe lattice and the feynman diagrams .
these loops _ do _ , however , play a role once antiferromagnetic couplings are introduced .
the calculations of @xcite , suggesting a spinglass phase in the ising model for sufficiently negative coupling , are supported by the non - trivial @xmath106 s that were observed in section 4 .
the quenched average over different feynman diagrams was obligatory in this case because of the non - self - averaging properties of @xmath106 .
the results seen on the feynman diagrams are strongly reminiscent of the bethe lattice with closed boundary conditions that was simulated in @xcite .
the form of the @xmath106 for the ising antiferromagnet suggests that we are closer to a mean field picture of spin glass behaviour than a droplet one on feynman diagrams , again much like the bethe lattice , and in accordance with earlier calculations for finite connectivity random lattices .
the critical exponents we extracted were also all in reasonable agreement with those in the mean field sk model .
in section 5 we looked briefly at the ising antiferromagnet on planar or fat graphs , finding that the distribution of @xmath106 was rather similar to that on a thin graph of a similar size .
the antiferromagnetic higher state potts models that we also briefly looked at in section 6 show interesting behaviour , and there are some intriguing links with colouring problems in graph theory .
the utility of feynman diagrams as a tool in analytical calculations with random graphs of finite connectivity is thus clear .
the close connection between the feynman diagrams and the bethe lattice is somewhat unexpected at first sight , but persists even for the antiferromagnet and closed versions of the bethe lattice .
both the ferromagnet and antiferromagnet have displayed essentially mean field behaviour on thin graphs .
the simulations of @xcite suggest that this is _ not _ the case for the ferromagnet on fat graphs , so it would be worth pursuing the simulations of the antiferromagnet on fat graphs to look for non mean field behaviour and exponents .
we have not explored issues such as the scaling of @xmath106 , the ultrametricity ( or otherwise ) of the distribution of states and the response to external fields for the antiferromagnet in the current work , but we hope to return to these in future .
there are also various direct generalizations that we have not had the time to explore in the current paper .
a first is to look at `` real '' spinglasses with either @xmath110 or gaussian couplings on the feynman diagrams to compare with previous bethe lattice results for such models .
a second obvious extension is to look at other models such as the @xmath158 model on thin graphs , for which some analytical results already exist in the unfrustrated case @xcite
. it might also be worthwhile to implement some of the more advanced simulation algorithms on the models described here .
there are clearly plenty of interesting questions to be addressed regarding quenched disorder on both thin and fat graphs .
the ferromagnetic simulations reported in this paper were carried out on workstations at heriot - watt university , lpthe orsay and the university of colorado .
the antiferromagnetic simulations were carried out on the front range consortium s 208-node intel paragon located at noaa / fsl in boulder .
some of the binder s cumulant analysis was carried out using programs written by a. krzywicki .
daj was supported at orsay by an eec human capital and mobility fellowship and an alliance grant and would like to thank a. krzywicki and h. hilhorst for various useful conversations there .
cfb is supported by doe under contract de - fg02 - 91er40672 , by nsf grand challenge applications group grant asc-9217394 and by nasa hpcc group grant nag5 - 2218 .
99 v. a. kazakov , phys . lett .
* a119 * ( 1986 ) 140 ; + d.v .
boulatov and v.a .
kazakov , phys . lett .
* b186 * ( 1987 ) 379 ; + z. burda and j. jurkiewicz , acta physica polonica * b20 * ( 1989 ) 949 .
e. brezin , c. itzykson , g. parisi and j.b .
zuber , commun . math
* 59 * ( 1978 ) 35 ; + m.l .
mehta , commun .
phys . * 79 * ( 1981 ) 327 .
z. burda and j. jurkiewicz , phys .
b214 * ( 1988 ) 425 .
knizhnik , a.m. polyakov and a.b .
zamolodchikov , mod .
* a3 * ( 1988 ) 819 ; + f. david , mod .
a3 * ( 1988 ) 1651 ; + j. distler and h. kawai , nucl
* b321 * ( 1989 ) 509 .
j. jurkiewicz , a. krzywicki , b. petersson and b. soderberg , phys . lett .
* b213 * ( 1988 ) 511 ; + r. ben - av , j. kinar and s. solomon , nucl . phys .
* b ( proc . suppl . )
20 * ( 1991 ) 711 ; + s.m .
catterall , j.b .
kogut and r.l .
renken , phys .
* d45 * ( 1992 ) 2957 ; + c. baillie and d. johnston , mod .
a7 * ( 1992 ) 1519 ; + j. ambjrn , b. durhuus , t. jonsson and g. thorleifsson , nucl . phys . *
b398 * ( 1992 ) 568 .
edwards and p.w .
anderson , j. phys .
* f5 * ( 1975 ) 965 ; + m. mezard , g. parisi and m. virasoro , `` spin glass theory and beyond '' world scientific , 1987 ; + k. binder and a.p .
young , rev .
phys . * 58 * ( 1986 ) 801 .
d. johnston , phys .
b277 * ( 1992 ) 405 ; + c. baillie , k. hawick and d. johnston , phys . lett . *
b328 * ( 1994 ) 251 .
c. bachas , c. de calan and p. petropoulos , `` quenched random graphs '' , ecole polytechnique preprint cpth - a264.1093 , hep - th 9405068 .
j. banavar , d. sherrington and n sourlas , j. phys . * a 20 * ( 1987 ) l1 ; + p. mottishaw , europhys
. lett . * 4 * ( 1987 ) 333 ; + c. de dominicis and y. goldschmidt , j. phys .
* a22 * ( 1989 ) l775 .
p - y lai and y. goldschmidt , j. phys .
* a23 * ( 1990 ) l775 .
h. a. bethe , proc .
soc . * a 150 * ( 1935 ) 552 ; + c. domb , advan .
* 9 * ( 1960 ) 145 ; + t. p. eggarter , phys . rev . * b9 * ( 1974 ) 2989 ; + e. muller - hartmann and j. zittartz , phys . rev .
lett . * 33 * ( 1974 ) 893 .
h. hilhorst , private communication .
j. binney , n. dowrick , a. fisher and m. newman , `` the theory of critical phenomena '' , o.u.p , oxford 1992 .
r. baxter , `` exactly solved models in statistical mechanics '' , academic press , london 1982 . l. de seze , j. phys . * c10 * ( 1977 ) l353 .
p - y lai and y. goldschmidt , j. phys . * a22 * ( 1989 ) 399 .
r. swendsen and j - s .
wang , phys .
* 57 * ( 1986 ) 2607 ; + b. berg and t. celik , phys .
* 69 * ( 1992 ) 2992 ; + e. marinari and g. parisi , europhys .
* 19 * ( 1992 ) 451 .
d. sherrington and s. kirkpatrick , phys .
* 35 * ( 1975 ) 35 .
g. parisi , phys .
* 50 * ( 1983 ) 1946 .
l. viana and a. bray , j. phys . * c 18 * ( 1985 ) 3037 .
d. fisher and d. huse , j. phys . *
a 20 * ( 1987 ) l997,l1005 .
j. ciria , g. parisi and f. ritort , j. phys . * a 26 * ( 1993 ) 6711,6731 .
b. derrida , phys .
* b24 * ( 1981 ) 2613 .
d. thouless , phys .
* 56 * ( 1986 ) 1082 ; + j. chayes , l. chayes , p sethna and d. thouless , comm
. math phys .
* 106 * ( 1986 ) 41 .
h. kawai , n. kawamoto and y. watakibi , nucl .
* b306 * ( 1990 ) 19 .
s. hikami , physica a204 ( 1994 ) , hep - th 9309005 .
d. gross , i. kanter and h. sompolinsky , phys .
rev . lett . * 55 * ( 1985 ) 304 .
y. goldschmidt , europhys .
* 6 * ( 1988 ) 7 .
r. wilson , `` introduction to graph theory '' ,
3rd edition longman 1985 p - y .
lai and y. goldschmidt , j. stat .
* 48 * ( 1987 ) 513 ; + b. bollobas and p. erdos , math .
* 80 * ( 1976 ) 419 .
k. appel and w. haken , bull .
* 82 * ( 1976 ) 711 .
e. lieb , phys . rev .
* 162 * ( 1963 ) 162 .
n. dorey and p. kurzepa , phys . lett . *
b295 * ( 1992 ) 51 . fig .
3.1 : : the crossover in binder s cumulant for the magnetization in the ferromagnetic model for three representative @xmath0 graphs . fig . 3.2 : : the specific heat in the ferromagnetic model , displaying mean field like behaviour .
fig . 4.1 : : the crossover in binder s cumulant for the overlap in the antiferromagnetic model on @xmath0 graphs with @xmath159 . fig . 4.2 : : the specific heat in the antiferromagnetic model , which shows a characteristic cusp . fig . 4.3 : : the linear susceptibility @xmath127 for the antiferromagnet .
4.4 : : the spin glass susceptibility @xmath129 for the antiferromagnet .
fig . 4.5 : : a histogram of @xmath106 at @xmath160 which is deep in the paramagnetic phase .
we have only plotted the range @xmath161 and histograms for @xmath162 for clarity .
4.6 : : a histogram of @xmath106 at @xmath163 , in the conjectured spin glass phase .
we have plotted three graph sizes to show the sharpening of the peak and the stability of the tail .
5.1 : : a histogram of @xmath106 at @xmath140 for the antiferromagnet on a @xmath0 planar ( fat ) graph with 250 vertices .
a histogram of @xmath106 at @xmath164 for a thin graph of similar size is also shown for comparison . | the ising model on `` thin '' graphs ( standard feynman diagrams ) displays several interesting properties . for ferromagnetic couplings
there is a mean field phase transition at the corresponding bethe lattice transition point . for antiferromagnetic couplings
the replica trick gives some evidence for a spin glass phase . in this paper
we investigate both the ferromagnetic and antiferromagnetic models with the aid of simulations .
we confirm the bethe lattice values of the critical points for the ferromagnetic model on @xmath0 and @xmath1 graphs and examine the putative spin glass phase in the antiferromagnetic model by looking at the overlap between replicas in a quenched ensemble of graphs .
we also compare the ising results with those for higher state potts models and ising models on `` fat '' graphs , such as those used in 2d gravity simulations .
+ + submitted to nucl . phys .
b. [ fs ] + + @xmath2 _ address : _ sept .
1993 - 1994 , _ permanent address : _ maths dept , heriot - watt university , edinburgh , scotland + 22.5 cm 0.05 cm 0.2 cm 0.2 cm |
superbursts are a recently discovered new mode of unstable thermo - nuclear burning on the surface of a neutron star .
these neutron stars reside in so - called low - mass x - ray binaries ( lmxbs ) in which it continuously receives fresh material from a donor star .
a large fraction of these lmxbs are known to exhibit short ( seconds to minutes ) so - called type i x - ray bursts which are due to unstable thermo - nuclear burning of h and/or he ( for reviews , see lewin et al .
@xcite ; strohmayer & bildsten @xcite ) .
these ` normal ' x - ray bursts appear as rapid ( @xmath2110s ) increases in the x - ray flux , followed by an exponential - like decline .
they recur with a frequency ( typically hours to days ) which is ( partly ) set by the supply rate of the fresh fuel .
the ( net ) burst x - ray spectra are well described by black - body emission from a compact object with @xmath210 km radius and color temperature of @xmath212kev .
the color temperature increases during the burst rise and decreases during the decay , reflecting the heating and subsequent cooling of the neutron star surface .
typical integrated burst energies are in the 10@xmath3 to 10@xmath4erg range .
the durations and output energies of superbursts are roughly a factor of a thousand more than that of the normal x - ray bursts .
they display , however , the usual relatively fast rise and exponential decay , as well as the hardening during the rise and subsequent softening during the decay , as seen in normal x - ray bursts .
also , the ( net ) superburst x - ray spectra are described in the same way as the normal x - ray bursts . for an overview of the superburst properties we refer to kuulkers @xcite , and references therein .
the current view is that the superbursts are due to the energy release of unstable burning of c , which is the left over from the stable and unstable burning of h and he ( cumming & bildsten @xcite ; strohmayer & brown @xcite ) .
additional energy may be released through photo - desintegration triggered by the superburst ( schatz et al .
@xcite ) .
the first superburst was discovered with the bepposax / wfcs to come from the x - ray burster 4u1735@xmath0444 ( cornelisse et al .
another superburst was independently found with the rxte / pca , originating from the x - ray burster 4u1820@xmath0303 ( strohmayer & brown @xcite ; see also ballantyne & strohmayer @xcite ) .
thereafter , six more events have been seen to occur in five other x - ray bursters with the bepposax / wfcs ( cornelisse et al .
@xcite ; kuulkers et al .
@xcite , see also in t zand et al .
@xcite ) , rxte / pca ( strohmayer & markwardt @xcite ) and rxte / asm ( wijnands @xcite ; kuulkers @xcite ) .
the fact that only eight such events have been found so far indicates that they must be rare .
two superbursts from 4u1636@xmath0536 were seen to occur @xmath14.7 yrs apart ( wijnands @xcite ) ; observational estimates of the recurrence times are on the order of a year ( see , e.g. , kuulkers @xcite ; in t zand et al .
@xcite ) .
4u1636@xmath0536 is a regular x - ray burster and a so - called ` atoll ' source ( e.g. , van der klis et al .
we here present preliminary results focussed on the x - ray behavior of the source during and around the superburst which occurred on 2001 feb 22 . a first account of the rxte / pca data during the 2001 superburst from 4u1636@xmath0536 is given by strohmayer & markwardt @xcite .
note that this is event was also seen by the rxte / asm ( wijnands @xcite ; see also below ) .
strohmayer & markwardt @xcite found a highly coherent 582hz pulsation during a small part ( @xmath2800s ) of the superburst .
the light curve ( see also strohmayer & markwardt @xcite ) , as well as the hardness curve , of the superburst are shown in fig . 1 . clearly , they resemble the curves usually seen for normal x - ray bursts , except for the duration of a few hours . a high time resolution ( 0.125s ) light curve of the superburst is shown by kuulkers @xcite , including data from the slew just before the third orbit of data on 4u1536@xmath0536 .
this light curve showed an increase in count rate of a few 1000cs@xmath5 about 125s before the precursor x - ray burst .
it was pointed out to us by tod strohmayer that this increase is actually due to the turn - on of pcu 3 ; this was not taken into account when doing the collimator response correction .
the superburst , therefore , did _ not _ start @xmath2125s before the precursor x - ray burst , but between @xmath2145s and @xmath246min before the precursor . a color - color diagram ( cd ) of all public data on 4u1636@xmath0536 between feb 28 , 1996 and april 10 , 2002 is shown in fig . 2 ( this is an extension of the data set used by di salvo et al
. @xcite ) .
the superburst ( shown in red ) started at the bottom part of the cd , i.e. , in the
so - called ` banana ' branch . during the rise to superburst peak
the source hardened , while during the decay it softened , with the hard colors even dropping below pre - superburst values . at the end of the observation the source resumed its pre - superburst colors again .
we investigated the broad - band timing behavior using the high - time resolution modes available during parts of the observation ( see strohmayer & markwardt @xcite ) .
no strong features are seen before and after the superburst , except for some very - low frequency noise .
this is as expected from previous analyses of the timing behavior at the corresponding portion of the cd ( see di salvo et al .
the high - timing data during the superburst , however , show a hint of a quasi - periodic oscillation near 1150hz .
a more careful analysis is in progress . using the standard 2 mode data we extracted x - ray spectra at 16s time resolution throughout the 2001 superburst .
as is common practice for normal x - ray bursts , we subtracted the average pre - superburst x - ray spectrum from the superburst spectra .
the residual spectra were modeled by pure black - body emission ( keeping the interstellar absorption , @xmath6 , fixed at 3@xmath710@xmath8 atoms@xmath9 , see e.g. , christian & swank @xcite ) . as the superburst flux increased the color temperature increased from @xmath11kev up to 2.35kev , while during the decay the color temperature decreased again .
the inferred black - body radius was more or less constant ( @xmath16 km at 6kpc ) .
the fits are , however , not perfect throughout the whole burst , with the strongest deviations seen when the superburst flux had dropped to about half its peak value .
the deviations are very similar to those seen during the superburst from 4u1820@xmath0303 , and may show the influence of the superburst emission on the accretion disk surrounding the neutron star ( strohmayer & brown @xcite ; ballantyne & strohmayer @xcite ) . the total integrated superburst flux is about 0.65@xmath710@xmath10erg ( at 6kpc ) . in fig . 3 ( bottom ) we show in black the wfc light curves acquired around the 2001 superburst of 4u1636@xmath0536 , as well as in gray the rxte / asm 1-day average and in red the rxte / pca light curves .
the bepposax / wfc observations did not cover the superburst ; the rxte / asm , however , did cover it ( wijnands @xcite ; see also below ) . in the top of the bottom panel figure we annotate with bars the times of normal x - ray bursts as seen by the bepposax / wfc ( black ) and the rxte / pca ( red ) .
clearly , the source showed normal x - ray bursts before the superburst , with an average rate of about 6 per day .
observations after the 2001 superburst are sparse ; the first bepposax / wfc and rxte / pca observations were @xmath223 days , respectively , @xmath242 days after the superburst . during both observations
a normal x - ray bursts was seen .
we note that the peak count rates during the normal x - ray bursts are a factor of @xmath21.52 higher than that reached during the precursor burst during the superburst .
this is consistent with an early ignition of a h / he layer by the superburst as it starts ( cumming @xcite ) . in the top panel of fig .
3 we show the long - term rxte / asm 3-day average light curve of 4u1636@xmath0536 . the source varies on time scales of days to months to years .
it may be interesting to note that this light curve resembles that of ks1731@xmath0260 during the first 4 year of the rxte mission .
apart from the 2001 event , a superburst was also seen by the rxte / asm on june 19 , 1996 ( wijnands @xcite ) . in fig .
4 we show the latter event on the left , whereas the 2001 event is shown on the right . during the inspection of the individual rxte / asm
dwell light curves ( see , e.g. , wijnands @xcite ) , we found , however , another candidate superburst .
this candidate , which occurred on june 26 , 1999 , is shown in the middle panels of fig .
although the peak of the light curve ( top ) consists of 1 data point only , the data points after that are on average higher than data before and after . to guide the eye ,
we have connected the data points ; the light curve of the candidate is consistent with a fast rise plus slower exponential decay light curve .
in the bottom panel we show the hardness curve in the same period ; the hardness values during the peak and during the decay , as well as their evolution , are very similar to those of the superbursts in 1996 and 2001 .
based on the similarities we regard the 1999 event as a real superburst ; if true , then the recurrence times between the 1996 and 1999 events and the 1999 and 2001 are @xmath12.9 yrs and @xmath11.75 yrs , respectively .
in this proceedings paper we have shown some preliminary results on the analysis of rxte / pca , rxte / asm and bepposax / wfc data obtained during and around the 2001 superburst seen from the x - ray burster 4u1636@xmath0536 .
we found that during the first observation after this superburst , i.e. , 23 days later , a normal x - ray burst occurred .
previously , we found that normal x - ray bursts seem to cease to occur for up to a month after the superbursts of serx-1 ( cornelisse et al .
@xcite ) and ks1731@xmath0260 ( kuulkers et al .
our observed upper limit on the duration of the cessation of normal x - ray bursts after the 2001 superburst puts strong constraints on the modeling of the thickness of the layer where the superburst originates as well as where the energy is deposited ( cumming & macbeth @xcite ) .
we also investigated the entire rxte / asm data set on 4u1636@xmath0536 and found evidence for a third superburst , which occurred in 1999 ( the first superburst occurred in 1996 ) .
the shortest observed superburst recurrence time is now @xmath11.75 yrs , which is more or less in agreement with previous estimates ( roughly one superburst per source per year ; kuulkers @xcite ; in t zand et al .
superburst recurrence times may constrain the original composition of the material which was deposited on the neutron star ( cumming @xcite ; see also cumming @xcite , and the discussion in kuulkers @xcite ) . 1 ballantyne , d.r . & strohmayer , t.e .
, apj letters , in press [ astro - ph/0401154 ] ( 2004 ) cornelisse , r. , et al .
, a&a , 357 , l21 ( 2000 ) cornelisse , r. , et al . , a&a , 382 , 174 ( 2002 ) christian , d.j . & swank , j.h . , apjs 109 , 177 cumming , a. , apj , 595 , 1077 ( 2003 ) cumming , a. , in the restless high - energy universe , e.p.j .
van den heuvel , j.j.m . in t zand & r.a.m.j .
wijers ( eds . ) , nucl .
physics b. suppl .
elsevier , in press [ astro - ph/0309626 ] ( 2003 ) cumming , a. & bildsten , l. , apj , 559 l127 ( 2001 ) cumming , a. , macbeth , j. , apj letters , in press [ astro - ph/0401317 ] ( 2004 ) di salvo , t. , et al . , a&a , 406 , 177 ( 2002 ) in t zand , j.j.m . , et al . ,
a&a , 411 , l487 ( 2003 ) in t zand , j.j.m . , et al . ,
these proceedings ( 2004 ) kuulkers , e. , a&a , 383 , l5 ( 2002 ) kuulkers , e. , in the restless high - energy universe , e.p.j .
van den heuvel , j.j.m . in t zand & r.a.m.j .
wijers ( eds . ) , nucl .
physics b. suppl .
elsevier , in press [ astro - ph/0310402 ] ( 2003 ) kuulkers , e. , et al . , a&a , 382 , 503 ( 2002 ) lewin , w.h.g .
, et al . , space sci .
, 62 , 223 ( 1993 ) schatz , h. , et al . , apj , 583 , l87 ( 2003 ) t. strohmayer , t. & bildsten , l. , in compact stellar x - ray sources , w.h.g . lewin and m. van der klis ( eds . ) , cambridge university press , in press [ astro - ph/0301544 ] ( 2003 ) strohmayer , t.e . & brown , e.f . , apj , 566 , 1045 ( 2002 ) strohmayer , t.e . & markwardt , c.b . ,
apj , 577 , 337 ( 2002 ) van straaten , s. , et al .
, apj , 596 , 1155 ( 2003 ) van der klis , m. , et al . ,
apj , 360 , l19 ( 1990 ) wijnands , r. , apj , 554 , l59 ( 2001 ) | preliminary results are reported on the spectral and timing properties of the spectacular 2001 superburst of 4u1636@xmath0536 as seen by the rxte / pca .
the ( broad - band ) power - spectral and hardness properties during the superburst are compared to those just before and after the superburst .
not all of the superburst emission can be fitted by pure black - body radiation .
we also gathered bepposax / wfc and rxte / asm data , as well as other rxte / pca data , obtained days to months before and after the superburst to investigate the normal x - ray burst behavior around the time of the superburst .
the first normal x - ray burst after the 2001 superburst was detected @xmath123 days later . during inspection of all the rxte / asm data we found a third superburst .
this superburst took place on june 26 , 1999 , which is @xmath12.9 yrs after the 1996 superburst and @xmath11.75 yrs before the 2001 superburst .
the above findings are the strongest constraints observed so far on the duration of the cessation of normal x - ray bursts after a superburst and the superburst recurrence times .
address = integral science operations centre , science operations and data systems division , research and scientific support department of esa , estec , postbus 299 , nl-2200 ag noordwijk , the netherlands address = sron national institute for space research , sorbonnelaan 2 , 3584 ca utrecht , the netherlands & utrecht university , po box 80000 , 3508 ta utrecht , the netherlands address = center for space research , massachusetts institute of technology , 77 massachusetts avenue , cambridge , ma 02139 , usa address = astronomical institute anton pannekoek " university of amsterdam and center for high - energy astrophysics , kruislaan 403 , 1098 sj amsterdam , the netherlands address = astronomical institute anton pannekoek " university of amsterdam and center for high - energy astrophysics , kruislaan 403 , 1098 sj amsterdam , the netherlands address = astronomical institute anton pannekoek " university of amsterdam and center for high - energy astrophysics , kruislaan 403 , 1098 sj amsterdam , the netherlands |
we refer the reader to @xcite for notation and terminology not explicitly defined in this paper .
an * antidirected path * in a digraph @xmath1 is a path in which the arcs alternate between forward and backward arcs .
the digraph @xmath1 is said to be * anticonnected * if it contains an antidirected path between @xmath5 and @xmath6 for every pair of distinct vertices @xmath0 of @xmath1 .
anticonnected digraphs were studied in @xcite , where several properties such as antihamiltonian connectivity have been considered .
we will show in theorem [ antipathcheck ] below that it is np - complete to decide whether a given digraph contains an antidirected path between given vertices .
our main purpose is to introduce a related connectivity property based on the concept of a * forward antidirected trail * , i. e. a walk with no arc repeated which begins and ends with a forward arc and in which the arcs alternate between forward and backward arcs .
a digraph @xmath1 is * antistrong * if it has at least three vertices and contains a forward antidirected @xmath2-trail for every pair of distinct vertices @xmath0 of @xmath1 .
we say that @xmath1 is * @xmath4-arc - antistrong * if it has at least three vertices and contains @xmath4 arc - disjoint forward antidirected @xmath2-trails for all distinct @xmath7 .
the paper is organized as follows .
first we show that , from an algorithmic point of view , anticonnectivity is not an easy concept to work with , since deciding whether a digraph contains an antidirected path between a given pair of vertices is np - complete
. then we move to the main topic of the paper , antistrong connectivity , and show that this relaxed version of anticonnectivity is easy to check algorithmically .
in fact , we show in section [ assec ] that there is a close relation between antistrong connectivity of a digraph @xmath1 and its so called bipartite representation @xmath8 , namely @xmath1 is antistrong if and only if @xmath8 is connected .
this allows us in section [ asaugsec ] to find the minimum number of new arcs we need to add to a digraph which is not antistrong so that the resulting digraph is antistrong .
furthermore , using the bipartite representation we show in section [ matroidsec ] that the arc - minimal antistrong spanning subdigraphs of a digraph @xmath1 form the bases of a matroid on the arc - set of @xmath1 .
more generally , we show that the subsets of @xmath9 which contain no closed antidirected trails are the independent sets of a matroid on @xmath9 . in section [ asorsec ]
we study the problem of deciding whether a given undirected graph has an antistrong orientation .
we show how to reduce this problem to a matroid problem and give a characterization of those graphs that have an antistrong orientation . for the convenience of readers who are not familiar with matroids , we also provide , as an appendix , a purely graph theoretical proof of the key step which is lemma [ lem : mat3 ] .
both proofs can be converted to polynomial time algorithms which either finds an antistrong orientation of the given input graph @xmath10 or produce a certificate which shows that @xmath10 has no such orientation . in section [ detachsec ]
we show that being orientable as an antistrong digraph can be expressed in terms of connected 2-detachments of graphs ( every vertex @xmath11 is replaced by two copies @xmath12 and every original edge @xmath13 becomes an edge between precisely one of the 4 possible pairs @xmath14 ) with the extra requirement that the 2-detachment is bipartite and contains no edge of the form @xmath15 or @xmath16 .
this imediately leads to a characterization of graphs having such a 2-detachment .
finally , in section [ nosepsec ] we show that one can decide in polynomial time whether a given digraph @xmath1 has a spanning antistrong subdigraph @xmath17 so that @xmath18 is connected in the underlying sense ( while it is np - hard to decide whether a given digraph contains a non - separating strong spanning subdigraph ) .
+ we conclude the paper with some remarks and open problems .
we follow the notation from @xcite . we denote the vertex set and arc set of a digraph @xmath1 by @xmath19 and @xmath20 , respectively and write @xmath21 where @xmath22 and @xmath23 .
unless otherwise specified , the numbers @xmath24 and @xmath25 will always be used to denote the number of vertices and arcs / edges , respectively , in the digraph or graph in question .
digraphs and graphs will not have parallel arcs or edges unless explicitly stated .
we will use the notation @xmath26 $ ] for the set of integers @xmath27 .
the * underlying graph * of a digraph @xmath1 , denoted @xmath28 , is obtained from @xmath1 by suppressing the orientation of each arc and deleting one copy from parallel edges arising from 2-cycles in @xmath1 .
a digraph @xmath1 is * connected * if @xmath29 is a connected graph .
when @xmath30 is an arc of @xmath1 we say that @xmath5 is the * tail * and @xmath6 the * head * of the arc . for a digraph @xmath21
the * out - degree * @xmath31 and * in - degree * @xmath32 of a vertex @xmath33 is the number of arcs in @xmath9 of the form @xmath30 or @xmath34 , respectively .
a vertex of in - degree zero or out - degree zero is a * source * or a * sink * , respectively .
an arc @xmath35 is * forward * in a path / trail / walk @xmath36 if we have @xmath37 for some @xmath38 .
it is * backward * if we have @xmath39 for some @xmath38 .
an * antidirected trail * in a digraph is a trail ( a walk with no arc repeated ) in which the arcs alternate between forward and backward arcs .
an * antidirected path * is an antidirected trail where no vertex is repeated .
an * antidirected cycle * is the digraph we obtain by identifying the end vertices of an antidirected path with an even number of arcs .
an antidirected trail @xmath40 is a * forward * antidirected trail if it starts _ and _ ends with a forward arc .
two vertices @xmath0 in a digraph @xmath1 are * anticonnected * if @xmath1 contains an antidirected path with end vertices @xmath5 and @xmath6 .
a digraph @xmath1 is * anticonnected * if it contains an antidirected path between @xmath5 and @xmath6 for every pair of distinct vertices @xmath0 of @xmath1 .
anticonnected digraphs were studied in @xcite , where several properties such as antihamiltonian connectivity have been considered .
it is easy to see that being anticonnected is not an equivalence relation on the vertex set of a digraph as we may have an antidirected path between @xmath5 and @xmath6 and between @xmath6 and @xmath41 but none between @xmath5 and @xmath41 .
in fact , as we shall show below in theorem [ antipathcheck ] , it is np - complete to decide whether there exists an antidirected @xmath2-path in a digraph @xmath1 . the purpose of this paper is to introduce a new notion , antistrong connectivity , which turns out to take polynomial time to check and show that this has connections to several important topics in combinatorial optimization .
a digraph @xmath1 is * antistrong * if it contains a forward antidirected @xmath2-trail for every choice of @xmath42 .
let us call @xmath1 * @xmath4-arc - antistrong * if it contains @xmath4 arc - disjoint forward antidirected @xmath2-trails for all @xmath7 .
+ * more to be written here *
it was shown in @xcite that every connected graph @xmath10 has an anticonnected orientation .
this can be seen by considering a breath first search tree rooted at some vertex @xmath43 .
let @xmath44 be the distance classes of @xmath10 .
orient all edges between @xmath43 and @xmath45 from @xmath43 to these vertices , orient all edges between @xmath46 and @xmath47 from @xmath47 to @xmath46 , orient all edges from @xmath48 to @xmath49 etc . finally , orient all the remaining , not yet oriented edges arbitrarily .
there are various results in the literature on antidirected cycles * it is np - complete to decide whether @xmath1 has an antidirected 2-factor @xcite . * every digraph with minimum semidegree ( min of in and out - degree ) at least @xmath50 has an antidirected 2-factor @xcite . * in @xcite there are given necessary and sufficient conditions under which a circulant digraph @xmath1 can be decomposed into @xmath4 pairwise arc - disjoint antidirected hamilton cycles , each induced by two jumps .
we will show that it is np - complete to decide if a digraph has an antidirected path between two given vertices .
we need the following result which is not new , as it follows from a result in @xcite on the vertex analogue , but we include a new and short proof for completeness
. [ avoidpairs ] it is np - complete to decide for a given graph @xmath51 , two specified vertices @xmath52 and pairs of distinct edges @xmath53 , all from @xmath54 , whether @xmath10 has an @xmath2-path which avoids at least one edge from each pair in @xmath55 . * proof : * we first slightly modify a very useful polynomial reduction , used in many papers such as @xcite , from 3-sat to a simple path problem and then show how to extend this to a reduction from 3-sat to the problem above . for simplicity
our proof uses multigraphs but it is easy to change to graphs
. let @xmath56 $ ] be the graph ( the variable gadget ) with vertices @xmath57 and the edges of the two @xmath58-paths @xmath59 .
+ let @xmath60 be an instance of 3-sat with variables @xmath61 and clauses @xmath62 .
the ordering of the clauses @xmath62 induces an ordering of the occurrences of a variable @xmath5 and its negation @xmath63 in these . with each variable @xmath64
we associate a copy of @xmath65 $ ] where @xmath64 occurs @xmath66 times and @xmath67 occurs @xmath68 times in the clauses of @xmath69 . identify end vertices of these graphs by setting @xmath70 for @xmath71 .
let @xmath72 and @xmath73 and denote by @xmath74 the resulting graph . in @xmath74
we respectively denote by @xmath75 and @xmath76 the vertices @xmath77 and @xmath78 in the copy of @xmath79 associated with the variable @xmath64 .
+ next , for each clause @xmath80 we associate this with 3 edges from @xmath74 as follows : assume @xmath80 contains variables @xmath81 ( negated or not ) . if @xmath82 is not negated in @xmath80 and this is the @xmath43th copy of @xmath82 ( in the order of the clauses that use @xmath82 ) , then we associate @xmath80 with the edge @xmath83 and if @xmath80 contains @xmath84 and this is the @xmath4th occurrence of @xmath84 , then we associate @xmath80 with the edge @xmath85 .
we make similar associations for the other two literals of @xmath80 .
thus for each clause @xmath80 we now have a set @xmath86 of three distinct edges @xmath87 from @xmath74 and @xmath88 for @xmath89 .
+ now it is easy to check that @xmath74 has an @xmath90-path which avoids at least one edge from each of the sets @xmath91 if and only if @xmath69 is satisfiable .
indeed , the @xmath90-path goes through the ` @xmath41-vertices ' of the copy of @xmath79 associated with @xmath64 if and only if @xmath64 is set to true to satisfy @xmath69 .
let us go back to the original problem .
let @xmath92 be the multigraph consisting of vertices @xmath93 and three edges ( denoted @xmath94 ) from @xmath95 to @xmath96 for @xmath97 .
let @xmath10 denote the multigraph we obtain from @xmath74 and @xmath92 by identifying @xmath98 and @xmath99 .
let @xmath100 and @xmath101 . finally , form three disjoint pairs of arcs @xmath102 between @xmath86 and @xmath103 for every @xmath104 .
+ by the observations above it is easy to check that @xmath10 has an @xmath2-path which avoids at least one arc from each of the forbidden pairs if and only if @xmath69 is satisfiable .
@xmath105 + [ antipathcheck ] it is np - complete to decide whether a given digraph contains an antidirected path between given vertices @xmath0 . * proof : * the following proof is due to anders yeo ( private communication , april 2014 ) .
let @xmath51 be a graph with two specified vertices @xmath52 and pairs of distinct edges @xmath53 , all from @xmath54 .
we will show how to construct a digraph @xmath106 with specified vertices @xmath107 such that @xmath106 contains an antidirected @xmath90-path if and only if @xmath10 has an @xmath2-path which avoids at least one edge from each pair in @xmath55 .
since the construction can be done in polynomial time this and theorem [ avoidpairs ] will imply the result .
let @xmath4 be the maximum number of pairs in @xmath108 involving the same edge from @xmath54 .
let @xmath109 be an antidirected path of length @xmath110 which starts with a forward arc ( and hence ends with a backward arc ) .
now construct @xmath106 as follows : start from @xmath10 and first replace every edge @xmath13 with a private copy @xmath111 of @xmath109 ( no internal vertices are common to two such paths ) .
then for each pair @xmath112 we identify one sink of @xmath113 with one source of @xmath114 so that the resulting vertex has in- and out - degree @xmath115 . by the choice of the length of @xmath109 we can identify in pairs , i. e. no three vertices will be identified .
note that all the original vertices of @xmath10 will be sources in @xmath106 .
the remaining ( new vertices ) will be called _ internal _ vertices . finally let @xmath116 and @xmath117 .
we claim that @xmath106 has an antidirected @xmath90-path if and only if @xmath10 has an @xmath2-path which uses at most one edge from each of the pairs in @xmath108 .
suppose first that @xmath118 is a path in @xmath10 which uses at most one edge from each of the pairs in @xmath108 .
then @xmath119 is an antidirected @xmath90-path in @xmath106 ( no vertex is repeated since the identifications above where only done for paths corresponding to pairs in @xmath120 ) .
conversely , suppose @xmath106 contains an antidirected @xmath90-path @xmath121 . by the way we identified vertex pairs according to @xmath108
, the internal vertices have in- and out - degree at most @xmath115 , and if an internal vertex is on two paths @xmath122 then it has both its in - neighbours on @xmath113 and both its out - neighbours on @xmath114 .
this implies that @xmath121 will either completely traverse a path @xmath113 or not touch any internal vertex of that path .
hence it can not traverse both @xmath113 and @xmath114 if @xmath112 , and it follows that if we delete all internal vertices of @xmath121 and add back the edges of @xmath10 corresponding to each of the traversed paths , we obtain an @xmath2-path in @xmath10 that uses at most one edge from each pair in @xmath108 .
it follows from our definition that every pair of vertices of an antistrong digraph is joined by a trail of odd length .
this immediately gives [ nobipas ] no bipartite digraph is antistrong . for every digraph @xmath1 we can associate an undirected bipartite graph which contains all the information we need to study antistrong connectivity .
the * bipartite representation * ( * ? ? ?
* page 19 ) of a digraph @xmath21 is the bipartite graph @xmath123 , where @xmath124 , @xmath125 and @xmath126 .
[ aschar ] let @xmath21 be a digraph with @xmath127 .
the following are equivalent . 1 .
@xmath1 is antistrong 2 .
@xmath8 is connected .
3 . for every choice of distinct vertices @xmath0 ,
the digraph @xmath1 contains both an antidirected @xmath2-trail @xmath128 of even length starting on a forward arc and an antidirected @xmath2-trail @xmath129 of even length starting on a backward arc .
* proof : * suppose ( a ) holds . then , following the edges corresponding to the arcs of a forward antidirected @xmath2-trail , @xmath8 contains an @xmath130-path for every pair of distinct vertices @xmath52 . now , if @xmath5 and @xmath6 are distinct vertices of @xmath1 , we choose a third vertex @xmath41 in @xmath1 ( @xmath131 and @xmath132 ) , and the union of an @xmath133-path and a @xmath134-path contains an @xmath135-path in @xmath8 .
similarly we obtain an @xmath136-path in @xmath8 for every pair of distinct vertices @xmath52 .
finally , for any @xmath33 , an @xmath137-path in @xmath8 can be found in the union of an @xmath135-path and a @xmath138-path , where @xmath6 is a vertex of @xmath1 distinct from @xmath5 . hence @xmath139 holds .
conversely , @xmath140 holds , since any @xmath130-path in @xmath8 corresponds to a forward antidirected @xmath2-path in @xmath1 which starts and ends with a forward arc .
+ now to prove @xmath141 , it suffices to remark that @xmath128 and @xmath129 correspond to an @xmath135-path and an @xmath136-path in @xmath8 , respectively . finally , to see that @xmath142 holds , it suffices to show that if ( c ) holds
, then @xmath8 contains an @xmath130-path for all @xmath52 ( possibly equal ) .
this follows by considering a neighbour @xmath143 of @xmath144 and a @xmath145-path in @xmath8 .
@xmath105 + proposition [ aschar ] immediately implies the next result .
[ checkas ] one can check in linear time whether a digraph is antistrong .
recall that a digraph is * @xmath146-strong * if it has at least @xmath147 vertices and it remains strong after deletion of any set of at most @xmath148 vertices .
the digraph obtained from three disjoint independent sets @xmath149 each of size @xmath4 by adding all arcs from @xmath150 to @xmath151 , from @xmath151 to @xmath152 , and from @xmath152 to @xmath150 is @xmath4-strong .
however , @xmath8 has three connected components .
this shows that no condition on the strong connectivity will guarantee that a digraph is antistrong .
+ recall that @xmath1 is @xmath4-arc - antistrong if it contains @xmath4 arc - disjoint forward antidirected @xmath2-trails for every ordered pair of distinct vertices @xmath0 .
we can check in time @xmath153 whether a digraph has @xmath4 arc - disjoint forward antidirected @xmath2-trails for given vertices @xmath0 , because they correspond to edge - disjoint @xmath130-paths in @xmath8 whose existence can be checked by using flows , see
e.g. ( * ? ? ?
* section 5.5 ) .
so we can check in polynomial time if a digraph is @xmath4-arc - antistrong .
if @xmath1 is @xmath154-arc - antistrong , then it contains @xmath4 arc - disjoint antistrong spanning subdigraphs .
* proof : * since @xmath1 is @xmath154-arc - antistrong , @xmath8 is @xmath154-edge - connected .
we can now use nash - williams theorem ( see ( * ? ? ?
* theorem 9.4.2 ) for instance ) to deduce that @xmath8 has @xmath4 edge - disjoint spanning trees .
proposition [ aschar ] now gives the required set of @xmath4 arc - disjoint antistrong spanning subdigraphs of @xmath1 .
@xmath105 + there exists a polynomial time algorithm which for a given digraph @xmath1 and a natural number @xmath4 either returns @xmath4 arc - disjoint spanning antistrong subdigraphs of @xmath1 or correctly answers that no such set exists .
* proof : * this follows from the fact that such subdigraphs exist if and only if @xmath8 has @xmath4 edge - disjoint spanning trees , and the existence of such trees can be checked via edmonds algorithm for matroid partition @xcite .
@xmath105 + the corresponding problem for containing two arc - disjoint _
strong _ spanning subdigraphs is np - complete ( see e.g. ( * ? ? ?
* theorem 13.10.1 ) ) .
[ twospstrongd ] it is np - complete to decide whether a digraph @xmath1 contains two spanning strong subdigraphs @xmath155 which are arc - disjoint .
note that every complete digraph on at least 3 vertices is antistrong .
hence it is natural to ask for the minimum number of new arcs one has to add to a digraph in order to make it antistrong .
[ augmenttoas ] there exists a polynomial time algorithm for finding , for a given digraph @xmath21 on at least 3 vertices , a minimum cardinality set of new arcs @xmath156 such that the digraph @xmath157 is antistrong .
* proof : * let @xmath1 be a digraph on @xmath158 vertices which is not antistrong . by proposition [ aschar ] , its bipartite representation @xmath8 is not connected .
first observe that in the bipartite representation each new arc added to @xmath1 will correspond to an arc from a vertex @xmath159 of @xmath160 to a vertex @xmath161 such that @xmath162 back in @xmath163 .
so we are looking for the minimum number of new edges of type @xmath164 with @xmath165 whose addition to @xmath8 makes it connected while preserving the bipartition @xmath166 .
note that , as long as @xmath158 , in which case @xmath8 has at least 6 vertices , we can always obtain a connected graph by adding edges that are legal according to the definition above .
so the number of edges we need is exactly the number of connected components of @xmath8 minus one . where @xmath43 is the rank function of the matroid @xmath167 which we define in section [ matroidsec ] .
] to find an optimal augmentation we add all missing edges between @xmath160 and @xmath168 to @xmath8 , except for those of the form @xmath169 and give the new edges cost 1 , while all original edges get cost 0 . now find a minimum weight spanning tree in the resulting weighted complete bipartite graph .
the edges of cost 1 correspond to an optimal augmenting set back in @xmath1 .
@xmath105 + the complexity of the analogous question for @xmath4-arc - antistrong connectivity is open .
+ [ addtogetkas ] given a digraph @xmath1 and a natural number @xmath4 , can we find in polynomial time a minimum cardinality set of new arcs whose addition to @xmath1 results in a digraph @xmath17 which is @xmath4-arc - antistrong ?
problem [ addtogetkas ] is easily seen to be equivalent to the following problem on edge - connectivity augmentation of bipartite graphs .
[ bipaug ] given a natural number @xmath4 and a bipartite graph @xmath170 with @xmath171 which admits a perfect matching @xmath172 in its bipartite complement , find a minimum cardinality set of new edges @xmath156 such that @xmath173 and @xmath174 is @xmath4-edge - connected and bipartite with the same bipartition as @xmath175 .
theorem [ augmenttoas ] can be extended to find the minimum number of new arcs whose addition to @xmath1 gives a digraph with @xmath4 arc - disjoint antistrong spanning subdigraphs @xmath176 , provided that @xmath19 is large enough to allow the existence of @xmath4 such subdigraphs .
note that since each @xmath177 needs at least @xmath178 arcs and we do not allow parallel arcs , we need @xmath24 to be large enough , in particular we must have @xmath179 . [ addtogetkdisjointas ] there exists a polynomial time algorithm for determining , for a given digraph @xmath1 on at least 3 vertices , whether one can add some edges to @xmath1 such that the resulting digraph is simple ( no parallel arcs ) and has @xmath4 arc - disjoint antistrong spanning subdigraphs . in the case
when such a set exists , the algorithm will return a minimum cardinality set of arcs @xmath180 such that @xmath181 contains @xmath4 arc - disjoint antistrong spanning subdigraphs .
* proof : * this follows from the fact that the minimum set of new arcs is exactly the minimum number of new edges , not of the form @xmath169 that we have to add to @xmath8 such that the resulting bipartite graph is simple and has @xmath4 edge - disjoint spanning trees .
this number can be found using matroid techniques as follows .
add all missing edges from @xmath160 to @xmath168 and give those of the form @xmath169 very large cost ( larger than @xmath182 ) and the other new edges cost 1 .
now , if the resulting complete bipartite digraph @xmath183 has @xmath4-edge - disjoint spanning trees of total cost less than @xmath182 , then the set of new edges added will form a minimum augmenting set and otherwise no solution exists .
recall from matroid theory that @xmath4 edge - disjoint spanning trees in @xmath183 correspond to @xmath4 edge - disjoint bases in the cycle matroid @xmath184 of @xmath183 which again corresponds to an independent set of size @xmath185 in the union @xmath186 .
this means that we can solve the problem by finding a minimum cost base @xmath175 of @xmath172 and then either return the arcs which correspond to edges of cost 1 in @xmath175 or decide that no solution exists when the cost of @xmath175 is more than @xmath187 .
we leave the details to the reader.@xmath105 +
having seen the equivalence between antistrong connectivity of digraph @xmath1 on @xmath24 vertices and connectivity of its bipartite representation @xmath8 ( see proposition [ aschar ] ) , and recalling from matroid theory that @xmath8 is connected if and only if the cycle matroid @xmath188 has rank @xmath189 , it is natural to ask how antistrong connectivity can be expressed as a matroid property on @xmath1 itself . for @xmath190 ,
we denote by @xmath191 and @xmath192 the numbers of vertices that are heads , respectively tails , of one or more arcs in @xmath156 .
recall that the independent sets of the cycle matroid @xmath193 of a graph @xmath51 are those subsets @xmath194 for which we have @xmath195 for all @xmath196 , where @xmath197 is the number of end vertices of the edges in @xmath198 .
inspired by this we define set @xmath199 of arcs in a digraph @xmath21 to be * independent * if @xmath200 a set @xmath201 is * dependent * if it is not independent .
[ goodisforest ] let @xmath21 be a digraph .
a subset @xmath202 is independent if and only if the corresponding edge set @xmath199 in @xmath8 forms a forest .
every inclusion - minimal dependent set @xmath201 corresponds to a cycle in @xmath8 and conversely .
* proof : * suppose @xmath202 is independent and consider the corresponding edge set @xmath203 in @xmath8 . if @xmath203 is not a forest , then some subset @xmath204 will be a cycle @xmath205 in @xmath8 with @xmath206 vertices in each of @xmath166 for some @xmath207
. the set @xmath208 corresponds to a set @xmath209 with @xmath210 , contradicting that @xmath199 is independent .
the other direction follows from the fact that every forest @xmath156 in @xmath8 spans at least @xmath211 vertices in @xmath8 and every subset of a forest is again a forest .
the last claim follows from the fact that every minimal set of edges which does not form a forest in @xmath8 forms a cycle in @xmath8 .
@xmath105 + the previous proposition implies that a set of arcs of a digraph is dependent if and only if it contains a closed trail of even length consisting of alternating forward and backward arcs .
we will refer to such a trail as a * closed antidirected trail * , or * cat * for short .
[ antimatroid ] let @xmath21 be a digraph and @xmath212 be the family of all independent sets of arcs in @xmath1 .
then @xmath213 is a graphic matroid with rank equal to the size of a largest collection of arcs containing no closed alternating trail .
* proof : * it follows immediately from proposition [ goodisforest ] that a set @xmath199 belongs to @xmath214 if and only if the corresponding edge set @xmath203 is independent in the cycle matroid on @xmath8 .
@xmath105 + a digraph @xmath1 is antistrong if and only if @xmath188 has rank @xmath215 .
* proof : * the rank of @xmath188 equals the size of a largest acyclic set of edges in @xmath8 .
this has size @xmath215 precisely when @xmath8 has a spanning tree @xmath92 . back in @xmath1 ,
the arcs corresponding to @xmath216 contain antidirected forward trails between any pair of distinct vertices .
recall that , by robbins theorem ( see e.g. ( * ? ? ? * theorem 1.6.1 ) ) a graph @xmath10 has a strongly connected orientation if and only if @xmath10 is 2-edge - connected . for antistrong orientations we have the following consequence of proposition [ aschar ] which implies that there is no lower bound to the ( edge- ) connectivity which guarantees an antistrong orientation of a graph .
[ nobipok ] no bipartite graph can be oriented as an antistrong digraph .
the purpose of this section is to characterize graphs which can be oriented as antistrong digraphs .
[ orient ] suppose @xmath51 and @xmath217 .
then @xmath10 has an antistrong orientation if and only if @xmath218 we derive theorem [ orient ] from the following characterization of graphs which can be oriented as digraphs with no closed antidirected trail ( cat ) .
[ catfree ] a graph @xmath51 has an orientation with no cat if and only if @xmath10 satisfies ( [ condition ] ) and ( [ bipcond ] ) . in particular
no @xmath24 vertex graph with at least @xmath219 edges and no @xmath24 vertex bipartite graph with at least @xmath178 edges admits a cat - free orientation .
it is not hard to see that theorem [ catfree ] implies theorem [ orient ] .
assume that theorem [ catfree ] holds and consider a graph @xmath51 with @xmath217 .
suppose that @xmath10 has an antistrong orientation @xmath1 .
then @xmath8 is connected by proposition [ aschar ] . as @xmath8 has @xmath220
edges it is a tree .
so @xmath1 is a cat - free orientation of @xmath10 and , by theorem [ catfree ] , conditions ( [ condition ] ) and ( [ bipcond ] ) hold for @xmath10 .
conversely , if ( [ condition ] ) and ( [ bipcond ] ) hold for @xmath10 , then @xmath10 has a cat - free orientation by theorem [ catfree ] , and we can deduce as above that this orientation is also an antistrong orientation of @xmath10 .
+ we next show that ( [ condition ] ) and ( [ bipcond ] ) are necessary conditions for a cat - free orientation .
for the necessity of ( [ condition ] ) , suppose that some nonempty subgraph @xmath92 has @xmath221 and that @xmath1 is any orientation of @xmath10
. then @xmath8 has at least @xmath222 edges between @xmath223 and @xmath224 , implying that it contains a cycle .
hence @xmath1 is not cat - free .
the necessity of ( [ bipcond ] ) can be seen as follows .
suppose @xmath92 is a bipartite subgraph on @xmath225 edges and let @xmath226 be an arbitrary orientation of @xmath92 .
since no bipartite graph has an antistrong orientation it follows that @xmath227 is not connected , and , as it has @xmath228 edges , it contains a cycle .
this corresponds to a cat in @xmath226 .
+ most of the remainder of this section is devoted to a proof of sufficiency in theorem [ catfree ] .
we first show that , for an arbitrary graph @xmath229 , the edge sets of all subgraphs @xmath10 of @xmath74 which satisfy ( [ condition ] ) and ( [ bipcond ] ) are the independent sets of a matroid on @xmath230 .
we then show that this matroid is the matroid union of the cycle matroid and the ` even bicircular matroid ' of @xmath74 ( defined below ) .
this allows us to partition the edge - set of a graph @xmath10 which satisfies ( [ condition ] ) and ( [ bipcond ] ) into a forest and an ` odd pseudoforest ' .
we then use this partition to define a cat - free orientation of @xmath10 .
we first recall some results from matroid theory .
we refer a reader unfamiliar with submodular functions and matroids to @xcite .
suppose @xmath54 is a set and @xmath231 is a submodular , nondecreasing set function which is nonnegative on @xmath232 .
edmonds @xcite , see ( * ? ? ?
* theorem 13.4.2 ) , showed that @xmath233 induces a matroid @xmath234 on @xmath54 in which @xmath235 is independent if @xmath236 for all @xmath237 .
the rank of a subset @xmath235 in @xmath234 is given by the min - max formula @xmath238 where the minimum is taken over all subpartitions @xmath239 of @xmath240 ( where a * subpartition * of @xmath240 is a collection of pairwise disjoint nonempty subsets of @xmath240 ) .
note that the matroid @xmath167 defined in the previous section is induced on the arc - set of the digraph @xmath1 by the set function @xmath241 . given a graph @xmath51 and @xmath235
let @xmath242 $ ] be the * subgraph induced * by @xmath240 i.e. the subgraph of @xmath10 with edge - set @xmath240 and vertex - set all vertices incident to @xmath240 .
let @xmath243 by putting @xmath244 equal to the number of vertices incident to @xmath240 , and @xmath245 equal to the number of bipartite components of @xmath242 $ ] .
it is well known that @xmath246 is submodular , nondecreasing , and nonnegative on @xmath247 and that @xmath248 is the cycle matroid of @xmath10 .
the function @xmath249 is also submodular , nondecreasing , and nonnegative on @xmath250 since it is the rank function of the matroid on @xmath54 whose independent sets are the edge sets of the * odd pseudoforests * of @xmath10 , i. e. subgraphs in which each connected component contains at most one cycle , and if such a cycle exists then it is odd , see ( * ? ? ?
* corollary 7d.3 ) .
we will refer to this matroid as the * even bicircular matroid * of @xmath10 .
the above mentioned properties of @xmath246 and @xmath249 imply that @xmath251 is submodular , nondecreasing , and nonnegative on @xmath252
. we will show that the independent sets in @xmath253 are the edge sets of the subgraphs which satisfy ( [ condition ] ) and ( [ bipcond ] ) .
[ lem : mat1 ] let @xmath51 be a graph and @xmath254 $ satisfies ( \ref{condition } ) and ( \ref{bipcond})}\}$ ] .
then @xmath255 is the family of independent sets of the matroid @xmath253 .
in addition , the rank of a subset @xmath256 in this matroid is @xmath257 where the minimum is taken over all subpartitions @xmath239 of @xmath240 .
we first suppose that some @xmath235 is not independent in @xmath253 .
then we may choose a nonempty @xmath258 with @xmath259 , and subject to this condition , such that @xmath260 is as small as possible .
the minimality of @xmath261 implies that @xmath262 $ ] is connected .
so @xmath263 if and only if @xmath92 is bipartite ( and 0 otherwise ) and we may now deduce that that @xmath264 $ ] fails to satisfy ( [ condition ] ) or ( [ bipcond ] ) . we next suppose that @xmath242 $ ] fails to satisfy ( [ condition ] ) or ( [ bipcond ] ) for some @xmath235 .
then there exists a nonempty subgraph @xmath92 of @xmath242 $ ] such that either @xmath265 , or @xmath92 is bipartite and @xmath266 .
then @xmath267 satisfies @xmath259 so @xmath240 is not independent in @xmath253 .
the expression for the rank function of @xmath253 follows immediately from ( [ eq : minmax ] ) .
@xmath105 + the * matroid union * of two matroids @xmath268 and @xmath269 on the same ground set @xmath54 is the matroid @xmath270 where @xmath271 .
suppose @xmath272 are submodular , nondecreasing , and nonnegative on @xmath273
. then @xmath274 will also be submodular , nondecreasing , and nonnegative on @xmath273 and hence will induce the matroid @xmath275 .
every independent set in @xmath276 is independent in @xmath275 , but the converse does not hold in general .
katoh and tanigawa ( * ? ? ?
* lemma 2.2 ) have shown that the equality @xmath277 does hold whenever the minimum in formula ( [ eq : minmax ] ) for the ranks @xmath278 and @xmath279 is attained for the same subpartition of @xmath240 , for all @xmath280 .
this allows us to deduce [ lem : mat2 ] for any graph @xmath51 , we have @xmath281 .
this follows from the above mentioned result of katoh and tanigawa , and the facts that @xmath282 and @xmath283 where @xmath239 is the partition of @xmath240 given by the connected components of @xmath242 $ ] ( since @xmath284 and @xmath285 are equal to the number of edges in a maximum forest and a maximum odd pseudoforest , respectively , in @xmath242 $ ] ) .
@xmath105 + lemma [ lem : mat1 ] and lemma [ lem : mat2 ] immediately give the following .
[ lem : mat3 ] let @xmath51 be a graph
. then @xmath10 satisfies ( [ condition ] ) and ( [ bipcond ] ) if and only if @xmath54 can be partitioned into a forest and an odd pseudoforest .
we provide an alternative graph theoretic proof of this lemma in the appendix .
we next show that every graph whose edge set can can be partitioned into a spanning tree and an odd pseudoforest has a cat - free orientation .
+ and a spanning odd pseudoforest @xmath109 ; @xmath40 governs the bipartition @xmath286 ( white / grey ) , its edges are drawn outside ( or on ) the disk spanned by the vertices .
the edges of @xmath109 are embedded in the interior of that disk , the root vertex is the encircled topmost one , the precious edge is the dashed one.,height=321 ] [ tree+forest ] let @xmath10 be the edge - disjoint union of a spanning tree @xmath40 and an odd pseudoforest @xmath109 .
may have parallel edges , but no more than two copies of any edge , in which case one copy is in @xmath40 and the other in @xmath109 ] then @xmath10 has a cat - free orientation . in addition , such an orientation can be constructed in linear time given @xmath40 and @xmath109 . *
proof : * let @xmath286 be the unique ( up to renaming the two sets ) bipartition of @xmath40 and orient all edges of @xmath40 from @xmath151 to @xmath150 . if @xmath109 has no edges we are done since there are no cycles in @xmath10 .
let @xmath288 be the connected components of @xmath109 .
we shall show that we can orient the edges of @xmath288 in such a way that none of the resulting arcs of these ( now oriented ) pseudoforests @xmath289 can belong to a closed antidirected trail .
clearly this will imply the lemma . for each @xmath290
we choose a root vertex @xmath291 of @xmath290 as follows .
if @xmath290 is a tree then we choose @xmath291 to be an arbitrary vertex of @xmath290 .
if @xmath290 contains an odd cycle @xmath80 then we choose @xmath291 to be a vertex of @xmath80 such that @xmath291 has as many neighbours on @xmath80 as possible in the same set of the bipartition @xmath292 as @xmath291 .
since @xmath80 is odd we may choose one such neighbour @xmath293 of @xmath291 .
we will refer to the edge @xmath294 as a * precious edge * of @xmath290 .
put @xmath295 if @xmath290 contains a cycle and otherwise put @xmath296 .
we orient the edges of @xmath297 as follows .
every edge of @xmath290 with one end in @xmath150 and the other in @xmath151 is oriented from @xmath150 to @xmath151 .
every edge @xmath13 of @xmath297 with @xmath298 is oriented towards @xmath291 in @xmath297 ( so if @xmath11 is closer to @xmath291 than @xmath299 in @xmath297 we orient the edge from @xmath299 to @xmath11 and otherwise we orient it from @xmath11 to @xmath299 , see figure [ twotreesfig ] ) .
every edge @xmath300 of @xmath297 with @xmath301 is oriented away from @xmath291 in @xmath297 . finally ,
if @xmath290 contains a precious edge @xmath294 , then we orient @xmath294 from @xmath291 to @xmath293 if @xmath302 , and from @xmath293 to @xmath291 if @xmath303 .
let @xmath1 denote the resulting orientation of @xmath10 .
the digraph @xmath1 can be constructed in linear time if we traverse each tree @xmath297 by a breath first search rooted at @xmath291 .
we use induction on @xmath304 to show that the above construction results in a cat - free digraph .
as noted above , this is true for the base case when @xmath305 .
suppose that @xmath306 and choose an edge @xmath13 in some @xmath290 according to the following criteria .
if @xmath290 is not a cycle then choose @xmath11 to be a vertex of degree one in @xmath290 distinct from @xmath291 and @xmath299 to be the neighbour of @xmath11 in @xmath290 .
if @xmath290 is an odd cycle then choose @xmath307 and @xmath308 .
we will show that @xmath13 belongs to no cat in @xmath1 . by symmetry
, we may suppose that @xmath309 .
we first consider the case when @xmath11 is a vertex of degree one in @xmath290 .
below @xmath310 denote the out - degree , respectively , the in - degree of the vertex @xmath11 in @xmath1 .
we have two possible subcases : * @xmath311 .
since @xmath309 , we oriented @xmath13 from @xmath11 to @xmath299 .
all the other edges incident to @xmath11 belong to @xmath40 and were oriented towards @xmath11
. then @xmath312 and the arc @xmath313 can not be part of a cat .
* @xmath314 . since @xmath309
, we oriented @xmath13 from @xmath11 to @xmath299 ( as @xmath299 is closer to @xmath291 than @xmath11 in @xmath297 ) .
as previously we have @xmath312 and the arc @xmath313 can not be part of a cat . since @xmath315 is cat - free by induction , @xmath1 is also cat - free .
we next consider the case when @xmath290 is an odd circuit .
let @xmath316 be the neighbour of @xmath291 in @xmath290 distinct from @xmath293 .
we again have two possible subcases : * @xmath317 .
since @xmath318 , we oriented the edge @xmath319 from @xmath316 to @xmath291 . then @xmath320 , and the arc @xmath294 can not be part of a cat . since @xmath321 is cat - free by induction , @xmath1 is also cat - free . *
let @xmath68 be the neighbour of @xmath293 in @xmath290 which is distinct from @xmath291 .
the choice of @xmath291 implies that @xmath323 , and hence that @xmath324 is oriented from @xmath293 to @xmath68 .
then @xmath325 , and the arc @xmath324 can not be part of a cat . since @xmath326 can be obtained by applying our construction to @xmath327 by choosing @xmath293 as the root vertex in @xmath328 rather than @xmath291 , it is cat - free by induction .
hence @xmath1 is also cat - free .
@xmath105 + * proof of theorem [ catfree ] ( sufficiency ) : * let @xmath51 be a graph satisfying ( [ condition ] ) and ( [ bipcond ] ) . by lemma [ lem : mat3 ] , @xmath54 can be partitioned into a forest @xmath156 and an odd pseudoforest @xmath109 . by adding a suitable set of edges to @xmath10
, we may assume that @xmath217 .
( this follows by considering the matroid @xmath329 on the edge set of the graph @xmath330 with vertex set @xmath163 in which all pairs of vertices are joined by two parallel edges .
it is easy to check that @xmath330 has an edge - disjoint forest and odd pseudoforest with a total of @xmath215 edges .
thus the rank of @xmath329 is @xmath215 .
since @xmath54 is an independent set in @xmath329 , it can be extended to an independent set with @xmath215 edges . ) the fact that @xmath217 implies that @xmath156 is a spanning tree of @xmath10
. we can now apply theorem [ tree+forest ] to deduce that @xmath10 has a cat - free orientation .
@xmath105 + we have seen that theorem [ catfree ] implies theorem [ orient ] , and hence that a graph @xmath51 has an antistrong orientation if and only if the rank of @xmath253 is equal to @xmath215 .
we can now apply the rank formula ( [ eq : minmax ] ) to characterize graphs which admit an antistrong orientation .
[ thm : char ] a graph @xmath51 has an antistrong orientation if and only if @xmath331 for all partitions @xmath332 of @xmath163 , where @xmath333 denotes the number of edges of @xmath10 between the different parts of @xmath332 and @xmath334 the number of parts of @xmath332 which induce bipartite subgraphs of @xmath10 . * proof : * suppose that @xmath10 has no antistrong orientation . then the rank of @xmath253 is less than @xmath215 so there exists a subpartition @xmath239 of @xmath54 such that @xmath335 by lemma [ lem : mat1 ] .
we may assume that @xmath239 has been chosen such that : 1 .
@xmath336 is as small as possible ; 2 .
subject to ( i ) , @xmath337 is as small as possible ; 3 .
subject to ( i ) and ( ii ) , @xmath338 is as large as possible .
let @xmath339 and let @xmath340 be the subgraph of @xmath10 induced by @xmath86 for all @xmath341 .
we will show that @xmath342 is a ( vertex-)induced connected subgraph of @xmath10 and that @xmath343 for all @xmath89 .
first , suppose that @xmath342 is disconnected for some @xmath341 .
then we have @xmath344 for some subgraphs @xmath345 and @xmath346 with @xmath347 .
let @xmath348 .
we have @xmath349 since @xmath350 and @xmath351 .
this implies that @xmath352 and contradicts ( i ) .
hence @xmath342 is connected and @xmath353 for all @xmath354 .
next , suppose that @xmath355 for some @xmath356 .
let @xmath357 .
we have @xmath358 since , if @xmath359 , then @xmath360 and @xmath361 , and , if @xmath362 , then @xmath363 and @xmath364 .
this implies that @xmath365 . since @xmath366 this contradicts ( i ) or ( ii ) .
hence @xmath367 for all @xmath89 .
finally , suppose that @xmath368 $ ] .
then some @xmath369 has both end vertices in @xmath86 .
let @xmath370 and @xmath371 .
this implies that @xmath365 . since @xmath372 and @xmath373 , this contradicts ( i ) or ( iii ) .
hence @xmath374 $ ] .
let @xmath332 be the partition of @xmath163 obtained from @xmath375 by adding the remaining vertices of @xmath10 as singletons . then @xmath376 and @xmath377 .
we can now use ( [ asorchar1 ] ) to deduce that @xmath378 for some partition @xmath379 of @xmath163 .
we may assume that @xmath380 $ ] is connected for all @xmath381 since we can replace @xmath382 by the vertex sets of the components of @xmath380 $ ] in @xmath332 and maintain this inequality .
let @xmath380=(v_i , e_i)$ ] for @xmath383 and @xmath384
. then @xmath385 and @xmath377 .
( note that a set @xmath386 with @xmath387 has no corresponding edge set in @xmath239 and contributes @xmath388 to the right hand side of the last equation . )
this implies that @xmath389 and hence @xmath10 has no antistrong orientation .
@xmath105 + + or + suppose , on the other hand , that @xmath10 has an antistrong orientation .
then @xmath10 contains an edge - disjoint spanning tree @xmath40 and odd pseudoforest @xmath109 with @xmath390| .
let @xmath379 be a partition of @xmath163 and let @xmath92 be the graph obtained by contracting each set @xmath382 to a single vertex @xmath391 and then adding a loop at @xmath391 whenever @xmath380 $ ] is not bipartite
. then there must be at least @xmath392 edges of @xmath40 and at least @xmath334 edges of @xmath109 joining the vertices of @xmath92 .
hence @xmath393 .
every @xmath394-edge - connected nonbipartite graph has an antistrong orientation .
* proof : * suppose @xmath51 is @xmath394-edge - connected and not bipartite and let @xmath332 be a partition of @xmath163 .
if @xmath395 then @xmath396 since @xmath10 is not bipartite , and if @xmath397 then latexmath:[$e({{\mathcal{q}}})\geq 2|{{\mathcal{q}}}|\geq @xmath394-edge - connected . hence @xmath10 has an antistrong orientation by theorem [ thm : char ] .
@xmath105 + [ cor:3trees - antistrong ] every nonbipartite graph with three edge - disjoint spanning trees has an antistrong orientation .
* proof : * we give two proofs of this corollary .
suppose @xmath51 is a nonbipartite graph with three edge - disjoint spanning trees and let @xmath332 be a partition of @xmath163 .
if @xmath395 then @xmath396 since @xmath10 is not bipartite , and if @xmath397 then @xmath399 since @xmath10 has three edge - disjoint spanning trees . since @xmath400 , @xmath401 and @xmath402 .
we could also remark that if @xmath403 , @xmath404 and @xmath405 denote three edge - disjoint spanning trees of @xmath10 , then there exists @xmath406 such that @xmath407 is not bipartite . then depending if @xmath408 or not , @xmath409 or @xmath410 is an edge - disjoint pair of a spanning odd pseudo - tree and a spanning tree of @xmath10 .
let @xmath92 denote this subgraph of @xmath10 . then using theorem [ tree+forest ]
, @xmath92 has a cat - free orientation which is also an antistrong orientation of @xmath92 since @xmath411 .
so @xmath10 has also an antistrong orientation .
@xmath105 + corollary [ cor:3trees - antistrong ] is tight in the sense that there exist graphs with many edge - disjoint trees , two spanning and the others missing just three vertices , which have no antistrong orientation .
consider the graph @xmath10 obtained by identifying one vertex of a complete bipartite graph @xmath412 and a complete graph @xmath413 .
then @xmath10 has no antistrong orientation . indeed , consider the partition @xmath332 of @xmath414 into four parts : the copy of @xmath412 , and one part for each remaining vertex of @xmath413 .
we have @xmath415 and then @xmath10 has no antistrong orientation by theorem [ thm : char ] . since @xmath416
, we can use edmonds matroid partition algorithm @xcite to determine the rank of @xmath253 in polynomial time , and hence determine whether @xmath10 has an antistrong orientation .
moreover , when such an orientation exists , we can use edmonds algorithm to construct an edge - disjoint spanning tree and pseudoforest with a total of @xmath215 edges , and then use the construction from the proof of theorem [ tree+forest ] to obtain the desired antistrong orientation in polynomial time .
this gives there exists a polynomial time algorithm which finds , for a given input graph @xmath10 , either an antistrong orientation @xmath1 of @xmath10 , or a certificate , in terms of a subpartition @xmath55 which violates ( [ asorchar ] ) , that @xmath10 has no such orientation .
we now show a connection between antistrong orientations of a graph @xmath10 and so - called detachments of @xmath10 .
we need only the special case of 2-detachments ( see e.g. @xcite for results on detachments ) .
a * 2-detachment * of a graph @xmath51 is any graph @xmath417 which can be obtained from @xmath10 by replacing every vertex @xmath418 with two new vertices @xmath12 and then for each original edge @xmath13 adding precisely one of the four edges @xmath14 to @xmath230 .
[ detach ] a graph @xmath51 has an antistrong orientation if and only if @xmath10 has a 2-detachment @xmath419 which is connected and bipartite with bipartition @xmath166 ( we call such a 2-detachment * good * ) . *
proof : * suppose @xmath10 has a good 2-detachment @xmath419
. then there are no edges of the form @xmath15 and no edges of the form @xmath16 . hence the orientation @xmath1 that we get by orienting the edges of the form @xmath164 from @xmath299 to @xmath11 will be an antistrong orientation of @xmath10 by proposition [ aschar ] .
conversely , if @xmath1 is an antistrong orientation of @xmath10 , then @xmath8 is a good 2-detachment of @xmath10 .
@xmath105 + we can now use theorem [ thm : char ] and the subsequent remark to deduce the following .
[ good2detach ] a graph @xmath51 has a good 2-detachment if and only if @xmath420 for all partitions @xmath332 of @xmath163 .
furthermore , there exists a polynomial time algorithm which returns such a @xmath115-detachment when it exists and otherwise returns a certificate , in terms of a partition violating ( [ detachcond ] ) , that no such detachment exists .
while there are polynomial time algorithms for checking the existence of two edge - disjoint spanning trees @xcite , or two arc - disjoint branchings ( spanning out - trees ) in a digraph ( see e.g. ( * ? ?
* corollary 9.3.2 ) ) , checking whether we can delete a strong spanning subdigraph and still have a connected digraph is difficult .
let @xmath28 denote the underlying undirected graph of a digraph @xmath1 .
@xcite [ nondisconstrong ] it is np - complete to decide whether a given digraph @xmath1 contains a spanning strong subdigraph @xmath92 such that @xmath421 is connected . if we replace `` strong '' by `` antistrong '' above , the problem becomes solvable in polynomial time .
[ deleteanticon ] we can decide in polynomial time for a given digraph @xmath21 on @xmath24 vertices whether @xmath1 contains a spanning antistrong subdigraph @xmath422 such that @xmath423 is connected .
* proof : * we may assume that @xmath1 is antistrong , since this can be checked in linear time by verifying that @xmath8 is connected .
let @xmath424 be the cycle matroid of of the underlying graph @xmath28 of @xmath1 and let @xmath425 be the matroid from section [ matroidsec ] whose bases are the antistrong sets consisting of @xmath178 arcs .
let @xmath426 be the union of the matroids @xmath427 , that is , a set @xmath150 of arcs is independent in @xmath172 if and only we can partition @xmath150 into @xmath428 such that @xmath429 is independent in @xmath430 .
for each of the matroids @xmath431 we can check in polynomial time whether a given subset @xmath150 of arcs is independent in @xmath432 and @xmath433 ( for @xmath432 we need to check that there is no cycle in @xmath434 $ ] and for @xmath433 we need to check that there is no cycle in the subgraph of @xmath435 $ ] induced by the edges @xmath436 corresponding to @xmath150 in @xmath8 ) .
thus it follows from edmonds algorithm for matroid partitioning @xcite that we can find a base of @xmath172 in polynomial time using the independence oracles of @xmath427 .
the desired digraph @xmath92 exists if and only if the size of a base in @xmath172 is @xmath437 .
@xmath105 + a similar proof gives the following .
we can decide in polynomial time whether a digraph @xmath1 contains @xmath438 arc - disjoint spanning subdigraphs @xmath439 such that @xmath176 are antistrong and @xmath440 are connected .
we saw in theorem [ twospstrongd ] that it is np - complete to decide whether a given digraph contains two arc - disjoint spanning strong subdigraphs .
we would be interested to know what happens if we modify the problem as follows .
can we decide in polynomial time whether a digraph @xmath1 contains arc - disjoint spanning subdigraphs @xmath155 such that @xmath441 is antistrong and @xmath442 is strongly connected ? inspired by theorem [ deleteanticon ] it is natural to ask the following intermediate question .
[ anti+2con ] can we decide in polynomial time whether a digraph @xmath1 contains arc - disjoint spanning subdigraphs @xmath155 such that @xmath441 is antistrong and @xmath443 is 2-edge - connected ?
the following conjecture was raised in @xcite .
@xcite there exists a natural number @xmath4 such that every @xmath4-arc - strong digraph has arc - disjoint strong spanning subdigraphs @xmath155 .
perhaps the following special case may be easier to study .
there exists a natural number @xmath4 such that every digraph @xmath1 which is both @xmath4-arc - strong and @xmath4-arc - antistrong has arc - disjoint strong spanning subdigraphs @xmath155 .
does there exist a polynomial algorithm for deciding whether a given undirected graph @xmath10 has an orientation @xmath1 which is both strong and antistrong ?
* * acknowledgement .
* bang - jensen and jackson wish to thank jan van den heuvel for stimulating discussions about antistrong connectivity .
they also thank the mittag - leffler institute for providing an excellent working environment .
10 j. bang - jensen and g. gutin , digraphs : theory , algorithms and applications , 2nd edition , springer verlag , london 2009 .
j. bang - jensen and a. yeo , decomposing @xmath4-arc - strong tournaments into strong spanning subdigraphs , _ combinatorica _ * 24 * ( 2004 ) , 331349 .
j. bang - jensen and a. yeo , arc - disjoint spanning sub(di)graphs in digraphs , _ theoret .
* 438 * ( 2012 ) , 4854 .
z. r. bogdanowicz , decomposition of circulants into antidirected hamilton cycles , _ applied mathematical sciences _ , * 6 * ( 2012 ) , 59355938 . g. chartrand , h. gavlas , m. schultz , and c. e. wall , anticonnected digraphs , _ utilitas math . _
* 51 * ( 1997 ) , 4154 .
a. a. diwan , j. b. frye , m. j. plantholt , and s. k. tipnis , a sufficient condition for the existence of an antidirected 2-factor in a directed graph , _ discrete math .
_ * 311 * ( 2010 ) , 25562562 .
j. edmonds , minimum partition of a matroid into independent sets , _
standards , sec .
b. _ * 69 * ( 1965 ) , 6772 .
j. edmonds , submodular functions , matroids , and certain polyhedra , in : combinatorial structures and their applications , eds .
r. guy , h. hanani , n. sauer , and j. schnheim , gordon and breach , new york 1970 , 6987 .
a. frank , connections in combinatorial optimization , oxford lecture series in mathematics and its applications * 38 * ( 2011 ) .
h. n. gabow , s. n. maheshwari , and l. j. osterweil , on two problems in the generation of program test paths , _ ieee trans .
software eng . se _
* 2 * ( 1976 ) , 227231 .
n. katoh and s. tanigawa , a rooted - forest partition with uniform vertex demand , _ j. comb
. optim . _
* 24 * ( 2012 ) , 6798 . c. st .
j. a. nash - williams , connected detachments of graphs and generalized euler trails ,
_ j. london math . soc . _ * 31 ( 2 ) * ( 1985 ) , 1729 .
w. whiteley , some matroids from discrete geometry , in : matroid theory , eds .
j.e bonin et al .
_ contemporary mathematics _ * 197 * , amer .
soc . , providence ( 1996 ) , 171311 .
t. zaslavski , signed graphs , _ disc .
_ , * 4 * ( 1982 ) , 4774 .
a. recski , matroid theory and its applications in electric network theory and in statics .
budapest , akademiai kiad ( 1989 ) xiii , 531 pp .
z.b zhang , d lou , bipartite graphs with a perfect matching and digraphs , arxiv preprint arxiv:1011.4359 , 2010 - arxiv.org .
in this appendix , we give a graph theoretical proof of lemma 6.6 , recalled below .
+ * lemma 6.6 * _ let @xmath51 be a graph
. then @xmath10 satisfies _ latexmath:[\[\begin{aligned } _ if and only if @xmath54 can be partitioned into a forest and an odd pseudoforest . _ + * proof : * recall that a * pseudoforest * is a graph in which each connected component contains at most one cycle , and it is called * odd * if it does not contain even cycles .
a theorem due to whiteley @xcite ( see also @xcite p.367 for a short proof based on edmonds branching theorem ) asserts that a graph satisfies condition ( 2 ) if and only if its edge set can be partitioned into a forest and a pseudoforest .
so let us denote by a * 2-decomposition * of a graph @xmath51 a pair @xmath445 of spanning subgraphs @xmath446 and @xmath447 such that @xmath448 is a partition of @xmath54 and @xmath449 is a forest of @xmath10 and @xmath450 is a pseudoforest of @xmath10 .
we will call any sub - structure edge , component , subgraph etc . of @xmath450 or of @xmath449 * red * or * black * , respectively , and for a subgraph @xmath92 of @xmath10 we denote by @xmath451 and @xmath452 the subgraph of @xmath92 induced by its red or black edges , respectively . without loss of generality
we may assume that @xmath10 is connected , and that @xmath449 is a spanning tree of @xmath10 ( otherwise we could move edges from @xmath450 to @xmath449 to make @xmath449 connected ) .
the * canonical bipartition * of a @xmath115-decomposition @xmath445 of a connected graph @xmath10 is the unique bipartition given by any @xmath115-colouring of @xmath449 .
moreover , an edge of @xmath450 which does not cross this bipartition is called ( as previously ) a * precious edge * in @xmath445 .
a 2-decomposition of a 2-decomposable graph is * nice * if every red cycle contains at least one precious edge .
+ first we establish the next claim .
* proof : * first observe that for any subgraph @xmath92 of @xmath10 with at least one black edge , we have @xmath453 . for any red subgraph @xmath92 with at least one edge
, we get @xmath454 . in particular ( [ condition ] )
holds for every 2-decomposable graph .
the necessity is quite clear .
indeed , consider a nice 2-decomposition @xmath445 of @xmath10 and assume that ( [ bipcond ] ) does not hold .
thus there exists @xmath92 a bipartite subgraph of @xmath10 with @xmath411 .
so equality holds in the previous computation and we have @xmath455 and @xmath456 .
in particular @xmath452 is a spanning tree of @xmath92 and @xmath451 contains at least one cycle @xmath205 . as @xmath445 is nice , @xmath205 contains a precious edge @xmath30 .
as @xmath452 is connected , there exists a black path @xmath109 from @xmath5 to @xmath6 and @xmath109 has even length because @xmath5 and @xmath6 belong to the same part of the canonical bipartition of @xmath445 .
so @xmath457 forms an odd cycle of @xmath92 , a contradiction .
now let us prove the sufficiency .
let @xmath445 be a 2-decomposition of @xmath10 .
a red component @xmath458 of the decomposition is * bad * if it is not a tree and its ( hence unique ) cycle does not contain any precious edges .
if we remove from a bad component @xmath458 all its precious edges , we obtain several connected components , one of which contains the cycle of @xmath458 .
we call this component the * core * of @xmath458 and denote it by @xmath459 . for convenience
we use @xmath459 below to denote both a vertex set and the red subgraph induced by these vertices .
note that @xmath460 $ ] is bipartite as @xmath459 contains no precious edge . a * sequence * of the decomposition
@xmath445 is a list @xmath461 of the cores of its bad red components in decreasing order of cardinality .
we will prove that this 2-decomposition @xmath445 is nice , that is , @xmath467 .
assume that it is not the case and consider @xmath468 the black components of @xmath469 $ ] ( that is , the connected components of @xmath470 $ ] ) .
if @xmath471 , then @xmath469 $ ] is connected in black , and as @xmath472 $ ] is unicyclic , the bipartite graph @xmath469 $ ] violates @xmath473 , a contradiction .
so we must have @xmath474
. now denote by @xmath475 the connected components of @xmath476 and construct a graph @xmath477 on @xmath478 by connecting two vertices of @xmath477 if there exists an edge in @xmath449 between the corresponding components . in other words , we contract the ( connected ) vertex sets @xmath479 in @xmath449 to single vertices .
so @xmath477 is a tree .
finally we consider @xmath40 the minimal subtree of @xmath477 containing @xmath480 . by definition
the leaves of @xmath40 are in @xmath481 and as @xmath474 , @xmath40 has at least two such leaves .
so we consider a leaf @xmath482 of @xmath40 which does not contain entirely the red cycle of @xmath483 ( this could occur even without violating @xmath473 if @xmath482 is not connected in red for instance ) .
we denote by @xmath484 the only neighbour of @xmath482 in @xmath40 .
now , we specify two edges , one black and one red in order to ` change their color ' and obtain a contradiction . first denote by @xmath13 the unique black edge between @xmath482 and @xmath484 .
we suppose that @xmath485 and @xmath486 ( so @xmath487 ) .
now we look at a 1-orientation of @xmath488 ( this is an orientation of @xmath488 in which every vertex has out - degree at most 1 ) and consider a maximal oriented red path leaving @xmath299 with all its vertices in @xmath482 .
as @xmath482 does not contain entirely the red cycle of @xmath483 , this path ends at a vertex @xmath489 which has a unique red out - neighbour @xmath490 for some @xmath491 .
we select this red edge @xmath30 .
+ notice that the unique black path from @xmath5 to @xmath6 contains the edge @xmath13 .
indeed the unique path from @xmath482 to @xmath492 in @xmath40 corresponds to the unique black path @xmath109 from @xmath482 to @xmath492 in @xmath10 . as @xmath482 is a leaf of @xmath40 , the first edge of @xmath109 is @xmath13 and as @xmath482 and @xmath492
are respectively connected in black , the unique black path from @xmath5 to @xmath6 contains @xmath109 and so it contains the edge @xmath13 .
this implies that @xmath493 is also a spanning tree of @xmath10 .
moreover , its bipartition is the same as the bipartition of @xmath449 .
indeed as @xmath30 is an edge lying inside the core of @xmath483 , it is not precious and @xmath494 is still bipartite and has the same bipartition as @xmath449 . removing @xmath13 does not affect the bipartition ( because @xmath493 is connected ) . to conclude , we focus on the red part of the new 2-decomposition @xmath495 . by construction ,
the component @xmath150 of @xmath496 containing @xmath299 ( and also @xmath5 ) is a red tree .
remark that @xmath497 if and only if @xmath30 is an edge of the cycle of @xmath483 . if @xmath11 does not belong to @xmath483 , then by adding @xmath13 we attached @xmath150 to another red component in @xmath498 . as @xmath150 contains at least one vertex , namely @xmath299 , @xmath466 has decreased , a contradiction to @xmath465 in the choice of @xmath445 ( or to @xmath463 if @xmath497 ) .
if @xmath11 belongs to @xmath483 but @xmath11 does not belong to @xmath150 ( in this case we have @xmath499 ) , then @xmath500 and by adding @xmath13 we attached @xmath150 to a vertex of @xmath501 .
once again , @xmath466 has decreased , a contradiction to @xmath465 in the choice of @xmath445 .
finally if @xmath11 belongs to @xmath150 then adding @xmath13 produces a new red unicyclic component .
however as the red path in @xmath450 from @xmath11 to @xmath299 starts in @xmath501 and ends in @xmath488 , it contains a precious edge .
so that newly created red unicyclic component is not bad , and @xmath466 has decreased .
hence , again , we either contradict @xmath465 , or @xmath463 if @xmath497 .
+ now to finish the proof , we will show how to go from a nice 2-decomposition of a connected graph to a decomposition into a spanning tree and an odd pseudoforest ( i.e. a pseudoforest in which every cycle has odd length ) .
let @xmath51 be a connected graph which admits a nice 2-decomposition and consider a nice 2-decomposition @xmath445 of @xmath10 with a minimum number of even red cycles .
we will show by contradiction that this decomposition has no even red cycle .
assume it is not the case and denote by @xmath502 the even red cycles of @xmath450 .
in each of these , select a precious edge @xmath503 and let @xmath504 .
exchanging two edges between @xmath449 and @xmath450 will modify the structure of @xmath449 , and some previously selected precious edges could become not precious any more . to avoid this we will find a vertex @xmath299 with the following property + @xmath505 there exists a component @xmath175 of @xmath506 such that one of the following hold : * @xmath507 contains only one element and this is not in the same component of @xmath450 as @xmath299 ( _ case a _ ) .
* @xmath508 contains exactly two elements and they are the end vertices of some @xmath509 ( _ case b _ ) .
first assume we have found such a vertex @xmath299 and let us see how to conclude , depending of which the two cases a or b we are in .
+ _ case a_. denote by @xmath64 the only element of @xmath510 and by @xmath511 the red component of @xmath450 containing @xmath64 .
as @xmath299 does not belong to @xmath511 , we can find an edge @xmath35 along the black path in @xmath449 from @xmath299 to @xmath64 such that @xmath512 and @xmath64 are in the same component of @xmath513 , @xmath512 belongs to @xmath511 and @xmath11 does not belong to @xmath511 .
so we exchange the colors of @xmath35 and @xmath514 .
the graph @xmath515 is a tree and when we add @xmath35 to @xmath450 we connect this tree to another component of @xmath450 .
the component of @xmath513 containing @xmath11 is a tree containing all the vertices of @xmath150 except @xmath64 .
so the precious edges @xmath516 with @xmath517 are still precious edges , and this is also the case in @xmath518 which is a spanning tree of @xmath10 .
so , we reduce the number of even red cycle of the nice 2-decomposition @xmath445 , a contradiction .
+ _ case b_. denote by @xmath64 and @xmath519 the two elements of @xmath520 and also by @xmath511 the red component of @xmath450 containing the precious edge @xmath514 .
if the black path @xmath109 in @xmath449 between @xmath64 and @xmath519 is not totally contained in @xmath511 then we can select a vertex @xmath159 not belonging to @xmath511 along this path and end up in the previous case with @xmath299 replaced by @xmath159 .
so @xmath109 is totally contained in @xmath511 .
then , as @xmath521 is an odd cycle ( because @xmath514 is precious ) , we can find along @xmath109 two consecutive vertices @xmath35 which are in the same part of the bipartition induced by the bipartite graph @xmath522 $ ] .
so we exchange @xmath514 and @xmath35 .
as previously @xmath523 is a spanning tree of @xmath10 such that all the edges @xmath516 with @xmath517 are still precious and @xmath35 is also precious .
the graph @xmath524 is now a pseudoforest , and we have reduced the number of even red cycles of the nice 2-decomposition @xmath445 , a contradiction .
finally , let us see how to find a vertex @xmath299 in @xmath10 which has property @xmath55 .
consider @xmath477 the minimal subtree of @xmath449 containing all the vertices of the set @xmath150 .
in particular all the leaves of @xmath477 are elements of @xmath150 .
then build the tree @xmath40 from @xmath477 by replacing iteratively each vertex of degree 2 in @xmath477 and not belonging to @xmath150 by an edge linking its two neighbours in @xmath477 .
the vertices of @xmath40 are now vertices of @xmath150 or have degree at least three in @xmath40 .
assume first that a leaf @xmath233 of @xmath40 has its neighbour @xmath525 in @xmath150 .
denote by @xmath175 the component of @xmath526 containing @xmath233 . by construction @xmath233
is the unique element of @xmath507 .
we select @xmath527 .
if @xmath233 and @xmath525 are in different components of @xmath450 then we are in _ case a _ , otherwise we are in _ case b_. + so we can assume that all the leaves of @xmath40 are neighbours of vertices of @xmath40 which are not in @xmath150 and have degree at least three in @xmath40 .
consider now a leaf @xmath525 of the tree obtained from @xmath40 by removing its leaves .
denote by @xmath528 the set of neighbours of @xmath525 in @xmath40 which are leaves of @xmath40 .
if @xmath529 and @xmath528 consists of the end vertices of some @xmath509 then we choose @xmath527 and are in _ case b_. otherwise , let @xmath511 be the component of @xmath450 containing @xmath525 and consider a vertex @xmath233 of @xmath528 not belonging to @xmath511 ( this exists as @xmath530 ) .
then we choose @xmath527 and @xmath175 to be the component of @xmath526 containing @xmath233 and we are in _ case a_. @xmath105 + | an antidirected trail in a digraph is a trail ( a walk with no arc repeated ) in which the arcs alternate between forward and backward arcs .
an antidirected path is an antidirected trail where no vertex is repeated .
we show that it is np - complete to decide whether two vertices @xmath0 in a digraph are connected by an antidirected path , while one can decide in linear time whether they are connected by an antidirected trail .
a digraph @xmath1 is antistrong if it contains an antidirected @xmath2-trail starting and ending with a forward arc for every choice of @xmath3 .
we show that antistrong connectivity can be decided in linear time .
we discuss relations between antistrong connectivity and other properties of a digraph and show that the arc - minimal antistrong spanning subgraphs of a digraph are the bases of a matroid on its arc - set .
we show that one can determine in polynomial time the minimum number of new arcs whose addition to @xmath1 makes the resulting digraph the arc - disjoint union of @xmath4 antistrong digraphs . in particular , we determine the minimum number of new arcs which need to be added to a digraph to make it antistrong .
we use results from matroid theory to characterize graphs which have an antistrong orientation and give a polynomial time algorithm for constructing such an orientation when it exists .
this immediately gives analogous results for graphs which have a connected bipartite 2-detachment .
finally , we study arc - decompositions of antistrong digraphs and pose several problems and conjectures .
+ * keywords : * antidirected path , bipartite representation , matroid , detachment , anticonnected digraph |
one of the key qualitative properties of a packing is whether it is jammed , that is , whether the particles are locked into place .
jamming is of obvious scientific importance if we are using the packing to model a granular material .
furthermore , it plays a central role in studying local optimality of packings , because one natural way to try to improve a packing is to deform it so as to open up more space .
jamming has been extensively studied for packings in euclidean space .
see , for example , torquato and stillinger @xcite and the references cited therein .
however , it has been less thoroughly investigated in other geometries . in this paper
, we investigate jamming for sphere packings in spherical geometry , that is , packings of caps on the surface of a sphere .
jamming has previously been studied for sphere packings in @xmath2 ( see tarnai and gspr @xcite ) , but there seems to have been little investigation in higher dimensions .
a packing of congruent spherical caps on the unit sphere @xmath3 in @xmath4 yields a _
spherical code _ ( that is , a finite subset of @xmath3 ) consisting of the centers of the caps .
the _ minimal distance _ of such a code is the smallest angular separation between distinct points in the code . in other words , the cosine of the minimal distance is the greatest inner product between distinct points in the code .
the _ packing radius _ is half the minimal distance , because spherical caps of this radius centered at the points of the code will not overlap , except tangentially .
a spherical code is _ optimal _ if its minimal distance is as large as possible , given the dimension of the code and the number of points it contains .
( note that this notion of optimality is different from requiring that no more caps of the same size can be added without causing overlap .
neither of these two notions implies the other . )
spherical codes arise naturally in many parts of mathematics and science ( see cohn @xcite for a more extensive discussion ) .
for example , in @xmath5 they model pores in pollen grains or colloidal particles adsorbing to the surface of a droplet in a emulsion formed by two immiscible liquids . in higher dimensions
, they can be used as error - correcting codes for a constant - power radio transmitter .
furthermore , many beautiful spherical codes arise in lie theory , discrete geometry , or the study of the sporadic finite simple groups .
a _ deformation _ of a spherical code is a continuous motion of the points such that the minimal distance never drops below its initial value .
a deformation is an _ unjamming _ if it does not simply consist of applying global isometries ( that is , the pairwise distances do not all remain constant ) .
a spherical code is called _ rigid _ or _ jammed _ if it has no unjamming .
it is called _ locally jammed _ if no single point can be continuously moved while all the others are held fixed .
for example , in the face - centered cubic packing of balls in @xmath5 , the _ kissing configuration _ ( that is , points of tangency on a given ball ) consists of the vertices of a cuboctahedron .
this code is locally jammed , but it is not in fact jammed ( see conway and sloane @xcite or below ) .
however , it can be deformed into an optimalspherical code , namely the vertices of a regular icosahedron , and the icosahedron is then a rigid code with a higher minimal distance than that of the cuboctahedron .
as this example shows , deforming a spherical code is one way to improve it .
some optimal codes are not jammed ; for example , the best five - point codes in @xmath2 consist of two antipodal points and three points orthogonal to them , and the three points can move freely as long as they remain separated by at least an angle of @xmath6 . furthermore ,
computer experiments suggest that an optimal code can have _ rattlers _ , that is , points not in contact with any other point , although no such case has ever been rigorously analyzed .
however , despite these issues , rigidity is a powerful criterion for understanding when acode can be improved .
note that whether a configuration is jammed depends on the ambient space .
for example , the vertices of a square are jammed in @xmath7 but not in @xmath2 . for infinite packings in euclidean space , there are more subtle distinctions between different types of jamming ( see bezdek , bezdek and connelly @xcite and torquato and stillinger @xcite ) based on what sorts of motions are allowed .
for example , are all but finitely many particles held fixed ?
are shearing motions allowed ? however , these issues do not arise for packings in compact spaces .
nevertheless , jamming seems to be a more subtle phenomenon on spheres than it is in euclidean space . in euclidean space
, there is an efficient algorithm to test for jamming ( see donev , torquato , stillinger and connelly @xcite ) but on spheres we do not know such an algorithm .
the difficulty is caused by curvature , which complicates certain arguments .
for example , in euclidean space every infinitesimal unjamming extends to an actual unjamming , as we will explain in , but the corresponding procedure does not work on spheres . .
the best lower bounds known for kissing numbers in up to thirty - two dimensions .
numbers in bold are known to be optimal ( see schtte and van der waerden @xcite , leventen @xcite , odlyzko and sloane @xcite and musin @xcite ) . [ cols="^,^,^,^",options="header " , ] we do not expect that these bounds are anywhere close to being optimal , and already in @xmath8 they are worse than the previous record : they give @xmath9 , while the record is @xmath10 ( @xmath11 would be required to break the record ) . however ,
our bounds improve upon all cases from dimension @xmath0 to @xmath1 .
we conclude with some open problems .
the most basic is whether there are spherical codes that span the ambient space and are jammed but not infinitesimally jammed .
we suspect that there are such codes , but we have not found one . a related question is how to test efficiently whether a code is jammed ( which would be possible if jamming and infinitesimal jamming were equivalent , although one could hope for even faster algorithms ) .
such an algorithm might be based on higher - order variants of infinitesimal jamming , but they are far more subtle than one might expect ( see connelly and servatius @xcite ) . our new kissing records in dimensions @xmath0 through @xmath1 can presumably be improved , and it would be very interesting to know how far these techniques can be pushed , or how to construct much better arrangements .
the ratio of the upper bounds from shtrom @xcite to the lower bounds proved here grows roughly like @xmath12 in @xmath13 dimensions for @xmath14 , so there is considerable room for improvement .
note that optimizing the sizes of the sets @xmath15 and @xmath16 from can be viewed as maximizing the size of cliques in highly symmetrical graphs .
finding large cliques is np - hard in general , but that does not settle the question of how well one can solve this problem in practice
. among kissing problems in low dimensions , dimensions @xmath17 through @xmath18 seem ripe for improvement , although it is unlikely that the approach we have used here can be made to work .
dimension @xmath19 is particularly interesting , and we would very much like to know whether the @xmath20 minimal vectors of the barnes wall lattice solve the kissing problem .
they are certainly not the unique solution , because we have found that at least one of the @xmath19dimensional packings constructed by conway and sloane @xcite has a different kissing configuration of the same size .
it seems unlikely that one could classify all the possibilities in @xmath21 , but it might be possible to extend the conjecturally exhaustive list from to higher dimensions , perhaps up through @xmath22 . in we gave a list of the best kissing configurations known in @xmath23 through @xmath22 , but we suspect even more such configurations remain to be found .
as mentioned in that section , the @xmath24 root system fits into the same framework .
in fact its three other competitors do too , in the following sense . in each case
, one can find a cross polytope contained in the configuration ( for example , by a randomized , greedy algorithm ) .
using it to define the coordinate system yields simple rational coordinates for each point . in the @xmath24 case , these coordinates come from a constant weight code . in the other three cases they do not , but the coordinates can be obtained by a systematic mutation of the @xmath24 case . perhaps one could obtain additional kissing configurations in nine through twelve dimensions via a similar construction .
the problem of rigidity naturally generalizes to many other ambient spaces , such as projective spaces , but this generalization presents new and interesting features . for example , a projective configuration is not determined by its pairwise distance matrix , and in complex projective space there are even continuous families of optimal codes with exactly the same pairwise distances ( see cohn and kumar @xcite )
. even in euclidean space , which is the most thoroughly studied case so far , there are unresolved questions . how can one test whether a periodic packing is jammed ?
if one restricts attention to packings consisting of @xmath25 translates of a lattice ( that is , those with @xmath25 particles per unit cell ) , then one can test for infinitesimal jamming and hence jamming ( see donev , torquato , stillinger and connelly @xcite ) .
however , the answer may depend on @xmath25 .
for example , the laminated lattice @xmath26 is jammed with @xmath27 ( that is , it can not be deformed as a lattice packing ) , but not with @xmath28 ( giving rise to the fluid diamond packings of conway and sloane @xcite ) .
we know of no bound for how large @xmath25 must be to detect a lack of rigidity .
finally , we conjecture that in all sufficiently high dimensions , there exist optimal kissing configurations with no contacts whatsoever ( that is , no pairs of points with inner product @xmath29 ) , so they are unjammed in the strongest possible sense .
this phenomenon occurs in @xmath5 , but we know of no higher - dimensional cases .
part of our motivation for making this conjecture is that we know of no large , jammed kissing configurations at all in high dimensions .
the @xmath30 root system is jammed for @xmath31 , but it contains only @xmath32 points , which is tiny compared with the exponential growth of the kissing number ( see , for example , conway and sloane @xcite ) .
are there exponentially large jammed kissing configurations in high dimensions ?
are there even any of greater than quadratic size ?
for example , is the kissing configuration of the barnes wall lattice @xmath33 in @xmath34 always jammed ?
( it has size asymptotic to @xmath35 , where @xmath36 is a constant @xcite . )
we thank kroly bezdek , noam elkies , michel goemans and achill schrmann for helpful discussions , patric stergrd for providing computer files for best s constant weight codes , and an anonymous referee for providing useful feedback on the manuscript .
abhinav kumar was supported in part by national science foundation grants dms-0757765 and dms-0952486 and by a grant from the solomon buchsbaum research fund , and he thanks princeton university for its hospitality .
yang jiao and salvatore torquato were supported in part by nsf grants dms-0804431 and dmr-0820341 .
as supplementary information for this paper , we have made available sixteen data files through the ` arxiv.org ` e - print archive , where it is paper number http://arxiv.org/abs/1102.5060[`arxiv:1102.5060 ` ] .
one can access these files by downloading the source files for the paper .
the data files can also be downloaded from the web page for this article ( http://dx.doi.org/10.2140/gt.2011.15.2235[doi:10.2140/gt.2011.15.2235 ] ) .
ten of them describe the kissing configurations enumerated in .
these files are each named after the corresponding configuration : ` 5-40a.txt ` , ` 5-40b.txt ` , ` 6-72a.txt ` through ` 6-72d.txt ` , and ` 7-126a.txt ` through ` 7-126d.txt ` .
the first line of the file specifies the number @xmath25 of points , and the second line specifies the dimension @xmath37 of the ambient euclidean space .
the third line consists of @xmath37 positive integers @xmath38 ( separated by spaces ) , which are the coefficients of the diagonal quadratic form used to measure distances .
finally , the remaining @xmath25 lines each give the @xmath37 coordinates of one of the points in the configuration ( again separated by spaces ) .
the scaling has been chosen so that all the coordinates will be integers , and the inner product between points @xmath39 and @xmath40 is defined by @xmath41 equivalently , if we use the standard inner product , then the @xmath42th coordinate must be scaled by @xmath43 , but phrasing it in terms of changing the inner product avoids the need to use irrational numbers as coordinates .
( there exist different coordinate systems that use only rational numbers , even with the standard inner product , but the coordinates used here are compatible with the fibering construction from . )
note that within each file , all the vectors have the same norm , but they are not unit vectors .
four of the files , namely ` b9-18.txt ` , ` b10-30.txt ` , ` b11-35.txt ` and ` b12-51.txt ` , describe the constant weight binary codes used to build the kissing configurations from in .
the file ` bn-n.txt ` contains all the codes of block length @xmath37 , size @xmath25 , constant weight @xmath44 and minimal distance @xmath44 , up to isomorphism .
( we thank patric stergrd for providing these codes . ) each codeword is given on a line by itself , with the binary digits separated by spaces , and there is a blank line between different codes .
the remaining two files , ` r.txt ` and ` s.txt ` , describe the @xmath45point configuration @xmath15 and the @xmath46point configuration @xmath16 from .
these files are in a slightly different format : they omit the first three lines ( @xmath25 , @xmath37 and @xmath38 ) .
instead , each line specifies the @xmath47 coordinates of one of the points in the coordinate system used by conway and sloane ( * ? ? ?
* , figure 4.12 ) , but with the irrational factor of @xmath48 omitted . * mr best * , _ @xmath49=35 $ ] or some new optimal constant weight codes _ , technical report zn 71/77 , mathematisch centrum , amsterdam ( 1977 ) available at + 0 http://repository.cwi.nl/search/fullrecord.php?publnr=7433[0 ]
* h cohn * , http://dx.doi.org/10.1142/9789814324359_0152[_order and disorder in energy minimization _ ] , from : `` proceedings of the international congress of mathematicians , hyderabad , august 1927 , 2010 , volume iv '' , hindustan book agency , new delhi ( 2010 ) 24162443 * r connelly * , http://dx.doi.org/10.1016/0001-8708(80)90037-7[_the rigidity of certain cabled frameworks and the second - order rigidity of arbitrarily triangulated convex surfaces _ ] , adv .
( 1980 ) 272299 * jh conway * , * nja sloane * , http://dx.doi.org/10.1017/s0305004100060746[_the coxeter
todd lattice , the mitchell group , and related sphere packings _ ] , math .
cambridge philos .
93 ( 1983 ) 421440 * a donev * , * s torquato * , * fh stillinger * , * r connelly * , http://dx.doi.org/10.1016/j.jcp.2003.11.022[_a linear programming algorithm to test for jamming in hard - sphere packings _ ] , j. comput . phys
. 197 ( 2004 ) 139166 * rl griess , jr * , http://dx.doi.org/10.1016/j.aim.2006.09.003[_corrections and additions to : `` pieces of @xmath52 : existence and uniqueness for barnes wall and ypsilanti lattices '' [ adv .
math . 196 ( 2005 )
147192 ] _ ] , adv . math . 211 ( 2007 ) 819824 * c muss * , http://dx.doi.org/10.1016/s0096-3003(97)00004-0[_the dimensional family approach in ( hyper)sphere packing : a typological study of new patterns , structures , and interdimensional functions _ ] , appl . math .
88 ( 1997 ) 126 * am odlyzko * , * nja sloane * , http://dx.doi.org/10.1016/0097-3165(79)90074-8[_new bounds on the number of unit spheres that can touch a unit sphere in @xmath37 dimensions _ ] , j. combin
. theory ser . a 26 ( 1979 )
210214 * s torquato * , * fh stillinger * , http://dx.doi.org/10.1021/jp011960q[_multiplicity of generating , selection , and classification procedures for jammed hard - particle packings _ ] , j. phys . chem .
b 105 ( 2001 ) 1184911853 * va zinoviev * , * t ericson * , _ new lower bounds for contact numbers in small dimensions _ , problemy peredachi informatsii 35 ( 1999 ) 311 ( russian ) english translation in problems inform .
transmission 35 ( 1999 ) 287294 | a packing of spherical caps on the surface of a sphere ( that is , a spherical code ) is called rigid or jammed if it is isolated within the space of packings . in other words , aside from applying a global isometry , the packing can not be deformed . in this paper
, we systematically study the rigidity of spherical codes , particularly kissing configurations .
one surprise is that the kissing configuration of the coxeter
todd lattice is not jammed , despite being locally jammed ( each individual cap is held in place if its neighbors are fixed ) ; in this respect , the coxeter
todd lattice is analogous to the face - centered cubic lattice in three dimensions .
by contrast , we find that many other packings have jammed kissing configurations , including the barnes wall lattice and all of the best kissing configurations known in four through twelve dimensions .
jamming seems to become much less common for large kissing configurations in higher dimensions , and in particular it fails for the best kissing configurations known in @xmath0 through @xmath1 dimensions .
motivated by this phenomenon , we find new kissing configurations in these dimensions , which improve on the records set in 1982 by the laminated lattices .
a packing of spherical caps on the surface of a sphere ( that is , a spherical code ) is called rigid or jammed if it is isolated within the space of packings . in other words , aside from applying a global isometry , the packing can not be deformed . in this paper
, we systematically study the rigidity of spherical codes , particularly kissing configurations .
one surprise is that the kissing configuration of the coxeter - todd lattice is not jammed , despite being locally jammed ( each individual cap is held in place if its neighbors are fixed ) ; in this respect , the coxeter - todd lattice is analogous to the face - centered cubic lattice in three dimensions .
by contrast , we find that many other packings have jammed kissing configurations , including the barnes - wall lattice and all of the best kissing configurations known in four through twelve dimensions .
jamming seems to become much less common for large kissing configurations in higher dimensions , and in particular it fails for the best kissing configurations known in 25 through 31 dimensions .
motivated by this phenomenon , we find new kissing configurations in these dimensions , which improve on the records set in 1982 by the laminated lattices . |
dynamical friction ( df ) arising from the gravitational interaction of a massive body with a background medium is considered to be one of the most powerful mechanisms responsible for the orbital decay of astronomical objects ( e.g. , @xcite and references therein ) .
df occurs not only in a collisionless background in which long - range , two - body interactions transfer momentum from an object in motion to the field particles , but also in a continuous gaseous medium where a perturber loses its momentum due to the gravitational drag force exerted by its own induced wake . the df in a gaseous background is likely to play an important role in the orbital decay of companions in common - envelope binaries , supermassive black holes ( smbhs ) at galaxy centers , massive galaxies in galaxy clusters , etc .
for instance , in a binary system , df causes a low - mass companion engulfed by an extended primary to spiral in toward the core of the primary , which in turn spins up the primary envelope , greatly affecting the subsequent evolution of the system ( e.g. , @xcite and references therein ) .
also , numerical simulations show that the df due to a gaseous medium expedites the growth of smbhs by mergers in colliding galaxies ( e.g. , @xcite ) , potentially explaining the ubiquity of smbhs at galactic nuclei ( e.g. , @xcite ) . using a time - dependent linear perturbation theory
, @xcite showed that the drag force on a perturber with mass @xmath3 moving at speed @xmath5 on a rectilinear trajectory through a uniform gaseous medium with density @xmath1 and sound speed @xmath2 is given by @xmath17 where @xmath18 is the mach number , and @xmath19 is the characteristic size of the perturber .
the key features of equation ( [ eq : f_pt_linear ] ) are that ( i ) the gaseous df force becomes identical to the classical formula of @xcite for the collisionless drag when @xmath20 , ( ii ) the gaseous df force is more efficient than the collisionless drag for @xmath21 , and ( iii ) the gaseous df force is nonvanishing even for subsonic perturbers with @xmath22 ( see also @xcite ) .
the last point improves the previous notion that the gaseous drag is absent in subsonic cases due to the front - back symmetry in the steady - state wakes ( e .. g , @xcite ) .
several recent studies showed that equation ( [ eq : f_pt_linear ] ) can be applicable , with some modifications , to more general cases such as in a radially stratified medium @xcite , for circular - orbit perturbers @xcite , for perturbers with relativistic speed @xcite or in accelerating motion @xcite , etc .
in particular , ( * ? ? ?
* hereafter kk07 ) used a semi - analytic method to show that equation ( [ eq : f_pt_linear ] ) is a good approximation to the df force on circular - orbit perturbers provided @xmath23 , where @xmath24 is the orbital radius .
while all the theoretical studies mentioned above consider low - mass perturbers and find various astrophysical applications ( e.g. , @xcite ) , there are some situations such as in orbital decay of smbhs or companions in common - envelope binaries , where perturbers have so large masses that the induced density wakes are in the nonlinear regime . using hydrodynamic simulations , (
* hereafter kk09 ) extended the work of @xcite to study nonlinear df force for a very massive perturber on a straight - line trajectory . by modeling a perturber using a plummer sphere with softening radius @xmath4
, kk09 found that a nonlinear supersonic wake is characterized by a detached bow shock and a hydrostatic envelope near the perturber .
the resulting drag force depends solely on the dimensionless parameter @xmath25 defined as @xmath26 and is given by @xmath27 while @xmath28 for @xmath29 .
the reduction of the nonlinear df force compared to the linear estimate is due to the presence of a hydrostatic envelope that makes the density distribution spherically symmetric near the perturber .
this clearly demonstrates that the nonlinear effect can be significant if a perturber is very massive . since astronomical objects usually follow curvilinear rather than straight - line trajectories , it is interesting to see whether equation ( [ eq : kk09 ] ) remains valid for perturbers on circular orbits . for this purpose ,
we in this paper take one step further from kk09 to study the nonlinear gravitational responses of gas to , and the associated drag force on , a very massive perturber moving on a circular orbit .
this work also extends the linear perturbation analyses of kk07 by considering the nonlinear effect on the density wakes .
the intensity of gravitational influence of a perturber to the background gas can be measured by the bondi radius @xmath30 . for perturbers on straight - line trajectories , the softening radius ( or the perturber size ) is the lone length scale with which the bondi radius can be compared .
the drag force , correspondingly , depends on @xmath4 and @xmath3 only through the dimensionless parameter @xmath31 : increasing @xmath4 is equivalent to decreasing @xmath3 .
equation ( [ eq : kk09 ] ) then predicts that the drag force on a highly nonlinear object with @xmath32 would be negligible , which results simply from the fact that @xmath4 is inseparable from @xmath3 . on the other hand , circular orbits with orbital radius @xmath24
naturally introduce an additional length scale , so that the dependence of the drag force on @xmath4 can be explored independently of the perturber mass .
while a perturber in real astronomical situations is likely to move through a background medium that is rotating and/or stratified , we in this work idealize the gaseous medium as being initially static and uniform .
this certainly introduces a few important caveats that should be noted from the outset . if the gaseous medium is supported primarily by rotation , as in protoplanetary disks , the perturber launches spiral waves at lindblad resonances that propagate away from it ( e.g. , @xcite and references therein ) , which is quite different from quasi - steady density wakes with long trailing tails produced in a non - rotating medium ( e.g. , kk07 ) .
if the medium is instead supported by thermal pressure , it should have a stratified density in the radial direction , as in intracluster media or common - envelope binaries . in this case
, the gradient in the background density profile can be ignored if the bondi radius is much smaller than _ both _ the pressure scale length _ and _ the orbital radius , that is , if @xmath33 and @xmath34 , where @xmath35 is the dynamical mass enclosed within @xmath24 and @xmath36 is the dimensionless perturber mass defined in
[ sec : code ] below . although the first condition is readily met , for example , in common - envelope binaries with planets or brown dwarfs as low - mass companions ( e.g. , @xcite ) , the second condition is not well satisfied especially for very massive perturbers with @xmath37 considered in this paper . in the latter case ,
the wakes and the associated df forces are likely to be affected by the density gradient of the background medium ( e.g. , see @xcite for collisionless cases ) .
neglecting the density stratification in the background medium also suppresses gas buoyancy , precluding the potential effects of convective motions and gravity modes in heating the medium by resonant excitations ( e.g. , @xcite ) and other processes .
in addition to the above assumptions , we ignore the orbital motion of the background gas with respect to the center of mass of the whole system , which can be a valid approximation only when @xmath38 .
if this condition is not satisfied , the neglect of the centrifugal and coriolis forces arising from the orbital motion of the background may affect the density wakes and the drag forces ( e.g. , @xcite ) . we also treat the gas using an adiabatic equation of state , which implicitly assumes that the orbital energy of the perturber is converted to heat as it spirals inward , and thus is valid only if radiative loss is negligible . by ignoring self - gravity
, we do not consider any back reaction of the gas on the perturber . given these limitations and constraints , we by no means attempt to apply the results of this work to real astronomical systems .
nevertheless , the idealized models considered in this paper help to isolate the effect of the perturber size or orbital radius on the nonlinear df force
. the results of this work will be particularly useful to justify large perturber sizes employed in recent hydrodynamic simulations , such as for smbhs at galactic nuclei ( e.g. , @xcite ) and companions in common - envelope binaries ( e.g. , @xcite ) , etc .
these simulations usually treat the perturber using a softened point mass , with its size inevitably limited by numerical resolution .
for instance , @xmath39-body / sph simulations for the orbital decay of smbhs take quite large values ( up to a few pc ) for @xmath4 ( e.g. , @xcite ) , although the realistic values are probably of order of the schwartzschild radius ( @xmath40 au for @xmath41 ) .
grid - based simulations of the common - envelope phase of binaries also represent a companion using a point mass with size set by the grid spacing , which is larger than the actual companion size by one or two orders of magnitude ( e.g. , @xcite ) .
if the drag force on circular - orbit perturbers depends on @xmath4 similarly to the linear - trajectory cases , simulations with such large @xmath4 would significantly underestimate the real decay time since the induced density wake would erroneously be in the linear regime .
if the drag force instead turns out insensitive to @xmath4 in circular - orbit cases , a large value of @xmath4 would be reasonable as long as it does not change the drag force much .
the paper is organized as follows : in [ sec : code ] , we describe numerical methods and models we adopt . in [ sec : linear ] , we run models for low - mass perturbers and compare the resulting drag forces with the analytical predictions to find a proper relationship between @xmath4 and @xmath19 . in [ sec : nonl ] , we present evolution and quasi - steady distributions of fully nonlinear density wakes and the associated drag forces on massive , circular - orbit perturbers . finally in
[ sec : diss ] , we summarize our findings and discuss their implications on the proper choice of the perturber size .
we consider a massive perturber moving on a circular orbit through a gaseous medium , and study the gravitational response of the gas to the perturber and the resulting drag force .
similar work for low - mass perturbers on circular orbits and massive perturbers on straight - line trajectories was presented in kk07 and kk09 , respectively .
since the df timescale is usually much longer than the orbital time , the idealized circular orbit is a reasonable approximation to real curvilinear orbits .
we assume that the background medium is unmagnetized , non - self - gravitating , and initially static and homogeneous with density @xmath1 and adiabatic sound speed @xmath2 .
we adopt an adiabatic equation of state with index @xmath42 . the perturber with mass @xmath3 moves with a constant angular velocity @xmath43 along a circle with radius @xmath6 on the @xmath44 plane ; the corresponding linear velocity is @xmath45 .
we represent the perturber using a plummer potential @xmath46 where @xmath4 is the softening radius and @xmath47 is the perturber location at time @xmath48 in the cartesian coordinates @xmath49 .
we take @xmath6 , @xmath2 , and @xmath50 as the units of length , velocity , and time , respectively , in our simulations .
then , our models are completely characterized by three dimensionless parameters : @xmath51 , @xmath18 , and @xmath52 which is the ratio of the bondi radius to the orbital radius .
note that @xmath15 is a new parameter introduced by the circular orbit , which is absent for perturbers on straight - line trajectories . for future purposes , we define a dimensionless parameter @xmath53 similarly to @xmath25 in equation ( [ etaa ] ) . to explore the parametric dependence of the drag force on @xmath15 , @xmath11 , and @xmath36 ( or equivalently @xmath31 ) , we run a total of 65 three - dimensional simulations with @xmath15 varying in the range of @xmath54@xmath55 , @xmath11 in @xmath564 , and @xmath57 from @xmath58 to @xmath59 ( @xmath31 from 0.01 to 80 ) .
our fiducial model has @xmath60 , @xmath61 , and @xmath62 .
we use the orbital time @xmath63 as the time unit of our presentation .
we integrate the basic equations of ideal hydrodynamics using a modified version of the zeus code @xcite , parallelized on a distributed - memory platform .
our simulations domain is a rectangular box which spans the region @xmath64 .
we construct a logarithmically spaced cartesian grid with @xmath65 zones , with spacings of @xmath66 , @xmath67 , and @xmath68 at @xmath69 , 1 , and 10 , respectively .
for some selected parameters , we have also run low - resolution simulations with @xmath70 zones and checked that the resulting drag forces agree , within a few percents , with those from the high - resolution models .
we adopt the outflow boundary conditions at all the boundaries and run the simulations typically up to @xmath71 .
the simulation outcomes are insensitive to the box size and the boundary conditions we adopt since it takes about @xmath72 for sound waves to travel across the simulation box , while the drag forces saturate typically within @xmath73 .
the time - dependent linear - perturbation analyses for the wakes of low - mass perturbers usually consider a point - mass object with vanishing @xmath4 in equation ( [ eq : plummer ] ) , which in turn requires to introduce the cut - off radius @xmath19 in the linear df force formula ( e.g , @xcite ; kk07 ) . on the other hand
, one has to adopt a non - zero value for the softening radius in numerical simulations , making the df force dependent on @xmath4 in the linear regime . to explore how the perturber mass affects the drag force in comparison with the linear theory , therefore
, it is necessary to find a proper relationship between @xmath4 and @xmath19 that makes the numerical results consistent with the analytic predictions when @xmath34 . for this purpose ,
we present in this section the results of numerical simulations for low - mass perturbers with @xmath74 ( @xmath75 ) and @xmath60 as functions of @xmath11 .
figure [ fig : lin_wake ] shows the snapshots of density wakes on the @xmath76@xmath77 plane at @xmath78 for @xmath79 , 1.5 , 2 , and 3 cases .
the perturber in each frame is moving along the white circle in the counterclockwise direction .
for subsonic perturbers , the circular orbit makes the density wakes curved along their orbits that would otherwise remain symmetric with respect to the line of motion . for supersonic models ,
the perturber is able to overtake its own wake , creating a strong trailing tail bounded by mach waves ( kk07 ) . as the mach number increases from unity , the opening angle of the head of the curved mach cone deceases , while the tail thickens .
the dotted lines for the supersonic models plot equation ( b2 ) of kk07 for the shape of the tail for @xmath80 , with @xmath81 being the azimuthal angle , suggesting that the wake tail at large radii resembles an archimedean spiral . ] , which are in excellent agreement with the outer boundaries of the tails produced in the simulations .
nevertheless , the finite perturber size in the latter makes the tail boundaries broader compared to the point - mass counterparts with @xmath82 . for a given density wake @xmath83 , it is straightforward to calculate the gravitational drag force on the perturber by evaluating the integral @xmath84 where @xmath85 is the mass distribution of the plummer sphere and @xmath86 is the unit vector in the azimuthal direction . figure [ fig : lin_df]_a _ shows the temporal evolution of @xmath87 for the cases shown in figure [ fig : lin_wake ] .
the friction force saturates to a constant value in less than @xmath88 ; the amplitudes of fluctuations in @xmath89 are less than @xmath90 of the mean values .
this is unlike in the linear - trajectory models where the df force on a supersonic perturber increases logarithmically with time , since the whole density wake growing in size with time is located behind the perturber .
the region of influence in the circular - orbit models expands at a sonic speed from the orbit center .
consequently , the far - field wake in these models is more or less spherically symmetric and thus does not contribute to the df force much .
figure [ fig : lin_df]_b _ plots as solid circles the steady - state drag forces against @xmath11 for low - mass perturbers with @xmath74 ( @xmath75 ) and @xmath60 .
the solid line represents the fit using equation ( [ eq : f_pt_linear ] ) with @xmath91 which is in good agreement with the numerical results . in what follows
, we will use equation ( [ eq : rmin ] ) as the cut - off radius when we compare the numerical df forces with the analytic estimates .
we now focus on the cases with very massive perturbers .
figure [ fig : den_sup ] plots the snapshots of logarithmic density on the @xmath76@xmath77 plane for our standard model with @xmath60 , @xmath61 , and @xmath62 ( @xmath92 ) . in each panel ,
the white line draws the perturber orbit , while the small circle in black marks the perturber location with its size corresponding to the softening radius @xmath4 .
the perturber is initially located at @xmath93 and orbits in the counterclockwise direction .
note that the size of the region displayed in figure [ fig : den_sup ] differs from panel to panel for clear presentation .
the evolution of the density wake at early phase is qualitatively similar to the linear - trajectory cases described in kk09 .
an introduction of the perturber at @xmath94 provides the background gas with strong perturbations that readily develop into a bow shock .
the incident gas along the line of motion is shocked and gathered near the perturber .
this sets up a strong pressure gradient in the region between the shock and the perturber , which in turn causes the shock to slowly move away from the perturber in the upstream direction until it reaches an equilibrium position .
some gas flowing with a non - zero impact parameter is gradually pulled by the perturber , adding to the density at the rear side ( fig .
[ fig : den_sup]@xmath95 ) . due to the strong gravity ,
the material piled up at the backside is pulled back toward the perturber , creating a strong counterstream along the line of motion as well as a surrounding vortex ring ( see kk09 ) . for the linear - trajectory cases
, kk09 has shown that the counterstream slows down due in part to the strong pressure gradient established near the perturber and in part to a buoyant expansion of the vortex ring in the lateral direction .
the interaction of the counterstream with the shocked gas causes the bow shock to move back and forth around an equilibrium position . during this time
, the vortex ring also undergoes an overstable oscillation whose amplitude grows secularly with time .
when the vortex ring moves out beyond a half of the hoyle - lyttleton radius @xmath96 , it becomes less gravitationally bound and soon swept downstream by the ram pressure of the incident gas , leaving a near - hydrostatic envelope surrounding the perturber .
this is how a density wake enters into a quasi - steady state in the linear - trajectory cases . for the circular - orbit cases
, however , the line of perturber motion keeps changing with time . with the perturber displaced from the direction of the counterstream ,
the counterstream is virtually unhindered and travels almost straight in the direction tangent to the orbit , producing the protruding region in the wake at @xmath97 shown in figure [ fig : den_sup]@xmath98 .
consequently , the vortex ring manifesting its presence by low density at @xmath99@xmath100 and @xmath101@xmath102 in figure [ fig : den_sup]@xmath103 that was carried by the counterstream decouples from the perturber , promptly leaving a quasi - static density distribution around the perturber .
note that the time to form a near - hydrostatic envelope is about @xmath104 , which is shorter than in the linear - trajectory cases by about an order of magnitude .
while the density distribution near the perturber is nearly spherical , a part of the wake outside the orbit is pushed radially outward by the counterstream , developing a trailing tail on the @xmath76@xmath77 plane . as the counterstream weakens , the low - density regions associated with the vortex ring first merge together and are then stretched along the orbit as the perturber continues the orbital motion ( fig .
[ fig : den_sup]@xmath105 ) . in the low - mass perturber cases with @xmath106 ,
the wake tail is created by the overlapping of the mach cone and the sonic sphere ( kk07 ) . in the nonlinear cases ,
however , it is the low - density regions produced by the counterstream that separate the tail from the rest of the wake . the black dotted curve in figure [ fig : den_sup]_f _ plots the shape of the linear tail , as in figure [ fig : lin_wake ] , which does not match the outer edge of the nonlinear tail in our fiducial model . instead ,
the latter is well matched by rotating the former by @xmath107 in the counterclockwise direction , shown as the dashed curve .
figure [ fig : den_vel ] shows the quasi - steady distributions of the density as well as the velocity field , seen in the inertial frame , on the @xmath108 , @xmath109 , and @xmath44 planes for our fiducial model at @xmath110 .
the colorbar labels the perturbed density in logarithmic scale .
the size of the short line segment shown in the upper left corner of each panel , corresponding to twice the sound speed , measures the amplitudes of the velocity vectors .
the density wake consists of a spherical envelope surrounding the perturber , a detached bow shock , and an extended low - density region at the rear side .
the envelope is almost hydrostatic , as evidenced by low - amplitude velocity vectors in the @xmath109 plane to which the perturber motion is perpendicular instantaneously .
the bow shock standing outside the orbit , in the @xmath44 plane , is just a curved version of that in the linear trajectory counterpart and extends to large radii @xmath111 in the orbital plane . on the other hand , the ( almost planner )
shock located inside the orbit gradually weakens at small @xmath112 and terminates at @xmath113@xmath114 since the speed of the density wake corotating with the perturber becomes subsonic there .
the low - density regions delineating the tail seen at @xmath78 ( fig .
[ fig : den_sup]_f _ ) are less apparent in figure [ fig : den_vel ] as they diffuse out over time by pressure gradients . while the velocity just ahead of the perturber is parallel to the direction of the instantaneous perturber motion , the direction of gas motions at the perturber location and the immediate behind it are inclined about @xmath115@xmath116 from the tangential direction .
this implies that the gas in the downstream side retains the information on the direction of perturber motion in the past .
figure [ fig : non_wake ] displays changes in the quasi - steady density wakes with varying @xmath11 on the orbital plane for models with @xmath117 and @xmath118 ( @xmath119 ) . in each panel , the perturber located at @xmath108 and @xmath109 is moving along the white circle in the counterclockwise direction . when @xmath120 , the density distribution is circular near the perturber and becomes eccentric away from it , with the minor axis parallel to the line of motion .
still , there exists a slight excess of the perturbed density in the rear side , giving rise to a non - vanishing ( but small ) df force .
compared to the @xmath120 model , the model with @xmath79 has a larger density excess at the backside . while the perturber is subsonic in this model , a strong gravitational pull is able to accelerate the background gas to a near - transonic speed , creating a weak shock outside the orbit .
the shape of the density wake in the @xmath121 model is similar to that for the @xmath122 model , although the former has a larger peak density and a larger distance of the detached bow shock . as @xmath11 increases further , both the detached shock distance and the size of the hydrostatic envelope decrease .
the ring - like low - density gap at @xmath123 in figure [ fig : non_wake]_d _ is a remnant of the counterstream that was stretched along the orbit . with large values of @xmath11 and @xmath36 ,
the counterstream in this model was so strong that the deep gap has not been filled in yet . in order to check whether the density wake near the perturber is in hydrostatic equilibrium , we plot in figure [ fig : cut_den ] the density profiles of steady - state wakes as functions of distance @xmath124 measured from the perturber location along the azimuthal direction in models with @xmath61 .
the solid lines in figure [ fig : cut_den]_a _ give the simulation results for perturbers with fixed @xmath60 and differing @xmath36 .
the sharp discontinuity in each curve is due to the presence of the bow shock .
the dotted lines draw the density distributions of the hydrostatic spheres @xmath125 \right\}^{1/(\gamma-1)},\ ] ] where @xmath126 and @xmath127 denote the density and adiabatic sound speed at @xmath128 ( i.e. , the perturber center ) , respectively ( kk09 ) .
these are in good agreement with the density profiles of the steady - state wakes in our simulations , suggesting that the envelopes are indeed almost hydrostatic .
figure [ fig : cut_den]_b _ shows how the density profile varies with @xmath15 . when @xmath118 , the nonlinear effect is not significant in models with @xmath129 since they have @xmath130 ; the density jump occurring at @xmath131 in the @xmath132 model is more like a mach wave rather than a shock .
as @xmath15 decreases ( or @xmath25 increases ) , the density wake becomes increasingly nonlinear , producing a hydrostatic envelope and a detached bow shock .
note that the shock location is unchanged as long as @xmath133 ( or @xmath134 ) .
when the density is highly nonlinear , the strong gravity associated with small @xmath4 increases the density only in the small regions within @xmath135 .
since this core region is spherically symmetric , the resultant df force becomes independent of @xmath15 provided @xmath134 .
we will show this directly in [ sec : df ] below .
unlike in the linear - trajectory cases where a perturber travels through an undisturbed medium , it is inevitable that a circular - orbit perturber enters the backside of its own wake before completing one orbit . in the circular - orbit cases , therefore , the density ahead of the bow shock differs from the undisturbed density @xmath1 and depends considerably on @xmath36 , as figure [ fig : cut_den]_a _ illustrates . the enhanced preshock density can alternatively be interpreted as arising from the response of the background gas to the orbit - averaged gravitational potential of the perturber @xmath136 , which is centered at the orbit center . in this sense ,
the enhanced preshock density @xmath137 acts as a modified background density to a circular - orbit perturber with the gravitational potential @xmath138 . to quantify @xmath137
, we take a spatial average of the preshock density in quasi - steady state over the region with @xmath139@xmath140 along the orbit , where @xmath141 denotes the location of the detached shock .
figure [ fig : rho1 ] plots the resulting @xmath137 for models with @xmath60 , with errorbars representing the standard deviations .
the solid line is the best fit @xmath142 showing that the change in the background density to circular - orbit perturbers is almost linearly proportional to @xmath36 . since @xmath137 is quite insensitive to @xmath11 , the @xmath11-dependence of @xmath137 in equation ( [ eq : rho1 ] ) can be ignored . in laboratory experiments and hydrodynamic theories ,
supersonic flows over a blunt - body generate a detached bow shock when the maximum angle allowed in the postshock flows is smaller than the nose angle ( e.g. , @xcite ) .
although a perturber in our models does not hold any defined surface and merely provides gravitational perturbations to the background gas , a spherical envelope formed around it nevertheless sends off pressure waves into the upstream direction , acting as an obstacle to the incoming gas .
since the nose - angle of a spherical body is @xmath143 , the shock should be detached in our nonlinear supersonic models .
figure [ fig : delta_time]_a _ plots the time evolution of the detached shock distance @xmath141 for some selected models with @xmath117 and @xmath61 .
note that @xmath141 initially increases rapidly and then saturates at @xmath144 to an equilibrium value with some fluctuations .
the quasi - steady value of @xmath141 becomes larger as the perturber mass increases . for massive perturbers moving on a straight - line trajectory
, kk09 found that the detached shock distance is given empirically by @xmath145 this can be interpreted by the balance between the postshock thermal energy ( @xmath146 ) and the gravitational potential energy ( @xmath147 ) at the shock location for strong shocks ( @xmath20 ) . to study how the circular obit affects the standoff distance of the shock ,
we take a time average of @xmath148 over @xmath149@xmath150 and plot @xmath151 in figure [ fig : delta ] as various symbols .
again , errorbars indicate the standard deviations of @xmath148 during this time interval .
the numerical results show that the detached shock distances in the circular orbits are in general smaller than those in the linear - trajectory cases . this is likely because the continuous change in the direction of perturber motion effectively reduces the forward momentum of the incident gas near the perturber , as illustrated by the spatially - varying velocity field in the wake shown in figure [ fig : den_vel ] . in this case , the ram pressure of the incoming flow is able to push the shock front toward the perturber , making @xmath141 smaller .
the modified background density and sound speed discussed in [ sec : wake ] might affect @xmath141 as well .
we fit our results using @xmath152 with @xmath153 , which is plotted as dotted lines . as expected , @xmath154 as @xmath155 corresponding to @xmath156 , although models with @xmath157 are weakly nonlinear and have @xmath141 comparable to or less than @xmath4 . since @xmath158 , equation ( [ eq : del_cir ] ) indicates that @xmath141 increases with @xmath159 , although its increasing rate slows down as @xmath160 where @xmath141 becomes comparable to @xmath24 .
for our nonlinear models , we calculate the df force on a perturber due to its own induced wake .
figure [ fig : delta_time]_b _ illustrates the temporal changes of the dimensionless drag force @xmath161 $ ] in models with @xmath61 and @xmath117 and varying @xmath36 .
the drag force reaches a quasi - steady value in less than one orbit , although it exhibits some fluctuations for large @xmath36 .
this is markedly different from the cases of straight - line trajectories where the drag force increases logarithmically with time .
the quasi - steady value of @xmath162 is almost independent of @xmath36 when @xmath163 , decreases with @xmath36 for @xmath164 , and becomes again insensitive to @xmath36 when @xmath165 .
the behavior of @xmath162 with @xmath36 is a consequence of the circular orbit combined with the nonlinear effect .
the size of a hydrostatic envelope in a nonlinear density wake , as measured by the detached shock distance , becomes larger with increasing @xmath36 , which makes the drag force smaller owing to the front - back symmetry in the wake . at the same time
, the circular orbit draws the background material toward the orbit center and increases the density in the preshock region , which in turn tends to increase the drag force . for models with @xmath166 ( so that @xmath167 ) , the nonlinear effect dominates and @xmath162 decreases with @xmath36 . when @xmath165 , on the other hand , the enhancement of the background density becomes significant enough to compensate for the nonlinear effect , making @xmath162 unchanged with @xmath36 .
as we will show below , the df force scales well with @xmath36 when @xmath168 is taken into account .
compared to the linear - trajectory cases , the circular orbit introduces an additional length scale , @xmath24 , so that it is interesting to explore the dependence of the drag force upon @xmath15 . figure [ fig : df_rs]_a _ plots as various symbols and errorbars the mean values and standard deviations of the drag forces averaged over @xmath1104 for models with @xmath61 .
note that the drag force is normalized by @xmath137 instead of @xmath1 . the dotted line marked with @xmath169 demarcates the parameter space into two parts such that the models shown in the lower - left region are fully nonlinear with @xmath170 , while those in the upper - right region are in the linear or weakly nonlinear regime .
the normalized drag force on weakly - nonlinear perturbers with @xmath171 is usually larger than that on low - mass perturbers with @xmath29 because the density wakes in the former are distributed closer to the perturber ( kk09 ) . in highly - nonlinear models with @xmath170 ( or @xmath172 ) ,
on the other hand , the df force becomes essentially independent of @xmath15 .
this is of course because the change in @xmath4 affects the wake only within a small region close to the perturber where the density is spherically symmetric , making a negligible contribution to the drag force ( see [ sec : wake ] ) .
figure [ fig : df_rs]_b _ plots @xmath173 versus @xmath25 for the same models shown in figure [ fig : df_rs]_a_. note again that @xmath1 in equation ( [ eq : f_pt_linear ] ) for @xmath174 is replaced by @xmath137 . for given @xmath36 (
so that increasing @xmath25 corresponds to decreasing @xmath15 ) , @xmath173 decreases logarithmically with increasing @xmath175 .
this is because @xmath176 for fixed @xmath36 and @xmath11 , leading to @xmath177 , while @xmath89 is constant for @xmath170 . for fixed @xmath15 ( so that increasing @xmath25 corresponds to increasing @xmath36 ) , @xmath178 , as illustrated by two dotted lines connecting the numerical results for @xmath60 and 0.025 , respectively .
figure [ fig : df_non]_a _ plots @xmath173 for all of our supersonic models with @xmath60 .
obviously , @xmath28 for low - mass perturbers with @xmath29 .
the dotted line is the best fit @xmath179 to the numerical results for nonlinear , circular - orbit perturbers , which is very similar to equation ( [ eq : kk09 ] ) for massive , linear - trajectory perturbers . in order to apply equation ( [ eq : linfit ] ) to estimate the decay timescale
, one needs to specify the softening radius which is not directly observable in most astronomical systems .
in addition , equation ( [ eq : linfit ] ) is applicable only when @xmath60 since @xmath89 on highly massive perturbers is independent of @xmath4 , while @xmath174 diverges as @xmath180 . for practical use in various situations , therefore , it is desirable to have an expression of @xmath89 that does not rely on @xmath4 . for this purpose ,
we plot in figure [ fig : df_non]_b _ the dimensionless df forces as a function of @xmath181 for all supersonic models with @xmath60 .
the dotted line is our fit to the nonlinear part using @xmath182 with @xmath137 given by equation ( [ eq : rho1 ] ) . for the models with @xmath14 , equations ( [ eq : f_pt_linear ] ) and ( [ eq : rmin ] ) with @xmath183 provide a good estimate of @xmath89 that depends on @xmath4
df due a gaseous medium may play a central role in removing angular momentum from astronomical objects in orbital motions , causing them to spiral in toward the orbit center . in this paper
, we use three - dimensional hydrodynamic simulations to explore the gravitational wake and the associated drag force on a very massive perturber with mass @xmath3 moving at speed @xmath5 on a circular orbit with radius @xmath24 .
this work extends our previous studies that considered low - mass , circular - orbit perturbers ( kk07 ) and high - mass , linear - trajectory perturbers ( kk09 ) .
the background medium is assumed to be uniform with density @xmath1 and the sound speed @xmath2 , adiabatic with index @xmath42 , non - self - gravitating , non - rotating , and unmagnetized .
the perturber is represented by a plummer sphere with softening radius @xmath4 .
our models are characterized completely by three dimensionless parameters : @xmath184 , @xmath185 , and @xmath186 .
we run a total of 65 models that differ in @xmath15 , @xmath11 , and @xmath36 , and obtain the dependence of the drag force on these parameters . for our supersonic models , the density wake of a massive perturber reaches a quasi - steady state typically in @xmath187 , about 10 times faster than the linear - trajectory cases . a quasi - steady density wake consists of a spherical envelope surrounding the perturber , a detached bow shock in the upstream direction , and an extended low - density region in the rear side .
the envelope is almost hydrostatic , providing a negligible contribution to the df force .
compared to the linear - trajectory cases , an orbit - averaged gravitational potential of a circular - orbit perturber is able to gather the background gas toward the orbit center , effectively increasing the background density ahead of the bow shock .
equation ( [ eq : rho1 ] ) describes the modified preshock density @xmath137 , showing that density enhancement in the preshock region is almost linearly proportional to @xmath36 and depends weakly on @xmath11 .
strong perturbations sent off from a massive perturber develop into a bow shock through which the upstream supersonic flow becomes subsonic .
since the density wake in a quasi - steady state forms a pattern that is corotating with the perturber at a constant angular speed , the incident flow outside the orbit is supersonic to a perturber with @xmath188 .
this naturally produces a bow shock outside the orbit that is simply a curved version of that in the linear - trajectory counterpart .
formation of a weak bow shock outside the orbit is possible even for a subsonic perturber with @xmath189 for which the incident flow is accelerated to near - transonic speeds . on the other hand , the motion of the steady - state wake is subsonic to the gas near the orbit center , even when @xmath188 .
this makes the bow shock gradually weaken at small radii , eventually terminating somewhere inside the orbit .
the detached shock distance measured along the tangential direction is generally smaller than in the liner - trajectory counterpart since the continuous change in the direction of the perturber motion effectively reduces the forward momentum of the wake ; equation ( [ eq : del_cir ] ) gives the algebraic fits to the numerical results for the detached shock distance .
unlike in the linear - trajectory cases where the drag force increases logarithmically with time , the df force in the circular orbits approaches a quasi - steady value in less than one orbit .
the quasi - steady drag force is essentially independent of @xmath15 as long as @xmath170 ( or @xmath172 ) since the change in @xmath4 modifies the density wake only within @xmath190 from the perturber center , where the wake is almost spherically symmetric and thus has a negligible contribution to the drag force . for sufficiently massive perturbers ,
the presence of a near - hydrostatic envelope in the density wake also makes the nonlinear drag force smaller than the linear estimate . since the choice of @xmath4 is uncertain in many practical applications , we fit the nonlinear drag force on a circular - orbit perturber using equation ( [ eq : nonlinfit ] ) that does not involve @xmath4 . that the nonlinear drag force in circular orbits is insensitive to the softening radius seems inconsistent with the results of kk09 for linear - trajectory perturbers .
but , it does not . for the latter case
, @xmath4 is in fact a lone length scale relative to which other length scales such as the bondi radius and the distance traveled by the perturber are measured .
it is then difficult to separate the effect of @xmath4 on the drag force from that of @xmath3 because the models depend on them only through the dimensionless parameter @xmath31 .
had we run simulations by using _
dimensional _ quantities in kk09 , we would probably have gotten the results that @xmath191 independent of @xmath4 for models with @xmath192 , similarly to the results of the present paper .
given that @xmath193 , the time - averaged values of @xmath173 , when expressed in terms of @xmath25 in kk09 , somehow picked up the power - law dependence on @xmath3 more clearly than the weak logarithmic dependence on @xmath4 .
the circular orbit breaks the degeneracy between @xmath4 and @xmath3 by introducing another length scale , @xmath24 , allowing to separate the effect on the df force of @xmath4 from that of @xmath3 .
the similarity between equation ( [ eq : linfit ] ) obtained for fixed @xmath15 in the current work and the result of kk09 suggests that the latter should be interpreted as the dependence of @xmath89 on @xmath3 for fixed @xmath4 .
now we compare the results of our idealized models with those from realistic simulations for the orbital decay of smbhs at galaxy centers ( e.g. , @xcite ; see also @xcite and references therein ) .
while these authors showed that smbhs inspiral rapidly in @xmath194 yrs due to the gaseous df force , they also found that the df force in supersonic models is @xmath195 times smaller , and depends on the black hole mass less sensitively , than the analytical predictions of @xcite formula ( e.g. , @xcite ) .
the discrepancies between the numerical and analytical results are most likely due to the nonlinear effect .
for instance , one of the models considered by @xcite has a black hole with mass @xmath196 and softening radius @xmath197 pc moving initially with speed @xmath198 on a circular orbit with @xmath199 pc through a medium with sound speed @xmath200 , corresponding to @xmath201 , @xmath202 , and @xmath117 .
for this model , equation ( [ eq : nonlinfit ] ) compared with equation ( [ eq : f_pt_linear ] ) would give @xmath203 , consistent with the results of @xcite .
since the orbital decay timescale is given by @xmath204 , equation ( [ eq : nonlinfit ] ) predicts @xmath205 for sufficiently massive perturber , which is more consistent with @xmath206 inferred from the results of @xcite than the linear prediction @xmath207 .
the difference between the results of the current paper and @xcite is presumably due to the effects of density stratification , rotation , self - gravity in the background medium , which are not considered in the present work .
another important issue regards the proper choice of the softening length of a point - mass perturber employed in numerical simulations for the gaseous df . as mentioned in
[ sec : intro ] , the limited numerical resolution commonly requires to take @xmath4 a few orders of magnitude larger than the realistic size .
are such large values of @xmath4 acceptable in calculating the orbital decay time accurately ? based on our numerical results ,
the answer is yes , provided @xmath170 ( or @xmath172 ) .
otherwise , a large value of @xmath4 would cause the gravity near the black hole to be reduced significantly .
the resulting density wake would then remain in the linear regime , making the drag force overestimated considerably . for the decay of smbhs , the models considered in @xcite have @xmath208 , well above the required lower limit .
some models with @xmath209 in @xcite and @xcite have @xmath210 , so that their choices of @xmath211 pc are marginally acceptable .
for the evolution of a main - sequence companion with @xmath212 in a common - envelope binary , @xcite took @xmath213 cm , about 10 times larger than the real size . since this corresponds to @xmath214
, our results suggest that the softening radius did not affect the decay time in their numerical simulations . finally
, we remark on the several assumptions made in the current work .
first , we consider a single perturber moving on a circular orbit .
if the resulting drag force ( eq . [ [ eq : nonlinfit ] ] ) is to be applied to the decay of double black holes , one has to take into account the drag force from the wake of its companion located at the other side of the orbit , as well .
@xcite found that for equal - mass perturbers , the ratio of the df force from the companion wake to that from its own induced wake is about @xmath215 for supersonic cases , depending on the mach number , when the perturbers have so low masses that the wakes are in the linear regime .
the nonlinear effect on massive perturbers has yet to be explored .
second , the current study assumes an adiabatic equation of state with index @xmath42 .
it is not well known what the most appropriate value of @xmath216 is in the background medium , but it would be @xmath42 if the gas is fully ionized or @xmath217 for preferentially molecular gas , under the assumption that radiative heating and cooling are unimportant .
the numerical results of @xcite indicate that the orbital decay is more effective for a softer equation of state , suggesting that the df force may depend sensitively on @xmath216 .
third , while we consider a static background medium with uniform density , it is more likely to be stratified in real situations . for a collisionless background
, @xcite found that a density gradient induces an additional drag force in the lateral direction of the perturber motion , amounting to about 10% of the drag in the backward direction .
we also have ignored the rotation , self - gravity , buoyancy , and turbulent motions of the background medium in the present paper .
it is interesting to see what effect each of these physical ingredients makes , which will direct our future research .
the author is grateful to e. c. ostriker and h.kim for helpful discussions and the anonymous referees for stimulating comments .
this work was supported by the national research foundation of korea ( nrf ) grant funded by the korean government ( mest ) , no .
2009 - 0063616 .
simulations were performed by using the supercomputing resource of the korea institute of science and technology information through the grant ksc-2009-s02 - 0008 .
adams , f. c. , ruden , s. p. , & shu , f. h. 1989 , , 347 , 959 barausse , e. 2007 , , 382 , 826 balbus s. a. , soker n. , 1990 , , 357 , 353 begelman , m. c. , blandford , r. d. , & rees , m. j. 1980 , , 287 , 307 binney , j. , & tremaine , s. 2008 , galactic dynamics , 2nd ed .
( princeton : princeton univ .
press ) chambers , j. e. 2009 , annu . rev .
earth planet .
2009 , 37 , 321 chandrasekhar , s. 1943 , , 97 , 255 colpi , m. , & dotti , m. 2009 , to appear on advanced science letters ; arxiv:0906.4339 conroy , c. , & ostriker , j. p. 2008
, , 681 , 151 cuadra , j. , armitage , p. j. , alexander , r. d. , & begelman , m. c. 2009 , , 393 , 1423 dokuchaev , v. p. 1964
, soviet astron . , 8 , 23 dotti , m. , colpi , m. , & haardt , f. 2006 , , 367 , 103 dotti , m. , colpi , m. , haardt , f. , & mayer , l. 2007 , , 379 , 956 edgar , r. 2004 , new astro .
, 48 , 843 el - zant , a. a. , kim , w .- t .
, & kamionkowski , m. 2004 , , 354 , 169 escala , a. , larson , r. b. , coppi , p. s. , & mardones , d. 2004 , , 607 , 765 escala , a. , larson , r. b. , coppi , p. s. , & mardones , d. 2005 , , 630 , 152 ferrarese , l. , & ford , h. c. 2005 , , 116 , 523 just , a. , & kegel , w. h. 1990 , , 232 , 447 just , a. , & pearrubia , j. 2005 , , 431 , 861 kim , h. , & kim , w .- t .
2007 , , 665 , 432 ( kk07 ) kim , h. , & kim , w .- t .
2009 , , 703 , 1278 ( kk09 ) kim , h. , kim , w .- t . , & snchez - salcedo , f. j. 2008 , , 679 , l33 kim , w .- t .
2007 , , 667 , l5 kim , w .- t .
, el - zant , a. a. , & kamionkowski , m. 2005 , , 632 , 157 liepmann , h. , & roshko , a. 1957 , elements of gasdynamics , galcit aeronautical series ( new york : wiley , 1957 ) lubow , s. h. , & ida , s. 2010 , to appear in exoplanets , ed .
s. seager , univ .
arizona press ; arxiv:1004.4137 lufkin , e. a. , balbus , s. a. , & hawley , j. f. 1995 , , 446 , 529 mastrodemos , n. , & morris , m. 1999 , , 523 , 357 maxted , p. f. l. , gnsicke , b. t. , burleigh , m. r. , southworth , j. , marsh , t. r. , napiwotzki , r. , nelemans , g. , & wood , p. l. 2009 , , 400 , 2012 mayer , l. , kazantzidis , s. , madau , p. , colpi , m. , quinn , t. , & wadsley , j. 2007 , science , 316 , 1874 menou , k. , haiman , z. , & narayanan , v. k. 2001 , , 558 , 535 milosavljevi , m. , & merritt , d. 2003 , , 596 , 860 narayan , r. 2000 , , 536 , 663 namouni , f. 2010 , , 401 , 319 nordhaus , j. , & blackman , e. g. 2006 , , 370 , 2004 ostriker , e. c. , shu , f. h. , & adams , f. c. 1992 , , 399 , 192 ostriker , e. c. 1999 , , 513 , 252 rephaeli , y. , & salpeter , e. e. 1980 , , 240 , 20 ricker , p. m. , & taam , r. e. 2008 , , 672 , l41 ruderman , m. a. , & spiegel , e. a. 1971 , , 165 , 1 ruffert , m. 1993 , , 280 , 141 sandquist , e. l. , taam , r. e. , chen , x. , bodenheimer , p. , & burkert , a. 1998 , , 500 , 909 snchez - salcedo , f. j. , & brandenburg , a. 2001 , , 322 , 67 shu , f. h. 1992 , the physics of astrophysics .
gas dynamics ( mill valley : univ .
science books ) stone , j. m. , & norman , m. l. 1992 , , 80 , 753 taam , r. e. , sandquist , e. l. 2000 , , 38 , 113 tanaka , h. , takeuchi , t. , & ward , w. r. 2002 , , 565 , 1257 villaver , e. , & livio , m. 2009 , , 709 , l81 ward , w. r. 1997 , icarus , 126 , 261 | we use three - dimensional hydrodynamic simulations to investigate the nonlinear gravitational responses of gas to , and the resulting drag forces on , very massive perturbers moving on circular orbits .
this work extends our previous studies that explored the cases of low - mass perturbers on circular orbits and massive perturbers on straight - line trajectories .
the background medium is assumed to be non - rotating , adiabatic with index @xmath0 , and uniform with density @xmath1 and sound speed @xmath2 .
we model the gravitating perturber using a plummer sphere with mass @xmath3 and softening radius @xmath4 in a uniform circular motion at speed @xmath5 and orbital radius @xmath6 , and run various models with differing @xmath7 , @xmath8 , and @xmath9 . a quasi - steady density wake of a supersonic model consists of a hydrostatic envelope surrounding the perturber , an upstream bow shock , and a trailing low - density region .
the continuous change in the direction of the perturber motion makes the detached shock distance reduced compared to the linear - trajectory cases , while the orbit - averaged gravity of the perturber gathers the gas toward the center of the orbit , modifying the background preshock density to @xmath10 depending weakly on @xmath11 . for sufficiently massive perturbers ,
the presence of a hydrostatic envelope makes the drag force smaller than the prediction of the linear perturbation theory , resulting in @xmath12 for @xmath13 ; the drag force for low - mass perturbers with @xmath14 agrees well with the linear prediction .
the nonlinear drag force becomes independent of @xmath15 as long as @xmath16 , which places an upper limit on the perturber size for accurate evaluation of the drag force in numerical simulations . |
the all sky monitor ( asm ) for x - ray observations has a long history ( holt and priedhorsky 1987 ) .
the asm on ariel 5 ( a british satellite ) was the first dedicated pioneer payload to observe x - ray novae and transients ( holt 1976 ) .
the ariel 5 asm , which has two sets of one - dimensional scanning pinhole cameras , discovered several novae and transient x - ray objects .
thereafter , the terms `` x - ray nova '' and `` x - ray transient '' have become well - known in x - ray astronomy .
x - ray instruments with a wide field of view ( fov ) as well as asm can not only detect x - ray novae and transients , but also monitor long - term x - ray variabilities of x - ray sources ( priedhorsky and holt 1987 ) .
although vela a & b satellites with a wide fov discovered the first gamma - ray burst ( grb ) before the ariel - asm ( klebesadel et al .
1973 ) , those detectors were not much suitable for monitoring grbs as well as x - ray novae and transients .
subsequently , dedicated grb monitors confirmed that a considerable number of mysterious grbs are produced in the universe .
grb monitors with a wide fov have advanced greatly since a wide field camera on beppo - sax discovered a grb afterglow ( costa et al .
1997 ) . since then , special grb satellites such as hete-2 ( ricker et al . 2002 ; shirasaki et al . 2004 ) and swift ( gehrels et al .
2004 ) have been realized .
the asm , however , advanced gradually as a supplemental payload .
ariel - asm with pin - hole cameras operated successfully for seven years with the main payload which made spectrum observations of bright x - ray sources ( holt 1976 ) . in 1987
the asm on the japanese satellite ginga ( tsunemi et al .
1989 ) succeeded the ariel - asm .
ginga - asm was able to observe the spectra of x - ray novae as it scanned the sky from 60 degrees through 360 degrees once every day with multi - slat collimators .
ginga - asm greatly advanced the science of black hole binaries with the discovery of nova - like black hole binaries ( tsunemi et al .
1989 ; kitamoto et al .
1992 ; kitamoto et al .
ginga - asm was operated successfully for 4.5 years , and some of transients discovered by asm were observed in detail by the main large area counters on ginga ( turner et al .
1989 ) . since 1996 , rxte - asm ( a nasa satellite ) has monitored known x - ray sources in addition to x - ray novae and transients ( levine et al .
systematic long - term data of galactic variable sources provide quasi - periodic properties of the accretion disc ( zdziarski et al . 2007a ; 2007b ) .
rxte - asm has provided much useful data for x - ray variable sources for 13 years , but the detection limit is around 10 mcrab . therefore , the main targets of rxte - asm were the galactic x - ray sources of which monitoring was suitable for performing detailed observations of these targets with an rxte prime instrument , a large - area proportional counter array ( remillard and mcclintock 2006 ) .
now we have long term light curves for bright x - ray sources from the above three asms .
some sources revealed periodic or quasi - periodic component of long time scale with combined analysis over the last 30 years ( paul et al .
future asms taking over the three asms can continue to investigate further long - term behaviour of x - ray sources .
although the above three asms have been in operation , they can not adequately monitor agns because of their low sensitivity .
thus far , galactic x - ray novae and transients data have been accumulated although the information is not completely clear .
no long term variability of agn has not been observed with sufficient data .
therefore , the most important role for future asms is to better the detection limit for agn monitoring . considering this historical situation ,
maxi was proposed as a payload of iss ( matsuoka et al . 1997 ) .
it will be one of the asms responsible for agn monitoring .
a slit hole camera with a large detection area meets this requirement .
since maxi has a composite structure of slit holes and slat collimators without a mirror system , the angular resolution is not good ( i.e. , the fwhm of slat collimators is 1.5 degrees ) , but the localization accuracy can be 0.1 degree for the sources of enough statistic ( e.g. , bright sources and sources with enough counts accumulated for optimum time ) .
a mirror - type asm , such as the lobster - eye asm , is promising for the future although its energy band is limited to soft x - rays ( priedhorsky et al . 1996 ) .
the maxi to be attached to the japanese experiment module ( jem ; kibo ) on the iss is ready to be launched . in this paper
we will present an overview of maxi and the new scientific contributions expected of it .
the first part of this paper will discuss the astronomical science expected of maxi .
the second part will describe the maxi instrumentations , and observational simulation .
the maxi mission enables the investigation of asms and surveying .
maxi will alert astronomers of grbs , x - ray novae , and flare - up increases of x - ray sources if they occur .
long - term data of x - ray sources will enable us to determine a special time scale of variability , e.g. , long - term periodic or quasi - periodic motions of x - ray sources .
maxi can promote multi - wavelength observation in collaboration with other space and ground observatories such as x - ray , infrared , and optical satellites , radio , and optical ground observatories .
maxi systematic observations of the variable activity of black hole binaries and agns are used to investigate how and where they produce their variable activities .
maxi provides unbiased x - ray source catalogues over all the sky .
monthly or biannual x - ray catalogues could contribute to the long - term study of variable behavior of agn for the first time .
a catalogue accumulated for two years could provide all - sky agns corresponding to a moderately deep survey to the entire sky .
this unbiased agn catalogue will far surpass the investigation of the distribution and evolution of agns for the heao-1 a2 catalogue which has been utilized for long time ( piccinotti et al .
maxi is also able to make an all sky x - ray map with soft x - rays and medium energy x - rays .
the soft x - ray map provides line features such as oxygen x - ray lines , which are useful in researching geo - coronal recombination lines ( fujimoto et al .
2007 ) as well as the evolution of hot gas in the galaxy ( mccammon and sanders 1990 ; tanaka and bleeker 1977 ) .
current swift ( gehrels et al . 2009
; burrow 2009 ) and integral ( ubertini et al . 2009 ) satellites with wide fovs of hard x - ray detectors have provided x - ray source catalogues including agns in addition to a considerable number of transients . since these results are complementary to those of maxi , it could be promising to science that these three asm missions will operate simultaneously .
furthermore , the recently developed gamma - ray large - area space telescope , the fermi gamma - ray space telescope ( abdo et al . 2008 ; thompson et al . 2009 ) , is a gamma - ray asm with an energy band complementary to that of maxi .
all - sky images with both asms may reveal new information .
more than 90% of the black hole candidates are x - ray transients or novae ( mcclintock and remillard 2007 ) . in the last 20 years
, 29 have been discovered , mostly with rxte - asm and partly with ginga - asm , while about 30% of them have been serendipitously discovered by pointing observatories ( negoro 2009 ) .
thus far , observed x - ray novae are located near the galactic center and near the solar system ; i.e. , their distances are within about 6 kpc .
distant x - ray novae appear weak , but maxi could detect these weak sources , since it can observe x - ray novae from distant regions about three times as far .
thus , it is expected that the nova discovery rate may increase by one order of magnitude .
x - ray light curves with energy spectra of x - ray novae have provided instability and phenomena deviating from the standard accretion disc model ( tanaka and shibazaki 1996 ) .
data of both low and high luminosities have resulted in the creation of a new theory concerning accretion discs ( mineshige et al .
1994 ) , but these samples are not enough .
we also expect that maxi could observe x - rays from classical novae .
multi - wavelength observations of classical novae could provide useful information about the bursting mechanism of classical novae . although classical novae are faint and soft in x - ray band , the durations are of months to years ( mukai , orio & della valle 2008 )
. thus maxi could detect some of them .
swift has discovered many grbs and has consequently enabled us to make rapid follow - up observations of them .
nevertheless , the energy band of the swift detector is 15 to 200 kev ( gehrels et al . 2002 ) .
maxi can cover the soft x - ray band of grbs , and this information is important to the understanding of x - ray rich grbs and x - ray flashes .
it is expected that the population of the x - ray rich grb is comparable to that of the grb with an ordinary energy band ( sakamoto et al . 2005 ) .
the instant fov of maxi is not large , maxi will be able to detect 3.5 prompt emissions and 2.5 x - ray afterglows of grb per year ( suzuki et al .
2009 ) .
maxi is also able to observe transient x - ray binary pulsars with high orbital eccentricity and transient low mass x - ray binaries with neutron stars .
furthermore , anomalous x - ray pulsars ( axps ) are still enigmatic objects that appear as soft gamma repeaters ( sgrs ) with enormous flux during a short time ( nakagawa et al .
2009 ; morii 2009 ) .
considering the recent swift discovery of a new sgr ( barthelmy et al .
2008 ; enoto et al .
2009 ) , we could even expect maxi to detect additional sgrs with weaker flux .
thus , maxi may reveal new frontier of axps ( i.e. , magnetars ) by detecting numerous new magnetar candidates ( nakagawa et al .
recently , special interest concerning long type i x - ray bursts has arisen with rare events of carbon - fueled super bursts and helium - fueled intermediately long bursts .
these bursts are promising in the investigation of the deeper neutron star envelope ( keek and int zand 2008 ; keek et al 2008 ) .
maxi may discover such rare occurrences of x - ray bursts .
most x - ray sources with compact objects are variable , due to the instability of the accretion disc and other reasons .
periodic or quasi - periodic long - term variability is sometimes created from interaction between the accretion disc and the binary system .
if there were a third object in the x - ray sources , we could expect some other periodicity in addition to binary period ( zdziarski 2007a )
. the third body might be useful for inspecting the accretion disc .
however , no one has discovered such a triple - body system in the x - ray sources
. a composite time scale of variability would help generate this knowledge and/or may create new knowledge . an unusual x - ray transient from the galactic center region ( smith et al . 1998 ) is progressing in a new class of recurrent and fast x - ray transient sources ( i.e. , sfxts ) .
these transients sometimes occur in high - mass x - ray binaries associated with super - giant companions .
it is promising for further investigation that maxi in addition to integral and swift monitors these objects with a wide fov(ebisawa 2009 ; ubertini et al .
generally , a time scale of agns variability is longer than that of x - ray binaries .
the time scale of agns is useful in understanding a complicated structure around the agns .
if some agn had formed a super - massive binary black hole system , we could expect some periodic variation over a year s time ( hayasaki et al .
2008 ) .
the knowledge of bursts , transients , and variable objects has progressed considerably with multi - wavelength observations .
research of grb afterglows has advanced greatly as a result of rapid follow - up observations in x - ray , optical , and radio bands ( gehrels et al .
2006 ) , but there are still just a few samples of short grbs and x - ray rich grbs .
coordinate observations of blazars in radio , infra - red , optical , x - ray , and ultra - high - energy ( tev ) gamma - rays have confirmed a synchrotron self - compton ( ssc ) model for highly variable periods ( kubo et al .
1998 ; kataoka et al . 1999 ; takahashi et al .
however , the emission mechanism of other agns as well as blazars has not yet been completely understood because of complicated correlations among multi - wavelength observations to help investigate this mechanism ( maoz et al .
tev gamma - ray observations by erenkov telescopes have progressed remarkably since tev gamma - rays from some blazars were detected ( petry et al .
1996 ; aharonian et al .
gev gamma - ray observations by fermi - glast are now available as an asm ( thompson et al .
maxi can promote further simultaneous multi - wavelength observations of galactic active objects as well as agns with gamma - rays ( gev & tev ) to radio bands ( fender 2009 ; madejsky et al .
. nominal beam size ( i.e. , angular resolution by fwhm ) of maxi is [email protected] deg@xmath5 .
it is estimated that a confusion limit of x - ray sources is 5@xmath110@xmath6 erg @xmath3 sec@xmath4 from recent log n - log s plot of x - ray sources ( ueda et al .
2003 ) ; i.e. , 0.2 mcrab in the energy band of 2 to 20 kev .
thus , we can set a detection limit of 0.2 mcrab as an ideal goal , although the period of its achievement depends on intrinsic background and systematic error . to estimate observation time we conducted a realistic maxi observational simulation .
maxi achieved a detection limit of 0.2 mcrab with a two - year observation ( ueda et al .
other simulations suggest that it is possible to detect 20 mcrab for one orbit ( 90 min . ) , 4.5 mcrab for one day , and 2 mcrab for one week ( hiroi et al .
2009 ; sugizaki et al .
some regions in the sky are covered by a bright region from the sun and are sometimes affected by the south atlantic anomaly ( saa ) . therefore , detection sensitivities are estimated under the best condition .
the detectability of maxi is not uniform in the entire sky , but slightly depends on the direction .
the maxi simulation indicates that we can obtain 30 - 40 agns every week ( ueda et al .
this sample is comparable to the number from heao 1-a2 ( piccinotti et al .
we are also able to estimate about 1000 agns with two - year observations .
maxi observes those agns in a harder energy band than observed by rosat .
thus , maxi could give an unbiased population ratio of type i and type ii agns .
furthermore , if weekly or monthly catalogues are created , we can discover the intensity variability for a considerable number of agns .
it is possible to detect a flare - up of blazars and then follow their light curves .
the time sequence of x - ray unbiased catalogues is very useful in researching the evolution of agns as well as their long - term variability . here
the next comment is noted .
although much deeper surveys of agns in the 2 - 10 kev band with asca ( e.g. , ueda et al . 1999 ) and chandra / xmm - newton ( brandt & hasinger 2005 and references therein ) have presented log n log s plots for researching the evolution of agns ,
they are limited to a small portion of the sky , and hence can not constrain that of bright agns with small surface densities .
the problem of the diffuse cosmic x - ray background ( cxb ) has a long history .
a recent deep survey of x - ray sources in a medium energy band of 2 to 10 kev ( ueda et al . 2003 ) suggests that the cxb may be due to the superposition of agns , as proposed theoretically ( morisawa et al .
it is still unknown how some kinds of agns are distributed in the universe .
a global cxb distribution is compared with that of optical agns and/or an infrared map .
thus we can investigate the evolution or emission mechanism of the all - sky agn distribution .
if we can obtain some difference of distribution in different energy bands , we can investigate the distribution of type i and type ii agns .
the diffuse emission of soft x - rays less than 1 kev is attributed to geo - coronal gas as well as hot bubbles from supernova remnants ( tanaka and bleeker 1977 ; mccammon and sanders 1990 ) .
rosat obtained a precise all sky map with a broad energy band of soft x - rays .
maxi can also obtain all - sky maps of soft x - rays but with better energy resolution ( e.g. , resolving oxygen k - line , and neon k - line ) .
recent suzaku observations revealed a strong contribution of recombination k - lines of oxygen and carbon in the geo - coronal region ( fujimoto et al .
maxi is able to observe the oxygen line with seasonal variation and solar activity .
therefore , maxi observations make it possible to discriminate between the contribution of geo - corona and that of supernova remnants .
thus , we could investigate geo - coronal science as well as element - evolution of supernova remnants from soft x - ray line observations .
the large - scale space station ( ss ) project started in 1985 with the collaboration of the usa , japan , canada , and the european space agency ( esa ) .
planning of the jem was also initiated at that time . in 1995 , russia began to take part in the ss project .
the project was subsequently reduced to its present scale and then renamed the international space station ( iss ) .
jem planning and designing have continued , but actual construction and basic tests of the engineering model began at that time .
jem consists of a pressurized module for micro - gravity experiments and an exposed facility ( ef ) .
lcc & gsc : gas slit camera & ssc : solid - state slit camera + + x - ray detector & & + x - ray energy range & 230 kev & 0.512 kev + total detection area & 5350 @xmath7 & 200 @xmath7 + energy resolution & 18 % ( 5.9 kev ) & @xmath8 150 ev ( 5.9 kev ) + field of view & 1.5 @xmath1 160 degrees & 1.5 @xmath1 90 degrees + slit
area for camera unit & 20.1 @xmath7 & 1.35 @xmath7 + detector position resolution & 1 mm & 0.025 mm ( pixel size ) + localization accuracy & 0.1 deg & 0.1 deg + absolute time resolution & 0.1 msec(minimum ) & 5.8 sec(nominal ) + weight & 160 kg & 11 kg + ( 85mm,)fig_maxi_overview_ueno_20090130a.eps iss rotates synchronously in its orbit so that one side always points towards the center of the earth and the opposite side views the sky .
therefore , the sky side of jem - ef surveys a great circle every iss orbit .
themes for jem - ef include the space science payload for astrophysics and earth observations , and space technology experiments such as robotics .
however , jem - ef does not provide a perfectly stable platform because of unknown factors of some attitude fluctuation .
payload size and weight are limited to the capacity of the jem - ef .
a payload can be suitable for survey observation , but not for pointing .
although a considerably wide fov is available , the iss structure and solar paddle partially block the view of jem - ef .
since the iss carries various experimental instruments and payloads , each instrument and payload has to accommodate many interfaces .
resources such as communication and power for each experiment or payload are also limited .
considering these problems maxi was proposed and finally accepted in 1997 by the national development space agency of japan ( nasda ) , now known as the japan aerospace exploration agency ( jaxa ) .
thus , maxi is the first astronomical payload for jem - ef on the iss . although the launch was scheduled for 2003 , the space shuttle , columbia accident , and the iss construction delay resulted in the postponement of the maxi launch by several years .
maxi science instruments consist of two types of x - ray cameras , the gas slit camera ( gsc ) and the solid - state slit camera ( ssc ) . these instruments and the support instruments on the maxi payload
are shown in figure 1 , and their characteristics are listed in table 1 .
the support instruments consist of a visual star camera ( vsc ) , a ring laser gyroscope ( rlg ) , a global positioning system ( gps ) and a loop heat pipe and radiation system ( lhprs ) .
the vsc and the rlg determine the directions of the gsc and the ssc as precisely as a few arc - seconds every second .
the gps attaches the absolute time as precisely as 0.1 msec to gsc photon data .
the lhprs is used for heat transportation and heat radiation from thermo - electric coolers ( peltier elements ) to cool the ccd .
the main role of the gsc is to perform as the x - ray asm , which has the best detectability than previous asms have for a time scale longer than hours .
all sky x - ray images obtained by gsc are also useful for new x - ray variable catalogues . on the other hand
the main mission of ssc is to make all sky x - ray maps for extended sources with better energy resolution than rosat maps , although the detection area and live time for discrete sources are less by 1/20 than those of gsc .
we also expect to detect the transients in soft x - ray band although the detectability of transients with ssc is also poorer by 1/20 than that of gsc .
maxi will be carried by the space shuttle , endeavour , along with the jem ( or kibo)- ef from kennedy space center in the middle of 2009 .
maxi will finally be mounted on jem - ef within two weeks of endeavour launch . at that time
all of kibo s modules will have been installed on iss .
after the basic arrangement of the structure and infrastructure of jem , maxi will conduct performance tests for about three months .
maxi has a nominal lifetime of two years , but the expected goal is five or more years to achieve long - term monitoring .
maxi data will be down linked through the low rate data link ( lrdl ; mil1553b ) , and the medium rate data link ( mrdl ; ethernet ) .
maxi is operated through the operation control system ( ocs ) at tsukuba space center ( tksc ) , jaxa .
maxi data are not only processed and analyzed at tksc , but they are also transferred to the institute of physical and chemical research ( riken ) maxi data facility .
general users of maxi can request the scientific data from riken ( a main port ) as well as from jaxa ( a sub - port ) whenever they desire . to search where x - ray novae , transients , or flaring phenomena appear suddenly in the sky
we will conduct automatic data analysis at tksc by using down linked data from lrdl .
if such a source is discovered , we will report this event to astronomers and dedicated users worldwide via the internet from tksc .
this nova alert system has been developed to achieve automatic alert ( negoro et al .
the maxi team will also maintain archival data at riken using the data from lrdl and mrdl , such as all sky x - ray images , x - ray light curves , and spectra of dedicated x - ray sources . in principle
all astronomical data of maxi will be available for public distribution ( kohama et al . 2009 ) .
( 45mm,)slit_camera_principle.eps ( 60mm,)gscpsf1998_s3.eps a specific object is observed by maxi with a slit camera for a limited time with every iss orbit .
the x - ray detector of the camera is sensitive to a one dimensional image through the slit , where the wide fov through the slit spans the sky perpendicular to the iss moving direction as shown in figure 2 .
the scanning image of an object is obtained with the triangular response of a slat collimator according to iss movement .
an intersection of the slit image and the triangular response image corresponds to a source location in the sky as a point spread function ( psf ) .
this is shown in figure 3 .
objects located along a great circle stay for 45 sec in the fov of maxi cameras , where the time of stay is the shortest in the maxi normal direction ( for a great circle ) . for objects in the slanted field from the great circle ,
they achieve slightly longer observation .
any target will come repeatedly in each field of the horizontal and zenithal cameras with every iss orbital period .
in this situation maxi can intermittently monitor a short time scale variability for bright x - ray sources such as x - ray pulsars and low mass x - ray binaries .
it can monitor their variability of weak sources on a time scale of 90 min or longer if we integrate the data .
maxi has two types of x - ray cameras : the gsc and the ssc which are described in the following sub - section .
( 85mm,)gsczu1.ps
gsc is the main x - ray camera and consists of six units of a conventional slit camera as shown in figure 1 and table 1 ( mihara et al . 2009 ) .
the gsc unit consists of two one - dimensional proportional counters ( produced by oxford instruments co. in finland ) , and slit & slat collimators as shown in figure 4 .
thus , twelve proportional counters have a 5350 cm@xmath5 detection area in total .
each counter has six cells of resistive carbon wires that are guarded by a veto - detector region in the bottom and on both sides ( mihara et al .
2001 ) , while the carbon wire divides the charge from the signal into both terminals for one - dimensional determination .
the x - ray detection efficiency of the proportional counter ( xe 1.4 atmospheres with 1 @xmath10 co@xmath11 ) is plotted against x - ray energy in figure 5 .
observational energy band of gsc is set to be 2 - 30 kev in standard operation mode where the detection efficiency for x - rays in this band is above 10 @xmath10 as seen in figure 5 .
the observation above xe k - edge of 34.6 kev is possible as a special mode by adjusting amplifier gain and high voltage .
although a response function above xe k - edge is complicated due to escape peak , the test for this function has been done using synchrotron x - ray beams .
the slat collimator response is 3.5 degrees in bottom - to - bottom , and a slit image corresponds to 1.5 degrees covering a wide field between @xmath12 40 degrees and @xmath13 40 degrees .
the proportional counters detect incident x - ray photons from vertical to @xmath14 40 degree slant direction
. two fovs of gsc - h and gsc - z are placed in horizontal ( forward ) and zenithal directions to compensate when the sky is unobservable to one fov or the other ( due to high radiation background , such as saa ) .
the fov of one camera is 80 degrees .
two camera units can cover 160 degrees , but the fov of each central camera of gsc - h and gsc - z overlaps with each half of the fov of both side cameras to even the exposures as shown in figure 6 which shows the product of effective area and dwell time according to one orbit scan .
the exposure time per orbit depends on the direction of a star by a factor of 1/cos@xmath15 , where @xmath15 is the angle of the star slanted from scanning a great circle .
both edges of 10 degrees in this figure are omitted , due to the shadow of the iss structure , where the scanning loss is 1.5@xmath10 .
the forward fov is tilted up by 6 degrees so as not to observe the earth even when the iss attitude changes .
thus , the fov of gsc - h can scan the sky with iss rotation without earth occultation . by making the observation time longer than 90 min
, we made the instrument simpler , and then optimized it to search in the agn discovery space for the longer - term agn variability ( @xmath16 1.5 hour ) .
considering this directional performance , we conducted various laboratory tests of all proportional counters . here
we explain some of the results .
the position resolution was tested at every 2 mm segment of two - dimension on the incident window for all proportional counters .
the data were taken using 0.1 mm pencil x - ray beams with x - ray energies of 4.6 , 8.0 , and 17.4 kev .
the total length of the carbon anode wire is 32 cm ; thus , the total resistance is 31 to 37 k@xmath17 .
the difference of resistance depends on the slight difference of wire diameters .
the energy response also differs for different anode wires .
we took comprehensive data useful for energy response and position response ( mihara et al .
2001 ; isobe et al . 2004 ) .
we also took the slat collimator response data using x - ray pencil beams ( morii et al .
the data - base , including all these data , is regarded as the response function at the time of data analysis . in actual data analysis
we will use each response function for the two dimensional surface of all carbon wires .
pulse height response depends on x - ray energy , where hard x - rays suffer the effect of anomalous gas amplification ( mihara et al .
2001 ) .
the simultaneous background for a localized source is measured in two regions separated from the source on the same detector in the fov direction .
another background is measured in two regions just before and after observing the source on the scanning path , where the two regions are separated from the source . in this case , both measurement locations of the background on the detector are the same as the measurement location on the detector of the source . the background is employed if there are no appreciable sources in these directions .
in addition , intrinsic background is necessary to obtain the cosmic diffuse background .
the intrinsic background is gradually changing on orbit and for time even in the same direction ( hayashida et al . 1989 ) .
there is the small portion covered by a window frame of each detector in which cosmic diffuse x - rays as well as source x - rays are never irradiated . a small portion of the north pole direction is also shadowed by iss structure .
although these intrinsic background count rates are not enough to measure for instant time , the data accumulated for days or more could be used to make a reasonable model of intrinsic background as obtained for ginga proportional counters ( hayashida et al .
1989 ) .
( 85mm,)fig7_mono.eps each unit of ssc - h / z consists of 16 ccds , where each ccd acts as a one dimensional position sensitive detector for the slat collimator and the slit hole similar to the gsc system presented in figure 1 and table 1 ( tomida et al . 2009 ) .
the ssc - h camera is tilted up 16 degrees so as not to view the upper atmosphere .
the response of the slat collimator scanning direction is 3 degrees for bottom - to - bottom , while a slit image corresponds to 1.5 degrees covering a wide fov of 90 degrees as shown in figure 7 .
one orbit scan of the ssc corresponds to 90 degrees x 360 degrees .
thus , it takes ssc 70 days to scan the entire sky depending on the precession of the iss orbital plane , except for the bright region around the sun .
thus , ssc will acquire actual all sky images every half year .
the x - ray ccd chip of ssc is produced by hamamatsu photonics k.k .. the ccd depletion layer is about 70 @xmath18 m , of which the x - ray detection efficiency is indicated in figure 8 . the observation energy band of ssc is set to be 0.5 - 12 kev in standard mode , but observations above 12 kev are possible in a special mode .
however , since ssc is characterized by a low energy band , the energy band may be slightly shifted to lower energy than 0.5 kev , depending on the temperature for thermal noise .
the ccd with 1024@xmath11024 pixels , and a pixel size of 24 @xmath18 m @xmath1 24 @xmath18 m is a two - dimensional array , but ssc requires only one - dimensional position information .
hence , multiple rows are summed in a serial register at the bottom of the imaging region , and the summed charges in the serial register are transferred to a read - out node .
eight , 16 , 32 or 64 summed rows can be selected in normal observation by commands .
the larger this number , the better the time resolution , and the better the angular resolution in the x - ray sky map .
the ccd is sensitive to particle radiations , but an irradiation test of the simulated radiation belt fluence suggests that the ccd can survive for the expected three - year mission ( miyata et al .
furthermore , it is possible to inject charges in the ccd before it becomes damaged due to irradiation ( miyata et al .
charge injection effectively restores the degraded performance of x - ray ccds ( tomida et al . 1997 ) . to achieve better energy resolution with a ccd , all ccds
are cooled to -60 degrees c using thermo - electric coolers ( peltier devices ) and the lhprs .
the maximum power of the thermo - electric cooler is 1 w / ccd .
the lhprs can automatically transfer heat from the peltier device and emit heat from the radiation panel .
performance tests in the laboratory have been conducted with satisfaction ( miyata et al .
2002 ; katayama et al .
the nominal energy resolution width of x - ray spectra is 150 ev for mn k x - rays ( 5.9 kev ) .
since the iss has a huge structure of @xmath0108 m and @xmath0420 tons , the attitude determined in a certain part of the iss may be slightly different from that in a distant part , due to shaking of the structure .
maxi is designed to determine a target location as precisely as less than 0.1 degree .
thus , it is necessary to determine the attitude of the maxi coordinate by less than 0.1 degrees every time . for this purpose
, maxi itself has a vsc and an rlg .
the vsc can observe three or more stars , and it can determine a maxi coordinate by 0.1 arc minute .
when the vsc is not available for measurement due to solar radiation , the rlg will extrapolate the attitude from a certain result of the vsc to the following result .
this attitude determination is performed automatically by onboard software including a kalman filter on the maxi data processor ( ueno et al .
2009 ; horike et al . 2009 ) .
the high voltage of a proportional counter of gsc will be reduced to 0 volt when solar radiation or saa radiation is extremely strong .
the command signal for this reduction will be issued by working of the radiation belt monitor ( rbm ) .
this high voltage reduction is also useful for setting program commands in advance because of the possible prediction of solar position and saa location .
the time precision of photon acquisition from gsc is 0.1 msec by referring to gps signal .
this precise absolute time is used for milli - second pulsars and burst acquisition analysis .
thus , only one gps is installed in maxi for precise time reference without unknown lag .
however , the time resolution of ssc is 5.8 sec in the normal observation mode .
a realistic simulation has been conducted with the recently developed maxi simulator , which generates fully simulated data of maxi instruments on the iss ( eguchi et al .
the simulator takes into account various conditions on the iss , i.e. , the occultation of the sky with solar panels , particle and intrinsic background , and response function of x - ray cameras .
the attitude data and absolute time are attached to each event data on the ground .
thus , we can plot each event with energy information in a certain direction in the sky .
integrated events from a certain direction for a certain time correspond to intensity , including background from the direction .
here we demonstrate some of the simulated results ( hiroi et al . 2009 ; sugizaki et al . 2009 ) .
( 50mm,)1orbit20mcrabsrc - contour - small.eps ( 50mm,)1week1mcrabsrc - contour - small.eps + ( 50mm,)1orbit20mcrabsrc - fit - small.eps ( 50mm,)1week1mcrabsrc - fit - small.eps first , we investigate the detectability of the gsc .
figure 9 illustrates one - orbit and one - week observations .
the results indicate that we can detect a 20 mcrab source in one - orbit at a confidence level of 5-sigma , and a 2 mcrab source in one - week at a level of 5-sigma .
these simulations are estimated for the energy band of 2 - 30 kev of the crab nebula spectrum .
this simulation is ideal , because the particle and intrinsic background of 10 counts sec@xmath4counter@xmath4 is steady , except for statistical fluctuation during the observation time , and this background is subtracted with statistical fluctuation from integrated events of the source . the spectrum of an intrinsic background is referred to the laboratory test result . in reality , we must consider the changeable background on the iss orbit and systematic errors for attitude determination and response functions .
the final detectability could be 0.2 mcrab for a two - year observation , which means a source confusion limit for the angular resolution of 1.5 degrees .
( 130mm,)gsc1orbit_raw_hsv.eps ( 130mm,)gsc30day_expcor_hsv.eps here it is noted again that the above 5-sigma detectability is under the best condition for known source .
the detectability depends on the intrinsic background or trapped particle background .
either of zenithal and horizontal cameras can observe most directional sources even if the iss passes through the saa .
this makes the detectability uniform for most direction except the direction around the sun if observations are repeated by orbits .
in fact , a deviation from uniformity of the detectability of one orbit becomes worse by about 65 @xmath10 for the sources observed on the saa path , while that of one day becomes uniform within 10 - 20 @xmath10 for any direction except the solar region .
we evaluated another detectability excluding observation in the region less than the cut - off rigidity of 8 ( e.g. , aurora region ) .
a situation of this detectability is not so much different from that of the aforementioned detectability .
lastly , it is noted that a detectability for unknown source is not significantly different from that for known sources .
one - sigma location error for unknown 5-sigma sources is about 0.2 degrees for both one day and one week observations .
second , we perform a simulation to derive the energy spectra of a source of 1 crab for a one - orbit and a one - day observation as shown in figure 10 .
this simulation is also ideal , because the background is steady only with statistical fluctuation and ideal determination of the attitude without an unknown systematic error is assumed .
nevertheless , it is concluded that reasonable spectra from a considerable number of sources are obtained by maxi .
another important goal of maxi is to make x - ray source catalogues for every period of concern .
we can make light curves with the time bin for various x - ray sources .
the light curves of most galactic x - ray sources are obtained with time bins of one orbit or one day . if the time bin is one month , it is possible to make light curves of a considerable number of agns brighter than 1 mcrab as shown on the lower panel in figure 11 .
it is not easy at a glance to discriminate between the weak sources and the cxb , but 1 - 2 mcrab sources are significantly detected as simulated in figure 9 ( b ) . if we integrate all event data from the sky for two years , we can create an x - ray source catalogue brighter than 0.2 mcrab at a confidence level of 5-sigma . in general
, since the time scale of variability of agns is longer than that of galactic sources , catalogues with time bins of one month and one year will provide new information about long - term variable sources such as the super massive binary black hole ( bbh ) .
actually , an all - sky image for every six - month observation is compatible with the result of heao-1 a2 .
maxi is the first payload to be attached to jem - ef ( kibo - ef ) of the iss .
it consists of two kinds of x - ray cameras : the gas slit camera ( gsc ) , which has one - dimensional position sensitive gas - proportional counters with a 2 to 30 kev x - rays energy band , and the solid - state slit camera ( ssc ) , which has x - ray ccds with a 0.5 to 12 kev x - rays energy band .
maxi provides an all - sky x - ray image acquisition capability under the best conditions for every iss orbit .
if maxi scans the sky for one week , it could make 2 mcrab x - ray all sky image excluding the bright region around the sun .
thus , maxi not only rapidly informs astronomers worldwide about x - ray novae and transients if they occur , but also observes long - term variability of galactic and extra - galactic x - ray sources . at the same time , maxi provides an x - ray source catalogues accumulated for various periods such as six months , one and more years .
all maxi instruments are ready for launch .
data processing and analysis software , including an alert system on the ground , are being developed by the mission team .
maxi will be attached to jem(kibo)-ef in mid-2009 .
anyone can participate in follow - up or multi - wavelength observations of maxi objects with ground optical and radio telescopes , and at space observatories .
please contact the present authors for more information and further collaboration .
finally , please refer to the proceedings of astrophysics with all - sky x - ray observations
3rd international maxi workshop ,
riken , wako , japan , june 10 - 12 , 2008 , ed .
, n.kawai , t.mihara , m.kohama and m.suzuki ( http://cosmic.riken.jp/maxi/astrows/ ) for further information on maxi instrumentations and preliminary guideline for maxi data .
+ maxi fabrication and testing are also supported by the staffs at the jem project office and the space utilization office .
we thank them for their valuable guidance and cooperation in maximizing the scientific output from maxi .
we also thank the former colleagues who developed and supported this long - term project of maxi , although these many colleagues ( staffs and graduate students ) have not been included as co - authors .
we also thank the following companies and institutions for developing maxi sub - systems : nec co. for the maxi bus system , meisei electric co. for the mission instruments , oxford instrum .
co. and japan i.t.s .
co. for the gas proportional counters , hamamatsu photonics k.k .
for the ccds , atk space ( former co. : swales aerospace ) for the lhprs , dtu ( technical university of denmark ) for the optical star sensor , the institute of aerospace technology in jaxa for the gps , and other cooperating companies and institutions . | the monitor of all sky x - ray image ( maxi ) mission is the first astronomical payload to be installed on the japanese experiment module - exposed facility ( jem - ef or kibo - ef ) on the international space station ( iss ) .
it is scheduled for launch in the middle of 2009 to monitor all - sky x - ray objects on every iss orbit .
it will be more powerful than any previous x - ray all sky monitor ( asm ) payloads , being able to monitor hundreds of active galactic nuclei ( agns ) .
a realistic simulation under optimal observation conditions suggests that maxi will provide all sky images of x - ray sources of @xmath020 mcrab ( @xmath07@xmath110@xmath2 ergs @xmath3 sec@xmath4 in the energy band of 2 - 30 kev ) from observation on one iss orbit ( 90 min ) , @xmath04.5 mcrab for one day , and @xmath0 2 mcrab for one week .
the final detectability of maxi could be @xmath00.2 mcrab for two years , which is comparable to the source confusion limit of the maxi field of view ( fov ) .
the maxi objectives are ( 1 ) to alert the community to x - ray novae and transient x - ray sources , ( 2 ) to monitor long - term variabilities of x - ray sources , ( 3 ) to stimulate multi - wavelength observations of variable objects , ( 4 ) to create unbiased x - ray source catalogues , and ( 5 ) to observe diffuse cosmic x - ray emissions , especially with better energy resolution for soft x - rays down to 0.5 kev .
maxi has two types of x - ray slit cameras with wide fovs and two kinds of x - ray detectors consisting of gas proportional counters covering the energy range of 2 to 30 kev and x - ray ccds covering the energy range of 0.5 to 12 kev .
both cameras scan all - sky x - ray images twice due to two different directional cameras every 90 minutes synchronized with the iss orbit .
the data are sent through the downlink between the iss and a ground station via data - relay satellites .
maxi will thus enable us to report x - ray novae or transients to astronomers worldwide within a few minutes .
the ground - based nova - alert system rapidly reports these events with 0.1 to 0.2 degree position accuracy to astronomers worldwide for further follow - up observations
. measurements of each source are combined on the ground . as a result
, we will be able to detect even a weak source such as an agn .
maxi is capable of creating source catalogues for specific periods to investigate the variabilities of x - ray sources . |
in classical mechanics , the problem of two particles , which interact with force depending only on the relative distance between them , is separated into two three - dimensional problems : the free - particle problem and one of a particle in a static potential field @xcite .
motion of a free particle , whose mass is equal to the sum of the masses of both particles , is a motion of the center of masses of the system and is uniform and rectilinear .
motion of a relative particle with so - called reduced mass occurs in the field with potential @xmath0 .
an analogous situation is also observed for the nonrelativistic wave equation of a system of two interacting particles , which was proposed by schrdinger already in the second communication on wave mechanics by the example of elastic rotator ( two - atom molecule ) @xcite .
the probabilistic interpretation of the square of the modulus of a wave function is possible only under the assumption that measurements of coordinates or momenta of various particles do not principally disturb one another even if there exists some interaction between particles @xcite .
this means that operators of coordinates and momenta of two particles commute with each other .
but in the theory of schrdinger , still operators of coordinates and momenta of various particles commute with one another that is equivalent to the lacking of any interference on measurement of the coordinate of one particle and the momentum of the other .
the last assertion is valid if the duration of measurement of the coordinate of a particle is much less than the duration of propagation of a light signal at distances of about the size of the system , or , what is the same , if the compton wave length is much less than the size of the system .
therefore , the schrdinger equation perfectly works in atomic physics and solid - state physics . but a direct application of the schrdinger equation to atomic nuclei seems not entirely correct because the compton wave length of a nucleon is of order of the size of an atomic nucleus itself .
in addition , a rigorous nonrelativistic statement requires to consider the interaction propagation velocity to be infinitely large that forces us to assume that operators of coordinates and momenta of various particles do not commute with one another .
as is known , classical equations of motion of a particle with mass @xmath1 in an external field @xmath0 follow from the hamilton function @xmath2 which depends on the coordinates @xmath3 of the particle and on the corresponding momentum @xmath4 . the total energy of the system @xmath5 with this classical system , we associate a quantum system , whose dynamic state is represented by the wave function @xmath6 defined in the configuration space .
the wave equation is deduced by the formal substitution of the quantities @xmath7 , @xmath3 , @xmath4 in both sides of relation ( [ e2 ] ) by the corresponding operators @xcite : @xmath8 here , @xmath9 , where @xmath10 is the universal constant introduced by planck .
it is meant that the result of action of both sides of equality ( [ e2 ] ) , considered as operators , on @xmath6 is the same .
the realization of this fact implies the schrdinger nonrelativistic equation for a particle in an external field @xmath0 : @xmath11 \psi ( { \bf r,}t)\ , . \label{e6}\ ] ] it is worth to emphasize that the operators @xmath12 and @xmath13 in ( [ e4 ] ) , ( [ e5 ] ) are written in the configuration space , and @xmath14 is the vector of position of the particle in the cartesian coordinate system .
operators of coordinate and momentum do not commute with each other : @xmath15 = i\hbar \,,\text{\quad } \left [ \hat{y},\,% \hat{p}_{y}\right ] = i\hbar \,,\text{\quad } \left [ \hat{z},\,\hat{p}_{z}% \right ] = i\hbar \ , , \label{e7}\ ] ] that leads to the heisenberg uncertainty relations @xmath16 where the quantities @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 , and @xmath22 are directly connected with corresponding measurements and present mean square deviations from the mean value . for example , we have @xmath23 for the coordinate @xmath24 , by definition , where @xmath25 is the mean value of the operator @xmath26 in the dynamic state defined by a wave function @xmath6 .
relations ( [ e8 ] ) assert that a particle can not be in states , in which its coordinate and momentum simultaneously take quite definite , exact values .
moreover , quantum theory accepts that an unpredictable and uncontrolled perturbation , undergone by a physical system in the process of measurement , is always finite and such that the heisenberg uncertainty relations ( [ e8 ] ) are satisfied @xcite .
hence , no experiment can lead to a simultaneous exact measurement of the coordinate and momentum of a particle .
for example , the measurement of the coordinate @xmath24 with accuracy @xmath27 in the well - known experiment with a microscope , considered by heisenberg , is accompanied by the uncontrolled transfer of momentum to the particle , which is characterized by the uncertainty @xmath28 in this case , limits of the accuracy of determination of a position are set always by the optical resolution stipulated by diffraction effects according to classical wave optics .
for example , it is known that the limit accuracy of an image @xmath17 for a microscope is defined by the formula @xmath29 where @xmath30 is the wave length of scattered light , which can differ from that of incident light , and @xmath31 is half the objective aperture . according to relation ( [ e11 ] ) , to increase the accuracy , it is profitable to have the wave length of scattered light as short as possible . but owing to the compton effect , the frequency of scattered light changes by a value defined by the conservation laws of energy and momentum .
this implies that , even in the limit @xmath32@xmath33 , the frequency of scattered emission @xmath34 can not exceed some finite value .
if @xmath4 is a momentum , @xmath35 is a velocity , and @xmath36 is the energy of a material particle prior to the process of scattering , then we have @xcite @xmath37 in this limit , which gives a maximum value for @xmath34 and a minimum one for @xmath38 . here , @xmath39 is the velocity of light in vacuum .
thus , to attain the maximum accuracy of determination of a position on the observation of the scattering of a quantum of light by using an optical instrument , we obtain the following expression : @xmath40 the duration of the process of measurement of a position , i.e. , the time , during which the interaction between a light quantum and a particle can occur , can in no way be less than the periods of oscillations of the incident and scattered emissions and should be of the order of @xmath41 @xmath42 if the size of the system is such that a characteristic time of flight for the system significantly exceeds @xmath43 , then one can say that the process of measurement of the particle coordinate with accuracy @xmath17 is followed by an impact on a particle with the force @xmath44 here , we assume that the momentum transferred to a particle under measurement of its coordinate is of order of the mean square deviation @xmath45 . on the measurement of the momentum of a particle with accuracy @xmath18 , it undergoes an impact with force @xmath46 consider now a system of two interacting particles , whose hamilton function is @xmath47 where @xmath48 and @xmath49 are cartesian coordinates of two particles with masses @xmath50 and @xmath51 , @xmath52 and @xmath53 are their corresponding momenta , and the potential energy depends only on the distance between particles . to derive the nonrelativistic schrdinger equation for this system , we proceed analogously to eq .
( [ e6 ] ) .
this classical system is put into correspondence to the quantum system , whose dynamic state is represented by the wave function @xmath54 defined in the configuration space .
the wave equation can be derived by the formal replacement of the quantities @xmath7 , @xmath48 , @xmath49 , @xmath52 , and @xmath53 on both sides of a relation analogous to ( [ e2 ] ) by the relevant operators @xmath55 then the well - known schrdinger nonrelativistic equation for a system of two interacting particles has the form @xmath56 \psi ( { \bf r}_{1}{\bf % , r}_{2}{\bf , } t)\ , . \label{e24}\ ] ] the operators @xmath57 , @xmath58 , @xmath59 , and @xmath60 are such that they satisfy the following commutation relations : @xmath61 = i\hbar \,,\text{\quad } \left [ \hat{y}_{k},\,\hat{p}_{ky}\right ] = i\hbar \,,\text{\quad } \left [ \hat{z}% _ { k},\,\hat{p}_{kz}\right ] = i\hbar \,,\text{\quad } k=1,\,2\ , .
\label{e25}\ ] ] all other possible commutation relations equal zero , including @xmath62 = 0\,,\text{\quad } \left [ \hat{y}% _ { k},\,\hat{p}_{ly}\right ] = 0\,,\text{\quad } \left [ \hat{z}_{k},\,\hat{p}% _ { lz}\right ] = 0\,,\text{\quad } k,\,l=1,\,2\,,\text{\quad } k\neq l\ , .
\label{e26}\ ] ] equalities ( [ e26 ] ) are based on the assumption that measurements of coordinates and momenta of different particles do not disturb one another in principle even in the presence of some interaction forces between particles @xcite .
that is , one supposes that the change in the force action of a particle on another one , caused by a measurement of the coordinate of the first , propagates with finite velocity . thus , to derive the schrdinger nonrelativistic equation for a two - particle system , one uses , on the one hand , the hamilton classical nonrelativistic function and , on the other hand , the implicit assumption about finiteness of the interaction propagation velocity . in the fully nonrelativistic quantum theory
, we must consider the interaction propagation velocity as infinitely large , which forces us to drop the requirement for the commutation relations ( [ e26 ] ) to hold .
having accepted this viewpoint , we will consider that , under a measurement of the coordinate of the first particle , there occurs the uncontrolled transfer of momentum not only to this particle but to the whole system since the particles are connected through the interaction potential , whose propagation velocity is infinitely large .
therefore , it is natural to demand that the commutator of the operator of the coordinate of the first particle with the operator of the total momentum of the system be equal to @xmath63 : @xmath64 = i\hbar \,,\text{\quad } % \left [ \hat{y}_{1},\,\hat{p}_{cy}\right ] = i\hbar \,,\text{\quad } % \left [ \hat{z}_{1},\,\hat{p}_{cz}\right ] = i\hbar \,\ , .
\label{e27}\ ] ] here , @xmath65 is the operator of the total momentum of the system .
the same should be true for the second particle : @xmath66 = i\hbar \,,\text{\quad } % \left [ \hat{y}_{2},\,\hat{p}_{cy}\right ] = i\hbar \,,\text{\quad } % \left [ \hat{z}_{2},\,\hat{p}_{cz}\right ] = i\hbar \ , .
\label{e28}\ ] ] note that relations ( [ e27 ] ) and ( [ e28 ] ) hold true also for a schrdinger nonrelativistic equation .
namely , they allow one to construct the operator of coordinates of the center of masses of the system . the commutator of the last with the operator of the total momentum of the system equals @xmath63 . on the contrary ,
the fulfillment of relations ( [ e25 ] ) is not obligatory for a system of interacting particles , and we intend to reject this requirement .
it is clear that , on measuring the coordinate of some particle with accuracy @xmath17 , the system undergoes an impact with force @xmath67 .
for example , the measurement of the coordinate of a nonrelativistic electron under observation of the scattering of a light quantum with an optical device with the greatest possible accuracy of order of the compton wave length @xmath68 is accompanied by impact with the force @xmath69 . for a proton with its compton wave length of the order of @xmath70 ,
the impact force is about @xmath71 .
the mean interaction force between particles in a hydrogen atom in the ground state is @xmath72 , and that for the bound state of the nucleus of deuterium is @xmath73 .
therefore , whereas we can neglect the interaction force @xmath74 between particles on measuring the coordinates of particles in an atom and consider the operators of coordinates and momenta of various particles to be commuting , it is not the case for an atomic nucleus , because the ratio of the interparticle interaction force to the impact one is of order @xmath75 .
in the general case , let @xmath76 = i\hbar \hat{f}_{1}\ , , \label{e29}\ ] ] where @xmath77 is some dimensionless hermitian operator .
then it follows from eq .
( [ e27 ] ) that @xmath78 = i\hbar ( 1-\hat{f}_{1})\ , .
\label{e30}\ ] ] by analogy , if @xmath79 = i\hbar \hat{f}_{2}\ , , \label{e31}\ ] ] then @xmath80 = i\hbar ( 1-\hat{f}_{2})\ , .
\label{e32}\ ] ] the dimensionless hermitian operators @xmath77 and @xmath81 depend generally on the interaction force between particles @xmath82 and on masses of the interacting particles @xmath50 and @xmath51 .
the operators @xmath83 and @xmath81 can not depend on a direction of the vector @xmath84 , since the commutation relations for the @xmath24 , @xmath85 , and @xmath86 components should be identical by analogy with ( [ e29])-([e32 ] ) , since there are no separated directions in the system , and independent variables in the cartesian coordinate system are fully equivalent . for this reason , the operators @xmath77 and @xmath81 are only functions of the absolute value of a force , i.e. , of @xmath87 : @xmath88 let us make permutation of @xmath50 and @xmath51
. then @xmath76 = i\hbar \hat{f}% _ { 1}(m_{2},m_{1},f_{12}^{2})\,,\text{\quad } \left [ \hat{x}_{2},\,\hat{p}_{1x}% \right ] = i\hbar \hat{f}_{2}(m_{2},m_{1},f_{12}^{2})\ , . \label{e34}\ ] ] compare ( [ e34 ] ) with ( [ e29 ] ) , ( [ e31 ] ) .
considering that the physical situation has not changed , we get @xmath89 thus , we have one unknown operator @xmath90 . for @xmath91
, @xmath77 must tend to zero since , in the absence of the second particle , the whole momentum transferred under the measurement of the coordinate @xmath92 falls namely this particle . if @xmath93 , then @xmath94 , i.e. , without any interaction forces between particles , the operators of coordinates and momenta of different particles commute among themselves .
the situation @xmath95 corresponds to the case where we have one particle of mass @xmath96 and mentally represent that it consists of two strongly bound particles with masses @xmath50 and @xmath51 .
therefore , the momentum , received under a measurement of some coordinate , is distributed proportionally to masses of particles .
this enables us to write down @xmath97 as @xmath98 .
here , @xmath99 is the system mass .
therefore , without loss of generality , we can present the operator @xmath97 as @xmath100 where @xmath101 is a new operator , which is assumed to be symmetric with respect to the masses of particles @xmath50 and @xmath51 . in
what follows , we will omit its explicit dependence on masses to shorten formulas , namely , @xmath102 . for @xmath103 , @xmath104 , and @xmath105 for @xmath106 . for the noncommuting operators @xmath107 and @xmath108
, the uncertainty relation has the form @xcite @xmath109 where @xmath110 is a quantum - mechanical average in the state @xmath54 . if we replace the operator @xmath101 in ( [ e36 ] ) by its averaged quantum - mechanical value @xmath111 , the uncertainty relation ( [ e37 ] ) does not change .
this makes it possible to write down a nonrelativistic wave equation for a two - particle system , since the operator @xmath77 is now a constant .
it is worth to say several words about the commutativity of the operators of coordinates and momenta of different particles between themselves : @xmath112 = 0\,,\text{\quad } \left [ \hat{p}% _ { 1x},\,\hat{p}_{2x}\right ] = 0\ , .
\label{e38}\ ] ] if we increase the accuracy of measurements of the coordinates @xmath92 , @xmath113 , the impact forces @xmath114 and @xmath115 also grow . for @xmath116 and @xmath117
, we have @xmath118 , @xmath119 .
therefore , we can neglect the interaction force @xmath120 between particles and consider the operators of coordinates of both particles as commuting .
analogously , on measuring the momenta @xmath121 , @xmath122 , the impact forces @xmath123 @xmath124 and @xmath125 tend to zero on increasing the accuracy of measurements .
for this reason , the operators @xmath126 and @xmath108 also can be considered as commuting .
we present now the commutation relations for all operators of coordinates and momenta in the two - body problem : @xmath127 = i\hbar \left ( 1-\frac{m_{2}}{m}% \varepsilon \right ) \ , ,
\label{e39 } \\ & & \left [ \hat{x}_{2},\,\hat{p}_{2x}\right ] = i\hbar \left ( 1-\frac{m_{1}}{m}% \varepsilon \right ) \ , , \label{e40 }
\\ & & \left [ \hat{x}_{1},\,\hat{p}_{2x}\right ] = i\hbar \frac{m_{2}}{m}% \varepsilon \ , , \label{e41 } \\ & & \left [ \hat{x}_{2},\,\hat{p}_{1x}\right ] = i\hbar \frac{m_{1}}{m}% \varepsilon \ , , \label{e42 } \\ & & \left [ \hat{x}_{1},\,\hat{x}_{2}\right ] = 0\ , , \label{e43 } \\ & & \left [ \hat{p}_{1x},\,\hat{p}_{2x}\right ] = 0\ , .
\label{e44}\end{aligned}\ ] ] for the @xmath85 and @xmath86 components , we have analogous relations .
we recall that @xmath128 is a quantum mechanical mean value of the operator @xmath129 in the state @xmath130 : @xmath131 we can construct now one of the possible representations for the operators of coordinates and momenta of a two - particle system : @xmath132 it is easily to verify that operators ( [ e46])-([e49 ] ) satisfy the commutation relations ( [ e39])-([e44 ] ) . in ( [ e46])-([e49 ] ) ,
@xmath3 and @xmath133 are independent operator variables .
the latter represents coordinates of the center of masses of the system : @xmath134 the operator of the total momentum of the system is @xmath135 by substituting operators ( [ e46])-([e49 ] ) to the hamilton function ( [ e18 ] ) , we get the nonrelativistic wave equation for a two - particle system : @xmath136\right ] \psi ( { \bf r , r,}t)\ , .
\label{e52}\ ] ] in this case , @xmath137 is defined according to ( [ e45 ] ) , and @xmath138 is the reduced mass of the system , @xmath139 . for the hamiltonian @xmath140 not depending explicitly on time , the substitution @xmath141 where @xmath142 depends on coordinates in the configuration space but not on time , implies the nonlinear system of integro - differential equations for stationary states of the two - particle system : @xmath143 \right ] \psi ( { \bf r , r)=}% e\psi ( { \bf r , r)\ , , } \label{e54 } \\ & & \varepsilon = \frac{\left\langle \psi , \hat{\varepsilon}(f_{12}^{2})\psi \right\rangle } { \left\langle \psi , \psi \right\rangle } \ , . \label{e55}\end{aligned}\ ] ] by the substitution @xmath144 , we can separate motions of the center of masses of the system as a whole . as a result
, we obtained the following nonlinear system of equations : @xmath145\right ] \varphi ( { \bf r)}=e\varphi ( { \bf r)\ , , } \label{e56 } \\ & & \varepsilon = \frac{\left\langle \varphi , \hat{\varepsilon}% ( f_{12}^{2})\varphi \right\rangle } { \left\langle \varphi , \varphi \right\rangle } .
\label{e57}\end{aligned}\ ] ] as in the schrdinger nonrelativistic theory , a wave function @xmath146 should be continuous together with its partial derivatives of the first order in the whole space and , in addition , be a bounded single - valued function of its arguments . as in the schrdinger theory , for particles interacting by means of a centrally symmetric potential , which depends only on the distance between particles
, the wave function @xmath147 can be represented in the following form : @xmath148 where @xmath149
are orthonormalized spherical functions
. then the function @xmath150 satisfies the following system of equations : @xmath151\right ] \chi _ { l}(r)=e\chi _ { l}(r)\,{\bf , } \label{e59 } \\ & & \varepsilon = \frac{\left\langle \chi _ { l},\hat{\varepsilon}% ( f_{12}^{2})\chi _ { l}\right\rangle } { \left\langle \chi _ { l},\chi _ { l}\right\rangle } \ , . \label{e60}\end{aligned}\ ] ] the quantity @xmath152 presents the share of the momentum transferred to the second particle on measuring the coordinate of the first . to construct the operator @xmath153 on the basis of classical mechanics
is an extremely difficult problem since one must be able to solve three - particle problems in the general form .
therefore , by taking into account the properties of the operator @xmath154 at @xmath103 @xmath155 and at @xmath156 @xmath157 , it is convenient to approximate @xmath158 in the first approximation by a gaussoid : @xmath159 where @xmath160 is a parameter with dimensionality inversely proportional to the square of force .
the explanation for the definition of this will be considered below .
the nonlinear system of the integro - differential equations ( [ e59 ] ) - ( [ e61 ] ) for a two - particle system has solutions for definite values of the energy @xmath7 , but solutions with different energies are not orthogonal to one another .
moreover , since the parameter @xmath137 is a quantum - mechanical mean in every quantum state , we may say that every quantum state has its own potential of interaction between particles .
the constant @xmath160 can be defined by analyzing the discrete spectrum of a hydrogenlike atom .
let two particles with masses @xmath50 ( electron ) and @xmath51 ( atomic nucleus ) be coupled through the coulomb potential @xmath161 , where @xmath162 is the charge of the atomic nucleus .
the nonlinear system of integro - differential equations for bound states ( [ e59 ] ) - ( [ e61 ] ) can be written in the following way : @xmath163 \chi _ { % { \rm nl}}(r)=e_{{\rm nl}}\chi _ { { \rm nl}}(r)\,{\bf , } \label{e62 } \\ & & \varepsilon _ { { \rm nl}}=1-\int_{0}^{\infty } \chi _ { { \rm nl}}^{2}(r)\exp \left ( -\omega _ { 0}\frac{z^{2}e^{4}}{r^{4}(1-\varepsilon _ { { \rm nl}})^{4}}% \right ) \,dr\ , , \label{e63 } \\ & & \int_{0}^{\infty } \chi _ { { \rm nl}}^{2}(r)\,dr=1\ , . \label{e64}\end{aligned}\ ] ] here , @xmath138 is the reduced mass of the system
. equations ( [ e62 ] ) and ( [ e64 ] ) define normalized radial functions of a hydrogenlike atom by the schrdinger theory .
their solutions for bound states are well known ( see , e.g. , @xcite ) : @xmath164 where @xmath165 ^{1/2}\left ( \frac{2z}{(1-\varepsilon _ { { \rm nl}})na_{0}}\right ) ^{l+3/2}\ , .
\label{e66}\ ] ] here , @xmath166 is the bohr radius , and @xmath167 is a degenerate hypergeometric function .
eigenvalues of energy are expressed as @xmath168 here , @xmath169 is the fine structure constant , @xmath170 and @xmath171 . by substituting @xmath172 into eq .
( [ e63 ] ) , we obtain the nonlinear equation for the determination of @xmath173 : @xmath174 where @xmath175 , @xmath176 ^{2}}}}\left [ { \displaystyle{% \frac{(n+l)!}{2n(n - l-1)!}}}\right ] $ ] and @xmath177 . for the ground state of a hydrogenlike atom
, we have @xmath178 for @xmath179 , the nonlinear eq .
( [ e69 ] ) has solutions if @xmath180 that is depicted in fig .
[ fig1 ] . from two solutions ,
that solution is considered as suitable which is located nearer to 1 .
the second solution should be omitted .
indeed , it corresponds to the case where @xmath137 tends to 1 as the parameter @xmath181 decreases to zero , which contradicts to the assumptions made above about the parameter of noncommutativity . for @xmath182 , eq .
( [ e69 ] ) has no solutions , which means it is impossible for a given bound state to exist . the system of eqs .
( [ e62])-([e64 ] ) describes a nonrelativistic motion of a particle with mass @xmath138 in the external field @xmath161 and possesses a critical value of the constant @xmath183 such that a given bound state can not exist if it is exceeded .
the relativistic equation for a hydrogenlike atom is the dirac equation @xcite for a particle with mass @xmath138 in the coulomb field , and it also has a critical constant of the ground state equal to @xmath184 .
it is reasonable to assume that these constants are the same .
this presents the possibility to determine the constant @xmath185 as @xmath186 in fig .
[ fig2 ] , we plot values of the binding energy for the ground state of a hydrogenlike atom , which is calculated by using the parameter @xmath185 defined in such a way .
there , we also present the analogous binding energies according to schrdinger and dirac .
it is seen that the results of our calculation occupy the intermediate position .
similar calculations can be easily performed for excited levels of hydrogenlike atoms . in this case ,
levels by schrdinger with a given @xmath187 split into @xmath188 close sublevels since the orbital quantum number @xmath189 can take @xmath187 values @xmath190 , i.e. , the degeneration is removed in this case . all levels with given @xmath187 and different @xmath189 are located under the corresponding schrdinger level .
the value of the nonrelativistic splitting is much less than that calculated by the dirac theory .
the parameter of noncommutativity for the operators of coordinates and momenta of different particles @xmath137 , presented in fig .
[ fig3 ] , increases with the quantum numbers @xmath187 and @xmath189 ( for the same @xmath162 ) .
that is , fully nonrelativistic solutions transfer to that of the schrdinger equation for large quantum numbers . as a peculiar feature of a fully nonrelativistic equation
, we indicate the presence of a critical value of the parameter @xmath191 for any energy level that is not revealed by the schrdinger nonrelativistic equation .
for example , if @xmath192 , @xmath193 for @xmath194 and @xmath195 for @xmath196 .
when the parameter @xmath197 grows , the mean distance between particles decreases . for the ground state of a hydrogenlike atom , it is defined as @xmath198 and takes the smallest value equal to @xmath199 when @xmath200 . at the same time
( i.e. , for @xmath201 ) , mean distances between particles significantly exceed the value @xmath202 for excited quantum states ( @xmath203 for @xmath204 and @xmath194 ; @xmath205 for @xmath192 and @xmath196 ) . if the parameter @xmath197 further grows , the nonlinear system of eqs .
( [ e62])-([e64 ] ) has no solutions for the state with @xmath206 , i.e. , the @xmath207 state can not exist , and the ground state is a state with @xmath192 and @xmath194 ( @xmath208 state ) or @xmath196 ( @xmath209 state ) .
it is possible if @xmath210 .
here we are faced with the essential difference from solutions of the schrdinger nonrelativistic equation , for which , as is well known , the ground state is always the @xmath211 state .
below , we present the quantum poisson brackets introduced by dirac @xcite as @xmath212 for @xmath213 , these brackets transfer to the classical poisson brackets , i.e. , we have the complete analogy between classical and quantum mechanics in this case . as is seen in fig .
[ fig3 ] , @xmath137 is remarkably different from zero for systems whose sizes are of the order of the compton wavelength of particles forming the system . in this case , there is no analogy with classical mechanics .
to what extent it would take place can be judged by comparing the proposed theory with experiment .
but it is already clear that we have obtained a considerably better agreement with the result of solving the relativistic dirac equation for the ground state of a hydrogenlike atom as compared with the schrdinger nonrelativistic theory .
the previous results can be easily generalized to a system consisting of @xmath214 particles interacting among themselves via two - particle forces .
let the operators of the coordinates and momenta of @xmath214 particles be @xmath215 , @xmath216 .
define the operators of coordinates and momentum of the center of masses of the system : @xmath217 here , @xmath218 is the mass of the whole system . by analogy with a two - particle problem , we require that the commutator of the operator of the coordinate for any particle with the operator of the total momentum of the system be equal to @xmath63 : @xmath219 = \left [ \hat{y}_{k},\,% \hat{p}_{\text{{\it c}}y}\right ] = \left [ \hat{z}_{k},\,\hat{p}_{\text{{\it c}% } z}\right ] = i\hbar \text{\quad } ( k=1,\,2\,,\,\ldots \,,n ) .
\label{e80}\ ] ] then , if @xmath220 = \left [ \hat{y}_{j},\hat{p}_{ky}% \right ] = \left [ \hat{z}_{j},\hat{p}_{kz}\right ] = i\hbar \frac{m_{k}}{m}% \varepsilon _ { jk}\quad ( \ j\neq k ) , \label{e81}\ ] ] we have @xmath221 = \left [ \hat{y}_{j},\hat{p}_{jy}% \right ] = \left [ \hat{z}_{j},\hat{p}_{jz}\right ] = i\hbar \left [ 1-\sum_{k=1}^{n}\frac{m_{k}}{m}\varepsilon _ { jk}\right ] \,,\text{\quad } % j=1,\,2,\,\ldots \text{\thinspace } , n . \label{e82}\ ] ] here , the parameter of noncommutativity of the operators of coordinates and momenta of different particles @xmath222 is symmetric with respect to a permutation of the indices @xmath223 and identically equals zero for @xmath224 by definition .
in addition , @xmath225 = \left [ \hat{y}_{j},\hat{y}_{k}\right ] = \left [ \hat{z}_{j},\hat{z}_{k}\right ] = 0\,,\text{\quad } \left [ \hat{p}_{jx},% \hat{p}_{kx}\right ] = \left [ \hat{p}_{jy},\hat{p}_{ky}\right ] = \left [ \hat{p}% _ { jz},\hat{p}_{kz}\right ] = 0\ , . \label{e84}\ ] ] one of the possible representations for the operators of coordinates and momenta of particles can be written as @xmath226 { \bf \nabla } _ { j}-i\hbar \sum_{k=1}^{n}\frac{m_{j}}{% m}\varepsilon _ { jk}{\bf \nabla } _ { k}\,,\text{\quad } j=1,\,2,\,\ldots \,,n\ , .
\label{e86}\end{aligned}\ ] ] here , we take coordinates of particles as independent variables since the corresponding operators mutually commute . in this case , a nonlinear system of equations for the nonrelativistic problem of @xmath214 particles has the form @xmath227 + \sum_{i=1}^{n}\sum_{j > i}^{n}v(\left| { \bf r}_{j}-{\bf r}% _ { i}\right| )
\right\ } \psi = e\psi \ , , \label{e87}\ ] ] where @xmath228 here , @xmath229 is defined according to ( [ e83 ] ) , and @xmath230 @xmath231 , and @xmath232 .
it can be shown that the introduction of the so - called jacobi coordinates provides the separation of motion of the center of masses as a whole .
the system of equations ( [ e87])-([e89 ] ) takes a particularly simple form in the important case of identical particles @xmath233 @xmath234 @xmath235 after the introduction of the so - called normed jacobi coordinates , @xmath236 in this case , after the separation of motion of the center of masses , we have @xmath237 here , @xmath238 represents the potential energy of the interaction of all particles expressed in terms of jacobi coordinates ( [ e90 ] ) . since @xmath239
, the mean value of the kinetic energy will be less than that according to schrdinger , and the energies of the bound states will be situated lower than the schrdinger ones .
the schrdinger equation for a system of interacting particles is not a strictly nonrelativistic equation because it is grounded on the implicit assumption about finiteness of the interaction propagation velocity .
the last means that if the commutator of operators of a coordinate and the corresponding momentum of a free particle is defined as @xmath15 = i\hbar \ , , \label{e94}\ ] ] this commutator for a system of coupled particles has the same value @xmath240.@xmath241however , in a nonrelativistic quantum system during measurement of the coordinate of a particle , a whole transferred momentum is distributed over all particles but is not transferred to only the measured one .
therefore , in a system of interacting particles , this commutator should have the form @xmath15 = i\hbar \delta \ , , \label{e95}\ ] ] where @xmath242 .
the rejection of the implicit assumption on finiteness of the propagation velocity of interactions implies the noncommutativity of the operators of coordinates and momenta of different particles .
but the operators of coordinates of all particles and operators of momenta of all particles mutually commute that allows one to use these collections as independent variables . a solution of this essentially nonlinear system exists for completely definite values of the energy of the system .
the wave functions corresponding to these energies , as a rule , are mutually nonorthogonal
. properties of solutions of the proposed nonlinear system of equations essentially differ from schrdinger solutions for systems , for which the compton wavelength of particles is comparable with the size of the system .
that is , the consideration of noncommutativity of the operators of coordinates and momenta of different particles is of importance for quantum mechanics of atoms with large charge of the nucleus @xmath243 and for phenomena of the physics of atomic nuclei , where the size of the system is of the order of the compton wavelength of particles which compose the system . | it is shown that the schrdinger nonrelativistic equation of a system of interacting particles is not a rigorously nonrelativistic equation since it is based on the implicit assumption of finiteness of the interaction propagation velocity . for a system of interacting particles , a fully nonrelativistic nonlinear system of integro - differential equations
is proposed . in the case where the size of the system of particles is of the same order as the compton wavelength associated with particles ,
certain essential differences are shown to exist as compared with traditional consequences of the nonrelativistic schrdinger equation . |
it has been known that stacked intrinsic josephson junctions ( ijjs ) form in naturally grown bi@xmath0sr@xmath0cacu@xmath0o@xmath1 ( bi-2212 ) single crystals.@xcite the collective josephson plasma oscillation , manifested by the electromagnetic resonant absorption in such josephson - coupled layered superconductors , has provided the key to understanding the superconducting state and the vortex phases@xcite forming in the materials .
two kinds of collective josephson plasma modes , _ longitudinal _ and _ transverse _ , exist in such a system of an _ infinite number _ of staked ijjs.@xcite the longitudinal plasma mode oscillation , excited in an externally applied @xmath4-axis - oscillating electric field , propagates along the @xmath4 axis of a stack of ijjs with the inter - layer phase difference of a junction being unform in a planar direction .
the resonant absorption of externally applied microwaves by the excitation of the longitudinal modes enables one to directly obtain the josephson plasma frequency , @xmath5 .
this , in turn , provides information on the interlayer josephson coupling strength and the @xmath4-axis superfluid density in bi-2212.@xcite by contrast , the transverse plasma modes can be excited by the oscillating magnetic fields in the @xmath6 plane .
the resulting induced current along the @xmath4 axis generates the oscillating electric field that resonates with the transverse josephson plasma oscillation . on the other hand , a system with a
_ finite number _ ( @xmath7 ) of stacked ijjs , described by linearized coupled sine - gordon differential equations with the inductive inter - junction coupling , exhibits _ collective transverse plasma _ ( ctp ) eigen modes , with the eigen frequencies lying between those of the longitudinal and the transverse modes found for a system of an infinite number of stacked junctions.@xcite the corresponding dispersion relation is expressed as @xmath8 here , @xmath9 @xmath10 is the @xmath4-axis ( @xmath6-plane ) london penetration depth , @xmath7 the number of the junctions in a stack , @xmath11=0.3 nm ( @xmath12=1.2 nm ) the thickness of a superconducting ( insulating ) layer , @xmath13(=@xmath14 ; @xmath15=the length of a stack ; @xmath16=the number of the vortices present in a junction ) the wave number of the plasma oscillation.@xcite the mode index @xmath17 runs from 1 to @xmath7 .
the ctp modes can be excited by the moving josephson vortex lattice forming in high magnetic fields , which are alternative solutions of the coupled sine - gordon equations.@xcite the temporal oscillation of the inter - layer phase difference due to the moving josephson vortex lattice , with the frequency matching with that of any ctp modes , excites a resonant plasma oscillation .
the josephson vortex lattice also transforms its spatial configuration along the @xmath4 axis in accordance with the @xmath4-axis standing - wave modes of the plasma oscillation .
the resonance of josephson vortex lattice to the ctp modes is revealed in the form of the multiple josephson vortex - flow branches in the tunneling current - voltage ( _ i - v _ ) characteristics of a stack of ijjs.@xcite experimental observation of the multiple ctp modes in a stack of a finite number of ijjs has often been attempted in a mesa or an s - shaped stack of ijjs fabricated on the surface of a large stack of ijjs , which is called a pedestal or a basal layer.@xcite since each superconducting bilayer in a stack of ijjs is much thinner than the @xmath4-axis london penetration depth @xmath9 , josephson vortices in a usual mesa or an s - shaped - stack structure are in a strong inductive coupling to those in the basal layer that contains a large number of ijjs by itself , although measurements are intended to be made on a finite number of stacks in the structure of interest .
thus , the stack of ijjs with a coupling to the basal layers corresponds to a system containing nearly an infinite number of junctions , which does not support the ctp mode formation along the @xmath4 axis .
this situation significantly affects the dynamics of the josephson vortices in the mesa or the s - shaped stack itself .
this may explain the usual failure of observing the multiple resonating modes in those structures . in this paper , we present the static and dynamic states of josephson vortices in stacks of ijjs , each sandwiched between two au electrodes without the basal layer , thus consisting of a genuinely finite number of ijjs . in the near - static state in a low tunneling bias current on a single josephson vortex - flow branch ,
pronounced magnetoresistance oscillations were observed with @xmath18 magnetic flux period .
this half - flux - quantum periodic resistance oscillations indicate that the josephson vortex lattice in the near - static state constitutes a triangular lattice along the @xmath4 axis .
in the dynamic state in a high tunneling bias current , on the other hand , a single josephson vortex - flow branch splits into multiple sub - branches .
detailed examination of the sub - branches for varying magnetic fields and temperatures revealed that the observed sub - branches were caused by the resonance between the josephson vortex lattices and the ctp modes .
the josephson vortex configuration in stacked ijjs of bi-2212 single crystals is determined by the magnetic flux density in the junctions and the junction size.@xcite josephson vortices are introduced to the insulating layer of a junction when its length @xmath15 is longer than the josephson penetration depth , @xmath19 @xmath20=@xmath21 $ ] , where @xmath22=@xmath23 is a flux quantum .
one can identify various characteristic magnetic fields in relation with the dynamics of josephson vortices [ see fig . 1 ] . for a stack of junctions of length @xmath15
josephson vortices start entering the junctions above a characteristic field,@xcite @xmath24 [ = @xmath25ln@xmath26 $ ] .
this field can be higher than the nominal value of one flux quantum threading a junction of length @xmath15 , that is @xmath27 ( = @xmath28 ) , especially for a junction comparable to @xmath29 .
even above @xmath27 the tunneling screening current prevents an external magnetic field from entering the junctions until it reaches @xmath24 .
thus , for a junction with its length in the range of @xmath30 [ @xmath31@xmath29ln@xmath32 , one can observe the usual oscillatory fraunhofer modulation of the tunneling critical current as a function of an external magnetic field of @xmath33 , although the junction size is larger than @xmath29 ( region 0).@xcite the josephson vortices entering the junctions for @xmath34 form a regular lattice as the magnetic field reaches the value@xcite @xmath35 ( = @xmath36 . in this field range ( region 1 ) , since the field - induced vortices exist in arbitrary configurations in a given magnetic field , disordered multiple vortex - flow subbranches are generated in the _ i - v _ characteristics . in this region , the number of vortices in a junction and the propagation velocities of different vortex configurations are not correlated with each other.@xcite thus , in this field range , the vortex - flow subbranches in the _ i - v _ characteristics are not related to vortex mode velocities@xcite or geometrical resonances.@xcite in an external field above @xmath35 ( region 2 ) , josephson vortices start forming a triangular lattice . recent observation of the oscillatory vortex - flow resistance arising from the interaction between the vortex lattice and the crystal boundary potential clearly confirmed the existence of the triangular vortex lattice@xcite above @xmath35 .
on the other hand , the resonance between the josephson vortex lattice and ctp modes is well known in this vortex dynamics system.@xcite the collective resonance revealed as voltage peaks ( in the case of the voltage bias ) or voltage jumps ( in the case of the current bias ) has been searched extensively both theoretically and experimentally.@xcite the collective resonance has been observed in the dense - vortex state corresponding to the field range of @xmath37 [ = @xmath38 , where the shortest inter - vortex spacing becomes comparable to the diameter of a josephson vortex @xmath39 ( region 3 ) .
thus , josephson vortices in regions 2 and 3 , although they form triangular lattices in both regions in a static state without a tunneling bias current ( fig .
1 ) , behave differently in the vortex - flow state in the presence of a finite tunneling bias current . in the dynamic state ,
josephson vortices in region 2 remains in a triangular lattice , while those in region 3 transform their lattice configuration to fit the plasma propagation modes in the stacked junctions.@xcite the non - josephson - like emission from the josephson - vortex motion@xcite and the shapiro resonance steps in the josephson vortex - flow state@xcite have been observed in this dense - vortex state .
these results confirm the existence of the coherent motion of the josephson vortex lattice through the entire thickness of stacked ( intrinsic ) josephson junctions . for the highest characteristic field
@xmath40 [ = @xmath41 ( @xmath424 t for bi-2212 ) ] , the shortest inter - vortex spacing becomes comparable to the josephson penetration depth,@xcite @xmath19 . in this region ( region 4 ) ,
the josephson current along the length of a junction distributes almost sinusoidally so that the collective resonance behavior between the vortex motion and the plasma modes is expected to be stronger than in a field lower than @xmath40 .
m. ( b ) schematics of sample and measurement configurations . ]
as - grown slightly overdoped bi-2212 single crystals were prepared by the conventional self - flux method .
we fabricated , using the double - side cleaving technique,@xcite three samples of ijjs sandwiched between two au electrodes at the top and the bottom of a stack of ijjs without the basal part [ fig .
this structure , which is in contrast to the usual mesa or the s - shaped - stack structure fabricated on the surface of a single crystal with a large basal layer , is more coincident with the model systems usually adopted in the theoretical analysis.@xcite the geometry with the basal layer eliminated by the double - side cleaving enables one to investigate the josephson vortex dynamics in coupled ijjs without the interference from the vortex motion in the basal layer .
furthermore , the top and the bottom au - film leads in a stack help the formation of the standing plasma oscillation modes along the @xmath4 axis , while forming a transmission guide of the plasma oscillations along the length of the stack .
the samples containing sandwiched stacks were prepared in the following way.@xcite a single crystal was first glued on a sapphire substrate using negative photoresist and was cleaved until an optically clean and flat surface was obtained .
then , a 100-nm - thick au film was thermally deposited on top of the crystal to protect the surface of the crystal and to make a good ohmic contact .
a mesa structure was then prepared by photolithographic patterning and ar - ion - beam etching .
the surface of the patterned mesa was fixed to another sapphire substrate using negative photoresist and the basal layer of the mesa was subsequently cleaved away .
this process , called the double - side - cleaving technique , was developed by wang _
et al_.@xcite a 100-nm - thick au film was again deposited on this freshly cleaved crystal surface , leaving a stack of ijjs sandwiched between two au electrodes
. a few-@xmath43m - long portions on both sides of the sandwiched stack were etched away subsequently [ see fig .
2(a ) ] to get the bottom au electrode exposed for @xmath4-axis transport measurements as schematically depicted in fig .
finally , a 300-nm - thick au - extension pad was attached on each of the au electrode .
the lateral sizes of three samples were 13.5 @xmath44 1.4 @xmath43m@xmath45 [ sp1 ] , 15 @xmath44 1.4 @xmath43m@xmath45 [ sp2 ] and 16 @xmath44 1.5 @xmath43m@xmath45 [ sp3 ] .
the magnetic field was aligned in parallel with the plane of junctions within the resolution of 0.01 degree to minimize the pinning of josephson vortices by the pancake vortices that can form in cuo@xmath46 bilayers due to the vertical component of a misaligned field .
the alignment was done , in a field of 4 t and at a temperature around 60 k , by tuning to the angle that gives the maximum josephson - vortex - flow resistance or the minimum pining of the josephson vortices by the pancake vortices generated by the angle misalignment.@xcite to minimize the external noise , transport measurements were performed with a low - pass filter connected to each electrode located at room temperature .
.[tab : tablei]sample parameters .
@xmath7 is the total number of intrinsic josephson junctions in a stack , @xmath19 the josephson penetration depth , @xmath47=@xmath48 ( @xmath11 and @xmath12 the thickness of the superconducting and the insulating layers , respectively ) , @xmath49 the maximum propagation velocity of vortices in region 3 , and @xmath50 the velocity of the fastest collective plasma mode estimated by the rcsj model . [ cols="^,^,^,^,^,^,^ " , ]
as a function of an external magnetic field @xmath3 for sp1 at @xmath2=55 k in bias currents from 1 @xmath43a to 10 @xmath43a at intervals of 1 @xmath43a .
( b ) temperature dependence of the josephson - vortex - flow sub - branches and quasiparticle branches of sp1 at @xmath3=3.9 t. inset of ( a ) : a single josephson vortex - flow branch and quasiparticle tunneling branches in @xmath3=1.2 t at @xmath2=55 k. inset of ( b ) : the temperature dependence of @xmath51(@xmath3 ) in a bias current of 3 @xmath43a . curves in the main panel and the inset of ( b ) are shifted vertically for clarity . ]
3(a ) shows the oscillation of josephson - vortex - flow resistance of sp1 [ @xmath7=35 ] as a function of external magnetic fields at @xmath2=55 k in bias currents from 1 @xmath43a to 10 @xmath43a at intervals of 1 @xmath43a .
the field period of the oscillation , @xmath52=510 g is one half of @xmath53 ( = @xmath54 ) through the junction area of length @xmath15=13.5 @xmath43 m and thickness @xmath55=1.5 nm . the oscillation of the josephson - vortex - flow resistance is known to be caused by the interaction between the triangular josephson vortex lattice along the @xmath4 axis and the vortex - flow boundary potential at the junction edges.@xcite the observation of the josephson - vortex - flow resistance oscillation in our sandwiched stack without the basal layers indicates that , as the s - shaped - stack geometry , our sandwiched geometry also supports the josephson - vortex triangular lattice in the steady state [ see the inset of fig .
the oscillation in fig .
3(a ) is reduced by increasing bias current as the bias current tilts the boundary potential.@xcite the oscillatory - magnetoresistance behavior is also suppressed completely at @xmath2 @xmath4280 k near @xmath56 ( = 92 k ) as in the inset of fig .
3(b ) , which is related to melting of the near - static josephson vortex lattice.@xcite the inset of fig .
3(a ) shows the single josephson - vortex - flow branch ( the low - bias region ) and quasiparticle branches ( the high - bias region ) in @xmath3=1.2 t at @xmath2=55 k. the field , belonging to region 2 in fig .
1 , corresponds to the position indicated by the arrow in the main panel . in this field region
, the single - curve josephson - vortex - flow branch caused by the triangular lattice persists even down to 4.2 k ( not shown ) . in fig .
3(b ) , _ i - v _ characteristics in a field of region 3 at @xmath2=25 k also show a single josephson vortex - flow branch corresponding to the triangular lattice .
table i shows the characteristic field value of @xmath47 of each sample .
contrary to the field range of region 2 , however , the single josephson - vortex - flow branch in region 3 evolves into the multiple josephson - vortex - flow sub - branches with decreasing temperature .
3(b ) shows the temperature dependence of the josephson - vortex - flow branches , where the multiple sub - branches start appearing at @xmath2@xmath4220 k and become clearer with decreasing temperature . at @xmath2=4.2
k , a single branch representing the triangular lattice below the current bias of @xmath57@xmath422 @xmath43a splits into multiple josephson - vortex - flow sub - branches for the higher current bias .
the josephson vortex flow , then , transits to the mccumber quasiparticle tunneling for @xmath57 above 8 @xmath43a . the multiple josephson - vortex - flow sub - branches rapidly suppress with increasing temperature above 4.2 k and disappear completely as the temperature reaches 25 k. on the contrary , the quasiparticle - tunneling branches are robust to the temperature variation and persist up to @xmath2=25 k. this contrasting temperature dependencies of two different regions strongly imply that the multiple branches in the two regions are of different physical origins .
= 4.2 k. the _ i - v _ characteristics in all the figures are shifted vertically for clarity .
a voltage @xmath58 denoted by a vertical dotted line in each set of curves demarcates the josephson vortex flow region ( jvfr ) from the mccumber quasiparticle - tunneling region ( qpr ) .
inset of ( a ) : ( i ) the triangular lattice and ( ii ) the rectangular lattice configurations of the josephson vortex lattice along the @xmath4 axis .
the contact resistance was subtracted numerically in the main panels of ( a ) and ( b ) . ]
we also observed the separation of the two characteristic multiple - branch regions in sp2 and sp3 .
4(a ) and 4(b ) show the magnetic field dependencies of _ i - v _ characteristics of two sandwiched stacks , sp2 and sp3 , respectively , at 4.2 k. the number of zero - field quasiparticle - tunneling branches of sp2 and sp3 were 22 and 45 , respectively ( not shown ) , which coincided with the total number of ijjs in the given stacks .
the differentiation of the multiple branches by the collective josephson vortex flow ( the low - bias region ) from the ones by the quasiparticle tunneling ( the high - bias region ) is based on the fact that the delimiting current , represented by the current values marked by the dotted vertical lines , stemmed from the zero - field josephson critical current .
nonetheless , a finite resistance is present in the region below the critical bias point because of the josephson vortex flow , although the region itself represents the dissipationless pair tunneling state .
the observed sub - branches for sp2 and sp3 have common features : ( i ) a single josephson - vortex - flow branch splits into multiple sub - branches for regions 3 and 4 , ( ii ) the number of the quasiparticle sub - branches is @xmath4218 and @xmath4242 , respectively , which is close to the number of the junctions in each sample , and ( iii ) the sub - branches in the josephson vortex - flow region become clearer and wider with increasing the transverse magnetic field ( in contrast to the case of quasiparticle sub - branches , which keep shrinking with field instead ) .
these three common properties of sub - branches are consistent with the prediction of the inductive coupling theory for the electrodynamics of stacked josephson junctions . since the speed of vortices , @xmath49 , is limited by that of the maximum electromagnetic wave in a junction , josephson - vortex - flow state changes to quasiparticle tunneling state near the maximum cut - off velocity or equivalently by the limiting voltage @xmath58 in the outermost sub - branch .
using the relation@xcite of @xmath59=@xmath60 , we could get @xmath49 as @xmath423.9@xmath4410@xmath61 m / s and @xmath424.2@xmath4410@xmath61 m / s for the corresponding delimiting bias point @xmath58 in sp2 and sp3 , respectively .
these values are similar to those obtained from a josephson - vortex - flow branch in mesa structures.@xcite in the case of the usual mesa or the s - shaped - stack structure with a basal layer , only a single josephson - vortex - flow branch used to be observed below the critical bias point , regardless of the temperature and the magnetic field intensity .
the velocity @xmath49 corresponding to the delimiting value on this single branch has been regarded as the slowest mode velocity , @xmath17=@xmath7 . as illustrated below , however , in our case without the basal layer @xmath49 , corresponding to @xmath58 in the outermost sub - branch , represents the propagation velocity of the @xmath17=1 plasma mode . in the resistively and
capacitively shunted junction ( rcsj ) model@xcite the junction capacitance is obtained by the relation , @xmath62=@xmath63 , where @xmath64 [ = @xmath65 is the mccumber parameter , @xmath66 is the normal - state junction resistance and @xmath67 is the return current . in sp2
, @xmath62 is obtained to be about 47 pf , with @xmath68(=0.2 ma ) , @xmath67(=4.8 @xmath43a ) and @xmath66(=10 @xmath69 ) .
the swihart velocity @xmath70=@xmath71 , estimated from the junction capacitance , is 3.6 @xmath44 10@xmath72 m / s , where @xmath9=200 nm and the junction area of @xmath73=21 @xmath43m@xmath45 .
the resulting fastest mode velocity @xmath50=3.7 @xmath44 10@xmath61 m / s is in remarkable agreement with the observed value of @xmath49 in sp2 .
experimental values of @xmath49 in table i are consistent with the calculated ones of @xmath50 for all three samples .
this indicates that the outermost sub - branch is the fastest velocity , @xmath17=1 , mode .
= 5 t with the current axis normalized by the critical bias current .
splitting of the low - bias josephson - vortex - flow curve into multiple sub - branches indicates the effectiveness of the inter - junction capacitive coupling .
( b ) josephson vortex - flow sub - branches for sp2 at @xmath3=4.35 and 4.5 t. the contact resistance was subtracted numerically in ( a ) and ( b ) .
inset of ( a ) : magnetic field dependence of the capacitive coupling parameter , @xmath74 in general , josephson - vortex resonance branches by the inter - layer inductive coupling have been predicted to generate current steps near the resonance voltages on the background of a single resistive branch.@xcite in the present study , however , jospehson - vortex resonance to the ctp modes constitutes multiple sub - branches over almost the entire bias range .
a recent theory for an interlayer inductive coupling hybridized by a capacitive coupling predicts that the resistive sub - branches corresponding to the collective resonance modes show linear _ i - v _ characteristics in the low - bias region with the varied slopes from one another.@xcite more spread in the slopes are predicted for a higher capacitive coupling .
5(a ) shows the josephson - vortex - flow sub - branches for sp2 at @xmath3=5 t. as predicted by the theory , all the sub - branches in the josephson vortex - flow regime exhibit linearities with varied slopes in the low bias range .
the extrapolation of the linear low - bias region converges to a single point on the current axis , although the inductive - capacitive hybrid coupling theory predicts they should converge to the origin .
the discrepancy indicates a finite pinning of josephson vortex motion due to the presence of any defects in the stacked junctions or pancake vortices due to any field misalignment , with the converging current point corresponding to the depinning current .
the inductive - capacitive hybrid coupling model indicates that the extent of the spread in the low - bias slope of the resonant branches is proportional to the strength of the capacitive coupling represented by the parameter @xmath75 as @xmath76=@xmath77\frac{v_c}{1 + 2\alpha[1-\mbox{cos}(n\pi / n)]}$ ] , where @xmath17=1,2, .. ,@xmath7 . here , @xmath78 is the depinning current of the josephson - vortex , @xmath68 is the tunneling critical josephson current at a given applied field . here
@xmath79 is the maximum mode voltage corresponding to a given bias current as represented by the curve denoted with an arrow in fig .
the gray region illustrates the best fit to the sub - branches in the low - bias region .
the best - fit value of @xmath75 turns out to be 0.45 with @xmath79=4.22 mv .
the best - fit values of @xmath75 were found to be almost insensitive to magnetic fields as @[email protected] for @xmath3=4@xmath425 t [ see the inset of fig .
5(a ) ] , which is consistent with the assumption of the capacitive coupling and reasonably close to the theoretical expectation@xcite of [email protected] . therefore , as shown in the upper inset of fig .
4(a ) , we consider that the first ( @xmath17=@xmath7 ; the leftmost ) and last ( @xmath17=1 ; the rightmost ) sub - branches correspond to ( i ) the triangular and ( ii ) the rectangular josephson vortex lattices along the @xmath4 axis , respectively .
5(b ) shows the change of josephson - vortex - flow sub - branches for sp2 when magnetic field increases from @xmath3=4.35 t to 4.5 t. the more changes take place in the voltages between the two magnetic fields for branches in the higher voltages ( branches more to the right ) .
this behavior can be compared to the dispersion relation of eq .
( 1 ) , which states that the lower - index modes are more dispersive than the higher - index modes . thus , the right - most branches among multiple josephson - vortex sub - branches in figs .
3@xmath816 are again considered to represent the square vortex lattice [ see the inset of fig .
4(a ) ] corresponding to the @xmath17=1 plasma excitation mode .
the maximum number of vortex - flow sub - branches counted for various fields in sp2 is 20 , which is two modes smaller than the number of quasiparticle branches ( or , equivalently , than the number of junctions ) .
recently , ryndyk@xcite proposed that the additional dissipation due to the charge - imbalance relaxation due to quasiparticle injection in a junction can prevent the structural transformation of josephson - vortex lattice along the @xmath4 axis . in our sample geometry quasiparticles
are injected directly from outside through the outermost two junctions , which may cause the non - equilibrium charge - imbalance effect in the two outermost cuo@xmath46 bilayer .
the resulting suppression of the configurational transformation of josephson - vortex lattice reduces the number of observable resonance modes . for inner cuo@xmath46 planes , on the other hand , with the pair tunneling for a bias below the critical current of the junction , the quasiparticle injection can be neglected .
= 5.2 t and ( b ) for sp3 in @xmath3=3 t. the curves are shifted vertically for clarity .
the insets of ( a ) and ( b ) show a single josephson vortex - flow branch above the characteristic temperature @xmath82 in two different magnetic fields for sp2 and sp3 , respectively . ]
6(a ) and ( b ) show the temperature dependencies of _ i - v _
characteristics for sp2 in @xmath3=5.2 t and for sp3 in @xmath3=3 t , respectively .
the temperature dependencies of the sub - branches in the josephson vortex - flow region and quasiparticle - tunneling region are much different from each other . although not clearly illustrated in the figures the quasiparticle branches survived up to temperatures close to @xmath56 .
the sub - branches in the josephson vortex - flow region , however , disappeared at a characteristic temperature @xmath82 , near 18 k and 28 k for sp2 and sp3 , respectively , leaving a single branch as sp1 in fig .
these characteristic temperatures are almost insensitive to the fields applied [ see insets of figs .
6(a ) and ( b ) ] . in the vortex flow state for the magnetic - field range of regions 3 and 4 the triangular vortex configuration and the multiple - mode configuration resonating with the collective ctp modes compete each other . at low - enough temperatures
, the triangular vortex configuration wins in the low - bias region , while the multiple - mode configuration predominates in the high - bias region . for @xmath2@xmath42@xmath82 , however , thermal fluctuation and quasiparticle dissipations prevent the josephson vortex lattice from resonating with the collective ctp modes .
then the vortices stay on the triangular - lattice configuration without a structural transformation even in the dynamic state.@xcite the value of @xmath58 seems to be insensitive to temperature near @xmath82 , which means that the triangular josephson vortex lattice can reach the maximum velocity corresponding to the fastest mode , @xmath17=1 , without the structural transformation in resonance with the ctp modes . in this branch above @xmath82 , therefore , the lattice structure may be the modulated triangular lattice but with the propagation velocity still locked to that of the fastest mode .
_ thus , the single branch above @xmath82 represents the triangular vortex lattice , while the well - defined outermost sub - branch below @xmath82 represents the rectangular lattice_. accordingly , in figs .
3(b ) and 6 , although the rightmost sub - branch below @xmath82 appears to directly transform into the single branch above @xmath82 , their characteristics correspond to very different josephson - vortex configurations .
as mentioned earlier , the single josephson vortex - flow branch without sub - branches has been observed in the usual mesa structure.@xcite in the usual mesa or the s - shaped - stack structure the structural transformation from the triangular lattice to the rectangular lattice in resonance to the ctp modes is hard to take place because of the strong coupling to the near - static triangular lattice in the basal layer even at sufficiently low temperatures . in this case , the velocity of the josephson vortex lattice may reach that of the fastest mode without the structural transformation.@xcite
the josephson vortex dynamics was comprehensively studied in compactly stacked intrinsic josephson junctions over a wide range of magnetic fields and temperatures .
measurements were made on three stacks of intrinsic josephson junctions , each sandwiched between two ( top and bottom ) normal - metallic electrodes , thus eliminating the interference from the basal layer(s ) .
characteristics for high external magnetic fields ( regions 3 and 4 defined in fig .
1 ) showed three states : ( i ) the near - static triangular josephson - vortex state with the formation of a single branch , ( ii ) the dynamic vortex state with multiple - mode sub - branches , and ( iii ) the mccumber quasiparticle - tunneling state with quasiparticle sub - branches . in the dynamic vortex state ,
each sub - branch corresponded to the different configuration of the josephson - vortex lattice along the @xmath4 axis from a triangular to a rectangular lattices.@xcite with increasing temperature , the sub - branches disappeared at @xmath82 , above which no configurational transformation of the josephson - vortex lattice can be assumed . due to the strong inductive coupling to many junctions in the basal layer(s ) , the mesa or the s - shaped - stack structure , although containing a finite number of junctions in the stack under study , behaves effectively as consisting of an infinite number of junctions along the @xmath4 axis .
thus , these structures can not support the formation of the multiple josephson - vortex - flow modes as predicted theoretically for a finite number of junctions .
although the @xmath4-axis boundary condition is not strictly imposed in these systems the transverse vortex motion is still restricted by the boundary conditions at the side edges .
this is the main reason why , in the mesa or the s - shaped - stack structure , clear magnetoresistance oscillations are observed by the interplay between the near - static triangular josephson - vortex configuration@xcite and the boundary potential in the systems , while the transverse josephson - vortex multiple modes are not obtained .
one should be noted that the multiple collective josephson vortex modes can be obtained only in a stack containing a finite number of ( intrinsic ) josephson junctions .
the triumph of this study is that our samples truly satisfied this condition of a finite number of junctions in a stack and , as a result , both the near - static josephson vortex state highly coupled to the edge potential and the collective transverse josephson - vortex state were identified in each junction .
this facilitated the detailed identification of the vortex configurations illustrated in fig . 1 .
the authors wish to acknowledge stimulating discussions with r. kleiner , h. b. wang , s. m. kim , m. machida , ju h. kim , and s. sakai .
this work was supported by the national research laboratory program administrated by korea science and engineering foundation ( kosef ) .
this paper was also supported by postech core research program . + * electronic address : [email protected] 00 r. kleiner , f. steinmeyer , g. kunkel , and p. mller , phys .
lett . * 68 * , 2394 ( 1992 ) o. k. c. tsui , n. p. ong , y. matsuda , y. f. yan , and j. b. peterson , phys .
lett . * 73 * , 724 ( 1994 ) ; l. n. bulaevskii , m. p. maley and m. tachiki , phys .
lett . * 74 * , 801 ( 1995 ) .
y. matsuda , m. b. gaifullin , k. kumagai , k. kadowaki , and t. mochiku , phys .
lett . * 75 * , 4512 ( 1995 ) .
m. tachiki , t. koyama and s. takahashi , phys .
b * 50 * , 7065 ( 1994 ) ; i. kakeya , k. kindo , k. kadowaki , s. takahashi , and t. mochiku , phys . rev .
b * 57 * , 3108 ( 1998 )
. m. b. gaifullin , y. matsuda , n. chikumoto , j. shimoyama , k. kishio , and r. yoshizaki , phys .
lett . * 83 * , 3928 ( 1999 ) .
r. kleiner , phys .
b. * 50 * , 6919 ( 1994 ) ; m. machida , t. koyama , a. tanaka , and m. tachiki , physica c * 330 * , 85 ( 2000 ) .
m. machida and m. tachiki , curr .
* 1 * , 341 ( 2001 ) . j. u. lee , p. guptasarma , d. hornbaker , a. el - kortas , d. hinks , and k. e. gray , appl . phys
. lett . * 71 * , 1412 ( 1997);y .- j .
doh , j. kim , h. s. chang , s. chang , h .- j .
lee , k. t. kim , w. lee , and j. h. choy , phys .
b * 63 * , 144523 ( 2001 ) . m .- h .
bae and h .- j .
lee , phys .
b * 70 * , 052506 ( 2004 ) .
g. hechtfischer , r. kleiner , k. schlenga , w. walkenhorst , p. mller , and h. l. johnson , phys .
b * 55 * , 14638 ( 1997 ) .
k. hirata , s. ooi , and t. mochiku , physica c * 362 * , 114 ( 2001 ) .
a. e. koshelev , phys .
b * 66 * , 224514 ( 2002 ) .
j. r. clem , m. w. coffey , and z. hao , phys . rev .
b * 44 * , 2732 ( 1991 ) . m. v. fistul and g.f . giuliani , physica c * 230 * , 9 ( 1994 ) . yu .
i. latyshev , j. e. nevelskaya , and p. monceau , phys .
* 77 * , 932 ( 1996 ) .
r. ikeda , j. phys .
* 71 * , 587 ( 2002 ) . v. m. krasnov , v. a. oboznov , v. v. ryazanov , n. mros , a. yurgens , and d. winkler , phys .
b * 61 * , 766 ( 2000 ) . s. sakai , p. bodin , and n. f. pedersen , j. appl .
phys * 73 * , 2411 ( 1993 ) . v. m. krasnov , n. mros , a. yurgens , and d. winkler , phys .
b * 59 * , 8463 ( 1999 ) .
s. ooi , t. mochiku , and k. hirata , phys .
lett . * 89 * , 247002 ( 2002 ) . for a sufficiently short stack of junctions
the boundary - potential effect induces a rectangular vortex lattice configuration in a steady state . for details ,
refer to m. machida , phys .
lett . * 96 * , 097002 ( 2006 ) .
a. e. koshelev and i. s aranson , phys .
lett . * 85 * , 3938 ( 2000 ) .
r. kleiner , t. gaber , g. hechtfischer , physica c * 362 * , 29 ( 2001 ) . g. hechtfischer , r. kleiner , a. v. ustinov , and p. mller , phys . rev . lett . *
79 * , 1365 ( 1997 ) .
i. latyshev , m. b. gaifullin , t. yamashita , m. machida , and y. matsuda , phys .
* 87 * , 247007 ( 2001 ) .
l. bulaevskii and j. r. clem , phys .
b * 44 * , 10234 ( 1991 ) .
h. b. wang , p. h. wu , and t. yamashita , appl .
lett * 78 * , 4010 ( 2001 ) .
bae , h .- j .
lee , j. kim , and k .-
kim , appl .
. lett . * 83 * , 2187 ( 2003 )
. s. m. kim , h. b. wang , t. hatano , s. urayama , s. kawakami , m. nagao , y. takano , t. yamashita , and k. lee , phys .
b * 72 * , 140504(r ) ( 2005 ) .
m. machida , phys .
lett . * 90 * , 037001 ( 2003 ) .
x. hu and m. tachiki , phys .
* 85 * , 2577 ( 2000 ) ; yu .
i. latyshev , v. n. pavlenko , a. p. orlov , and x. hu , jetp lett .
* 82 * , 232 ( 2005 ) .
a. irie , y. hirai , and g. oya , appl .
lett . * 72 * , 2159 ( 1998 ) .
ju h. kim and j. pokharel , physica c * 384 * , 425 ( 2003 ) .
d. a. ryndyk , v. i. pozdnjakova , i. a. shereshevskii , and n. k. vdovicheva , , phys .
b * 64 * , 052508 ( 2001 ) . | we report the experimental confirmation of the collective transverse plasma modes excited by the josephson vortex lattice in stacks of intrinsic josephson junctions in bi@xmath0sr@xmath0cacu@xmath0o@xmath1 single crystals .
the excitation was confirmed by analyzing the temperature ( @xmath2 ) and magnetic field ( @xmath3 ) dependencies of the multiple sub - branches in the josephson - vortex - flow region of the current - voltage characteristics of the system . in the near - static josephson vortex state for a low tunneling bias current ,
pronounced magnetoresistance oscillations were observed , which represented a triangular - lattice vortex configuration along the @xmath4 axis . in the dynamic vortex state in a sufficiently high magnetic field and for a high bias current ,
splitting of a single josephson vortex - flow branch into multiple sub - branches was observed .
detailed examination of the sub - branches for varying @xmath3 field reveals that sub - branches represent the different modes of the josephson - vortex lattice along the @xmath4 axis , with varied configuration from a triangular to a rectangular lattices .
these multiple sub - branches merge to a single curve at a characteristic temperature , above which no dynamical structural transitions of the josephson vortex lattice is expected . |
it is now more than decade that the properties of the new state of matter called quark - gluon plasma ( qgp ) are studied with the ultra - relativistic heavy - ion collisions at rhic and the lhc . among many expected results
there are also a few outstanding surprises .
in particular , broad analysis of the particle correlations @xcite in the experimental data revealed that , in contrast to the earlier expectations based on the asymptotic freedom property of quantum chromodynamics ( qcd ) , the qgp is a strongly coupled system with the lowest viscosity in nature .
moreover , it is surprisingly well described within the effective framework of relativistic dissipative fluid dynamics .
unfortunately , predictive power of the latter , required for the ongoing precision studies , is highly limited due to the requirement of the precise knowledge of its initial conditions .
in particular , one has only qualitative information on the amplitude of the viscous corrections just after the collision , which , apart from resulting in possible quantitative differences in the predicted results , may question applicability of the viscous fluid dynamics in such modelling in general @xcite .
thus , the study of the approach of the created qgp toward the local thermal equilibrium ( _ thermalization _ ) is nowadays of highest interest for the field of heavy - ion physics @xcite . in this proceedings contribution
we briefly review our main results of the study of the qgp thermalization rate within the framework of the color - flux - tube model proposed in ref .
@xcite . based on the detailed analysis of the transverse - momentum spectra of the produced partons
@xcite we show that it is difficult to fully thermalize the system within a short time ( @xmath0 fm ) unless the shear viscosity to entropy density ratio is that of ads / cft lower bound , @xmath1 @xcite .
however , at the same time , for @xmath2 the observed deviations from equilibrium state are quite well described within dissipative corrections of viscous hydrodynamics .
the latter observation confirms the fast hydrodynamization of the qgp first proposed in ref .
based on the saturation physics @xcite , due to the ultra - relativistic energies , most of particles produced in midrapidity of heavy - ion collisions originate from the decay of purely longitudinal color fields spanned by the receding color - charged nuclei . in refs . @xcite it was proposed to describe this physics situation within the simplified model of initial color electric flux tube @xcite treated in the abelian dominance approximation @xcite which subsequently decays into quarks and gluons according to the schwinger tunneling mechanism @xcite . within this
framework phase - space densities of quarks , antiquarks and charged gluons are described by the following kinetic boltzmann - vlasov - type equations @xmath3 respectively . in eqs .
( [ kineaq ] ) partons couple to the mean color field @xmath4 through the charges @xmath5 ( for quarks ) , @xmath6 ( for antiquarks ) , and @xmath7 ( for gluons ) @xcite , where the color indices @xmath8 run from 1 to 3 .
the first terms on the right hand side of eqs .
( [ kineaq ] ) denote the particle production due to the schwinger tunnelling @xcite .
in addition to the usual treatment we include the subsequent collisions between produced particles through the collisional kernels treated in the relaxation time approximation ( rta ) @xcite @xmath9 where the relaxation time is expressed according to the anderson - witting formula @xmath10 @xcite . in the case of boost - invariant and transversely homogeneous system , also called the bjorken symmetry @xcite , the solutions of eqs .
( [ kineaq ] ) depend only on the proper time , @xmath11 , transverse momentum , @xmath12 , and the boost - invariant variable , @xmath13 , i.e. , @xmath14 , and may be found analytically @xcite .
in order to study approach of the system toward the local thermal equilibrium one typically analyses the deviations of pressures from the equilibrium pressure , see ref . @xcite .
alternatively , one can also study the @xmath12 spectra of the produced partons , and their approach toward the equilibrium spectrum @xcite .
for that purpose we use cooper - frye formula @xcite , which for the bjorken symmetry @xcite considered herein reduces to the following form @xmath15 where @xmath16 arbitrarily sets the transverse size in such a way that the integral ( [ cfbi ] ) gives the production per unit area .
we assume that the phase - space distribution @xmath17 in eq .
( [ cfbi ] ) is given by the sum of the quark , antiquark and gluon distribution functions satisfying transport equations ( [ kineaq ] ) , @xmath18 where @xmath19 is the number of flavours .
for the results presented herein we use the solutions from ref .
@xcite obtained for the initial field strength corresponding to the lhc case .
in the left panels of fig . [ sp1 ] we present the partonic @xmath12 spectra obtained using eqs .
( [ cfbi ] ) and ( [ f ] ) for two values of the shear viscosity to entropy density ratio , @xmath20 ( top ) and @xmath21 ( bottom ) , and various freeze - out proper times .
we observe that , except for the low-@xmath12 range , in all cases the spectra have approximately thermal - like ( exponential ) shape .
this property , sometimes called _ apparent thermalization _
, is , however , mainly related to the specific production mechanism of partons , which in our case is the schwinger tunnelling in the oscillating color field @xcite . in order to judge
if the thermalization is really achieved we propose to study the corresponding _ negative inverse logarithmic slopes _ , @xmath22 , of the spectra , @xmath23^{-1}.\label{slope}\ ] ] + + in the right panels of fig .
[ sp1 ] we compare the slopes @xmath22 of the spectra from the left panels with the corresponding slopes of the spectra of the viscous fluid , @xmath24 having the same energy density @xmath25 . in eq .
( [ lvisc ] ) @xmath26 and @xmath27 , with @xmath28 and @xmath29 being the longitudinal and transverse pressure , respectively . in the special case of the system in local thermal equlibrium , @xmath30 ,
the formula ( [ lvisc ] ) reduces to the one for the perfect fluid , @xmath31 from the fig .
[ sp1 ] we observe that , although the spectra look approximately thermal for all @xmath32 , only in the case of @xmath20 the slopes @xmath22 agree with the equilibrium ones @xmath33 within a short proper - time ( @xmath0 fm ) .
this suggest that the system undergo fast thermalization only in the case of maximal possible coupling . on the other hand ,
the discrepancies at medium and large @xmath12 observed for @xmath34 ( bottom panel ) are very well addressed within the dissipative corrections of the viscous fluid dynamics , see @xmath35 .
it means that although the fast thermalization seems to be limited to @xmath20 , the system may still undergo fast hydrodynamization if @xmath36 .
the thermalization rate of the quark - gluon plasma produced from the initial color electric flux tubes through the schwiner mechanism is studied .
once produced the partons are interacting with each other according to the collisional kernels treated in rta .
the detailed comparison of the resulting transverse - momentum spectra of partons with the viscous hydrodynamics ones is done .
obtained results suggest that , although the early thermalization may be limited to strongly coupled systems ( @xmath37 ) , as long as @xmath36 the early hydrodynamization of the plasma in such reactions may be achieved .
r.r . was supported by polish national science center grant no .
dec-2012/07/d / st2/02125 . p. bozek and w. broniowski , _ phys . rev
. lett . _ * 109 * , 062301 ( 2012 ) .
p. bozek , w. broniowski and a. olszewski , _ phys .
rev . _ * c92 * , no . 5 , 054913 ( 2015 ) . w. florkowski , r. ryblewski and m. strickland , _ phys .
rev . _ * c88 * , 024903 ( 2013 ) .
g. s. denicol , w. florkowski , r. ryblewski and m. strickland , _ phys .
* c90 * , no . 4 , 044905 ( 2014 ) .
a. jaiswal , r. ryblewski and m. strickland , _ phys .
rev . _ * c90 * , no . 4 , 044908 ( 2014 ) . m. nopoush , m. strickland , r. ryblewski , d. bazow , u. heinz and m. martinez , _ phys . rev . _ * c92 * , no . 4 , 044912 ( 2015 ) .
s. mrowczynski , _ phys .
rev . _ * c49 * , 2191 ( 1994 ) .
a. rebhan , m. strickland and m. attems , _ phys .
rev . _ * d78 * , 045023 ( 2008 ) .
a. rebhan , p. romatschke and m. strickland , _ jhep _ * 0509 * , 041 ( 2005 ) .
t. epelbaum and f. gelis , _ phys .
lett . _ * 111 * , 232301 ( 2013 ) .
j. berges , k. boguslavski , s. schlichting and r. venugopalan , _ phys .
rev . _ * d89 * , no . 7 , 074011 ( 2014 ) .
a. kurkela and y. zhu , _ phys .
lett . _ * 115 * , no .
18 , 182301 ( 2015 ) .
m. chesler and l. g. yaffe , _ phys .
lett . _ * 106 * , 021601 ( 2011 ) .
s. caron - huot , p. m. chesler and d. teaney , _ phys .
rev . _ * d84 * , 026012 ( 2011 ) .
m. p. heller , r. a. janik and p. witaszczyk , _ phys .
lett . _ * 108 * , 201602 ( 2012 ) .
w. van der schee , p. romatschke and s. pratt , _ phys .
lett . _ * 111 * , no .
22 , 222302 ( 2013 ) .
r. ryblewski and w. florkowski , _ phys .
* d88 * , 034028 ( 2013 ) .
j. jankowski , g. plewa and m. spalinski , _ jhep _ * 1412 * , 105 ( 2014 ) .
m. ruggieri , a. puglisi , l. oliva , s. plumari , f. scardina and v. greco , _ phys .
_ * c92 * , 064904 ( 2015 ) .
r. ryblewski , arxiv:1512.04117 [ nucl - th ] .
p. kovtun , d. t. son and a. o. starinets , _ phys .
lett . _ * 94 * , 111601 ( 2005 ) .
t. lappi and l. mclerran , _ nucl .
* a772 * , 200 ( 2006 ) .
a. casher , h. neuberger and s. nussinov , _ phys .
_ * d20 * , 179 ( 1979 ) .
a. bialas and w. czyz , _ phys .
rev . _ * d30 * , 2371 ( 1984 ) .
a. bialas and w. czyz , _ acta phys .
polon . _ * b17 * , 635 ( 1986 ) .
a. bialas , w. czyz , a. dyrek and w. florkowski , _ nucl .
phys . _ * b296 * , 611 ( 1988 ) . j. s. schwinger , _ phys . rev . _ * 82 * , 664 ( 1951 ) .
k. huang , singapore , singapore : world scientific ( 1992 ) 333p .
l. bhatnagar , e. p.
gross and m. krook , _ phys .
rev . _ * 94 * , 511 ( 1954 ) .
j. l. anderson and h. r. witting , _ physica _ * 74 * , 466488 ( 1974 )
. j. d. bjorken , _ phys .
rev . _ * d27 * , 140 ( 1983 ) .
f. cooper and g. frye , _ phys .
_ * d10 * , 186 ( 1974 ) .
a. bialas , _ phys .
_ * b466 * , 301 ( 1999 ) .
w. florkowski , _ acta phys .
polon . _ * b35 * , 799 ( 2004 ) . | the study of transverse - momentum spectra of quarks and gluons produced by the color electric flux tube decaying through the schwinger tunneling mechanism is reviewed .
the hints for a fast hydrodynamization in the ultra - relativistic heavy - ion collisions are found . |
the electric field extracted from the electromagnetic form - factors of the proton is , in the short range ( where it differs significantly from the coulomb field ) , not a correction to the latter .
this short range part is handled as a perturbation in calculating the energy correction which is possible due to the rapid fall - off of finite size contribution outside the proton .
however , the magnitude of the electric field reaches a value up to 10@xmath0 mev@xmath1 which is not only extremely high , but also dangerously close to allowing electron - positron pair production . on the other hand , quantum corrections to maxwell s equations known as the euler - heisenberg theory yield a different picture of the same physical phenomenon inside the hydrogen atom :
the field becomes significantly smaller falling below the pair production threshold .
we outline the theory which leads to this effect .
the electromagnetic properties of the nucleons are encoded in their two form - factors , @xmath2 and @xmath3 . in calculating the elastic amplitude of the scattering process , @xmath4
, these form factors enter one of the vertices by making the replacement @xmath5 the standard procedure to extract an interaction potential from the elastic amplitude is to ( i ) expand the latter in terms proportional to powers of @xmath6 ( which leads to the non - relativistic amplitude @xmath7 ) and ( ii ) take the fourier transform , i.e , @xmath8 . in the case of the electromagnetic interaction ,
this leads to the breit equation @xcite and if we include the proton form - factors we explicitly take into account in the breit equation the finite size of the proton in @xmath9 @xcite . in terms of the so - called electric ( @xmath10 ) and magnetic ( @xmath11 ) sachs form - factors @xmath12 and @xmath13 the result of the scalar interaction potential in momentum space can be written as @xmath14\ , .\ ] ] in principle , the breit equation gives also the fine and hyperfine structure of the interaction which have to be added to the term above .
we do not spell it out explicitly here since we will not make use of it . for details
we refer to @xcite .
we can identify the first term as the coulomb term ( modified due to finite size ) @xmath15 , the second term is called the proton darwin @xmath16 and the third the electron darwin term @xmath17 .
we identify the electric potential of the proton from the breit equation as terms which are ( i ) scalar ( no spin and no momentum terms which appear in the fine and hyperfine part ) and ( ii ) independent from the properties of the test particle . by this token
the proton darwin term can be part of the electric potential , but the electron darwin not .
this has some consequences for the electric field at small distances inside the proton and also for the charge distribution .
hence we write @xmath18 where @xmath19 is the proton electric potential . as an example
let us take the dipole parametrization @xmath20 .
then we obtain ( see also figure 1 ) @xmath21 since the main issues which we want to emphasize will remain mostly insensitive to the parametrization of the electric form - factors , we shall use the dipole parameterization where the darwin terms can also be calculated analytically .
the electric field @xmath22 calculated from @xmath23 ( which includes the form - factors ) can be now as strong as 10@xmath0 mev@xmath1 .
it is reasonable to ask if this is indeed the case .
we will give a partial answer in the following two sections starting with a discussion of the the non - linear euler - heisenberg theory .
.,width=226,height=226 ] [ mo ]
at the born level , photons do not interact with each other . however , the box diagram with four external photon ( light - light scattering ) legs is an example of how quantum mechanics effectively introduces a four photon vertex .
one can introduce this effect into the lagrangian language which is known as the euler - heisenberg theory . in accordance with the gauge invariance we have to formulate the theory in terms of
the two gauge invariant quantities @xmath24 then , we can expect that in the first approximation ( weak field approximation ) the new part of the lagrangian @xmath25 is quadratic in these invariants .
indeed , euler and heisenberg obtained @xcite @xmath26 with @xmath27 our requirement of weak fields means that in @xmath28 ( @xmath29 ) the new term must be small as compared to one or @xmath30 and @xmath31 .
it then follows that @xmath32 .
the new inhomogeneous maxwell s equations now read : @xmath33 with @xmath34 \end{aligned}\ ] ] and @xmath35\ , .
\end{aligned}\ ] ] in the electrostatic case , putting @xmath36 and neglecting all time derivatives we can write @xmath37 where @xmath38 is the electric field as calculated using the maxwellian theory with @xmath39 ( we treat the charge distribution as a given source ) . hence , assuming spherical symmetry we end up with an algebraic equation @xmath40 with @xmath41 .
the solution is always in the form @xmath42 $ ] .
one can extend the above theory to two - loops @xcite in which case the algebraic equation will become a quintic polynomial equation of the form @xmath43 where @xmath44 .
we can apply the results of the euler - heisenberg theory in the electrostatic case for @xmath45 where @xmath22 is the electric field calculated using the electromagnetic form - factors .
the results are shown in figure 2 . as the electric field calculated from the form - factors is not a correction to the @xmath46 coulomb field at small distances
so is the result from euler - heisenberg not a correction to @xmath47 at small distances .
since the significant changes happen only at short distances , both modifications ( the finite size and the light - light modifications ) will induce only small corrections to the energy levels .
the energy content of the electric field is given by @xmath48=\int \mathbf{e}^2 d^3x$ ] .
for @xmath49 we get @xmath50\simeq 1 \ , \ , mev\ ] ] which , given the uncertainties in the form - factors , is close to the pair production threshold of @xmath51 .
on the other hand we obtain @xmath52 \simeq 0.26\ , \ , mev\ ] ] in the one - loop euler - heisenberg theory and @xmath53 \simeq 0.076\ , \ , mev\ ] ] in its two - loops version .
both results are safely below the pair production threshold .
perturbations in the hydrogen atom can be viewed as additional contributions to the coulomb potential suppressed by higher powers of the fine structure constant @xmath54 ( for instance the uehling potential @xcite ) or as very short range modifications of the coulomb potential .
the best paradigm is given by the finite size corrections to the coulomb part .
in the short range these modifications can not be considered as perturbations of @xmath55 , but their short range character qualifies them as perturbative contributions in the hydrogen atom .
a similar situation occurs when we calculate the full non - linear contributions of light - light scattering . in the short range , where the non - linear effects are important , the modifications of the euler - heisenberg theory are not perturbative contributions to the coulomb potential , but as in the case of the finite size effects they will contribute perturbatively to the observables of the hydrogen atom . moreover , the electric field extracted from the electromagnetic form - factors via the maxwellian theory is too strong as its energy content is close to the pair production threshold .
in contrast to this , the electric field in the euler - heisenberg theory is much smaller in the short range and its energy comes out to be safely below the pair production threshold .
how this affects the electron darwin term and the hydrogen observables remains to be explored @xcite .
99 g. breit , phys . rev . * 34 * , 553 ( 1929 ) v. b. berestetskii , e m : lifshitz and l. o. pitaevskii , _ quantum electrodynamics _ , 2nd edition , oxford , butterworth - heinemann 2007 f. garcia daza , n. g. kelkar and m. nowakowski , j. phys .
* g39 * , 035103 ( 2012 ) n. g. kelkar , f. garcia daza and m. nowakowski , nucl . phys . *
b864 * , 382 ( 2012 ) ; d. bedoya fierro , n.g .
kelkar and m. nowakowski , jhep * 1509 * , 215 ( 2015 ) , arxiv:1410.4228 w. heisenberg and h. euler , z. phys .
* 98 * , 714 ( 1936 ) b. krs and m. g. schmidt , eur .
j. * c6 * ( 1999 ) 175 e. a. uehling , phys . rev . *
48 * , 55 ( 1935 ) m. nowakowski , n. g. kelkar , d. bedoya fierro and a. d. bermudez manjarres , _
static field configurations in the euler - heisenberg theory _
( under preparation ) . | we apply the non - linear euler - heisenberg theory to calculate the electric field inside the hydrogen atom .
we will demonstrate that the electric field calculated in the euler - heisenberg theory can be much smaller than the corresponding field emerging from the maxwellian theory . in the hydrogen atom
this happens only at very small distances .
this effect reduces the large electric field inside the hydrogen atom calculated from the electromagnetic form - factors via the maxwell equations .
the energy content of the field is below the pair production threshold .
* inside the hydrogen atom * m. nowakowski , n. g. kelkar , d. bedoya fierro and a. d.
bermudez manjarres + departamento de fisica , universidad de los andes + cra .
1e , 18a-10 , bogot , colombia |
the analysis of the screening in @xmath0plasma performed keeping the analogy with the classical treatment of an electric plasma @xcite gave as result a slight difference in the screening behavior with distance , with respect to the more standard , exponential decay @xcite .
now this analysis is pushed further by introducing the effect of the real gluons which are certainly present in such situations .
once the relevance of the gluons in the quark - quark correlation functions is taken into account one in lead , unavoidably , to consider also the quark - gluon correlations and this is performed in the present paper . while a definite quark population is taken as a initial condition of the problem , the gluon population is considered of thermal origin and is given therefore in terms of a bose - einstein density , together with this population also the creation of @xmath0pair is considered , with a corresponding fermi - dirac density .
being interested in the large distance effects , that part of the interaction of real gluons with quark is considered , which correspond to a @xmath1channel exchange of a virtual gluon , so that it has , suitably redefining color and spin factors , the same form of quark - quark interaction .
the result is a general behavior of the correlation function not very different from that previously found in a pure @xmath0plasma , since the analytical form is however awkward a numerical investigation has been performed for both situations without gluons and with gluons .
the behavior differs little from the exponential , in both case the possibility of small oscillations is found , but only when the correlation function is already very small .
a discussion of the range of values for which the model can be meaningful and also a comparison with some different treatments and results , in particular with the formulation using the termal green functions are briefly presented @xcite .
three short appendices contains some details of the needed calculations .
as it has been explained in the introduction the basic equation for the correlation functions have the same general features as for the pure @xmath2 populations , but we must distinguish two possible correlation , the two - fermion case and the fermion - boson case . in principle
we should look also at the two - boson correlation function , since however our main interest is in the two - fermion correlations while keeping into account the boson - fermion correlation , that enter directly in the equation for the two - fermion function we neglect the effect of the pure bosonic correlation : in so doing we are forced to bound our discussion to situation where the fermion density is dominant , having in mind the nucleus - nucleus collision there are certainly situations , of not too large temperature , where this condition is satisfied .
the ways in which the usual determination of the shielding has been adapted to the case of non commuting charges have been presented in the provios paper and will not be discussed here , only the integrodifferential equation for the two body correlation is reproduced @xmath3\bigg ] \label{2-correlation}\end{aligned}\ ] ] where the variables @xmath4 embody both the space variables @xmath5 and the color indices and the integration over the auxiliary variable @xmath6 takes care of the non commutativity of the charges .
now a further derivative with respect to @xmath7 is calculated , then one performs the fourier transform with respect to the space variables and the laplace transform with respect to the inverse temperature with the result : @xmath8 \label{2-correlation - transform}\ ] ] the check ( as @xmath9 ) means both the fourier and the laplace transform ; the variable @xmath10 is the fourier - conjugated of @xmath5 and @xmath11 is the laplace - conjugated of the inverse temperature @xmath12 , @xmath13 .
the factors @xmath14 and @xmath15 are matrices so their order is relevant . in order to proceed
is is necessary to specify the colour structure of the quarks and of the gluons , so the indices will now must now be displayed and the precise meaning of @xmath15 and @xmath9 is specified .
it is not convenient to work with two set of indices , for quark and gluon respectively , so the color charge of the gluon is represented by a traceless tensor with triplet indices , like @xmath16 with @xmath17 .
the term @xmath15 is specified in this way : the interactions @xmath18 and @xmath19 are given by [ app.a ] @xmath20\ ] ] the interaction @xmath2 is given by @xmath21 ; the interaction @xmath22 is given by @xmath23\ ] ] the interaction @xmath24 is given by @xmath25 .
the term @xmath9 is indicated with different symbols following its quark and gluon content and is decomposed according to the color structure : for @xmath18 and @xmath19 it is called respectively @xmath26 and @xmath27 and decomposed in terms of _ triplet _ and _ sextet _ for @xmath2 it is called @xmath28 and decomposed into _ singlet _ and _ octet _ and for @xmath22 or @xmath24 it is called @xmath29 or @xmath30 and decomposed into _ triplet , _ _ sextet _ and
_ 15-plet _ by means of the following projectors : @xmath31 & { } ^6\pi_{a , c}^{b , d}&=\frac{1}{2}\left[\delta_a^b\,\delta_c^d+\,\delta_a^d\,\delta_c^b\right]\nonumber\\
{ } ^1\pi_{a , c}^{b , d}&=\frac{1}{3}\,\delta_a^d\,\delta_a^d & { } ^8\pi_{a , c}^{b , d}&=\left[\delta_a^b\,\delta_c^d-\frac{1}{3}\,\delta_a^d\,\delta_c^b\right]&\end{aligned}\ ] ] @xmath32+\frac{1}{24}\,\delta_{a}^{b}\,\delta_{c}^{g}\,\delta_{f}^{d}\nonumber\\ { } ^6p_{a , cf}^{b , dg } & = \frac{1}{2}\,\left[\delta_{a}^{b}\,\delta_{c}^{d}-\delta_{a}^{d}\,\delta_{c}^{b}\right]\,\delta_{f}^{g}-\frac{1}{4}\,\left[\delta_{a}^{g}\,\delta_{c}^{d}\,\delta_{f}^{b } + \delta_{a}^{b}\,\delta_{c}^{g}\,\delta_{f}^{d } -\delta_{a}^{g}\,\delta_{c}^{b}\,\delta_{f}^{d } -\delta_{a}^{d}\,\delta_{c}^{g}\,\delta_{f}^{b}\right]\nonumber\\ { } ^{15}p_{a , cf}^{b , dg } & = \frac{1}{2}\,\left[\delta_{a}^{b}\,\delta_{c}^{d}+\delta_{a}^{d}\,\delta_{c}^{b}\right]\,\delta_{f}^{g}-\frac{1}{8}\,\left[\delta_{a}^{g}\,\delta_{c}^{d}\,\delta_{f}^{b } + \delta_{a}^{b}\,\delta_{c}^{g}\,\delta_{f}^{d } + \delta_{a}^{g}\,\delta_{c}^{b}\,\delta_{f}^{d } + \delta_{a}^{d}\,\delta_{c}^{g}\,\delta_{f}^{b}\right]\end{aligned}\ ] ] the correct normalization of the projector is easily verified @xmath33 for the different amplitudes , the equations corresponding to the general form of eq.([2-correlation - transform ] ) are @xmath34\\ -k^2\,q_{a , c}^{b , d } & = \frac{4\,\pi\,\alpha}{s^2}\,i_{a , c}^{b , d}+\frac{4\,\pi\,\alpha}{s}\,\left[\left(q_{a , f}^{b , g}\,i_{g , c}^{f , d}+i_{a , f}^{b , g}\,q_{g , c}^{f , d})\right)\,\rho\right.\\ & + \left.\left(m_{a , f}^{b , g}\,(-i_{g , c}^{f , d})+(-i_{a , f}^{b , g})\,\bar m_{g , c}^{f , d})\right)\,\bar \rho+\left(b_{a , fm}^{b , gl}\,j_{c , gl}^{d , fm}+j_{a , fm}^{b , gl}\ , b_{c , gl}^{d , fm}\right)\,\gamma\right]\\ -k^2\,b_{a , cf}^{b , dg } & = \frac{4\,\pi\,\alpha}{s^2}\,j_{a , cf}^{b , dg}+\frac{4\,\pi\,\alpha}{s}\,\left[\left(q_{a , l}^{b , m}\,j_{m , cf}^{l , dg}+i_{a , l}^{b , m}\,b_{m , cf}^{l , dg}\right)\,\rho\right.\\ & + \left.\left(m_{a , l}^{b , m}\,(-j_{m , cf}^{l , dg})+(-i_{a , l}^{b , m})\,\bar b_{m , cf}^{l , dg})\right)\,\bar \rho \right]\label{color - eq } \end{split}\end{aligned}\ ] ] the coefficients @xmath35 give the densities of quarks , antiquarks ans gluons ; the equations for @xmath27 and @xmath30 can be obtained by interchange @xmath36 , @xmath37 and @xmath38 in the equations for @xmath26 and @xmath29 .
it is possible to give a graphical representation to eq.([color - eq ] ) , see fig.[color - eq - graph ] , through the rules set in fig.[rules - equation ] ; it is consistent with the colour structure to use lower indices for incoming quarks and outgoing antiquarks and upper indices for incoming antiquarks and outgoing quarks .
since the gluon colour structure is given in the spinorial version [ app.a ] , the gluon is represented in the form of a @xmath2 pair , where up - arrow and down - arrow are used respectively for quarks and antiquarks .
we recall that the interaction term @xmath39 is pure imaginary ; in the last of eq.([color - eq ] ) it can be seen that @xmath29 depends linearly on j , in the equations for @xmath28 and @xmath26 , however , @xmath29 and @xmath39 appear always together and always multiplied by @xmath40 , so it is useful to perform the substitutions @xmath41 ; in this way in the last equation we get a trivial factor @xmath42 , in the other two we get a change of sign in the terms that multiplies @xmath43 . in the whole set of eq ( 5 ) this amounts to eliminate the `` prime '' in @xmath39 and @xmath29 and to change the sign @xmath44 .
when we perform the decomposition @xmath45 we find contracting the interaction term @xmath39 with the projectors @xmath46 that the rhs of equations ( [ color - eq ] ) for the ( @xmath47 ) systems contain only the @xmath48 tensor ; in fact in any case an octet is exchanged in the t - channel , in this way the relations @xmath49 hold as in the case without gluons and we get the equations : @xmath50=0\nonumber\\ & k^2\,f_3-\frac{8\,\pi\,\alpha}{3\,s^2}+\frac{4\,\pi\,\alpha}{s}\left[f_3\,\rho+ f_1\,\left(\frac{\bar \rho}{2}\right)-2\left(b_3+b_6\right)\,\gamma\right]=0\label{f1-f3-f3bar } \\ & k^2\,\bar f_3-\frac{8\,\pi\,\alpha}{3\,s^2}+\frac{4\,\pi\,\alpha}{s}\left[\bar f_3\,\bar \rho+ f_1\,\left(\frac{\rho}{2}\right)+2\left(\bar b_3+\bar b_6\right)\,\gamma\right]=0,\nonumber\end{aligned}\ ] ] where @xmath51 and @xmath52 have been substituted by means of the following relations : @xmath53 the system ( [ f1-f3-f3bar ] ) yields immediately @xmath54 and is reduced to a two - equation system : @xmath55+\bar f_3\,\left[\frac{2\,\pi\,\alpha\,\bar \rho}{s}\right]-\frac{8\,\pi\,\alpha}{3\,s^2}-\frac{8\,\pi\,\alpha}{s}\,\gamma\,(b_3+b_6)=0\nonumber\\ & f_3\,\left[\frac{2\,\pi\,\alpha\ , \rho}{2\,s}\right]+ \bar f_3\left[k^2+\frac{4\,\pi\,\alpha}{s}\,\left(\bar \rho+\frac { \rho}{2}\right)\right]-\frac{8\,\pi\,\alpha}{3\,s^2}+\frac{8\,\pi\,\alpha}{s}\,\gamma\,(\bar b_3+\bar b_6)=0.\nonumber\end{aligned}\ ] ] from the previous linear system it results that @xmath56 and @xmath57 can be expressed in terms of @xmath58 and @xmath59 , as follows : @xmath60 where @xmath61 and @xmath62
. furthermore we get @xmath63 @xmath64 projecting the equations satisfied by @xmath29 and @xmath30 with @xmath65 the following equations are obtained : @xmath66=0\label{eq : b3}\\ & k^2\,b_6-\frac{2\,\pi\,\alpha}{s^2}+\frac{2\,\pi\alpha}{s}\,\left[\frac{3}{4}\,f_3\,\rho+\frac{3}{8}\,f_1\,\bar \rho+(b_6\,\rho-\bar b_6\,\bar \rho)\right]=0\label{eq : b6}\ ] ] the equation for @xmath67 can obtained by the interchange @xmath68 for @xmath69 , @xmath70 and @xmath71 in the equations for @xmath72 .
after some tedious calculations , the following two equations for @xmath58 and @xmath59 variable are obtained : @xmath73\,(b_3+b_6)+\left[\frac{2\,\pi\,\alpha}{s}\,\left(3\,f_3^{(2)}\,\rho+\frac{3}{2}\,f_1^{(2)}\,\bar \rho-\bar \rho\right)\right]\,(\bar b_3+\bar b_6)\nonumber \\\quad&-\frac{8\,\pi\,\alpha}{s^2}+\frac{6\,\pi\,\alpha}{s}\,f_3^{(0)}\,(\rho+\bar \rho)=0\nonumber\\ & \left[\frac{2\,\pi\,\alpha}{s}\,\left(3\,\bar f_3^{(1)}\,\bar \rho+\frac{3}{2}\,f_1^{(1)}\ , \rho+\rho\right)\right]\,(b_3+b_6)+\left[-k^2+\frac{2\,\pi\,\alpha}{s}\,\left(3\,\bar f_3^{(2)}\,\bar \rho+\frac{3}{2}\,f_1^{(2)}\ , \rho-\bar \rho\right)\right]\,(\bar b_3+\bar b_6)\nonumber \\\quad&-\frac{8\,\pi\,\alpha}{s^2}+\frac{6\,\pi\,\alpha}{2\,s}\,f_3^{(0)}\,(\rho+\bar \rho)=0\nonumber\end{aligned}\ ] ] solving the previous linear system we have : @xmath74 substituting the previous expressions in eq.([eq : f3-f3bar ] ) we have : @xmath75 from eq.([eq : b3],[eq : b6],[eq : b15],[f3-solution ] ) it can be easily verified that : @xmath76 all the @xmath77correlations may be expressed in term of @xmath56 , so we evaluate its laplace antitrasform with the following result : @xmath78\ , \left[1-e^{-\frac{3\,\pi\,n\,\alpha\,\beta}{k^2 } } \left(\cos\left(\frac{3\,\pi\,n\,\alpha\,\beta\,y } { k^2}\right)- \frac{1 - 12\,\gamma /n}{3\,y\,(1 + 12\,\gamma /n)}\ , \sin\left(\frac{3\,\pi\,n\,\alpha\,\beta\,y}{k^2}\right)\right)\right ] \label{f3laplace}\end{aligned}\ ] ] where @xmath79 .
the inversion of the fourier transform leads to an expression that is very little transparent , it is reported for completeness in [ app.c ] ; the correlation function in real space has been computed numerically , by means of standard integration subroutines @xcite , starting form the following integral @xmath80 in fig.[qq - correlation ] and fig.[qq - gg - correlation ] the correlation function without and with gluons for different choices of temperature , at fixed coupling constant @xmath81 and initial quark density @xmath82 , is shown . in both the cases it can be noted that the damping of the correlation , if not exponential , differs very little from that behavior ; there is also a long distance @xmath5 region ( with @xmath83 at @xmath84 ) , where the correlation begin to oscillate , but the amplitude of these oscillations is very small , being present when the correlation has been already much damped .
the previous plots give a good overall description of the behavior of the correlation function , but do not show immediately how these functions differ from a yukawa shape @xmath85 ; in order to give a more complete description of this property in fig.[qq - gg - correlation - lenght ] the expression @xmath86 is plotted .
if the yukawa shape was exact , we would find a straight line ; these plots confirm ( obviously ) that the gluons make the damping weaker , moreover the correlation function becomes slightly farther from a yukawa shape .
it can be seen that the effect associated to the gluon presence is to produce , at fixed temperature , a perceptible increase in the value of the correlation and consequently a displacement towards higher @xmath5 values of the region where the correlation function begin to oscillate . for a detailed analysis of the values of energies , quark , anti - quark and gluon densities
, we refer to the next section and [ app.b ] .
the inclusion of the gluon in the effect of mutual screening in a @xmath87plasma refines a previous analysis given in term of a pure quark - antiquark population , and it confirms the results . in particular
the correlation length and the shape of the damping is still the same for @xmath18 and @xmath2 , so that it must reflect in the same way in the meson and baryon production .
an item that requires some condideration is the comparison with other ways of dealing with the same phenomenon .
the treatment given in term of strong coupling on the lattice is difficult to compare with @xcite because it is very different since the beginning , more similar are the treaments in terms of thermal green function as in @xcite . in that case one extracts the contribution of the pole of the propagator in momentum space so that a precise yukawa - like decay of the correlation function is certainly produced .
the treatment here presented is purely static , so it seems that it lacks of some effects that are present in the other treatment , does it contains also someting more ?
perhaps what is more is seen looking at the structure of the coupled equation ( [ color - eq ] ) : the interaction of one particle takes place with other particles which are already correlated , so that at least at the level of two - fermion distribution a self consistent treatment of the correlations is performed .
it must be noted that the dependence of the shielding effect on the temperature has two origins : one is the kinematical effect that is present in every plasma - like system , the other is dynamical and typical of a relativistic system since it comes from the thermal production of particles , both gluons and @xmath2 pairs .
the two processes have not the same role because initially there are more quarks than antiquarks , however , looking at the evolution of the chemical potential fig.[betamu ] , we see that there is an equivalent temperature , not extremely high , at which the effects of the initial condition become very small .
one could also ask what the answer would be in the opposite limit , low temperature and high density , which might be realized inside a neutron star , but in this situation the pauli principle would be very relevant and one should start from the consideration of a degenerate relativistic plasma , what is not attempted in this paper . as explained previously only the quark - gluon interaction corresponding to a gluon exchange in @xmath1channel is considered . for this amplitude tha color structure
is usually given in mixed form @xcite as @xmath88 where @xmath89 and @xmath90 .
this form is not convenient here and will be transformed in a purely spinorial version . using the fundamental commutator @xmath91=i f_{abc}t_c$ ]
we express the interaction as commutator ; then defining the matrices with spinorial indices @xmath92\ ] ] the form of the interactions used in this paper for @xmath18 and for @xmath22 is @xmath20\qquad\qquad j_{a , cf}^{b , dg}=\frac{1}{2\,i } \left[\delta_a^d\,\delta_c^b\,\delta_f^g- \,\delta_a^g\,\delta_c^d\,\delta_f^b\right]\ ] ] note that the antisymmetry of @xmath39 with respect to the exchange @xmath93 together with the factor @xmath94 ensures the hermiticity of the interaction . when we take care of the fact that there can be a thermal production of gluons we can not ignore the concurrent production of quark - antiquark pairs .
we investigate briefly the problem in this form .
assume the baryonic density , @xmath95 the quark density minus the antiquark density , as given and find the simultaneous production of gluons and of quark - antiquark pairs .
the expression for the baryonic density is simple when the quark mass can be neglected , or gives at most a small correction , in this case for the density we get : @xmath96 \label{initial - density}\ ] ] in particular the antiquark density is given by @xmath97\ ] ] where @xmath98 and for the weight we get @xmath99 resulting from a factor 2 for the spin , a factor 3 for the color , a factor 2 for the flavor because we consider only @xmath100 and @xmath101 quarks . from eq ( [ initial - density ] )
it follows : @xmath102 from this relation we see the conditions qualitatively well known in which the mass term is negligible : either @xmath103 or @xmath104 , form now on we shall assume that at least one of these conditions is fulfilled and the mass term is dropped . solving the equation for @xmath105 we find @xmath106^{\frac{2}{3}}-3^{\frac{1}{3}}\,\pi^2}{3^{\frac{2}{3 } } \,\left[\left(\frac{27\,\pi^2\,x}{g_f}\right ) \left(1+\sqrt{1+\left(\frac{\pi\,g_f}{9\,\sqrt 3\ , x}\right)^2 } \right)\right]^{\frac{1}{3}}}\ ] ] with @xmath107 . in this way we get the actual expression of @xmath108 in terms of @xmath109 and @xmath12 ,
the total density which appears in the previous formulae is given by @xmath110 .
it is now possible to give a semi - quantitative analysis of the different densities which are relevant for our problem : we start assigning a density @xmath109 which we take equal to 2 @xmath111 . in fig.[betamu ] , fig.[ndensity ] and fig.[gammanratio ] the values of @xmath112 , @xmath113 and the ratio @xmath114 as functions of @xmath115 are shown .
we have worked out three cases : the first is chosen so that the density of antiquarks is much less than @xmath109 , we take @xmath116 , this gives @xmath117 and for the corresponding gluon density @xmath118 , in the second case we take @xmath119 , this gives @xmath120 and @xmath121 , in the third one @xmath122 is considered , which produces @xmath123 and @xmath124 , which is the saturation
value for @xmath125 ratio . in these expression the density of gluons has be taken as the free - boson thermal density , which amounts to : @xmath126 here @xmath127 a factor 2 for the spin and a factor 8 for the color .
since we have @xmath128 for @xmath129 it results that in that limit @xmath130 and this limit is numerically reached already at @xmath131 mev .
the inversion of the fourier transform @xmath132 implies a standard angular integration and then an integral over the radial coordinates that gives @xmath133\end{aligned}\ ] ] where the generalized hypergeometric functions are used @xmath134 and the coefficient are @xmath135 the function @xmath136 is real because it is the sum of two terms with their complex conjugate .
note that for very low values of @xmath115 , @xmath137 could become imaginary , in this case all the addenda would be separately real .
99 g. calucci , eur .
j. c * 36 * , 221 - 226 ( 204 ) .
birse , c.w .
kao , g.c .
nayak , screening and antiscreening in an anisotropic qed and qcd plasma [ hep - ph/0304209 ] ; o. philipsen , what mediates the longest correlation length in the qcd plasma ? [ hep - ph/0301128 ] .
o. philipsen , phys . lett .
b * 521 * , 273 ( 2001 ) .
j. kogut , l. susskind , phys .
d * 11 * , 395 ( 1975 ) .
m. le bellac , thermal field theory , ( cambridge u.p .
, cambridge ( 1996 ) ) .
e. leader , e. predazzi , an introduction to gauge theories and modern particle physics ( cambridge u.p .
, cambridge 1996 ) , app.2 . g. p. lepage , j. comp . phys .
* 27 * , 192 ( 1978 ) . | the color screening in a qcd plasma , that was studied in a formulation making evident similarities and differences with the electric case , is continued by taking into account the contributions of real gluons .
the results , which include a numerical analysis not previously performed , show a damping of the correlation function which , if not exponential , does not differ very much from that form . the role of the temperature , which affect both the population and the dynamics of the quark - gluon system , is found to be relevant . |
here we calculate the ground - state physical properties from the variational method , including the energy @xmath38 , the mean photon number @xmath116 , the coupling correlation @xmath117 and the spin flipping ( tunneling ) strength @xmath118 .
as introduced in the main part of the paper , the wavefunction for the reformulated hamiltonian has the following form @xmath119 where @xmath120 is the parity .
we adopt the variational trial wavefunction as a superposition of the polaron and the antipolaron @xmath121 where @xmath122 with @xmath123 being the @xmath124th eigenstate of the standard quantum harmonic oscillator with frequency @xmath125 . in this work we focus on the ground state so that @xmath126 and @xmath127 .
the displacement of the bare potential @xmath30 in the single - well energy @xmath128 @xmath129 is driven by the interaction @xmath20 and for simplicity we have assumed the unit @xmath130 . note that the polaron ( antipolaron ) has a displacement in the same ( opposite ) direction as ( to ) that of the bare potential @xmath131 .
the interplay of the interaction and the tunneling leads to the deformation of the wavepacket : the frequency of the polaron ( antipolaron ) is renormalized by @xmath132 ( @xmath133 ) and the displacement by @xmath134 ( @xmath135 ) , respectively .
the weights of the polaron and the antipolaron are subject to the normalization condition @xmath136 .
these deformation parameters , independently @xmath137 , are optimized by minimization of the total energy formulated in the following .
the energy can be directly obtained as @xmath138 where @xmath139 contribute to the single - well energy and the tunneling energy , respectively . here , we have defined @xmath140 for @xmath141 , while @xmath142 is a constant energy .
explicit formulas for these quantities are readily available . in this section
we give the result for the ground state @xmath143 , \label{haaa}\\ h_{\beta \beta } ^{+ } & = & \frac 12\omega \left [ \frac 12(\xi _ \beta + \xi _ \beta ^{-1})+\left ( 1-\zeta _ \beta \right ) ^2g^{\prime 2}\right ] , \label{habb}\\ h_{\alpha \beta } ^{+ } & = & \frac 12\omega \left [ \left ( 1-\xi _ \alpha ^2\right ) \left\langle \hat x_\alpha ^2\right\rangle _ { \alpha \beta } + \left ( 1-\zeta _ \alpha \right ) \left\langle \hat x_\alpha \right\rangle _ { \alpha \beta } 2g^{\prime } \right .
\nonumber\\ & & \left .
\alpha s_{\alpha \beta } + \left ( 1-\zeta _ \alpha \right ) ^2g^{\prime 2}s_{\alpha \beta } \right ] , \label{haab}\end{aligned}\ ] ] where @xmath144 \label{x2aab}\end{aligned}\ ] ] and @xmath145 are given by the function @xmath146 ^{1/4}.\end{aligned}\ ] ] from the relation @xmath147 and the symmetric relation @xmath148 , the mean photon number simply reads as @xmath149 where @xmath150 for the ground state @xmath151 , \\ h_{\beta \beta } ^0 & = & \frac 12\omega \left [ \frac 12(\xi _ \beta ^{-1}+\xi _
\beta ) + 2\zeta _
\beta ^2g^{\prime 2}\right ] , \\ h_{\alpha \beta }
^0 & = & \frac 12\omega \left [ \left ( 1-\xi _ \alpha ^2\right ) \left\langle
\hat x_\alpha ^2\right\rangle _ { \alpha \beta } -\zeta _
\alpha \left\langle \hat x_\alpha \right\rangle _ { \alpha \beta } 2g^{\prime } \right.\nonumber \\ & & \left.+\xi _ \alpha s_{\alpha \beta } + \zeta _ \alpha ^2g^{\prime 2}s_{\alpha \beta } \right ] , \end{aligned}\ ] ] and @xmath152 , @xmath153 are given by and .
now we calculate the coupling correlation @xmath156 .
since @xmath157 , we have @xmath158 where @xmath159 and @xmath160 the strength of spin flipping or tunneling , @xmath161 , is simply @xmath162 which has been formulated in . .
@xmath64 is taken as an example . *
a * , the ground state energy . * b * , the mean photon number . * c * , the correlation function @xmath163 . * d * , the tunneling strength @xmath164 .
the orange circles denote the numerically exact results as a benchmark , the red dashed - lines are calculated in adiabatic approximation(aa ) @xcite or generalized rotating - wave approximation ( grwa ) @xcite , and the blue lines are our results obtained by the present variational method . ]
the most widely - used approximations in the literature are the rotating - wave approximation ( rwa)@xcite , adiabatic approximation(aa ) @xcite , generalized rotating - wave approximation ( grwa ) @xcite and generalized variational method ( gvm ) @xcite , each working in some specific parameter regime .
the rwa neglects the counter - rotating terms in the interaction , valid in regime @xmath165 under near - resonance ( @xmath166 ) condition .
the aa and the grwa have the same groundstate , working for @xmath167 or negative detuning ( @xmath168 ) regime .
the gvm works for @xmath169 .
recently a mean - photon - number dependent variational method was proposed to cover validity regimes of both the gvm and the grwa @xcite . however , all the approximations collapse when the ratio of @xmath170 is getting small , e.g. below around @xmath171 ( see ref . [ ] )
. an improved variational method by including the antipolaron@xcite also finds breakdowns at @xmath172
. it would be favorable to have a variational method that always preserves a high accuracy in varying all parameters which might facilitate and even deepen the physical analysis . indeed , our variational wavefunction yields such accuracy requirements . as an illustration , in fig .
[ fig - phyquan ] we compare with the exact numerics on the the ground state energy , mean photon number , coupling correlation and tunneling strength , at the example @xmath173 ( one can find other examples for comparison at @xmath174 , @xmath175 , @xmath176 , @xmath171 for another physical quantity @xmath177 in fig . [ fig - scaling ] and fig . [ fig - sij ] ) . as a comparison ,
the results obtained by the aa or the grwa are also shown
. clearly , our results are completely consistent with the exact ones in the whole parameter regime .
( black line ) or @xmath178 ( red line ) , an incorrect cusp behavior appears in the energy @xmath38 and the spin flipping ( tunneling ) strength @xmath118 has a spurious jump around @xmath49 at @xmath179 , in contradiction with the smooth crossover in the exact numerics ( orange circles ) .
the blue lines are our results in full minimal parameters which reproduce accurately the exact ones . ] and
the orange dots , empty blue diamonds and empty red squares represent the results of the exact numerics , our method and the grwa , respectively . ]
it may be worthwhile to have further discussion on the physical necessity of the variational parameters .
an unnecessary reduction of our parameters , on the one hand , will not lead to much of a reduction in the computational cost as the calculation in full parameters is actually quite easy to carry out , on the other hand , however , the price of physical loss would be too high . as discussed in the main part of the paper , our variational parameters physically correspond to the deformations of the polaron and the antipolaron with displacement and frequency renormalizations , which is justified by the behavior of the effective potential . in the subtle energy competitions of the potential of harmonic oscillator ,
the interaction and the tunneling , both the polaron and the antipolaron can adapt themselves via the variations of their displacements , frequencies and weights .
thus , corresponding to the key physical degree of freedom of the polaron and the antipolaron , the five variational parameters , @xmath181 , are the minimal necessary parameters to capture the true physics of the behavior of the polaron and the antipolaron , subject to the normalization of the wavefunction . therefore ,
reducing the parameters would lead to mismatch of the physical degree of freedom and thus give rise to unreliable results , the physical properties may deviate not only quantitatively but also qualitatively . for an example , assuming @xmath182 or imposing @xmath178 can reduce the parameter number by 2 .
however , as shown in fig.[fig - paramtnecessity ] , without mentioning the quantitative deviations , an incorrect cusp behavior appears in the energy @xmath38 at low frequencies as illustrated at @xmath179 , and even worse , a spurious jump emerges in the tunneling ( spin flipping ) strength @xmath118 around @xmath49 .
the other physical quantities , such as the mean photon number @xmath116 , the coupling correlation @xmath117 also have a false discontinuity similar to @xmath118 .
both the cusp and the discontinuity are qualitatively in contradiction with the smooth crossover in the exact numerics ( orange circles ) .
nevertheless , our results using the full minimal variational parameters ( blue line ) reproduce accurately the exact results in the entire regime of the coupling strengths at different frequencies .
moreover , in the cases of reduced parameters , some important underlying physics would also be missing , such as the scaling relation of the displacement and frequency renormalizations as we revealed in the main text ( see also appendix [ app - c ] ) .
our method can also be useful for the excited states . as a first simple extension
the variational energy of the excited state can be obtained by replacing expressions - with @xmath183 , \\ h_{\beta \beta } ^{+ } & = & \frac \omega 2\left [ \left ( n+\frac 12\right ) \left ( \xi
_ \beta + \xi _ \beta ^{-1}\right ) + \left ( 1-\zeta _ \beta
\right ) ^2g^{\prime 2}\right ] , \\
h_{\alpha \beta } ^{+ } & = & \frac \omega 2\ \left [ \left ( 1-\xi _ \alpha ^2\right ) \left\langle
\hat x_\alpha ^2{}\right\rangle _ { \alpha \beta } + \left ( 1-\zeta _ \alpha \right ) \left\langle \hat x_\alpha \right\rangle _ { \alpha \beta } 2g^{\prime } \right .
\nonumber \\ & & + \left .
( 2n+1)\xi _ \alpha s_{\alpha \beta } + \left ( 1-\zeta _ \alpha \right ) ^2g^{\prime 2}s_{\alpha \beta } \right],\end{aligned}\ ] ] where both @xmath184 and @xmath185 can be included by a unified function @xmath186 here the function @xmath187 is defined by @xmath188^j\sum_{p=0}^{\min [ n , j]}\sum_{q=0}^{\min [ n , j - p]}\frac{\left ( -i\right ) ^{j - p - q}a^pb^q}{p!q!\left ( j - p - q\right ) ! } \sqrt{\frac{2^{p+q}}{\left ( n - p\right ) !
\left ( n - q\right ) ! } } h_{j - p - q}\left(\frac 12ab^2c\right)\tilde s_{n - p , n - q } , \\
\tilde s_{k , k^{\prime } } & = & \sum_{r=0}^{\min [ k , k^{\prime } ] } c_{kk^{\prime } r}h_{k - r}\left(\frac{ab^2c}{2\sqrt{1-a^2}}\right)h_{k^{\prime } -r}\left(-\frac{a^2bc}{2% \sqrt{1-b^2}}\right ) , \\ c_{kk^{\prime } r } & = & \sqrt{\frac{ab}{2^{k+k^{\prime } } k!k^{\prime } ! } } % e^{-\left ( abc\right ) ^2/4}\frac{k!k^{\prime } ! \left ( 2ab\right ) ^r\left ( 1-a^2\right ) ^{(k - r)/2}\left ( 1-b^2\right ) ^{(k^{\prime } -r)/2}}{\left ( k - r\right ) ! \left ( k^{\prime } -r\right ) ! r!}.\end{aligned}\ ] ] and the factors @xmath189 depend on the variational parameters @xmath190 for the other group of overlap in the tunneling term @xmath191 , one can also formulate using @xmath192 with the corresponding replacement @xmath193 , but there is sign variation @xmath194 . here @xmath124 is the level number of the standard quantum harmonic oscillator and @xmath195 is the standard hermite polynomials .
it is worthwhile to see that this simple extension for the excited states has already yielded some considerable improvements in strong couplings as illustrated in fig.[fig - excited ] for a number of lowest energy levels . with the above expressions , one
may further analytically construct an improved extension of the variational energy for overall coupling range by imposing the deformed polaron and antipolaron in the grwa form of wavefunction . on the other hand ,
the dynamics of the system also can be calculated in terms of @xmath196 which provides the intra - overlap and inter - overlap of the deformed polarons and antipolarons with different oscillator quantum number @xmath197 . since here
the focus is the ground state which , as we show in the present work , already has rich underlying physics to be uncovered , we shall present a more detailed method description and systematical discussion for the excited state properties in our future work .
in the variation of the coupling strength , the system undergoes a phase - transition - like changeover around @xmath198 . in the super - strong tunneling or low - frequency limit ,
i.e. @xmath199 this changeover is very sharp , it behaves more like a phase transition , as discussed by ashhab @xcite . in the other cases it behaves like a crossover .
we can get more insights into this transition - like behavior from the profile deformation of the wavepacket .
the increase of the coupling strength is splitting the wavepacket into the polaron and the antipolaron , while the tunneling is trying to keep them as close as possible in the groundstate . before a full
splitting the system remains in a quadpolaron state with four channels of tunneling , @xmath200 , @xmath201 , @xmath202 , and @xmath203 , while after the splitting the system enters a bipolaron state with only two tunneling channels , @xmath204 and @xmath205 , surviving .
here we have labeled the tunneling channels by the overlaps @xmath206 , defined in , to which the corresponding tunneling energies are proportional .
we show the tunneling channel difference for these two regimes in fig.[fig - sij ] a - c at various frequencies .
one can also see that the change in the tunneling channel number is universal for different frequencies .
thus , the two regimes distinguished by wavepacket splitting are essentially different in the quantum states .
therefore , the coupling strength at which the splitting really starts can be used to formulate @xmath49 . , @xmath207 , @xmath208 and @xmath209 ( @xmath210 ) as functions of the coupling strength @xmath211 .
the dashed lines roughly separate the quadpolaron ( @xmath212 ) regime and the bipolaron regime ( @xmath213 ) , the former has four channels of tunnelings while the latter has two channels . *
d - f * , the frequency renormalization factors @xmath214 and @xmath215 .
* g - i * , the scaling quantity @xmath98 .
the results from our variational method ( solid lines ) almost reproduce the ones from exact numerics ( orange circles ) for all values of coupling at different frequencies .
the boundary of the quadpolaron and bipolaron regimes is associated with the minimum of @xmath57 and the maximum of @xmath98 .
the black triangles mark the positions for @xmath216 which becomes farther away from @xmath217 as @xmath18 increases .
* a , d * @xmath77 .
* b , e * , @xmath218 .
* c , f * , @xmath219 . ]
+ 0.0 cm we adopt the value of @xmath49 at the point where the distance between the polaron and the antipolaron is equal to their total radii , @xmath220 where we take the radii by @xmath221 at which the value of the corresponding wavepacket is becoming small @xmath222 for both @xmath223 . note that both sides of the above equation are essentially averaging over the polaron and the antipolaron , thus assuming symmetric polaron and antipolaron , i.e. @xmath224 and @xmath225 , would be a reasonable approximation as far as @xmath49 is concerned . under this constraint
the explicit results for the deformation parameters are available for the well - separated polaron and antipolaron from the energy minimization formulated in appendix [ app - a ] , reading as @xmath226 where the critical point @xmath227 is obtained in the semiclassical approximation at @xmath228@xcite we stress that we limit the application of this approximation to the estimation of @xmath49 , while for other properties one should fall back upon asymmetric polaron and antipolaron for higher accuracy . actually ,
as mentioned in appendix [ app - a ] , imposing symmetric polaron and antipolaron would lead to a spurious discontinuous behavior of physical properties , such as the tunneling strength , around @xmath49 at low frequencies , while in reality it should be smooth as predicted by asymmetric polaron and antipolaron in agreement with exact numerics .
also , in the strong coupling regime the displacement asymmetry of the polaron and the antipolaron actually plays an important role in inducing the amplitude - squeezing effect ( @xmath229 ) which extends the wavepackets of the polaron and the antipolaron to increase their overlap , thus enhancing the tunneling . without the asymmetry
there would be no squeezing beyond @xmath49 , as indicated by , since the symmetric polaron and antipolaron in up and down spins would completely coincide , with an already - maximum overlap .
in fact , as uncovered in the main text , there is a hidden relation between the squeezing and the displacement , which is also discussed in detail in appendix [ app - c ] .
substitution of into leads us to a simple analytic expression @xmath230 it is easy to check @xmath231 in the slow oscillator limit @xmath232 . besides the transition - like changeover in this low frequency limit
, our @xmath49 is also providing a valid coupling scale for the quadpolaron / bipolaron changeover at finite frequencies , which can be seen from fig.[fig - sij ] where the quadpolaron regime and bipolaron regime adjoin each other really around @xmath49 . a more quantitative way to identify the transition - like point
is , as shown by fig.[fig - sij ] * d - i * , the minimum point of the frequency renormalization factor or the maximum point of the scaling quantity introduced in appendix [ app - c ] .
still , one sees that it is well approximated by @xmath49 in . at this point , it is worthwhile to further discuss the scale of the coupling strength , the criterion for which is actually a bit controversial in the literature@xcite .
although the terms for the coupling strengths were given in relation to the validity of the rwa as well as the progress of experimental accessibility , essentially the frequency @xmath125 has been conventionally taken as the evaluation scale : @xmath233 for the weak coupling regime , @xmath234 for the strong coupling , @xmath235 for the ultrastrong coupling regime@xcite , @xmath236 for deep strong coupling regime@xcite .
on the other hand , it should be noticed that recently it has been proposed @xcite that the strength scale should be modified to be the semiclassical critical point @xmath237 .
still , as afore - mentioned , @xmath237 is obtained in low - frequency semiclassical limit , while the situations at finite frequencies would be different .
the controversy essentially comes from the fact that a consensus on the nature of the interaction - induced variation in different frequencies is still lacking . here
, our expression of @xmath49 in is obtained by the observation that it is the wavepacket splitting that makes the essential change in increasing the coupling strength , which controls the final effective coupling tunneling strength and leads to transition ( in low frequency limit ) or crossover ( at finite frequencies ) of the quadpolaron / bipolaron states .
we believe that @xmath49 is a more universal scale valid for all frequencies , as indicated by fig .
[ fig - sij ] . under these considerations ,
we simply divide the coupling strength into weak , intermediate and strong regimes under the conditions that @xmath20 is smaller than , comparable to and larger than @xmath49 , respectively . as a reference
, we compare the different scales for the coupling strength used in the literature in fig.[fig - different - scales ] . and
@xmath238 , which has been reached by experiments in rapid progress@xcite . the conventional deep strong coupling regime ( the light - cyan area )
is surrounded by @xmath238 and @xmath239 ( gray dash - dot lines ) , into which investigations have been entering@xcite .
the black dotted line denotes the semiclassical critical - like point in low frequency limit , @xmath237@xcite , proposed as a different scale of coupling strength@xcite , while the blue solid line schematically represents the quadpolaron / bipolaron boundary @xmath49 as a novel scale generalized for the whole range of frequencies .
thus , the coupling strength is divided into weak , intermediate and strong regimes which correspond to that @xmath20 is smaller than , comparable to and larger than @xmath49 , respectively .
the orange - shaded window edged by the dash lines opens for the overweighted antipolaron discussed in our paper . ] + 0.0 cm
when the coupling strength grows beyond @xmath49 , we find that the squeezing factor @xmath76 and the displacement factor @xmath240 begin to collapse into the same values and scale with each other in the further evolution , i.e. @xmath241 this hidden scaling relation can be more explicitly formulated at low frequencies
. note that the parameters can be extracted from the energy minimization @xmath242 in the bipolaron regime , only the polaron - antipolaron tunneling remains so the overlaps @xmath91 and @xmath87 are vanishing , but @xmath243 is finite .
in such a situation , controlling the polaron - antipolaron center of mass , @xmath244 , can be decoupled from the relative motion in tunneling and squeezing , which enables us to extract the weight of the polaron , @xmath245 to obtain analytical results we assume a low frequency which enables a small-@xmath125 expansion and leads us to @xmath246 where @xmath132 ( @xmath247 ) takes the sign @xmath27 ( @xmath10 ) . in the small-@xmath248 limit , @xmath76 and @xmath74 collapse into their average @xmath249 and @xmath250 which are equal : @xmath251 up to @xmath252 order .
we can see the scaling relation from the approximate analytic results : ( i ) in low - frequency limit , one sees that @xmath253 holds , up to an @xmath252 order correction which is negligible for small @xmath125 .
( ii ) for higher frequencies , the @xmath252 terms in @xmath254 and @xmath255 become almost the same due to @xmath256 , since in bipolaron regime we have @xmath257 ( e.g. , for @xmath258 , @xmath259 while @xmath260 is negligible beyond the crossover range . ) .
these analytic considerations account for the scaling relation as we showed in the main text for different frequencies . to test the scaling relation , we shall propose a physical quantity that may be either measured experimentally or verified by exact numerics . on one hand , applying the above expansion to the photon number and neglecting
the difference of @xmath261 and @xmath262 leads us to an expression of @xmath263 as a function of @xmath264 and @xmath265@xmath266 on the other hand , the same approximation yields @xmath267 thus , considering the ratio @xmath268 , we introduce the following phyiscal quantity @xmath269 where we have defined @xmath270 . in the bipolaron regime with strong couplings ,
this quantity becomes the scaling ratio , @xmath271 . in this regime , if the scaling relation holds , the value of @xmath177 will be equal to 1 .
indeed , this scaling relation is confirmed by the exact numerics , as shown in the main text . in the quadpolaron regime with intermediate or weak couplings , not only the scaling relation is violated but also the relation between @xmath177 and @xmath268 is breaking down , @xmath272 .
nevertheless , we find that , besides the bipolaron regime having the value @xmath273 , the quadpolaron with four strong channels of tunneling is located in the range @xmath274 and the state with decaying left - right tunnelings ( @xmath200 , @xmath201 ) falls in a range @xmath275 , as one can see in fig.[fig - sij]a - c , g - i . in this sense , according to the values and behavior of @xmath177 , one can distinguish the quantum states of the bipolaron , quadpolaron and their changeover , respectively .
, in the strong coupling regime . as in this regime within the same spin component there is no overlap between the polaron ( @xmath61 , labelled by the blue arrow ) and antipolaron ( @xmath60 , labelled by the orange arrow ) in the two wells , to have both finite weights for the polaron and the antipolaron the two sub - well energies have to be degenerate , i.e. , @xmath276 .
here @xmath277 is the position of local minimum potential and @xmath278 ( scaled by @xmath279 ) , @xmath75 . ]
+ 0.0 cm .
@xmath280 is taken . @xmath57 and
@xmath58 almost take the same values in the strong coupling regime above the @xmath49 .
the inset shows their tiny differences by a zoom - in plot . ]
apart from variational method on energy minimization introduced in appendix [ app - a ] , an alternative way to see the scaling relation is to investigate the effective potential . as we discussed in the main text
, the eigenequation actually can be written in a single particle form @xmath281 where we have assumed that the particle mass @xmath282 and the total potential is composed of the bare potential @xmath283 and an additional effective potential @xmath284 induced by the tunneling , @xmath285 with @xmath286 and we have considered the ground state with @xmath287 . in the strong coupling regime , the total potential exhibits an obvious two - well structure , with a larger barrier separating the wells , as shown in fig .
[ fig - degeneracy ] . in the lowest order , the two wells can be considered as a local harmonic potential . actually , an expansion around the local minimum point @xmath288 of the potential leads to @xmath289 where @xmath290 with @xmath291 , and @xmath292 , @xmath293 respectively for the polaron and the antipolaron .
the coefficients are defined by @xmath294 first , the approximation of local harmonic potential requires @xmath295 the condition-1 ensures the potential minimum point at @xmath296 , while the condition-2 indicates the same renormalized frequency @xmath297 of the local harmonic potential as that of the wavefunction of the harmonic oscillator .
furthermore , since effectively there is no single - particle inter - site hopping ( if regarding the wells as two sites ) in the presence of the large barrier in the strong coupling regime , to have finite weights for both the polaron and the antipolaron in the single - particle effective potential the local energies of the two wells need to be degenerate @xmath298 where @xmath299 is the energy of the local harmonic oscillator scaled by @xmath279 and @xmath300 corresponds to the reference energy .
taking the variational wavefunction ( [ wf-1 ] ) , in the strong coupling regime we have @xmath301 while @xmath302 in the strong coupling regime .
now one can ( i ) control the displacement renormalization @xmath240 to satisfy condition-1 so that the linear term @xmath303 is eliminated and the minimum is located at @xmath304 , ( ii ) tune the frequency renormalization @xmath76 to fulfill condition-2 so that both the local potential and the wavefunction self - consistently shares the same frequency @xmath305 , ( iii ) balance the weight ratio of @xmath306 to meet condition-3 so that the two wells have degenerate local energies to self - consistently guarantee the finiteness of the weights @xmath307 and @xmath308 . at this point
, we see that the degree of the basic deformation factors introduced for our variational wavefunction is consistent with the minimum requirements of self - consistence conditions . from conditions-1,2,3 we also refind the scaling relation as illustrated by fig.[fig - scalingfromveff ] , which might provide some alternative insights for the scaling relation that we obtained from the energy minimization in last subsection
. still , we should mention there is a small difference between the two ways , since the above consideration from the effective potential is based on the lowest order expansion which guarantees only the local potential itself to be harmonic without taking care of the energy , while the energy minimization ensures only the most favorable energy but the effective potential @xmath309 includes higher order terms beyond the harmonic potential approximation . despite the small difference ,
both approaches lead to the scaling relation . with the scaling relation at hand
, it might provide some more insight to discuss the quantum state changeover from the symmetry point of view .
generally , for all eigenstates , the hamiltonian has the parity symmetry , @xmath310 which involves simultaneous reversion of the spin and the space . specifically for the ground state that we are focusing on in this work
, one could find extra symmetries .
in fact , in the low frequency limit the photon number vanishes for the ground state below @xmath49 , as indicated by fig.[fig - phyquan]b ( this is more obvious for lower frequencies ) , so that additionally the total parity symmetry can be decomposed into separate spin reversal symmetry @xmath311 and oscillator spatial reversal symmetry @xmath312 .
these additional symmetries are broken beyond @xmath49 due to the emergence of a number of photons , so that there is a subsymmetry transition when the system goes across @xmath49 .
still , these spin and spatial subsymmetries are considered from the weak coupling side and become less valid at finite frequencies due to a nonvanishing photon number .
nevertheless , our finding of the scaling relation provides compensation but from the strong coupling side .
actually , as shown in fig.4 , the physical quantity we proposed , @xmath98 , demonstrates an invariant behavior beyond @xmath49 , which confirms the scaling relation and thus the symmetric aspect between the displacement and frequency renormalizations in this regime .
note that , as shown in last subsection , in this bipolaron regime the remaining left - left and right - right tunnelings ( i.e. polaron - antipolaron inter - tunnelings ) render both the polaron and the antipolaron to have finite weights , while to preserve finite weights as a quantum effect in the absence of left - right tunneling channels ( i.e. intra - polaron and intra - antipolaron tunnelings ) the polaron and the antipolaron have to maintain the displacement - frequency scaling relation . in other words ,
this displacement - frequency symmetry arises only in the absence of the left - right channels , and conversely , there will be no left - right channels if the symmetry is preserved there .
going from the bipolaron regime to the quadpolaron regime , this symmetry will be broken in the presence of the left - right tunneling channels . in this sense , besides the afore - mentioned parity subsymmetry breaking originating from the weak coupling side in the low frequency limit , for both low frequency limit and finite frequencies there is another hidden symmetry - breaking - like behavior in the changeover of the two quantum states stemming from the strong coupling side .
thus , it is interesting to see a deeper nature of the interaction that not only induces the bipolaron / quadpolaron quantum state changeover but also brings about the symmetry breaking .
in the two - mode case .
the parameters used are @xmath313 and @xmath314 .
our variational method ( solid line ) is in good agreement with exact numerics(dots ) . ]
our ground - state phase diagram obtained for the rabi model might also provide some insights or implications for the spin - boson model @xcite which is a multiple - mode version of the rabi model and has wide relevance to other fields , including the kondo model @xcite and the ising spin chain @xcite in condensed matter .
on the one hand , the bipolaron / quadpolaron changeover in the rabi model can provide insights for localized / delocalized transition in the spin - boson model .
in fact , the spin - boson model exhibits different behaviors in the ohmic , super - ohmic , and sub - ohmic spectra , which actually have different weights of distributions for low and high frequency modes .
note that , as discussed in our work on the nature of the interaction - induced variation , the bipolaron and the quadpolaron states respectively have blockaded and enhanced left - right tunnelings , which is closely related to the situation of the localized and delocalized states involved in the spin - boson model . as indicated by our ground - state phase diagram and the obtained @xmath49 expression ,
the same coupling could be located in different regimes depending on whether frequency is low or high .
our ground - state phase diagram and @xmath49 expression might provide a primary reference and some insights into the different behaviours of the ohmic , super - ohmic , and sub - ohmic spectra , since the distribution weights of low and high frequencies would make different contributions to blockaded and enhanced tunneling , thus affecting the competition in the quantum phase transition of the localized and delocalized states .
on the other hand , the overweighted antipolaron region might have some implication for the coherence - incoherence transition in the spin - boson model .
it has been found that , within the delocalized phase of the spin - boson model , there is possibly another coherence - incoherence transition @xcite for which the nature is still not very clear .
interestingly , in our ground - state phase diagram of the rabi model , within the strong - tunneling regime in the quadpolaron state , there is also an underlying particular region characterized by an unexpected overweighted antipolaron , the possible implication and relation of the overweighted antipolaron regime in the rabi model and the coherence - incoherence transition in the spin - boson model might be worthwhile exploring . since in the present work our focus is on the single - mode rabi model , we would like to leave the investigations of these possible implications for the spin - boson model to some future works .
nevertheless , in the following we provide some indication of the method and variational wavefunction .
the basic variational physical ingredients introduced in the single - mode case also should apply for the multiple - mode case .
the treatments are readily extendable from the single - mode case .
the hamiltonian including @xmath315 modes of harmonic oscillators can be written as @xmath316 we propose the variational trial wavefunction as @xmath317=\alpha \prod_{k=1}^m\varphi _
\alpha ^k+\beta \prod_{k=1}^m\varphi _ \beta ^k,\ ] ] where @xmath318 ( @xmath319 ) is the @xmath115th mode polaron ( antipolaron ) under the direct extension @xmath320 .
the energy is simply that of the single - mode energy in with the overlap integrals replaced by the product of all modes .
we find the same accuracy as in the single - mode case , as illustrated in fig.[fig-2modes ] by an example of the two - mode case , for which exact numerics are available for comparison .
one can also include a bias term @xmath321 with a broken parity for @xmath322 at different spin @xmath323 .
the multiple - mode case deserves special investigations in detail which we shall discuss in future works .
chen , c. wang , s. he , t. liu , and k .- l .
wang , phys .
a * 86 * , 023822 ( 2012 ) .
f. alexander wolf , m. kollar , and d. braak , phys .
a * 85 * , 053817 ( 2012 ) .
b * 89 * , 121108(r ) ( 2014 ) .
s. bera , a. nazir , a. w. chin , h. u. baranger , and s. florens , phys .
b * 90 * , 075110 ( 2014 ) .
a. j. leggett , s. chakravarty , a. t. dorsey , m. p. a. fisher , a. garg , and w. zwerger , rev .
* 59 * , 1 ( 1987 ) .
f. guinea , v. hakim , and a. muramatsu , phys .
b * 32 * , 4410 ( 1985 ) . a. winter , h. rieger , m. vojta , and r. bulla , phys . rev .
lett . * 102 * , 030601 ( 2009 ) .
tong , j .- h . an , h .- g .
luo , and c. h. oh , phys . rev .
b * 84 * , 174301 ( 2011 ) . | the rabi model plays a fundamental role in understanding light - matter interaction .
it reduces to the jaynes - cummings model via the rotating - wave approximation , which is applicable only to the cases of near resonance and weak coupling .
however , recent experimental breakthroughs in upgrading light - matter coupling order require understanding the physics of the full quantum rabi model ( qrm ) .
despite the fact that its integrability and energy spectra have been exactly obtained , the challenge to formulate an exact wavefunction in a general case still hinders physical exploration of the qrm . here
we unveil a ground - state phase diagram of the qrm , consisting of a quadpolaron and a bipolaron as well as their changeover in the weak- , strong- and intermediate - coupling regimes , respectively .
an unexpected overweighted antipolaron is revealed in the quadpolaron state , and a hidden scaling behavior relevant to symmetry breaking is found in the bipolaron state .
an experimentally accessible parameter is proposed to test these states , which might provide novel insights into the nature of the light - matter interaction for all regimes of the coupling strengths . _
introduction. _ in the past decade , it has been witnessed that the exploration of fundamental quantum physics in light - matter coupling systems has significantly evolved toward the ( ultra-)strong coupling regime @xcite .
for example , in 2004 , the strong coupling of a single microwave photon to a superconducting qubit was realized experimentally by using circuit quantum electrodynamics @xcite . in 2010 ,
this coupling rate was enhanced to reach a considerable fraction up to 12% of the cavity transition frequency@xcite .
even with such small fractions the system has already entered into so - called ultrastrong - coupling limit@xcite . in this situation ,
the well - known jaynes - cummings ( jc ) model@xcite is no longer applicable because the jc model is valid only in the cases of near resonance and weak coupling @xcite .
indeed , the experimentally observed anticrossing in the cavity transmission spectra @xcite was due to counter - rotating terms , which are dropped in the jc model as a rotating - wave approximation .
in addition , experimental observation of the bloch - siegert shift @xcite also requires taking into account the counter - rotating terms in the description of the jc model .
thus the importance of the counter - rotating terms raises requests to comprehend the behavior of a full quantum rabi model @xcite ( qrm ) for all regimes of the coupling strengths @xcite .
remarkably , an important progress in the study of the qrm in the past years is the proof of its integrability @xcite . as a result , its energy spectra
have been exactly obtained @xcite .
however , to calculate the dynamics of the system , correlation functions , and even other simpler physical observables , it is not enough to know only the exact eigenvalues , but the wavefunctions ( e.g. , the exact eigenstates ) are desirable .
based on series expansions of the eigenstates in terms of known basis sets , it was realized that a standard calculation with double precision , sufficient to compute the spectrum , fails for the eigenstates@xcite .
therefore , the challenge to formulate an exact wave function in a general case still hampers access to a full understanding of the qrm . in this work ,
by deforming the polaron and antipolaron @xcite we propose a novel variational wavefunction ansatz to extract the ground state physics of the qrm .
it is found that this ansatz is valid with high accuracy in all regimes of the coupling strengths .
thus a ground state phase diagram of the qrm is constructed .
the nature of the system variation , by increasing the coupling strength from weak to strong , becomes transparent in the ground - state phase diagram with a quantum state changeover from quadpolaron to bipolaron , around a novel critical - like coupling scale analytically extracted .
in particular , an unexpected overweighted antipolaron is revealed in the quadpolaron state and a hidden scaling behavior is found in the bipolaron state .
moreover , we propose an experimentally accessible parameter to test these states . for perspective
, we also extend this ansatz to the multiple - mode case , which is expected to be useful to understand the physics of the spin - boson model @xcite .
, the original harmonic oscillator ( @xmath0 ) is coupled with two levels denoted by @xmath1 and @xmath2 to form two polarons ( associated with @xmath3 ) with the left(@xmath4 ) and right(@xmath5 ) displacement @xmath6 . *
b * , when the tunneling @xmath7 is switched on , the left- and right - polarons provide an effective potential for each other @xmath8 ( @xmath9 represents the parity , here we focus on the ground state with
@xmath10 " parity ) , which induces an antipolaron , as shown in c. * c * , the potential of the left - polaron deforms from @xmath11 to @xmath12 .
the size of @xmath1 indicates the weight of the polaron ( blue ) and antipolaron ( orange ) , respectively , in the same and opposite directions of the potential displacement .
the situation is symmetric for the right - polaron .
* d*-*f * , typical deformed potentials in the weak ( @xmath13 ) , intermediate ( @xmath14 ) , and strong coupling ( @xmath15 ) cases .
there exist four tunneling channels between @xmath1 and @xmath2 states , as shown in * e * , forming a _ quadpolaron _ state . in the strong coupling case
, the tunneling between left and right states decays until it is vanishingly small due to the large potential barriers between them , yielding a _ bipolaron _ state in * f*. ] _ the model and effective potential. _ the qrm @xcite describes a quantum two - level system coupled to a single bosonic mode or quantized harmonic oscillator , which is a paradigm for interacting quantum systems ranging from quantum optics @xcite to quantum information @xcite to condensed matter @xcite .
the model hamiltonian reads @xmath16 where @xmath17 is the bosonic creation ( annihilation ) operator with frequency @xmath18 and @xmath19 is the pauli matrix with level splitting @xmath7 .
the last term describes the interaction with coupling strength @xmath20 . in terms of the quantum harmonic oscillator with dimensionless formalism@xcite @xmath21 , @xmath22 , where @xmath23 and @xmath24 are the position and momentum operators , respectively , the model can be rewritten as @xmath25 where @xmath26 and @xmath27(@xmath10 ) labels the up @xmath1 ( down @xmath2 ) spin in the @xmath28 direction , respectively .
@xmath29 , with @xmath30 and @xmath31 , while @xmath32 is a constant energy .
apparently , @xmath33 define two bare polarons @xcite in the sense that the harmonic oscillator is bound by @xmath34 due to the coupling @xmath35 , as shown in fig .
[ fig1]a .
these two polarons form two bare potential wells but the existence of the level splitting @xmath7 ( resulting in the tunneling between these two wells@xcite ) makes the model difficult to solve analytically .
let us begin with the wave - function @xmath36 satisfying the schrdinger equation @xmath37 with the eigenenergy @xmath38 .
due to the fact that the model possesses the parity symmetry , namely , @xmath39=0 $ ] with @xmath40 , @xmath36 should take the form of @xmath41 , where @xmath42 will be given below and @xmath43 @xmath44 for positive ( negative ) parity .
without loss of generality , here we consider the ground state , with negative parity .
the schrdinger equation becomes @xmath45 where @xmath46 is an additional effective potential originating from the tunneling , as shown in the lower panel of fig .
[ fig1]b .
the additional potential will deform the bare potential and as a result creates a subwell in the opposite direction of the bare potential @xmath47 , as illustrated in fig .
[ fig1]c .
the subwell induces an _ antipolaron _ as a quantum effect . the above analysis from potential subwell verifies the existence of antipolaron from wavefunction expansion@xcite .
thus , the polaron and antipolaron constitute the basic ingredients of the ground - state wavefunction . _ deformed polaron and antipolaron. _ with the concept of polaron and antipolaron in hand
, the competition between different energy scales @xmath48 and @xmath35 involved in the qrm will inevitably lead to deformations of the polaron and antipolaron .
physically , they can deform predominantly in two possible ways : the position is shifted and the frequency is renormalized , which will introduce four independent variational parameters given below .
explicit deformation depends on the coupling strength once the tunneling is fixed , as shown in fig .
[ fig1 ] d - f from weak to strong couplings according to a critical - like coupling strength @xmath49 .
thus a trial variational wave - function for @xmath50 takes the superposition of the deformed polaron ( @xmath51 ) and antipolaron ( @xmath52 ) , @xmath53 where @xmath54 and @xmath55 , with @xmath56 being the ground - state of standard harmonic oscillator with frequency @xmath18 . here
@xmath57 ( @xmath58 ) , with @xmath59 and @xmath60 , describes the renormalization for frequency ( displacement ) independently for the polaron and the antipoalron , while the coefficients of @xmath61 and @xmath60 denote their weights , subject to the normalization condition @xmath62 .
we stress that in contrast to the direct expansion on basis series without frequency renormalization @xcite , we design our trial wavefunction based on the dominant physics of deformation .
it turns out that our variational wavefunction is capable of providing a reliable analysis on the qrm in the whole parameter regime , ranging from weak to strong couplings , as shown for several physical quantities for the ground - state including the energy , the mean photon number , the coupling correlation and the tunneling strength in appendix [ app - a ] .
obviously the remarkable agreement between our results and the exact ones roots in the fact that our trial wavefunction correctly captures the basic physics , as illustrated by the accurate wavefunction profiles compared to the exact numerical ones for various couplings in fig.[fig - mechanism]a .
the variational wavefunction , with its concise physical ingredients and its accuracy , in turn facilitates unveiling more underlying physics . , at @xmath63 respectively , for weak ( green ) , intermediate(navy ) , and strong(red ) couplings , with @xmath64 .
the symbols denote the numerical exact results .
the spin - down wavefunction is given by @xmath65(not shown ) . *
b * and * c * , @xmath61- and @xmath60-components of @xmath66 , which correspond to the polaron(blue ) and antipolaron(orange ) , respectively , for the intermediate @xmath14 and weak @xmath67 coupling cases .
* d * , the overlaps between different polarons and/or antipolarons without the weights , @xmath68 with @xmath69 .
it is clear that @xmath70 .
* e * , schematic illustration of the physics for the overweighted antipolaron . when decreasing the coupling strength @xmath35 , the potential provided by the left - displaced oscillator for the antipolaron gets lower , so the tunneling energy gain from large @xmath71 in * d * overwhelms the potential cost , which favors a larger weight of antipolaron .
* f * , the overweighted antipolaron with a larger weight than the polaron . ] _ quadpolaron / bipolaron quantum state changeover. _ from fig.[fig - mechanism]a - c one sees that when increasing the coupling , the wavepacket splits into visible polaron and antipolaron ( see animated plots in supplementary material for more vivid evolutions of potentials and wavepackets ) . before the full splitting , there are significant tunnelings in all the four channels between the polarons and antipolarons , as schematically shown in fig .
[ fig1]e .
thus , in this sense we call this state a _ quadpolaron_. after the splitting , only two same - side channels of tunneling survives while the left - right channels are blocked gradually by the increasing barrier , as sketched in fig .
[ fig1]f .
this state is termed here as a _
bipolaron_. despite the evolution from a transition - like feature in the low frequency limit to a crossover behavior in finite frequencies for the changover between quadpolaron and bipolaron states , the nature of the afore - mentioned splitting is essentially the same .
this enables us to obtain an analytic coupling scale ( see appendix [ app - b ] ) , @xmath72 , which generalizes the low frequency - limit result@xcite @xmath73 and correctly captures the quantum state changeover between quadpolaron and bipolaron for the whole range of frequencies . .
@xmath74 ( @xmath75 ) is the displacement renormalization and @xmath76 is the frequency renormalization .
@xmath61 and @xmath60 denote the weights of the polaron and the antipolaron , respectively , in the variational groundstate wavefunction . * a * and * d * , @xmath77 .
* b * and * e * , @xmath78 .
* c * and * f * , @xmath79 . ] _ quadpolaron asymmetry and overweighted antipolaron in the regime of @xmath80. _ we find that the polaron and antipolaron in the quadpolaron state have asymmetric displacements , which leads to a subtle competition depending on the frequency @xmath81 .
figure [ fig - parameters ] shows three types of distinct behaviors of the variational parameters in three different frequency regimes : high frequency ( @xmath82 ) , intermediate frequency(@xmath83 $ ] ) , and low frequency ( @xmath84 ) .
the result is understandable due to the fact that the antipolaron always has a higher potential energy owing to its opposite direction to the displacement of @xmath3 . roughly speaking , at a high frequency
, the antipolaron should have a lower weight than the polaron ( @xmath85 ) since the antipolaron is suppressed by the high potential . at a low frequency , the polaron benefits from both potential and tunneling energies .
however , competition becomes subtle at an intermediate frequency as each of these different energy scales may only favor either the polaron or the antipolaron respectively , which may lead to overweighted antipolaron , as shown in fig.[fig - parameters]e .
below we give a more explicit analysis . actually , the four channel tunneling energies in the quadpolaron are proportional to the overlaps of the polarons and antipolarons , @xmath86 , @xmath87 , @xmath88 and @xmath89 , respectively , as shown in fig.[fig - mechanism]d
. the mixture terms @xmath90 and @xmath89 do not affect the weight competition between the polaron and antipolaron , while @xmath91 and @xmath87 yield imbalances . indeed
, the antipolarons have larger overlap than the polarons , i.e. @xmath92 ( see fig . [ fig - mechanism]d ) .
this is because the antipolarons in up and down spins are closer to each other than the polarons in order to reduce their higher potential energy , as indicated in fig .
[ fig - mechanism]e and quantitatively shown by @xmath93 in fig .
[ fig - parameters]a and [ fig - parameters]b .
therefore , as far as the tunneling is concerned , it would tend to have more weight of antipolarons to gain a maximum tunneling .
when the intermediate frequency reduces the cost of potential energy for such tendency , a larger antipolaron weight might finally occur , as in fig .
[ fig - mechanism]f , leading to an unexpected overweighted antipolaron .
we find that this really occurs as demonstrated in fig .
[ fig - parameters]e where a weight reversion appears at the crossing of @xmath61 and @xmath60 for a weaker coupling . at the low frequency
, the harmonic potential becomes very flat , the polarons may get closer than antipolarons , as indicated by @xmath94 in fig.[fig - parameters]c in the weak coupling regime . in this case , @xmath95 is greater than @xmath96 so that polarons have favorable energies in both potential and tunneling . thus the polaron regains its priority in weight . as a function of the coupling strength @xmath20 .
* a * , our results compared with exact numerics at @xmath97 as an example . *
b * , @xmath98 for different values of the ratios @xmath81 .
the scaling relation @xmath99 is tested by @xmath100 beyond @xmath49 . ]
-0.cm _ bipolaron and hidden scaling behavior in the regime of @xmath101. _ in the bipolaron regime , the remaining tunneling in channels @xmath102 and @xmath103 leads to intriguing physics , showing a deeper nature of the interaction in the symmetry breaking aspect . indeed ,
fig.[fig - parameters]a - c show that in this regime the frequency factor @xmath76 and the displacement factor @xmath74 collapse into the same value , i.e @xmath99 .
in fact , due to vanishing photon number below @xmath49 at low frequency limit , the parity @xmath104 can be decomposed into separate spin and spatial reversal sub - symmetries which are broken beyond @xmath49 .
however , further seeking the symmetry breaking character from these sub - symmetries would fail at finite frequencies due to emergence of a finite number of photons below @xmath49 .
nevertheless , the @xmath58-@xmath57 symmetric aspect revealed here provides a compensation , from the beyond-@xmath49 side instead but valid also for finite frequencies . to test this scaling behavior ,
we propose an experimentally accessible quantity , @xmath105 , where @xmath106 , which becomes the scaling ratio @xmath107 ( see appendix [ app - c ] ) for their average @xmath108 and @xmath109 and thus equals to one above @xmath49 , as shown in fig .
[ fig - scaling ] for various frequencies .
the experimental measurement of @xmath98 thus provides a possible way to distinguish the states of bipolaron and quadpolaron as well as their changeover . ) and bipolaron ( @xmath110 ) as well as their crossover near the minimum of @xmath57 ( red solid line ) or the maximum of @xmath98 around analytic @xmath49 .
the quadpolaron regime is further divided into the normal ( @xmath111 ) and overweighted antipolaron ( @xmath112 ) regimes . the dashed and dot - dashed lines have been obtained numerically from the cross points as shown in fig .
[ fig - parameters ] ( e ) and ( f ) .
the color density for @xmath98 further distinguishes the characters of the different regimes and their changeovers . ]
-0.0 cm _ ground - state phase diagram. _ the above discussions on polaron - antipolaron competition can be summarized into a ground - state phase diagram unveiled , as shown in fig .
[ fig - diagram ] . the ground - state with different channels of tunneling
is identified as a quadpolaron when @xmath80 and as a bipolaron when @xmath101 .
an overweighted antipolaron is hidden in the quadpolaron regime , while a scaling relation between the displacement and frequency renormalizations is revealed in the bipolaron regime .
note that the polaron and antipolaron structures might be detected by optomechanics@xcite and @xmath98 is experimentally measurable .
the diagram may provide a renewed panorama for deeper theoretical investigations and may raise more challenges for experiments .
_ perspective in multiple modes. _ the basic physics in the qrm has a profound implication for the spin - boson model@xcite , which is a multiple - mode version of the qrm .
the essential variational ingredients remain similar .
the trial wavefunction can be written as @xmath113=\alpha \prod_{k=1}^m\varphi _
\alpha^k+\beta \prod_{k=1}^m\varphi _
\beta^k $ ] with the extension @xmath114 for the @xmath115th mode .
we illustrate the same accuracy by a two - mode case in appendix [ app - d ] . _
acknowledgements. _ we thank jun - hong an for helpful discussions .
work at csrc and lanzhou university was supported by national natural science foundation of china , pcsirt ( grant no .
irt1251 ) , national 973 " projects of china . z .- j.y
. also acknowledges partial financial support from the future and emerging technologies ( fet ) programme within the seventh framework programme for research of the european commission , under fet - open grant number : 618083 ( cntqc ) . |
the vitali theorem has been widely studied in the literature . for a historical survey ,
see for instance @xcite and the bibliography therein .
this theorem is a well - known milestone in integration theory , directly related with @xmath0 convergence and weak compactness .
however it has also many applications in several branches of mathematics , for example in reconstruction of signals , integral and discrete operators ( see @xcite ) .
the vitali theorem has been investigated even in the context of vector lattices with respect to different types of convergence ( see @xcite and its bibliography ) and in abstract convergence groups , not necessarily endowed with an order structure ( see @xcite ) . also the modular spaces
( see for instance @xcite ) are a rich field of research .
they are a natural generalization of the lebesgue spaces @xmath0 and contain as particular cases the orlicz and the musielak - orlicz spaces ( see e.g. @xcite ) . in modular spaces
the vitali theorem is often used in the problem of approximating a real - valued function @xmath1 by means of urysohn - type or sampling operators @xmath2 , ( see also @xcite ) .
these operators are particularly useful in order to approximate a continuous or analog signal by means of discrete samples , and therefore they are widely applied for example in reconstructing images and videos . in the classical literature ,
a signal @xmath1 is considered as a real - valued function , defined on a ( bounded or not ) time interval . in these situations
, usually it happens that the original signal or image is unknown , and it must be reconstructed from some sampled values of the type @xmath3 , where @xmath4 , and @xmath5 is a fixed rate .
so it would be advisable to consider the studied signal , for example , as a map defined on a ( finite or not ) time interval and taking values in a suitable space of random variables ( for more details related with these topics see also @xcite and the literature therein ) . in this context
, it is natural to consider the tool of _ vector lattices_. moreover , also domains with infinite measure , like for instance an infinite time interval , can be treated . in this paper
we consider vector lattices with convergences generated by filters ( see also @xcite ) , which generalize classical ones . for example , when the filter involved is the class of all subsets of @xmath6 having asymptotic density one , we obtain the so - called statistical convergence .
we introduce the theory of modular spaces in the vector lattice context : integration theory can be considered as a particular case of the theory of modulars , even in our setting , so we give a general version of the vitali and lebesgue dominated convergence theorems for modular convergence , with applications to integrals for vector lattice - valued functions , defined on sets of possibly infinite measure . in section 4
we deal with moment kernels in the vector lattice setting , in order to construct a suitable sequence of convolution type operators ; moreover in subsection [ bm ] , among the applications , we give convergence results for the it integral with respect to brownian motion .
finally , the last section is devoted to the proof of the main theoretical result ( theorem [ unico ] ) of section 4 .
we begin with recalling some basic properties of vector lattices and filter convergence . a vector lattice ( also called riesz space ) @xmath7 is said to be _ dedekind complete _ iff every nonempty subset @xmath8 of @xmath7 , bounded from above , has a lattice supremum in @xmath7 , denoted by @xmath9 .
+ from now on @xmath7 denotes a dedekind complete vector lattice , @xmath10 is the set of all strictly positive elements of @xmath7 , and we set @xmath11 . for all @xmath12 , let @xmath13 .
we add to @xmath7 an extra element @xmath14 , extending order and operations in a natural way , set @xmath15 , @xmath16 and assume by convention @xmath17 .
+ a sequence @xmath18 in @xmath7 is called _ @xmath19-sequence _ iff it is decreasing and @xmath20 .
+ a _ strong unit _ of @xmath7 is an element @xmath21 , such that for every @xmath12 there is a positive real number @xmath22 with @xmath23 .
+ let @xmath24 , @xmath25 , @xmath7 be three dedekind complete vector lattices .
we say that @xmath26 is a _ product triple iff a map @xmath27 ( _ product ) is given , satisfying natural conditions of compatibility , see ( * ? ? ?
* assumption 2.1 ) and @xcite . from now on we suppose that : _ _ * @xmath28 is a product triple , where @xmath29 are endowed with strong units @xmath21 , @xmath30 respectively .
let @xmath31 be any fixed countable set .
a class @xmath32 of subsets of @xmath31 is called a _ filter _ of @xmath31 iff @xmath33 , @xmath34
whenever @xmath35 and for each @xmath36 and @xmath37 we get @xmath38 .
we denote by @xmath39 the filter of all cofinite subsets of @xmath31 .
a filter of @xmath31 is said to be _
free _ iff it contains @xmath39 .
an example of free filter ( different from @xmath39 ) is the family @xmath40 of all subsets of @xmath6 , having asymptotic density one . in this paper
we only deal with free filters .
+ we now give the notion of filter convergence in the vector lattice setting ( see also @xcite ) . [ filterconv ] let @xmath41 be any filter of @xmath31 .
a sequence @xmath42 in @xmath7 _
@xmath43-converges to @xmath12 _ ( @xmath44 )
iff there exists an @xmath19-sequence @xmath45 in @xmath7 such that for all @xmath46 the set @xmath47 is an element of @xmath41 . + a sequence @xmath42 in @xmath7 _
@xmath48-converges to @xmath49 _ ( @xmath50 )
iff there exist a positive element @xmath51 and an @xmath19-sequence @xmath52 such that for all @xmath46 the set @xmath53 is an element of @xmath41 .
+ clearly a sequence @xmath54 in @xmath7 _
@xmath19_-converges to @xmath12 ( in the classical sense ) iff it @xmath55-converges to @xmath56 , and similarly with @xmath57-convergence .
let @xmath58 be any infinite set , @xmath59 be a algebra and @xmath60 be a finitely additive measure .
we shall denote by @xmath61 the family of all sets @xmath62 with @xmath63 .
for every @xmath64 and @xmath65 , we shall denote by @xmath66 the function whose values @xmath67 are @xmath68 when @xmath69 and 0 otherwise .
we now introduce the modulars in the setting of vector lattices ( for the classical case and a related literature , see e.g. @xcite ) .
let @xmath70 be a linear sublattice of @xmath71 , such that * @xmath72 for every @xmath73 ; * if @xmath74 and @xmath64 , then @xmath75 .
a functional @xmath76 is said to be a _
modular _ on @xmath70 iff it satisfies the following properties : ( @xmath77 ) : : @xmath78 ; ( @xmath79 ) : : @xmath80 for every @xmath74 ; ( @xmath81 ) : : @xmath82 for every @xmath1 , @xmath83 and for any non negative real numbers @xmath84 with @xmath85 .
we say that : ( @xmath86 ) : : a modular @xmath87 is _ monotone _ iff @xmath88 for every @xmath1 , @xmath83 with @xmath89 . in this case , if @xmath74 , then @xmath90 and hence @xmath91 .
( @xmath92 ) : : a modular @xmath87 is _ finite _ iff for every @xmath93 and every @xmath19-sequence @xmath94 in @xmath95 , the sequence @xmath96 is @xmath57-convergent to @xmath97 .
( for the case @xmath98 , see also @xcite ) .
we now prove the following [ propertyglobalfiniteness ] let @xmath99 be a product triple satisfying * ( @xmath100 ) * , and assume that @xmath87 is a finite modular , @xmath101 is a sequence in @xmath61 and @xmath94 is any @xmath19-sequence in @xmath102 . then there exists a strictly increasing mapping @xmath103 from @xmath6 to @xmath6 , with @xmath104 * proof : * fix a sequence @xmath101 in @xmath61 , and any @xmath19-sequence @xmath94 in @xmath105 . by finiteness of @xmath87
, we get @xmath106 this means that a sequence @xmath107 exists in @xmath10 with the following property : for every @xmath108 , @xmath109 there is @xmath110 such that @xmath111 is an @xmath19-sequence for all @xmath108 , and @xmath112 for every @xmath113 there is a positive real number @xmath114 such that @xmath115 .
therefore , from ( [ unoo ] ) we also have @xmath116 now , for each integer @xmath108 , a corresponding integer @xmath117 can be found , such that the sequence @xmath118 is strictly increasing and @xmath119 the assertion follows from ( [ duee ] ) and ( [ tree ] ) .
@xmath120 + we now give the concept of ( equi)-absolute continuity in the context of modulars and filter convergence : this is one of the crucial tools in the vitali theorem ( for a classical modular version , see ( * ? ? ?
* theorem 2.1 ) ) .
( @xmath121 ) : : we say that @xmath74 is _
@xmath43-absolutely continuous with respect to the modular @xmath87 _ ( or in short _ absolutely continuous _ )
iff there is a positive real constant @xmath122 , satisfying the following properties : ( @xmath123 ) : : for each @xmath19-sequence @xmath124 in @xmath125 there exists an @xmath19-sequence @xmath126 in @xmath10 such that for all @xmath127 and whenever @xmath128 we get @xmath129 ; ( @xmath130 ) : : there is an @xmath19-sequence @xmath131 in @xmath10 such that to each @xmath132 there corresponds a set @xmath133 with @xmath134 .
( @xmath135 ) : : given a modular @xmath87 and any free filter @xmath41 of @xmath31 , we say that a sequence @xmath136 , @xmath137 , is _ @xmath87-@xmath41-equi - absolutely continuous _ , or in short _ equi - absolutely continuous , _ iff there is @xmath138 , satisfying the following two conditions : + ( @xmath139 ) ; ; for every @xmath19-sequence @xmath124 in @xmath140 there are an @xmath19-sequence @xmath126 in @xmath10 and a sequence @xmath141 in @xmath41 , with @xmath142 whenever @xmath143 and @xmath144 , @xmath145 ; ( @xmath146 ) ; ; there are an @xmath19-sequence @xmath147 in @xmath10 and a sequence @xmath148 such that @xmath149 in @xcite some sufficient conditions are given for ( equi)-absolute continuity with respect to modulars .
as we already pointed out in the introduction , the vitali theorem has a fundamental importance in approximation theory , also when @xmath150 : several applications can be found in @xcite and the literature therein .
together with equi - absolute continuity , also some notions of convergence are needed , so we recall the concepts of filter uniform convergence and convergence in measure ( see also @xcite ) .
let @xmath41 be any fixed free filter of @xmath31 .
[ u - misura ] ( [ u - misura].1 ) : : a sequence of functions @xmath151 in @xmath71 _ _
@xmath48_-converges uniformly _ ( shortly _ converges uniformly _ ) to @xmath1 , iff there exists an @xmath19-sequence @xmath94 in @xmath95 with @xmath152 in this case we write : @xmath153 ( * [ u - misura].2 ) * : : given a sequence @xmath154 in @xmath71 and @xmath155 , we say that @xmath154 _ @xmath48-converges in measure _ ( shortly _ @xmath156-converges _ ) to @xmath1 , iff there are two @xmath19-sequences , @xmath94 in @xmath157 , @xmath124 in @xmath140 , and a double sequence @xmath158 in @xmath159 such that + * @xmath160 for every @xmath137 and @xmath127 , and * @xmath161 for all @xmath46 .
it is easy to see that ( * [ u - misura].1 ) * implies ( * [ u - misura].2)*. if @xmath162 @xmath156-converge to @xmath163 respectively , then @xmath164 @xmath156-converges to @xmath165 and @xmath166 @xmath156-converges to @xmath167 , ( see also @xcite ) .
moreover , when @xmath168 and the convergence involved is the usual one , these notions coincide with the classical ones .
+ we now prove a vitali - type theorem , which links the theories of modulars and vector lattices , in the context of @xmath43-convergence .
[ trediecibis ] let @xmath87 be a monotone and finite modular , and @xmath169 be a fixed free filter of @xmath31 .
if @xmath154 is a sequence in @xmath70 , @xmath156-convergent to @xmath97 and equi - absolutely continuous , then there is a positive number @xmath122 with @xmath170 * proof : * let @xmath138 be related with equi - absolute continuity , and @xmath147 in @xmath10 be an @xmath19-sequence associated with ( @xmath146 ) ; let moreover @xmath124 in @xmath140 , @xmath94 in @xmath95 be related with @xmath156-convergence of the sequence @xmath154 to @xmath97 , and @xmath126 in @xmath10 be associated with @xmath124 and ( @xmath139 ) .
choose any sequence @xmath101 in @xmath61 , and any corresponding sequence @xmath171 in @xmath32 , satisfying ( [ equiac2 ] ) .
without loss of generality , we assume that the @xmath172 s are positive . for
every @xmath173 , @xmath46 , let @xmath174 be such that @xmath175 , according to @xmath156-convergence .
for every @xmath173 , @xmath108 , @xmath46 and @xmath176 we have @xmath177 by the properties of the modular @xmath87 we obtain ( with obvious meaning of the notations ) : @xmath178 by ( [ equiac2 ] ) we get @xmath179 moreover , since @xmath180 by convergence in measure , then , thanks to property @xmath181 of equi - absolute continuity , for each @xmath127 there exists an element @xmath182 , with @xmath183 in order to estimate @xmath184 , observe that by proposition [ propertyglobalfiniteness ] there is a strictly increasing mapping @xmath103 from @xmath6 to @xmath6 , with @xmath185 now , for all @xmath173 , @xmath132 and @xmath186 we see that @xmath187 and , consequently , @xmath188 thus from ( [ matematicaa ] ) , ( [ iuno ] ) , ( [ itre ] ) and ( [ idue ] ) it follows that for every @xmath108 there are @xmath189 , @xmath190 such that @xmath191 whenever @xmath192 , and so , since @xmath193 is an @xmath19-sequence , we conclude @xmath194 that is the assertion . @xmath120 + an easy consequence of theorem [ trediecibis ] is the following [ lebesgue ] let @xmath7 , @xmath87 , @xmath41 be as in theorem [ trediecibis ] and assume that @xmath154 is a sequence in @xmath70 , @xmath156-convergent to @xmath97 .
suppose that there are an absolutely continuous function @xmath195 in @xmath70 and an element @xmath196 , such that @xmath197 for all @xmath198 and @xmath199 .
then there is a positive real number @xmath122 with @xmath200 .
we give here some applications of the results obtained above in different fields of mathematics , as integration theory , moment operators , stochastic processes and brownian motion .
let @xmath58 , @xmath159 and @xmath156 be as above .
we now extend the integration theory investigated in @xcite to the case in which @xmath156 can assume the value @xmath14 , in the setting of @xmath43-convergence .
+ a function @xmath201 is said to be _ simple _ iff @xmath202 is a finite set and @xmath203 for every @xmath204 .
the space of all simple functions is denoted by @xmath205 .
let @xmath206 be the set of all simple functions @xmath207 vanishing outside a set of @xmath156-finite measure .
if @xmath208 , we denote its elementary integral with @xmath209 , as usual .
[ additivita ] it is clear that the functional @xmath210 defined as @xmath211 is a monotone modular , and it is also linear and additive on positive functions and constants . also finiteness of @xmath212 is easy to see : indeed , let @xmath73 and @xmath94 be any @xmath19-sequence in @xmath157 . according to our definitions , @xmath213 for all @xmath145 : since @xmath214 is a positive element in @xmath7 , it is obvious that the condition of finiteness is satisfied .
[ integrabilitainfinita ] a positive function @xmath155 is _ integrable _ iff there exist an equi - absolutely continuous sequence of functions @xmath215 in @xmath206 , @xmath156-convergent to @xmath1 , and a map @xmath216 , with @xmath217 ( the sequence @xmath218 is said to be _ defining _ ) .
note that , in this case , we get @xmath219 if @xmath155 , then we say that @xmath1 is _ integrable _ iff @xmath220 and @xmath221 are integrable , where @xmath222 , @xmath223 , @xmath199 . in this case
, we set @xmath224 similarly as in ( * ? ? ?
* proposition 3.11 ) , we now prove that the integral in definition [ integrabilitainfinita ] is well - defined .
[ int - inf ] in the setting described above , let @xmath225 be as in ( [ zvezda ] ) .
then the quantity @xmath226 is independent of the choice of the defining sequence @xmath218 .
* proof : * let @xmath227 , @xmath228 , be two defining sequences for @xmath1 , and put @xmath229 since @xmath230 , @xmath228 , are equi - absolutely continuous , then @xmath231 is too . as @xmath232 and @xmath233 @xmath156-converge to @xmath1 ,
@xmath231 @xmath156-converges to @xmath97 .
thus , by theorem [ trediecibis ] applied to the modular @xmath210 defined in remark [ additivita ] , we obtain : @xmath234 therefore for all @xmath64 and @xmath235 we get : @xmath236 by letting @xmath237 tend to infinity , we obtain @xmath238 , namely @xmath239 for all @xmath64 . @xmath120 + [ additivitaestesa ] let @xmath240 be the space of all integrable functions in the sense of definition [ integrabilitainfinita ] .
it is clear that in this way the elementary integral , defined in @xmath206 , can be uniquely extended to a linear monotone functional @xmath241 defined on the whole of @xmath240 . in particular , it turns out that , fixed a measurable subset @xmath242 and a mapping @xmath243 , then @xmath1 is integrable in the whole of @xmath58 iff both mappings @xmath244 and @xmath245 are ( and of course the integral is additive in this case ) .
[ perdopo ] if @xmath246 is defined by @xmath247 , where @xmath248 is fixed and @xmath249 is lebesgue integrable , then @xmath1 belongs to @xmath250 and @xmath251 .
let @xmath252 be a vector lattice with a strong unit @xmath21 , @xmath253 , @xmath254 , @xmath255 be the @xmath256-algebra of all measurable subsets of @xmath257 , @xmath258 be the lebesgue measure and @xmath259 . from now on
we consider functions @xmath260 , and the so - called _ moment kernels _ , @xmath261 , defined by @xmath2620,1 [ } ( w),\,\ , n \in \mathbb{n } , w \in
\mathbb{r}^+.\ ] ] let @xmath263 be the set defined by @xmath264 where @xmath265 for every @xmath266 .
we consider the modular @xmath267 . + in order to define and study the moment operators we need the following notions and results .
[ order - uc ] we say that @xmath268 is _ uniformly continuous _
iff there are a positive element @xmath269 and two @xmath19-sequences @xmath124 and @xmath270 in @xmath95 such that @xmath271 for all @xmath272 , @xmath273 and @xmath274 satisfying @xmath275 .
we remark here that any uniformly continuous mapping @xmath1 is _ locally bounded _ ,
i.e. the set @xmath276 is bounded in @xmath7 , for any bounded interval @xmath277 . [ order - lip]we say that @xmath268 is _ lipschitz _
, iff there is @xmath278 with @xmath279 for all @xmath272 , @xmath273 . if @xmath1 is lipschitz , then @xmath1 is uniformly continuous
. let @xmath280 be a connected set , and @xmath281 .
we say that @xmath1 is _ uniformly differentiable on @xmath282 _ iff there exist a bounded function @xmath283 and two @xmath19-sequences , @xmath124 and @xmath270 in @xmath95 , with @xmath284 for every @xmath46 .
in this case we say that @xmath195 is the _ uniform derivative _ of @xmath1 in @xmath282 .
note that in general the concept of uniform derivative is a strict strengthening of the classical one . however if @xmath285 in the classical sense , then @xmath1 is uniformly differentiable and @xmath286 .
+ we see that uniform differentiability and uniform continuity in the vector lattice setting satisfy analogous properties as the corresponding classical ones .
[ uc - prodotto ] let @xmath287 \to { \bf x}$ ] , @xmath288 \to \mathbb{r}$ ] .
then we have : ( [ uc - prodotto].1 ) : : if @xmath289 are uniformly continuous , then @xmath290 is uniformly continuous too ; ( [ uc - prodotto].2 ) : : if @xmath1 is uniformly differentiable , then @xmath1 is uniformly continuous ; ( [ uc - prodotto].3 ) : : if @xmath289 are uniformly differentiable , with uniform derivatives @xmath291 , @xmath292 respectively , then @xmath290 is uniformly differentiable too , and its uniform derivative is @xmath293 ; ( [ uc - prodotto].4 ) : : if @xmath1 is uniformly continuous , then @xmath1 is bounded and belongs to @xmath250 .
moreover its integral function @xmath294 is uniformly differentiable in @xmath295 $ ] , with uniform derivative @xmath1 .
* proof : * the proofs of statements ( [ uc - prodotto].@xmath296 ) for @xmath297 are straightforward .
we prove now ( [ uc - prodotto].4 ) .
+ boundedness has already been pointed out .
so we prove just integrability .
choose @xmath124 and @xmath270 in @xmath95 according to uniform continuity of @xmath1 .
for each @xmath46 , let @xmath298 be a division of @xmath295 $ ] , with mesh @xmath299 , and take @xmath300 in such a way that @xmath300 is a refinement of @xmath301 .
now , set @xmath302 so , there exists a suitable positive real number @xmath303 such that @xmath304 for all @xmath145 and @xmath305 $ ] . for every @xmath109 and @xmath306
$ ] we have @xmath307 , and so the sequence @xmath308 converges uniformly , and then also in @xmath156-measure to @xmath1 . since @xmath295 $ ] has finite measure , and @xmath309 for every borel set @xmath310 $ ] and every integer @xmath311 , the sequence @xmath308 is equi - absolutely continuous . moreover , since @xmath312 for all @xmath311 , @xmath313 , we can obtain a mapping @xmath314 as in definition [ integrabilitainfinita ] . indeed ,
if for all @xmath64 we set @xmath315 then we easily get @xmath316 and , for all @xmath64 and @xmath46 @xmath317 this clearly implies that @xmath318 for all @xmath64 , and moreover @xmath319 for all @xmath311 , which proves formula ( [ zvezda ] ) and hence integrability of @xmath1 . + finally , the proof of differentiability of @xmath320 , @xmath321 $ ] is straightforward .
@xmath322 + we are now in position to introduce the moment operators .
+ for every @xmath323 ( see ( [ integrale - cappuccio ] ) ) let us consider the operators @xmath324 defined by @xmath325 in what follows , we shall prove that the sequence @xmath326 satisfies the hypotheses of theorem [ trediecibis ] , and therefore it is modularly @xmath327-convergent to @xmath97 .
the proof of this result and of the next theorem [ unico ] will be given in subsection [ prova ] .
[ unico ] let @xmath1 be a function with compact support @xmath328 $ ] with @xmath329 and belonging to @xmath263 .
then we have the following : if @xmath1 is bounded , and @xmath330 , then the functions @xmath331 are lipschitz and integrable for every @xmath332 ; moreover , if @xmath1 is uniformly continuous , we have @xmath333 if @xmath1 is uniformly continuous , then @xmath2 converges uniformly to @xmath1 ; + if @xmath1 is uniformly continuous and @xmath334 , then @xmath335 for every @xmath336 , @xmath337 is uniformly differentiable and its uniform derivative satisfies the following relation : @xmath338 replacing @xmath237 with a real positive constant @xmath339 , we can see that , also in our general setting , the linear differential equation @xmath340 ( with @xmath1 uniformly continuous ) can be solved by setting @xmath341 where @xmath22 is an arbitrary constant value .
a first application can be formulated in terms of weak convergence of the derivative mappings @xmath342 .
let @xmath343 be any uniformly continuous map , vanishing outside the interval @xmath344 $ ] , with @xmath334 , and fix any @xmath345 mapping @xmath346 , with compact support .
for each uniformly continuous mapping @xmath347 , define @xmath348 it is easy to see that @xmath87 is a monotone and real - valued modular , and that @xmath349 thanks to ( [ unico].3.1 ) . from this
we deduce that @xmath350 and also , integrating by parts : @xmath351 ( where the last integral is intended in the riemann - stieltjes sense ) .
thus we can conclude that , though the mappings @xmath342 do not converge in general , we always have @xmath352 for all @xmath345 mappings @xmath5 , with compact support .
+ in order to obtain a more concrete application in stochastic integration , we shall assume that @xmath353 ( with @xmath354 ) is the standard brownian motion defined on a probability space @xmath355 .
of course , we can consider @xmath356 as a mapping from @xmath357 $ ] into @xmath358 .
since @xmath359 has not a strong unit in general , in order to apply the previous theory we shall suitably restrict it : indeed , thanks to the well - known maximum principle ( see e.g. @xcite ) , we see that there is a suitable positive element @xmath360 such that @xmath361 for all @xmath362 $ ] .
moreover , we remark that , ( see e.g. @xcite and @xcite ) , there exists a positive random variable @xmath363 in @xmath359 such that @xmath364 holds , as soon as @xmath365 $ ] .
thus , taking @xmath7 as the ( complete ) subspace of @xmath359 generated by all elements dominated by some real multiple of @xmath366 , we see that @xmath7 has a strong unit ( i.e. @xmath366 ) , that @xmath367 is @xmath7-valued and is also a uniformly continuous mapping from @xmath357 $ ] to @xmath7 , in the sense of our definition . + in order to establish the next results ,
we introduce a definition .
[ processiy ] let @xmath368\to { \bf x}$ ] be any stochastic process , predictable with respect to @xmath356 , and with continuous trajectories .
if we assume that @xmath369}e(y(t)^2)=k<+\infty}$ ] , then it is well - known that @xmath370 is integrable in the it s sense , with respect to @xmath356 ( see e.g. @xcite ) .
any process @xmath370 of this kind will be called a _ regular _ process
. the it integral of @xmath370 will be denoted as usual with @xmath371 .
[ integralepiccolo ] if the process @xmath370 is regular , and @xmath372}e(y(t)^2)=k$ ] , then @xmath373 * proof * : let @xmath374 be any decomposition of @xmath357 $ ] obtained by means of the points @xmath375 , and consider the it sum : @xmath376 we easily see that @xmath377 , since the increment @xmath378 is independent of @xmath379 for all @xmath380 .
similarly , we deduce that @xmath381=0 $ ] as soon as @xmath382 .
so , @xmath383 ^ 2)= \sum_{i=0}^{n-1 } e(y(t_i)^2)e[(b(t_{i+1})-b(t_i))^2]\leq \\ & \leq & k\sum_{i=0}^{n-1}(t_{i+1}-t_i)=kt.\end{aligned}\ ] ] since the sums @xmath384 are norm - convergent to the it integral @xmath385 , the assertion is obvious .
@xmath322 + the following corollary is an immediate consequence .
[ primaconvergenza ] let @xmath386 be any sequence of regular stochastic processes .
assume that the processes @xmath387 converge to a regular process @xmath370 uniformly in @xmath359 , i.e. for every real @xmath388 an integer @xmath389 exists , such that @xmath390}\|y_n(t)-y(t)\|_2\leq \varepsilon\ ] ] for all @xmath391 .
then the it integrals @xmath392 converge in @xmath359 to the it integral we shall apply these results to the process @xmath393 defined as follows : @xmath394\\ ( t - t)(b(t)-b(a ) ) , & t\in [ a , t ] \end{array}\right.\ ] ] where @xmath395 and @xmath70 are fixed positive numbers , @xmath396 .
( so @xmath1 is a process similar to the well - known _
brownian bridge _ ) .
[ lastfinale ] let @xmath397 be the process defined above .
then @xmath1 is clearly @xmath7-valued and uniformly continuous , and we have @xmath398 ( we remark that the integral on the left - hand side is in the riemann - stieltjes sense , since the mappings @xmath399 are more regular than @xmath1 ) .
* proof * : it is clear that @xmath1 is predictable ; also @xmath399 is for all @xmath237 , since @xmath400 , so the values of @xmath401 depend only on the values of @xmath402 for @xmath403 ; then it is sufficient to apply [ integralepiccolo ] to the sequence @xmath404 , which converges uniformly to 0 .
@xmath120 + we note that , in the previous results , @xmath19-convergence is not mentioned , since we do nt know whether the integrals @xmath405 are dominated by an element in @xmath359 .
however , this problem can be solved by replacing the sequence @xmath406 with the following one : @xmath407 where @xmath408 is any positive mapping larger than @xmath409 .
indeed , @xmath386 is obviously dominated , and we always have @xmath410 which proves @xmath359-convergence of @xmath386 to the it integral of @xmath1 with respect to @xmath356 .
@xmath120 + the following pictures show a comparison among some trajectories of the mapping @xmath1 and the corresponding ones of the mappings @xmath411
. 0.48 0.48
let @xmath412 denote a fixed real positive number such that @xmath413 } |f(t)|\leq m_f e$ ] and set @xmath414 .
+ * proof of ( [ unico].1 ) * we first show that @xmath411 is lipschitz and then we will prove integrability of @xmath411 on @xmath415 . for each @xmath416 we can write : @xmath417 we now estimate the quantity @xmath418 .
let @xmath419 , @xmath420 , with @xmath421 . if @xmath422 , then @xmath423 .
+ if @xmath424 , then without loss of generality we can suppose @xmath425 .
we get : @xmath426 .\end{aligned}\ ] ] then @xmath427 for a suitable @xmath428 with @xmath429 , thanks to the lagrange theorem applied to the real function @xmath430 .
so , we see that @xmath431 thus proving that @xmath399 is lipschitz . + we now prove integrability .
let @xmath332 be fixed .
first of all observe that for every @xmath432 we have @xmath433 , where @xmath434 is the integral of @xmath435 in @xmath436 $ ] .
since @xmath437 , the real - valued mapping @xmath438 is clearly integrable in @xmath439 , and therefore it is easy to see that @xmath440 is , in the same halfline , thanks to remark [ perdopo ] .
moreover , since @xmath440 is uniformly continuous , it is also integrable in @xmath436 $ ] , in view of ( [ uc - prodotto].4 ) .
therefore we get integrability on the whole of @xmath441 ( see remark [ additivitaestesa ] ) .
let us now turn to prove ( [ consfubini ] ) , in the case that @xmath1 is uniformly continuous . in order to compute the integral of @xmath440 , @xmath437 ,
we first point out that @xmath442 where @xmath443 is the constant above .
now we evaluate the integral in @xmath295 $ ] . to this aim
, we can observe that , by uniform continuity of @xmath1 and of the mapping @xmath444 in @xmath295 $ ] , there exist an @xmath19-sequence @xmath124 in @xmath157 and a strictly increasing sequence @xmath445 in @xmath446 , such that @xmath447 for all @xmath145 and @xmath448 $ ] satisfying @xmath449 .
+ now , fix @xmath145 and let @xmath450 be the division of @xmath295 $ ] obtained with precisely @xmath451 subintervals of the same length , i.e. @xmath452 . by additivity of the integral we have @xmath453 ds.\ ] ] for each fixed @xmath454 let us denote by @xmath455 the maximum index such that @xmath456 , and set also @xmath457 .
we have then : @xmath458 thus we see that @xmath459 ds|\leq \\ & \leq & \sum_{j=0}^{h_p-1}\int_{s_j}^{s_{j+1}}\frac{1}{s^n } ( \frac { 2 nb^nm_f}{h_p}+b\sigma_p)e ds= ( \frac { 2 nb^nm_f}{h_p}+b\sigma_p)e\int_a^{b}\frac{1}{s^{n}}ds= \\ & = & \frac{b^{n-1}-a^{n-1}}{n-1 } ( \frac{2 nb^nm_f}{h_p}+ b\sigma_p)e.\end{aligned}\ ] ] letting @xmath460 , we can see that @xmath461 is an @xmath19-sequence in @xmath105 , and summarize the previous result in the following way : @xmath462 now , by changing order in the summations , we get @xmath463 ds+ ... \right)= \\ & = & \frac{n}{n-1}\frac{b - a}{h_p } \left ( \int_{s_0}^{b}(n-1)s^{-n}s_0^{n-1}f(s_0)ds+\int_{s_1}^{b}(n-1)s^{-n}s_1^{n-1}f(s_1)+ \ldots + \right . \\ & + & \left .
\int_{s_{h_p-1}}^{b}(n-1)s^{-n}s_{{h_p-1}}^{n-1}f(s_{h_p-1 } ) ds \right)= \frac{n}{n-1}\frac{b - a}{h_p } \left ( \left [ f(s_0)-\frac{s_0^{n-1}f(s_0)}{b^{n-1}}\right]+ \right .
\\ \end{aligned}\ ] ] @xmath464 + \ldots + \left[f(s_{h_p-1})-\frac{s_{h_p-1}^{n-1}f(s_{h_p-1})}{b^{n-1}}\right ] \right)=\\ & = & \frac{n}{n-1}\sum_{j=0}^{h_p-1}f(s_j)(s_{j+1}-s_j)-\frac{n}{n-1}\sum_{j=0}^{h_p-1}\frac{s_j^{n-1}f(s_j)}{b^{n-1}}(s_{j+1}-s_j).\end{aligned}\ ] ] now , since @xmath465 and @xmath466 we obtain @xmath467 where @xmath468 . in conclusion , from ( [ kn ] ) , ( [ appro1 ] ) and ( [ appro2 ] ) we deduce that @xmath469 holds , for all @xmath145 .
letting @xmath311 tend to @xmath14 , we obtain finally @xmath470 .2 cm * proof of ( [ unico].2 ) * let @xmath1 be as in the hypothesis .
for every @xmath336 and @xmath416 , we have @xmath471 since @xmath1 is uniformly continuous , there exist two @xmath19-sequences @xmath124 and @xmath472 in @xmath4730,1[$ ] such that @xmath474 as soon as @xmath475 holds , for all positive integers @xmath311 .
now , for every @xmath145 we get @xmath476 now , for all @xmath477 we see that @xmath478 while for @xmath479 we get @xmath480 in any case , we find : @xmath481 hence @xmath482 holds , for all @xmath311 : this obviously implies that @xmath483.@xmath484 + * proof of ( [ unico].3 ) * by ( [ unico].1 ) @xmath485 for every @xmath336 , by ( [ unico].2 ) @xmath326 @xmath156-converges to zero and so it is enough to prove equi - absolute continuity of @xmath326 with respect to the modular @xmath486 , in order to apply theorem [ trediecibis ] . from uniform convergence of @xmath411 to @xmath1 we see that , setting @xmath487 @xmath488 is an @xmath19-sequence
. then @xmath489\ , 1_b ) & = & \int_b \bigl|t_n f ( s ) - f(s)\bigr| \ , ds \leq \zeta_n \lambda(b)\end{aligned}\ ] ] for every @xmath237 and whenever @xmath490 .
this clearly implies @xmath139 .
we now turn to ( @xmath146 ) . for each @xmath113 , let @xmath491 $ ] . as we already pointed out , when @xmath492 it holds : @xmath493
so , for @xmath108 , @xmath336 we have @xmath489 \,1_{\mathbb{r}^+ \setminus b_m } ) \leq m_f b^n e\int_{2b}^{+\infty } \frac{1}{s^n } ds \leq m _ f \ b^n \ \frac{1}{(2b)^{n-1}}e= \frac{m_f b}{2^{n-1}}e.\end{aligned}\ ] ] this clearly means that @xmath494 \,1_{\mathbb{r}^+ \setminus b_m } ) = 0 $ ] uniformly with respect to @xmath108 , and hence ( @xmath146 ) .
so , all hypotheses of the vitali theorem are satisfied and the assertion follows from theorem [ trediecibis ] .
@xmath484 + we shall use the following expression for @xmath411 : @xmath495 we first observe that the integrand @xmath496 \to { \bf x}$ ] satisfies ( [ uc - prodotto].1 ) and hence its integral function is uniformly differentiable by ( [ uc - prodotto].4 ) .
so @xmath411 is the product of two functions which satisfy ( [ uc - prodotto].3 ) , i.e. @xmath497 and @xmath498 . by proposition [ uc - prodotto ]
we obtain : @xmath499 formula ( [ recupero ] ) shows that @xmath1 is uniquely determined by each of the functions @xmath500 s , in particular @xmath501 holds for all @xmath502 , and therefore all functions @xmath500 s can be easily obtained from one another .
@xmath484 + 99 l. angeloni - g. vinti , _ convergence and rate of approximation for linear integral operators in @xmath503-spaces in multidimensional setting _ , j. math .
* 349 * ( 2 ) ( 2009 ) , 317334 . l. angeloni - g. vinti , _ approximation with respect to goffman - serrin variation by means of non - convolution type integral operators _ , numer . funct
. anal and optim . *
31 * ( 2010 ) , 519548 , doi : 10.1080/01630563.2010.490549 .
l. angeloni - g. vinti , _ approximation in variation by homothetic operators in multidimensional setting _ ,
( 2012 ) , 121 . c. bardaro - a. boccuto - x. dimitriou - i. mantellini , _ modular filter convergence theorems for abstract sampling - type operators _ , appl .
( 2012 ) , doi:10.1080/00036811.2012.738480 . c. bardaro - p. l. butzer - r. l. stens - g. vinti , _ approximation of the whittaker sampling series in terms of an average modulus of smoothness covering discontinuous signals _ , j. math .
* 316 * ( 2006 ) , 269 - 306 . c. bardaro - p. l. butzer - r. l. stens - g. vinti , _ kantorovich - type generalized sampling series in the setting of orlicz spaces _
, sampling theory in signal and image processing * 6 * ( 2006 ) , 29 - 52 . c. bardaro - p. l. butzer - r. l. stens - g. vinti , _ prediction by samples from the past with error estimates covering discontinuous signals _ , ieee trans .
theory * 56 * ( 1 ) ( 2010 ) , 614-633 . c. bardaro - i. mantellini
, _ uniform modular integrability and convergence properties for a class of urysohn integral operators in function spaces , _ math .
slovaca * 56 * ( 4 ) ( 2006 ) , 465 - 482 . c. bardaro - i. mantellini , _ approximation properties in abstract modular spaces for a class of general sampling type operators , _ appl . anal .
* 85 * ( 4 ) ( 2006 ) , 383 - 413 . c. bardaro - j. musielak - g. vinti ,
nonlinear integral operators and applications .
de gruyter series in nonlinear analysis and applications , 9 .
walter de gruyter & co. , berlin , 2003 .
isbn : 3 - 11 - 017551 - 7 c. bardaro - g. vinti , _ integral operators on vector measures and applications to the moment kernel _ , rend . mat .
* 8 * ( 1988 ) , 149 - 164 . c. bardaro - g. vinti , _ modular convergence in generalized orlicz spaces for moment type operators _ , appl
* 32 * ( 1989 ) , 265 - 276 . c. bardaro - g. vinti , _ on convergence of moment operators with respect to the @xmath504- variation _ , appl
* 41 * ( 1991 ) , 247 - 256 .
a. boccuto - d. candeloro , _ integral and ideals in riesz spaces _ , inform .
* 179 * ( 2009 ) , 647 - 660 .
a. boccuto - d. candeloro , _ differential calculus in vector lattice sand applications to g - calculus _ , mediterranean j. math .
* 8 * ( 3 ) ( 2011 ) , 315 - 329 .
a. boccuto - x. dimitriou , _ modular filter convergence theorems for urysohn integral operators and applications _ , acta math .
( 2013 ) , doi:10.1007/s10114 - 013 - 1443 - 6 .
a. boccuto - x. dimitriou , _ modular convergence theorems for integral operators in the context of filter exhaustiveness and applications _
, mediterranean j. math . * 10 * ( 2 ) ( 2013 ) , 823842 , doi:10.1007/s00009 - 012 - 0199-z a. boccuto - x. dimitriou - n. papanastassiou , _ schur lemma and limit theorems in lattice groups with respect to filters , _ math .
slovaca * 62 * ( 6 ) ( 2012 ) , 1145 - 1166 .
a. boccuto - a. m. minotti - a. r. sambucini , _ set - valued kurzweil - henstock integral in riesz space setting _ , panamerican mathematical journal * 23 * ( 1 ) ( 2013 ) , 5774 .
a. boccuto - a. r. sambucini , _ abstract integration in convergence groups , _ atti sem .
modena * 46 * ( 1998 ) , 315 - 333 .
p. breiman , probability .
addison - wesley publishing company , reading , mass .- london - don mills , ont . , 1968 .
l. butzer - r. j. nessel , fourier analysis and approximation , pure appl . math . * 40 * , academic press , new york , london , 1971 .
j. r. choksi , _
vitali s convergence theorem on term by term integration , _
lenseignement mathmatique * 47 * ( 2001 ) , 269 - 285 .
d. costarelli - g. vinti _ approximation by multivariate generalized sampling kantorovich operator in the setting of orlicz spaces _ ,
viii * 9 * ( 4 ) ( 2011 ) , 445 - 468 .
a. m. garsia - e. rodemich - h. rumsey jr . , _ a real variable lemma and the continuity of paths of some gaussian processes _ , indiana univ .
* 20 * ( 1970/71 ) , 565 - 578 .
r. b. holmes , _ mathematical foundations of signal processing , _ siam review * 21 * ( 3 ) ( 1979 ) , 361 - 388
. w. m. kozlowski , modular function spaces , pure appl .
, marcel dekker , new york , 1988 . j. musielak , orlicz spaces and modular spaces , lecture notes in math . * 1034 * , springer - verlag , new york , 1983 .
d. nualart , _ fractional brownian motion : stochastic calculus and applications _ ,
international congress of mathematicians , madrid , spain , ( ems ) vol . 3 , * 74 * ( 2006 ) 1541 - 1562 m. m. rao - z. d. ren , theory of orlicz spaces , marcel dekker , new york , 1991 .
s. r. s. varadhan , diffusion processes , shandbhag and rao eds . ,
handbook of statistics * 19 * , ( 2001 ) , 853 - 872 .
g. vinti , _ a general approximation result for nonlinear integral operators and applications to signal processing , _ appl .
79 * ( 2001 ) , 217 - 238 .
g. vinti - l. zampogni , _
approximation by means of nonlinear kantorovich sampling type operators in orlicz spaces _ , j. approx .
theory * 161 * 2 ( 2009 ) , 511528 .
g. vinti - l. zampogni , _ a unifying approach to convergence of linear sampling type operators in orlicz spaces _ , adv . diff .
equations * 16 * ( 2011 ) , 573600 . | a vitali - type theorem for vector lattice - valued modulars with respect to filter convergence is proved .
some applications are given to modular convergence theorems for moment operators in the vector lattice setting , and also for the brownian motion and related stochastic processes .
[ section ] [ teorema]lemma [ teorema]proposition [ teorema]corollary [ teorema]remark [ teorema]remark [ teorema]remarks [ teorema]definition [ teorema]condition [ teorema]notations [ teorema]examples [ teorema]example [ teorema]definitions [ teorema]definitions [ teorema]assumptions vector lattice , filter convergence , modular , vitali convergence theorem , moment operator , brownian motion , it integral + 28b15 , 41a35 , 46g10 |
let @xmath0 be a self - adjoint possibly unbounded operator on a separable hilbert space @xmath1 such that the spectrum of @xmath0 is separated into two disjoint components , that is , @xmath2 let @xmath3 be a bounded self - adjoint operator on @xmath1 .
it is well known ( see , e.g. , ( * ? ? ?
* theorem v.4.10 ) ) that the spectrum of the perturbed self - adjoint operator @xmath4 is confined in the closed @xmath5-neighbourhood of the spectrum of the unperturbed operator @xmath0 , that is , @xmath6 where @xmath7 denotes the open @xmath5-neighbourhood of @xmath8 . in particular , if @xmath9 then the spectrum of the operator @xmath4 is likewise separated into two disjoint components @xmath10 and @xmath11 , where @xmath12 therefore , under condition , the two components of the spectrum of @xmath4 can be interpreted as perturbations of the corresponding original spectral components @xmath13 and @xmath14 of @xmath8 . clearly , the condition is sharp in the sense that if @xmath15 , the spectrum of the perturbed operator @xmath4 may not have separated components at all .
the effect of the additive perturbation @xmath3 on the spectral subspaces for @xmath0 is studied in terms of the corresponding spectral projections .
let @xmath16 and @xmath17 denote the spectral projections for @xmath0 and @xmath4 associated with the borel sets @xmath13 and @xmath18 , respectively .
it is well known that @xmath19 since the corresponding inequality holds for every difference of orthogonal projections in @xmath1 , see , e.g. , ( * ? ? ? * section 34 ) . moreover ,
if @xmath20 then the spectral projections @xmath16 and @xmath17 are unitarily equivalent , see , e.g. , ( * ? ? ?
* theorem i.6.32 ) . in this sense , if inequality holds , the spectral subspace @xmath21 can be understood as a rotation of the unperturbed spectral subspace @xmath22 .
the quantity @xmath23 serves as a measure for this rotation and is called the _ maximal angle _ between the spectral subspaces @xmath22 and @xmath21 .
a short survey on the concept of the maximal angle between closed subspaces of a hilbert space can be found in ( * ? ? ?
* section 2 ) ; see also @xcite , ( * ? ? ?
* theorem 2.2 ) , ( * ? ? ?
* section 2 ) , and references therein .
it is a natural question whether the bound is sufficient for inequality to hold , or if one has to impose a stronger bound on the norm of the perturbation @xmath3 in order to ensure .
basically , the following two problems arise : 1 . what is the best possible constant @xmath24 $ ] such that @xmath25 2 .
which function @xmath26 is best possible in the estimate @xmath27 both the constant @xmath28 and the function @xmath29 are supposed to be universal in the sense that they are independent of the operators @xmath0 and @xmath3 .
note that we have made no assumptions on the disposition of the spectral components @xmath13 and @xmath14 other than . if , for example , @xmath13 and @xmath14 are additionally assumed to be subordinated , that is , @xmath30 or vice versa , or if one of the two sets lies in a finite gap of the other one , then the corresponding best possible constant in problem ( i ) is known to be @xmath31 , and the best possible function @xmath29 in problem ( ii ) is given by @xmath32 , see , e.g. , ( * ? ? ?
* lemma 2.3 ) and ( * ? ? ?
* theorem 5.1 ) ; see also ( * ? ? ?
* remark 2.9 ) .
however , under the sole assumption , both problems are still unsolved .
it has been conjectured that @xmath33 ( see @xcite ; cf . also @xcite and @xcite ) , but there is no proof available for that yet .
so far , only lower bounds on the optimal constant @xmath34 and upper bounds on the best possible function @xmath29 can be given .
for example , in ( * ? ?
* theorem 1 ) it was shown that @xmath35 and @xmath36 in ( * ? ? ?
* theorem 6.1 ) this result was strengthened to @xmath37 and @xmath38 recently , albeverio and motovilov have shown in ( * ? ? ? * theorem 5.4 ) that @xmath39 and @xmath40 where @xmath41 \frac{1}{2}\arcsin\bigl(\frac{4\pi}{\pi^2 + 4}\bigr ) + \frac{1}{2}\arcsin\bigl(\pi\,\frac{(\pi^2 + 4)x-4 } { \pi^2 - 4}\bigr ) & \text{for}\quad \frac{4}{\pi^2 + 4 } < x
\le \frac{8\pi^2}{(\pi^2 + 4)^2}\,,\\[0.2 cm ] \arcsin\bigl(\frac{4\pi}{\pi^2 + 4}\bigr ) + \frac{1}{2}\arcsin\bigl(\pi\,\frac{(\pi^2 + 4)^2x-8\pi^2 } { ( \pi^2 - 4)^2}\bigr ) & \text{for}\quad \frac{8\pi^2}{(\pi^2 + 4)^2 } < x \le
c_*\ , . \end{cases}\ ] ] it should be noted that the first two results and were originally formulated in @xcite and @xcite , respectively , only for the case where the operator @xmath0 is assumed to be bounded .
however , both results admit an immediate , straightforward generalization to the case where the operator @xmath0 is allowed to be unbounded , see , e.g. , ( * ? ? ? * proposition 3.4 and theorem 3.5 ) . the aim of the present work is to sharpen the estimate .
more precisely , our main result is as follows .
[ thm : mainresult ] let @xmath0 be a self - adjoint operator on a separable hilbert space @xmath1 such that the spectrum of @xmath0 is separated into two disjoint components , that is , @xmath42 let @xmath3 be a bounded self - adjoint operator on @xmath1 satisfying @xmath43 with @xmath44 then , the spectral projections @xmath16 and @xmath17 for the self - adjoint operators @xmath0 and @xmath4 associated with @xmath13 and the open @xmath45-neighbourhood @xmath18 of @xmath13 , respectively , satisfy the estimate @xmath46 where the function @xmath47\to\bigl[0,\frac{\pi}{2}\bigr]$ ] is given by @xmath48 \arcsin\bigl(\sqrt{\frac{2\pi^2x-4}{\pi^2 - 4}}\,\bigr ) & \text { for } \quad \frac{4}{\pi^2 + 4 } < x < 4\,\frac{\pi^2 - 2 } { \pi^4}\,,\\[0.15 cm ] \arcsin\bigl(\frac{\pi}{2}(1-\sqrt{1 - 2x}\,)\bigr ) & \text { for }
\quad 4\,\frac{\pi^2 - 2}{\pi^4 } \le x \le \kappa\,,\\[0.15 cm ] \frac{3}{2}\arcsin\bigl(\frac{\pi}{2}(1-\sqrt[\leftroot{4}3]{1 - 2x}\,)\bigr ) & \text { for } \quad \kappa
< x \le c_\mathrm{crit}\ , . \end{cases}\ ] ] here , @xmath49 is the unique solution to the equation @xmath50{1 - 2\kappa}\,\bigr)\bigr)\ ] ] in the interval @xmath51 $ ] .
the function @xmath52 is strictly increasing , continuous on @xmath53 $ ] , and continuously differentiable on @xmath54 .
numerical calculations give @xmath55 the estimate in theorem [ thm : mainresult ] remains valid if the constant @xmath56 in the definition of the function @xmath52 is replaced by any other constant within the interval @xmath57 , see remark [ rem : kapparepl ] below .
however , the particular choice ensures that the function @xmath52 is continuous and as small as possible .
in particular , we have @xmath58 for @xmath59 and @xmath60 where @xmath61 and @xmath62 are given by and respectively , see remark [ rem : estoptimality ] below . from theorem [ thm : mainresult ]
we immediately deduce that @xmath63 and @xmath64 both are the best respective bounds for the two problems ( i ) and ( ii ) known so far .
the paper is organized as follows : in section [ sec : optprob ] , based on the triangle inequality for the maximal angle and a suitable a priori rotation bound for small perturbations ( see proposition [ prop : genrotbound ] ) , we formulate a constrained optimization problem , whose solution provides an estimating function for the maximal angle between the corresponding spectral subspaces , see definition [ def : optprob ] , proposition [ prop : mainestimate ] , and theorem [ thm : soloptprob ] . in this way ,
the approach by albeverio and motovilov in @xcite is optimized and , in particular , a proof of theorem [ thm : mainresult ] is obtained . the explicit solution to the optimization problem
is given in theorem [ thm : soloptprob ] , which is proved in section [ sec : soloptprob ] .
the technique used there involves variational methods and may also be useful for solving optimization problems of a similar structure .
finally , appendix [ app : sec : inequalities ] is devoted to some elementary inequalities used in section [ sec : soloptprob ] .
in this section , we formulate a constrained optimization problem , whose solution provides an estimate on the maximal angle between the spectral subspaces associated with isolated parts of the spectrum of the corresponding perturbed and unperturbed operators , respectively .
in particular , this yields a proof of theorem [ thm : mainresult ] .
we make the following notational setup .
[ app : hyphyp ] let @xmath0 be as in theorem [ thm : mainresult ] , and let @xmath65 be a bounded self - adjoint operator on the hilbert space @xmath1
. for @xmath66 , introduce @xmath67 , @xmath68 , and denote by @xmath69 the spectral projection for @xmath70 associated with the open @xmath45-neighbourhood @xmath18 of @xmath13 .
under hypothesis [ app : hyphyp ] , one has @xmath71 for @xmath66 .
taking into account the inclusion , the spectrum of each @xmath70 is likewise separated into two disjoint components , that is , @xmath72 where @xmath73 in particular , one has @xmath74 moreover , the mapping @xmath75 is norm continuous , see , e.g. , ( * ? ? ?
* theorem 3.5 ) ; cf .
also the forthcoming estimate . for arbitrary @xmath76
, we can consider @xmath77 as a perturbation of @xmath78 .
taking into account the a priori bound , we then observe that @xmath79 furthermore , it follows from and the inclusion that @xmath80 is exactly the part of @xmath81 that is contained in the open @xmath82-neighbourhood of @xmath83 , that is , @xmath84 let @xmath85 be arbitrary , and let @xmath86 with @xmath87 be a finite partition of the interval @xmath88 $ ] .
define @xmath89 recall that the mapping @xmath90 given by @xmath91 defines a metric on the set of orthogonal projections in @xmath1 , see @xcite , and also ( * ? ? ?
* lemma 2.15 ) and @xcite . using the triangle inequality for this metric
, we obtain @xmath92 considering @xmath93 as a perturbation of @xmath94 , it is clear from and that each summand of the right - hand side of can be treated in the same way as the maximal angle in the general situation discussed in section [ sec : intro ] .
for example , combining with the bound yields @xmath95 where we have taken into account that @xmath96 and that @xmath97 if @xmath98 .
obviously , the estimates and hold for arbitrary finite partitions of the interval @xmath88 $ ] . in particular , if partitions with arbitrarily small mesh size are considered , then , as a result of @xmath99 , the norm of each corresponding projector difference in is arbitrarily small as well .
at the same time , the corresponding riemann sums @xmath100 are arbitrarily close to the integral @xmath101 .
since @xmath102 as @xmath103 , we conclude from and that @xmath104 once the bound has been generalized to the case where the operator @xmath0 is allowed to be unbounded , this argument is an easy and straightforward way to prove the bound .
albeverio and motovilov demonstrated in @xcite that a stronger result can be obtained from .
they considered a specific finite partition of the interval @xmath88 $ ] and used a suitable a priori bound ( see ( * ? ? ?
* corollary 4.3 and remark 4.4 ) ) to estimate the corresponding summands of the right - hand side of .
this a priori bound , which is related to the davis - kahan @xmath105 theorem from @xcite , is used in the present work as well .
we therefore state the corresponding result in the following proposition for future reference .
it should be noted that our formulation of the statement slightly differs from the original one in @xcite . a justification of this modification , as well as a deeper discussion on the material including an alternative , straightforward proof of the original result ( * ? ? ?
* corollary 4.3 ) , can be found in @xcite .
[ prop : genrotbound ] let @xmath0 and @xmath3 be as in theorem [ thm : mainresult ] .
if @xmath106 , then the spectral projections @xmath107 and @xmath17 for the self - adjoint operators @xmath0 and @xmath4 associated with the borel sets @xmath13 and @xmath18 , respectively , satisfy the estimate @xmath108 the estimate given by proposition [ prop : genrotbound ] is universal in the sense that the estimating function @xmath109 depends neither on the unperturbed operator @xmath0 nor on the perturbation @xmath3 .
moreover , for perturbations @xmath3 satisfying @xmath110 , this a priori bound on the maximal angle between the corresponding spectral subspaces is the strongest one available so far , cf .
* remark 5.5 ) .
assume that the given partition of the interval @xmath88 $ ] additionally satisfies @xmath111 in this case , it follows from , , , and proposition [ prop : genrotbound ] that @xmath112 along with a specific choice of the partition of the interval @xmath88 $ ] , estimate is the essence of the approach by albeverio and motovilov in @xcite . in the present work ,
we optimize the choice of the partition of the interval @xmath88 $ ] , so that for every fixed parameter @xmath113 the right - hand side of inequality is minimized .
an equivalent and more convenient reformulation of this approach is to maximize the parameter @xmath113 in estimate over all possible choices of the parameters @xmath114 and @xmath115 for which the right - hand side of takes a fixed value .
obviously , we can generalize estimate to the case where the finite sequence @xmath116 is allowed to be just increasing and not necessarily strictly increasing .
altogether , this motivates the following considerations .
[ def : params ] for @xmath87 define @xmath117 and let @xmath118 .
every finite partition of the interval @xmath88 $ ] that satisfies condition is related to a sequence in @xmath119 in the obvious way .
conversely , the following lemma allows to regain the finite partition of the interval @xmath88 $ ] from this sequence .
[ lem : seq ] 1 . for every @xmath120 the mapping @xmath121\ni t\mapsto t+x(1 - 2t)$ ] is strictly increasing .
2 . for every @xmath122
the sequence @xmath123 given by the recursion @xmath124 is increasing and satisfies @xmath125 for all @xmath126 .
moreover , one has @xmath127 for @xmath128 if @xmath129 . in particular , @xmath130 is eventually constant .
the proof of claim ( a ) is straightforward and is hence omitted . for the proof of ( b ) , let @xmath122 be arbitrary and let @xmath123 be given by . observe that @xmath131 and that ( a ) implies that @xmath132 thus , the two - sided estimate @xmath125 holds for all @xmath126 by induction . in particular
, it follows that @xmath133 for all @xmath126 , that is , the sequence @xmath130 is increasing .
let @xmath87 such that @xmath129 .
since @xmath134 for @xmath128 , it follows from the definition of @xmath130 that @xmath135 for @xmath128 , that is , @xmath127 for @xmath128 .
it follows from part ( b ) of the preceding lemma that for every @xmath136 the sequence @xmath130 given by yields a finite partition of the interval @xmath88 $ ] with @xmath137 . in this respect ,
the approach to optimize the parameter @xmath113 in with a fixed right - hand side can now be formalized in the following way .
[ def : optprob ] let @xmath138 denote the ( non - linear ) operator that maps every sequence in @xmath119 to the corresponding increasing and eventually constant sequence given by the recursion .
moreover , let @xmath139\to\bigl[0,\frac{\pi}{4}\bigr]$ ] be given by @xmath140 finally , for @xmath141 $ ] define @xmath142 and @xmath143 where @xmath144 with @xmath145 .
for every fixed @xmath141 $ ] , it is easy to verify that indeed @xmath146 .
moreover , one has @xmath147 by part ( b ) of lemma [ lem : seq ] , and @xmath148 holds if and only if @xmath149 . in order to compute @xmath150 for @xmath151 $ ]
, we have to maximize @xmath152 over @xmath153 .
this constrained optimization problem plays the central role in the approach presented in this work .
the following proposition shows how this optimization problem is related to the problem of estimating the maximal angle between the corresponding spectral subspaces .
[ prop : mainestimate ] assume hypothesis [ app : hyphyp ] .
let @xmath154\ni\theta\mapsto s(\theta)\in\bigl[0,s\bigl(\frac{\pi}{2}\bigr)\bigr]\subset\bigl[0,\frac{1}{2}\bigr]$ ] be a continuous , strictly increasing ( hence invertible ) mapping with @xmath155 then @xmath156 since the mapping @xmath157 is invertible , it suffices to show the inequality @xmath158 considering @xmath159 , the case @xmath149 in inequality is obvious .
let @xmath160 .
in particular , one has @xmath161 . for arbitrary @xmath113 with @xmath162
choose @xmath163 such that @xmath164 .
denote @xmath165 .
since @xmath166 for all @xmath126 by part ( b ) of lemma [ lem : seq ] , it follows from the definition of @xmath130 that @xmath167 moreover , considering @xmath168 , there is @xmath169 such that @xmath170 . in particular
, one has @xmath171 using the triangle inequality for the metric @xmath90 given by , it follows from , , , , and proposition [ prop : genrotbound ] that @xmath172 that is , @xmath173 since the mapping @xmath174 is norm continuous and @xmath175 , estimate also holds for @xmath176 .
this shows and , hence , completes the proof .
it turns out that the mapping @xmath154\ni\theta\mapsto t(\theta)$ ] is continuous and strictly increasing .
it therefore satisfies the hypotheses of proposition [ prop : mainestimate ] . in this respect , it remains to compute @xmath150 for @xmath141 $ ] in order to prove theorem [ thm : mainresult ] .
this is done in section [ sec : soloptprob ] below .
for convenience , the following theorem states the corresponding result in advance .
[ thm : soloptprob ] in the interval @xmath177 $ ] the equation @xmath178 has a unique solution @xmath179 .
moreover , the quantity @xmath150 given in has the representation @xmath180 \frac{2}{\pi^2}+\frac{\pi^2 - 4}{2\pi^2}\sin^2\theta & \text { for } \quad \arctan\bigl(\frac{2}{\pi } \bigr)<\theta<\arcsin\bigl(\frac{2}{\pi}\bigr)\,,\\[0.1 cm ] \frac{1}{2}-\frac{1}{2}\bigl(1-\frac{2}{\pi}\sin\theta\bigr)^2 & \text { for } \quad \arcsin\bigl(\frac{2}{\pi}\bigr)\le\theta\le\vartheta\,,\\[0.1 cm ] \frac{1}{2}-\frac{1}{2}\bigl(1-\frac{2}{\pi}\sin\bigl(\frac{2\theta}{3}\bigr)\bigr)^3 & \text { for } \quad \vartheta <
\theta \le \frac{\pi}{2}\ , . \end{cases}\ ] ] the mapping @xmath154\ni\theta\mapsto t(\theta)$ ] is strictly increasing , continuous on @xmath154 $ ] , and continuous differentiable on @xmath181 . theorem [ thm : mainresult ] is now a straightforward consequence of proposition [ prop : mainestimate ] and theorem [ thm : soloptprob ] . according to theorem [ thm : soloptprob ] , the mapping @xmath154\ni\theta\mapsto t(\theta)$ ] is strictly increasing and continuous .
hence , its range is the whole interval @xmath53 $ ] , where @xmath182 is given by @xmath183 .
let @xmath184\to\bigl[0,\frac{\pi}{2}\bigr]$ ] denote the inverse of this mapping .
obviously , the function @xmath52 is also strictly increasing and continuous .
moreover , using representation , it is easy to verify that @xmath52 is explicitly given by . in particular , the constant @xmath185 is the unique solution to equation in the interval @xmath51 $ ] .
furthermore , the function @xmath52 is continuously differentiable on @xmath54 since the mapping @xmath186 is continuously differentiable on @xmath181 .
let @xmath3 be a bounded self - adjoint operator on @xmath1 satisfying @xmath187 .
the case @xmath188 is obvious .
assume that @xmath65 .
then , @xmath189 , @xmath68 , and @xmath69 for @xmath66 satisfy hypothesis [ app : hyphyp ] .
moreover , one has @xmath190 with @xmath191 .
applying proposition [ prop : mainestimate ] to the mapping @xmath186 finally gives @xmath192 which completes the proof .
[ rem : kapparepl ] numerical evaluations give @xmath193 and @xmath194 .
however , the estimate remains valid if the constant @xmath56 in the explicit representation for the function @xmath52 is replaced by any other constant within the interval @xmath57 .
this can be seen by applying proposition [ prop : mainestimate ] to each of the two mappings @xmath195 these mappings indeed satisfy the hypotheses of proposition [ prop : mainestimate ] .
both are obviously continuous and strictly increasing , and , by particular choices of @xmath196 , it is easy to see from the considerations in section [ sec : soloptprob ] that they are less or equal to @xmath150 , see equation below .
the statement of theorem [ thm : soloptprob ] actually goes beyond that of theorem [ thm : mainresult ] . as a matter of fact , instead of equality in
, it would be sufficient for the proof of theorem [ thm : mainresult ] to have that the right - hand side of is just less or equal to @xmath150 .
this , in turn , is rather easy to establish by particular choices of @xmath196 , see lemma [ lem : critpoints ] and the proof of lemma [ lem : twoparams ] below .
however , theorem [ thm : soloptprob ] states that the right - hand side of provides an exact representation for @xmath150 , and most of the considerations in section [ sec : soloptprob ] are required to show this stronger result . as a consequence ,
the bound from theorem [ thm : mainresult ] is optimal within the framework of the approach by estimate .
in fact , the following observation shows that a bound substantially stronger than the one from proposition [ prop : genrotbound ] is required , at least for small perturbations , in order to improve on theorem [ thm : mainresult ] .
one can modify the approach by replacing the term @xmath197 by @xmath198 and relaxing the condition to @xmath199 . yet
, it follows from theorem [ thm : soloptprob ] that the corresponding optimization procedure leads to exactly the same result .
this can be seen from the fact that each @xmath198 is of the form of the right - hand side of ( cf .
the computation of @xmath150 in section [ sec : soloptprob ] below ) , so that we are actually dealing with essentially the same optimization problem . in this sense , the function @xmath52 is a fixed point in the approach presented here .
we close this section with a comparison of theorem [ thm : mainresult ] with the strongest previously known result by albeverio and motovilov from @xcite .
[ rem : estoptimality ] one has @xmath200 for @xmath201 , and the inequality @xmath202 holds for all @xmath203 , where @xmath204 and @xmath205\to\bigl[0,\frac{\pi}{2}\bigr]$ ] are given by and , respectively . indeed , it follows from the computation of @xmath150 in section [ sec : soloptprob ] ( see remark [ rem : amvss ] below ) that @xmath206 since the function @xmath207\to\bigl[0,\frac{\pi}{2}\bigr]$ ] is strictly increasing , this implies that @xmath208
we split the proof of theorem [ thm : soloptprob ] into several steps .
we first reduce the problem of computing @xmath150 to the problem of solving suitable finite - dimensional constrained optimization problems , see equations and .
the corresponding critical points are then characterized in lemma [ lem : critpoints ] using lagrange multipliers . the crucial tool to reduce the set of relevant critical points
is provided by lemma [ lem : paramsubst ] .
finally , the finite - dimensional optimization problems are solved in lemmas [ lem : twoparams ] , [ lem : threeparams ] , and [ prop : soloptprob ] . throughout this section
, we make use of the notations introduced in definitions [ def : params ] and [ def : optprob ] .
in addition , we fix the following notations . for @xmath87 and @xmath141 $ ] define @xmath209
. moreover , let @xmath210 and set @xmath211 if @xmath212 . as a result of @xmath213 , we have @xmath214 for every @xmath87 .
let @xmath151 $ ] be arbitrary .
since @xmath215 , we obtain @xmath216 moreover , we observe that @xmath217 in fact , we show below that @xmath218 for every @xmath219 , so that @xmath220 , see lemma [ prop : soloptprob ] .
let @xmath221 be arbitrary and let @xmath222 .
denote @xmath165 .
it follows from part ( b ) of lemma [ lem : seq ] that @xmath223 .
moreover , we have @xmath224 since @xmath225 , this implies that @xmath226 in particular , we obtain the explicit representation @xmath227 an immediate conclusion of representation is the following statement . [
lem : parampermut ] for @xmath222 the value of @xmath152 does not depend on the order of the entries @xmath228 .
another implication of representation is the fact that @xmath223 can be considered as a continuous function of the variables @xmath228 .
since the set @xmath229 is compact as a closed bounded subset of an @xmath230-dimensional subspace of @xmath231 , we deduce that @xmath232 can be written as @xmath233 hence , @xmath232 is determined by a finite - dimensional constrained optimization problem , which can be studied by use of lagrange multipliers . taking into account the definition of the set @xmath229
, it follows from equation and representation that there is some point @xmath234^{n+1}$ ] such that @xmath235 where @xmath236 for @xmath237 .
in particular , if @xmath238 , then the method of lagrange multipliers gives a constant @xmath239 , @xmath240 , with @xmath241 hence , in this case , for every @xmath242 we obtain @xmath243 this leads to the following characterization of critical points of the mapping @xmath244 on @xmath229 .
[ lem : critpoints ] for @xmath245 and @xmath151 $ ] let @xmath246 with @xmath247 .
assume that @xmath248 .
if , in addition , @xmath249 and @xmath250 , then either one has @xmath251 so that @xmath252 or there is @xmath253 with @xmath254 in the latter case , @xmath255 and @xmath256 satisfy @xmath257 where @xmath258 let @xmath249 and @xmath250 . in particular , one has @xmath238 .
hence , it follows from that @xmath259 does not depend on @xmath260 . if @xmath261 , then all @xmath115 coincide and one has @xmath262 , that is , @xmath263 . inserting this into representation yields equation .
now assume that @xmath264 .
a straightforward calculation shows that @xmath265 is the only critical point of the mapping @xmath266\ni
x\mapsto \frac{\sqrt{1-\pi^2x^2}}{1 - 2x}\,,\ ] ] cf .
[ fig : functionplot ] .
the image of this point is @xmath267 .
moreover , @xmath268 and @xmath269 are mapped to @xmath270 , and @xmath271 is mapped to @xmath268 . in particular , every value in the interval @xmath272 has exactly two preimages under the mapping , and all the other values in the range @xmath273 $ ] have only one preimage .
since @xmath264 by assumption , it follows from that @xmath274 has two preimages .
hence , @xmath275 and @xmath276 .
furthermore , there is @xmath253 with @xmath277 and @xmath278 .
this proves and .
finally , the relations follow from the fact that the equation @xmath279 can be rewritten as @xmath280 the preceding lemma is one of the main ingredients for solving the constrained optimization problem that defines the quantity @xmath232 in .
however , it is still a hard task to compute @xmath232 from the corresponding critical points .
especially the case in lemma [ lem : critpoints ] is difficult to handle and needs careful treatment .
an efficient computation of @xmath232 therefore requires a technique that allows to narrow down the set of relevant critical points .
the following result provides an adequate tool for this and is thus crucial for the remaining considerations .
the idea behind this approach may also prove useful for solving similar optimization problems .
[ lem : paramsubst ] for @xmath245 and @xmath151 $ ] let @xmath246 .
if @xmath247 , then for every @xmath260 one has @xmath281 suppose that @xmath247 .
the case @xmath282 in the claim obviously agrees with this hypothesis .
let @xmath242 be arbitrary and denote @xmath165 .
it follows from part ( b ) of lemma [ lem : seq ] that @xmath283 . in particular , one has @xmath284 since @xmath285 .
assume that @xmath286 , and let @xmath287 with @xmath288 .
denote @xmath289 and @xmath290 .
again by part ( b ) of lemma [ lem : seq ] , one has @xmath291 and @xmath292 .
taking into account part ( a ) of lemma [ lem : seq ] and the definition of the operator @xmath293 , one obtains that @xmath294 iterating this estimate eventually gives @xmath295 , which contradicts the case @xmath282 from above .
thus , @xmath296 as claimed .
lemma [ lem : paramsubst ] states that if a sequence @xmath297 solves the optimization problem for @xmath232 , then every truncation of @xmath298 solves the corresponding reduced optimization problem .
this allows to exclude many sequences in @xmath229 from the considerations once the optimization problem is understood for small @xmath114 .
the number of parameters in can thereby be reduced considerably .
the following lemma demonstrates this technique .
it implies that the condition @xmath249 in lemma [ lem : critpoints ] is always satisfied except for one single case , which can be treated separately .
[ lem : lambda_0 ] for @xmath245 and @xmath151 $ ] let @xmath246 with @xmath247 and @xmath248 .
if @xmath299 , then @xmath300 and @xmath301 .
let @xmath299 and define @xmath302 .
it is obvious that @xmath303 is equivalent to @xmath304 .
assume that @xmath305 . clearly , one has @xmath306 and @xmath307 . taking into account representation , for
@xmath308 one computes @xmath309 since @xmath310 , it follows from part ( [ app : lem1:a ] ) of lemma [ app1:lem1 ] that @xmath311 where the last inequality is due to representation .
this is a contradiction to lemma [ lem : paramsubst ] .
hence , @xmath304 and , in particular , @xmath312 .
obviously , one has @xmath313 , so that @xmath314 .
taking into account that @xmath315 , it follows from representations and that @xmath316 since @xmath317 by hypothesis , this implies that @xmath301 .
we are now able to solve the finite - dimensional constrained optimization problem in for every @xmath141 $ ] and @xmath221 .
we start with the case @xmath301 .
[ lem : twoparams ] the quantity @xmath318 has the representation @xmath319 \frac{2}{\pi^2}+\frac{\pi^2 - 4}{2\pi^2}\sin^2\theta & \text { for } \quad
\arctan\bigl(\frac{2}{\pi}\bigr)<\theta<\arcsin\bigl(\frac{2}{\pi}\bigr)\,,\\[0.1 cm ] \frac{1}{2}-\frac{1}{2}\bigl(1-\frac{2}{\pi}\sin\theta\bigr)^2 & \text { for } \quad \arcsin\bigl(\frac{2}{\pi}\bigr)\le\theta\le\frac{\pi}{2}\ , .
\end{cases}\ ] ] in particular , if @xmath320 and @xmath321 with @xmath322 , then the strict inequality @xmath323 holds .
the mapping @xmath154\ni\theta\mapsto t_1(\theta)$ ] is strictly increasing , continuous on @xmath154 $ ] , and continuously differentiable on @xmath324 .
since @xmath325 , the representation is obviously correct for @xmath149 . for @xmath300 one has @xmath313 , so that @xmath326 by representation
. this also agrees with the claim .
now let @xmath160 be arbitrary .
obviously , one has @xmath327 if @xmath328 , and @xmath329 if @xmath330 .
hence , @xmath331 and @xmath332 if @xmath330 . by lemmas [ lem : parampermut ] , [ lem : critpoints ] , and [ lem : lambda_0 ]
there are only two sequences in @xmath333 that need to be considered in order to compute @xmath318 .
one of them is given by @xmath334 with @xmath335 .
for this sequence , representation yields @xmath336 the other sequence in @xmath333 that needs to be considered is @xmath337 with @xmath255 and @xmath338 satisfying @xmath339 and @xmath340 where @xmath341 it turns out shortly that this sequence @xmath298 exists if and only if @xmath342 . using representation and the relations in , one obtains @xmath343 the objective is to rewrite the right - hand side of in terms of @xmath344 .
it follows from @xmath345 and the relations and that @xmath346 taking into account that @xmath347 , equation can be rewritten as @xmath348 in turn , this gives @xmath349 that is , @xmath350 we show that the second case in does not occur . since @xmath351 , by equation one has @xmath352 , which implies that @xmath353
moreover , combining relations and , @xmath338 can be expressed in terms of @xmath255 alone .
hence , by equation the quantity @xmath344 can be written as a continuous function of the sole variable @xmath354 . taking the limit @xmath355 in equation
then implies that @xmath356 and , therefore , @xmath357 .
this yields @xmath358 for every @xmath359 by continuity , that is , the sequence @xmath298 can exist only if @xmath358 .
taking into account that @xmath274 satisfies @xmath360 , it now follows from that the sequence @xmath298 exists if and only if @xmath342 and , in this case , one has @xmath361
combining equations and finally gives @xmath362 for @xmath342 .
as a result of lemmas [ lem : parampermut ] , [ lem : critpoints ] , and [ lem : lambda_0 ] , the quantities , , and are the only possible values for @xmath318 , and we have to determine which of them is the greatest . the easiest case is @xmath330 since then is the only possibility for @xmath318 .
the quantity is relevant only if @xmath363 . in this case
, it follows from parts ( [ app : lem1:b ] ) and ( [ app : lem1:c ] ) of lemma [ app1:lem1 ] that gives the greatest value of the three possibilities and , hence , is the correct term for @xmath318 here . for @xmath364 , by part ( [ app : lem1:d ] ) of lemma [ app1:lem1 ] the quantity is greater than .
therefore , @xmath318 is given by in this case .
finally , consider the case @xmath365 . since @xmath366
, it follows from part ( [ app : lem1:e ] ) of lemma [ app1:lem1 ] that is greater than and , hence , coincides with @xmath318 .
this completes the computation of @xmath318 for @xmath141 $ ] .
in particular , it follows from the discussion of the two cases @xmath367 and @xmath342 that @xmath368 is always strictly less than @xmath318 if @xmath320 .
the piecewise defined mapping @xmath154\ni\theta\mapsto t_1(\theta)$ ] is continuously differentiable on each of the corresponding subintervals .
it remains to prove that the mapping is continuous and continuously differentiable at the points @xmath369 and @xmath370 .
taking into account that @xmath371 for @xmath372 , the continuity is straightforward to verify .
the continuous differentiability follows from the relations @xmath373 and @xmath374 where the latter is due to @xmath375 this completes the proof .
so far , lemma [ lem : paramsubst ] has been used only to obtain lemma [ lem : lambda_0 ] .
its whole strength becomes apparent in connection with lemma [ lem : parampermut ] .
this is demonstrated in the following corollary to lemma [ lem : twoparams ] , which states that in the sequences with @xmath376 do not need to be considered .
[ cor : critpoints : ln ] in the case in lemma [ lem : critpoints ] one has @xmath377 .
the case @xmath301 is obvious . for @xmath219 let @xmath378 with @xmath379 for some @xmath376 . in particular
, one has @xmath380 , which implies that @xmath381 .
hence , it follows from lemma [ lem : twoparams ] that @xmath382 by lemmas [ lem : parampermut ] and [ lem : paramsubst ] one concludes that @xmath383 this leaves @xmath377 as the only possibility in .
we now turn to the computation of @xmath384 for @xmath141 $ ] .
[ lem : threeparams ] in the interval @xmath177 $ ] the equation @xmath385 has a unique solution @xmath179 .
moreover , the quantity @xmath384 has the representation @xmath386
\dfrac{1}{2}-\dfrac{1}{2}\biggl(1-\dfrac{2}{\pi}\sin\bigl(\dfrac{2\theta}{3}\bigr)\biggr)^3 & \text { for } \quad \vartheta<\theta\le\frac{\pi}{2}\ , . \end{cases}\ ] ] in particular , one has @xmath387 if @xmath388 , and the strict inequality @xmath389 holds for @xmath151 $ ] and @xmath390 with @xmath391 .
the mapping @xmath154\ni\theta\mapsto t_2(\theta)$ ] is strictly increasing , continuous on @xmath154 $ ] , and continuously differentiable on @xmath181 . since @xmath392 , the case @xmath149 in the representation for @xmath384 is obvious .
let @xmath151 $ ] be arbitrary .
it follows from lemmas [ lem : parampermut ] , [ lem : critpoints ] , and [ lem : lambda_0 ] and corollary [ cor : critpoints : ln ] that there are only two sequences in @xmath393 that need to be considered in order to compute @xmath384 .
one of them is @xmath394 with @xmath395 . for this sequence representation yields @xmath396
the other sequence in @xmath393 that needs to be considered is @xmath397 , where @xmath398 and @xmath255 and @xmath399 are given by and .
using representation , one obtains @xmath400 according to lemma [ app : lem3 ] , this sequence @xmath298 can exist only if @xmath344 satisfies the two - sided estimate @xmath401 .
however , if @xmath298 exists , combining lemma [ app : lem3 ] with equations and yields @xmath402 therefore , in order to compute @xmath384 for @xmath151 $ ] , it remains to compare with @xmath318 .
in particular , for every sequence @xmath403 with @xmath391 the strict inequality @xmath404 holds . according to lemma [ app : lem2 ] , there is a unique @xmath179 such that @xmath405 and @xmath406 these inequalities imply that @xmath407 is the unique solution to equation in the interval @xmath177 $ ] .
moreover , taking into account lemma [ lem : twoparams ] , equation , and the inequality @xmath408 , it follows that @xmath409 if and only if @xmath388 .
this proves the claimed representation for @xmath384 . by lemma
[ lem : twoparams ] and the choice of @xmath407 it is obvious that the mapping @xmath154\ni\theta\mapsto t_2(\theta)$ ] is strictly increasing , continuous on @xmath154 $ ] , and continuously differentiable on @xmath181 . in order to prove theorem [ thm : soloptprob ] , it remains to show that @xmath150 coincides with @xmath384 .
[ prop : soloptprob ] for every @xmath141 $ ] and @xmath219 one has @xmath410 . since @xmath411 ,
the case @xmath149 is obvious .
let @xmath151 $ ] be arbitrary . as a result of equation
, it suffices to show that @xmath218 for all @xmath412 .
let @xmath412 and let @xmath413 .
the objective is to show that @xmath414 .
first , assume that @xmath415 .
we examine the two cases @xmath416 and @xmath417 . if @xmath416 , then @xmath418 . in this case
, it follows from lemma [ lem : twoparams ] that @xmath419 with @xmath420 .
hence , by lemma [ lem : paramsubst ] one has @xmath414 .
if @xmath417 , then @xmath421 which is possible only if @xmath422 , that is , @xmath423 . in this case , one has @xmath424 .
taking into account representation , it follows from lemma [ app : lem4 ] that @xmath425 so , one concludes that @xmath414 again .
now , assume that @xmath413 satisfies @xmath426 . since
, in particular , @xmath427 , lemma [ lem : threeparams ] implies that @xmath428 it follows from lemmas [ lem : parampermut ] and [ lem : paramsubst ] that @xmath429 that is , @xmath430 once again . hence , by lemmas [ lem : parampermut ] , [ lem : critpoints ] , and [ lem : lambda_0 ] and corollary [ cor : critpoints : ln ] the inequality @xmath431 holds for all @xmath432 , which implies that @xmath433 .
now the claim follows by induction .
we close this section with the following observation , which , together with remark [ rem : estoptimality ] above , shows that the estimate from theorem [ thm : mainresult ] is indeed stronger than the previously known estimates .
[ rem : amvss ] it follows from the previous considerations that @xmath434 where @xmath204 and @xmath205\to\bigl[0,\frac{\pi}{2}\bigr]$ ] are given by and , respectively .
indeed , let @xmath435 $ ] be arbitrary and set @xmath436 .
define @xmath437 by @xmath438 and by @xmath439 using representation , a straightforward calculation shows that in both cases one has @xmath440 if @xmath441 ( cf .
remark [ rem : kapparepl ] ) , that is , @xmath442 , then it follows from lemma [ lem : threeparams ] that @xmath443 .
if @xmath444 , then the inequality @xmath445 holds since , in this case , @xmath298 is none of the critical points from lemma [ lem : critpoints ] .
so , in either case one has @xmath446 .
[ app1:lem1 ] the following inequalities hold : 1 .
@xmath447 @xmath448,[app : lem1:a ] 2 .
@xmath449 @xmath450,[app : lem1:b ] 3 .
@xmath451 @xmath452,[app : lem1:c ] 4 .
@xmath453 @xmath454,[app : lem1:d ] 5 .
@xmath455 @xmath456.[app : lem1:e ] one has @xmath457 which is strictly positive if and only if @xmath458 a straightforward analysis shows that the last inequality holds for @xmath459 , which proves ( [ app : lem1:a ] ) . for @xmath460 one
has @xmath461 and @xmath462 .
thus , the inequality in ( b ) becomes an equality for @xmath463 .
therefore , in order to show ( b ) , it suffices to show that the corresponding estimate holds for the derivatives of both sides of the inequality , that is , @xmath464 this inequality is equivalent to @xmath465 for @xmath466 , which , in turn , follows from @xmath467 .
this implies ( [ app : lem1:b ] ) .
the claim ( [ app : lem1:c ] ) follows immediately from @xmath468 finally , observe that @xmath469 for @xmath470 , the right - hand side of is positive if and only if @xmath471 is less than @xmath271 .
this is the case if and only if @xmath472 , which proves ( [ app : lem1:d ] ) .
the proof of claim ( [ app : lem1:e ] ) is analogous .
[ app : lem2 ] there is a unique @xmath179 such that @xmath473 and @xmath474 define @xmath475 by @xmath476 obviously , the claim is equivalent to the existence of @xmath179 such that @xmath477 for @xmath478 and @xmath479 for @xmath480 .
observe that @xmath481 for @xmath482 .
in particular , @xmath483 is strictly decreasing on the interval @xmath154 $ ] .
moreover , @xmath484 is strictly increasing on @xmath154 $ ] , so that the inequality @xmath485 holds on @xmath154 $ ] .
one computes @xmath486 where @xmath487 the polynomial @xmath488 is strictly increasing on @xmath489 $ ] and has exactly one root in the interval @xmath490 . combining this with equation ,
one obtains that @xmath491 has a unique zero in @xmath324 and that @xmath491 changes its sign from minus to plus there . observing that @xmath492 and @xmath493
, this yields @xmath494 on @xmath495 , that is , @xmath496 is strictly decreasing on @xmath154 $ ] .
moreover , it is easy to verify that @xmath497 , so that @xmath498 on @xmath499 $ ] . since @xmath500 on @xmath154
$ ] as stated above , it follows that @xmath501 on @xmath499 $ ] , that is , @xmath502 is strictly increasing on @xmath499 $ ] .
recall that @xmath483 and @xmath496 are both decreasing functions on @xmath154 $ ] . observing the inequality @xmath503 , one deduces that @xmath504\,.\ ] ] moreover , one has @xmath505 . combining this with and the fact that @xmath502 is strictly increasing on @xmath499 $ ] , one concludes that @xmath502 has a unique zero in the interval @xmath324 and that @xmath502 changes its sign from minus to plus there .
since @xmath506 and @xmath507 , it follows that @xmath508 has a unique zero in @xmath324 , where it changes its sign from minus to plus . finally , observing that @xmath509 and @xmath510 , in the same way one arrives at the conclusion that @xmath511 has a unique zero @xmath512 such that @xmath477 for @xmath478 and @xmath479 for @xmath513 . as a result of @xmath514
, one has @xmath408 .
[ app : lem3 ] for @xmath515 let @xmath516 then , @xmath517 satisfies the inequalities @xmath518 and @xmath519 one has @xmath520 , @xmath521 , and @xmath522 ( cf .
lemma [ lem : critpoints ] ) .
moreover , taking into account that @xmath523 by , one computes @xmath524 observe that @xmath525 and @xmath526 as @xmath527 , and that @xmath528 and @xmath529 as @xmath530 . with this and taking into account , it is convenient to consider @xmath531 , @xmath532 , and @xmath533 as continuous functions of the variable @xmath534 $ ] .
straightforward calculations show that @xmath535 so that @xmath536 taking into account that @xmath537 , that is , @xmath538 , this leads to @xmath539 in particular , @xmath540 is the only critical point of @xmath344 in the interval @xmath541 and @xmath542 changes its sign from plus to minus there .
moreover , using @xmath543 and @xmath544 , one has @xmath545 , so that @xmath546 since @xmath547 , this proves the two - sided inequality .
further calculations show that @xmath548 where @xmath549 the polynomial @xmath488 is strictly negative on the interval @xmath550 $ ] , so that @xmath542 is strictly decreasing .
define @xmath551\to{\mathbb{r}}$ ] by @xmath552 the claim is equivalent to the inequality @xmath553 for @xmath554 .
since @xmath543 and , hence , @xmath555 , one has @xmath556 .
moreover , a numerical evaluation gives @xmath557 .
therefore , in order to prove @xmath553 for @xmath554 , it suffices to show that @xmath511 has exactly one critical point in the interval @xmath558 and that @xmath511 takes its maximum there . using and taking into account that @xmath559 , one computes @xmath560 hence , for @xmath554 one obtains @xmath561 where @xmath562\to{\mathbb{r}}$ ] are given by @xmath563
suppose that the difference @xmath564 is strictly negative for all @xmath515 .
in this case , @xmath508 and @xmath542 have the same zeros on @xmath541 , and @xmath565 and @xmath566 have the same sign for all @xmath515 . combining this with
, one concludes that @xmath540 is the only critical point of @xmath511 in the interval @xmath541 and that @xmath511 takes its maximum in this point .
hence , it remains to show that the difference @xmath567 is indeed strictly negative on @xmath541 .
since @xmath568 , @xmath543 , and @xmath555 , it is easy to verify that @xmath569 and @xmath570 . therefore , it suffices to show that @xmath571 holds on the whole interval @xmath541 .
one computes @xmath572 where @xmath573 a further analysis shows that @xmath574 , which is a polynomial of degree @xmath575 , has exactly one root in the interval @xmath550 $ ] and that @xmath574 changes its sign from minus to plus there .
moreover , @xmath576 takes a positive value in this root of @xmath574 , so that @xmath577 on @xmath550 $ ] , that is , @xmath578 is strictly increasing on this interval . since @xmath579
, one concludes that @xmath580 on @xmath550 $ ] .
it follows from that @xmath581 on @xmath541 , so that @xmath582 is strictly decreasing .
in particular , one has @xmath583 on @xmath558 .
a straightforward calculation yields @xmath584 where @xmath585 .
the polynomial @xmath586 is positive and strictly decreasing on the interval @xmath587 $ ] .
moreover , taking into account , one has @xmath588 . combining this with equation ,
one deduces that @xmath589 has the opposite sign of @xmath566 for all @xmath554 . in particular , by it follows that @xmath590 if @xmath591 . since @xmath592 on @xmath541 , this implies that @xmath593 for @xmath594 . if @xmath595 , then one has @xmath596 .
in particular , @xmath344 is strictly increasing on @xmath597 .
recall , that @xmath542 is strictly decreasing by . combining all this with equation again ,
one deduces that on the interval @xmath597 the function @xmath598 can be expressed as a product of three positive , strictly decreasing terms .
hence , on this interval @xmath599 is negative and strictly increasing .
recall that @xmath600 on @xmath541 , which now implies that @xmath601 since the inequality @xmath602 has already been shown for @xmath591 , one concludes that @xmath571 holds on the whole interval @xmath541 .
this completes the proof .
[ app : lem4 ] one has @xmath603 the proof is similar to the one of lemma [ app : lem2 ] .
define @xmath475 by @xmath604 obviously , the claim is equivalent to the inequality @xmath477 for @xmath605 .
observe that @xmath606 for @xmath482 .
in particular , @xmath484 is strictly increasing on @xmath154 $ ] and satisfies @xmath607 .
one computes @xmath608 where @xmath609 the polynomial @xmath488 is strictly increasing on @xmath610 $ ] and has exactly one root in the interval @xmath611 . combining this with equation ,
one obtains that @xmath491 has a unique zero in the interval @xmath324 and that @xmath491 changes its sign from minus to plus there .
moreover , it is easy to verify that @xmath612 .
hence , one has @xmath613 on @xmath154 $ ] . since @xmath614 on @xmath154
$ ] as stated above , this implies that @xmath501 on @xmath154 $ ] , that is , @xmath502 is strictly increasing on @xmath154 $ ] . with @xmath505 and
@xmath615 one deduces that @xmath502 has a unique zero in @xmath324 and that @xmath502 changes its sign from minus to plus there . since @xmath506 and @xmath616
, it follows that @xmath508 has a unique zero in @xmath324 , where it changes its sign from minus to plus . finally , observing that @xmath509 and @xmath617 , one concludes that @xmath477 for @xmath605 .
the author is indebted to his ph.d .
advisor vadim kostrykin for introducing him to this field of research and fruitful discussions .
the author would also like to thank andr hnel for a helpful conversation . [
10 ] n. i. akhiezer , i. m. glazman , _ theory of linear operators in hilbert space _ , dover publications , new york ( 1993 ) .
s. albeverio , a. k. motovilov , _ sharpening the norm bound in the subspace perturbation theory _
, complex anal .
oper . theory * 7 * ( 2013 ) , 13891416 .
l. g. brown , _ the rectifiable metric on the set of closed subspaces of hilbert space _ ,
* 337 * ( 1993 ) , 279289 . c. davis , _ the rotation of eigenvectors by a perturbation _ , j. math .
* 6 * ( 1963 ) , 159173 . c. davis , w. m. kahan , _ the rotation of eigenvectors by a perturbation .
, siam j. numer .
7 * ( 1970 ) , 146
. t. kato , _ perturbation theory for linear operators _ , springer - verlag , berlin heidelberg ( 1966 ) .
v. kostrykin , k. a. makarov , a. k. motovilov , _ on a subspace perturbation problem _ , proc .
* 131 * ( 2003 ) , 34693476 .
v. kostrykin , k. a. makarov , a. k. motovilov , _ existence and uniqueness of solutions to the operator riccati equation . a geometric approach _ ,
contemp . math . *
327 * , amer .
( 2003 ) , 181198 .
v. kostrykin , k. a. makarov , a. k. motovilov , _ perturbation of spectra and spectral subspaces _ ,
* 359 * ( 2007 ) , 7789 .
k. a. makarov , a. seelmann , _ metric properties of the set of orthogonal projections and their applications to operator perturbation theory _ , e - print arxiv:1007.1575 [ math.sp ] ( 2010 ) .
a. seelmann , _ notes on the @xmath105 theorem _ , e - print arxiv:1310.2036 [ math.sp ] ( 2013 ) . | the problem of variation of spectral subspaces for linear self - adjoint operators under an additive bounded perturbation is considered .
the aim is to find the best possible upper bound on the norm of the difference of two spectral projections associated with isolated parts of the spectrum of the perturbed and unperturbed operators . in the approach presented here , a constrained optimization problem on a specific set of parameters
is formulated , whose solution yields an estimate on the arcsine of the norm of the difference of the corresponding spectral projections .
the problem is solved explicitly .
this optimizes the approach by albeverio and motovilov in [ complex anal .
oper .
theory * 7 * ( 2013 ) , 13891416 ] .
in particular , the resulting estimate is stronger than the one obtained there . |
recent observations seem to indicate that the universe is undergoing a period of accelerated expansion @xcite . whereas cosmologists initially introduced a cosmological constant in order to explain this ,
a range of different models have emerged , amongst which quintessence has been particularly prominent in the literature @xcite .
it is defined as a scalar field rolling down its potential and presently dominating the dynamics of the universe .
an important class of quintessence models are known as tracking models @xcite , where the late - time evolution of the field has an attractor behaviour rendering its evolution fairly independent of initial conditions .
in contrast to a cosmological constant , which is by definition perfectly homogeneous , the quintessence field can , and indeed must , have perturbations .
the evolution of perturbations in quintessence models have been studied by many authors @xcite . in this paper
we carry out an exhaustive and elegant analysis of those in the large - scale approximation .
we model the contents of the universe by a perfect fluid with equation of state @xmath0 and a scalar field @xmath1 with potential @xmath2 .
we assume a flat universe throughout .
before studying the perturbations , we recall some results for the homogeneous background evolution . the geometry of the universe is described by a flat robertson
walker metric @xmath3 the einstein equations @xmath4 relate the matter components to the geometry .
the indices `` f '' and `` @xmath1 '' always refer to the perfect fluid and the quintessence field respectively , and dots are time derivatives .
we will use a prime to denote a derivative with respect to @xmath5 .
the evolution of the fluid is straightforward , with its energy density scaling as @xmath6 , where @xmath0 is the ratio of pressure to energy density of the fluid .
the quintessence field follows the euler lagrange equation @xmath7 and its equation of state is @xmath8 depending on the precise model and on the choice of initial conditions , the quintessence dynamics can feature up to five main regimes , which were classified by brax et al .
@xcite and which appear in a sequential order . during the first three the quintessence field is sub - dominant .
the `` kinetic '' regime is characterized by the domination of the kinetic energy which scales as @xmath9 . in the `` transition '' and `` potential
'' regimes the potential energy dominates and the energy density remains constant .
the sound speed of the quintessence field is defined by @xmath10 and is equal to @xmath11 or @xmath12 respectively during those regimes . during the `` tracker '' regime
the quintessence field approximately mimics the behaviour of the fluid , and usually its energy density is still sub - dominant when tracking begins .
if @xmath13 there is perfect tracking , even if the quintessence field is not sub - dominant . finally the field enters its `` domination '' phase , during which @xmath14 tends to @xmath15 in most cases .
four useful parameters can be defined to describe these five regimes .
they are the quintessence density parameter @xmath16 , the equation of state conveniently parametrized as @xmath17 , the speed of sound @xmath18 as defined in eq .
( [ csdef ] ) , and one further parameter relating to the speed of sound defined as @xmath19 in order to simplify the notation , we define the vector @xmath20 . in fig .
[ regimes ] , we show an example of the evolution of those parameters . we take the realistic case @xmath21 during radiation domination and @xmath22 during matter domination , and we use an inverse power - law quintessence potential .
we clearly see the five different regimes , and also the transition between radiation domination and matter domination which in this case occurs during the tracking regime . in table
[ val_reg ] we give the values of the parameters in the general case for each regime .
.values of the four parameters @xmath16 , @xmath23 , @xmath18 , and @xmath24 during the different regimes . [ cols="<,^,^,^,^ " , ]
in order to describe the perturbations we choose the conformal newtonian gauge @xcite . as long as there is no anisotropic stress , the perturbed metric is @xmath25 where in this case @xmath26 is equal to the gauge - invariant potential defined as in ref .
we work in fourier space and compute the first - order perturbed einstein equations @xmath27 where the perturbed fluid pressure @xmath28 and the perturbations in the quintessence energy density and pressure are given by @xmath29 the perturbed euler lagrange equation and fluid conservation equation lead to @xmath30 where @xmath31 and @xmath32 gives the fluid velocity . from now on we use @xmath33 as a time variable and study the evolution of @xmath26 , @xmath34 , @xmath35 and @xmath36 in the long - wavelength limit ( @xmath37 ) .
we define the vector @xmath38 , and considerable algebra leads to the expression @xmath39 where the matrix @xmath40 is given by @xmath41 in order to display the above expression in compact form we used the variables @xmath42 and , but the matrix depends only on the five independent parameters @xmath43 , @xmath16 , @xmath23 , @xmath18 and @xmath24 .
recall that @xmath43 and @xmath23 take on values between @xmath44 and @xmath45 .
the general evolution is of course very complicated , but as seen in fig .
[ regimes ] , during each of the five regimes described above the coefficients within the matrix take on constant values , and this allows us to study the main features of the perturbation evolution .
we are interested in the eigenvalues @xmath46 of the matrix @xmath47 and their corresponding eigenvectors @xmath48 .
the solution then takes the form @xmath49 where the @xmath50 are constants given by the initial conditions . before considering the general case ,
we restrict ourselves to adiabatic perturbations . for perturbations to be adiabatic , they must share a common perturbation according to the prescription @xmath51 which ensures that all matter perturbations vanish on uniform - density hypersurfaces . note that the quintessence pressure perturbation , as well as its density perturbation , must satisfy the adiabatic condition ( for the perfect fluid adiabaticity of its pressure perturbation is automatically guaranteed by its equation of state ) .
these conditions can be rewritten as @xmath52 it is well known that initially adiabatic perturbations remain purely adiabatic , and indeed it is not difficult to check that these conditions are conserved through evolution by our equations .
we can therefore reduce the dynamical system in the adiabatic case to a system of two first - order equations .
we define the vector @xmath53 , and using eqs .
( [ lin_sys ] ) and ( [ adiabatic ] ) we find @xmath54 where the matrix @xmath55 is given by @xmath56 where @xmath57 .
the eigenvalues and eigenvectors are given in the upper part of table [ results ] .
we see , as is well known , that there are always a constant and a decaying adiabatic mode , the former giving the late - time solution @xmath58 , where @xmath59 .
lcl & & + + @xmath60 & & @xmath61 + @xmath62 & & @xmath63 + & & + & & + + @xmath60 & & @xmath64 + @xmath65 & & @xmath66 + @xmath67 & & @xmath68 + @xmath69 & & @xmath70 + & & + + @xmath60 & & @xmath71 + @xmath65 & & @xmath72 + @xmath73 & & @xmath68 + @xmath74 & & @xmath70 + & & + + @xmath60 & & @xmath71 + @xmath65 & & @xmath72 + @xmath73 & & @xmath68 + @xmath75 & & @xmath76 + & & + + @xmath60 & & @xmath77 + @xmath65 & & @xmath78 + @xmath79 & & @xmath80 + @xmath81 & & @xmath82 + & & + + @xmath60 & & @xmath83 + @xmath65 & & @xmath84 + @xmath85 & & @xmath86 + @xmath87 & & @xmath88 + & & + + @xmath60 & & @xmath89 + @xmath90 & & @xmath78 + @xmath73 & & @xmath91 + @xmath92 & & @xmath93 + & & + we now return to the full set of perturbation equations eq .
( [ lin_sys ] ) , continuing to consider the regimes in each of which the coefficients of the matrix @xmath47 remain constant .
we summarize our main results in the lower part of table [ results ] .
for the perfect tracker regime the eigenvectors @xmath94 and @xmath95 have not been given , since they are long and complicated formulas which are anyway not very relevant . in order to simplify some expressions
we have used the variables @xmath96 for each regime the adiabatic modes can easily be identified as the first two entries in the table .
now let us analyze the two other , non - adiabatic , modes .
in the kinetic case there are a growing mode , for which @xmath97 and thus @xmath98 ( since @xmath99 ) , and another constant mode corresponding to @xmath100 . during the transition and potential regimes
the former growing mode becomes constant and for each regime the fourth mode is decaying .
therefore , before entering the tracker the quintessence field may feature large non - adiabatic perturbations . as long as the universe
is dominated by the fluid , they are isocurvature perturbations . in the usual tracker case
the last two eigenvalues may have an imaginary part , leading to oscillations , and their real part can either be negative or positive according to the value of @xmath43 and @xmath23 .
the regimes of the eigenvalues are shown in fig .
[ eigenvalues ] .
however , since the quintessence field has to dominate at the present epoch , @xmath101 must decrease more slowly than @xmath102 , and hence @xmath103 .
as one can see in fig .
[ eigenvalues ] , this implies @xmath104 . in the perfect tracker case one
easily sees that the last two modes decay , possibly oscillating as they do . as long
as all the other modes are decaying , during the tracker regime the constant adiabatic mode @xmath105 is an attractor . as a result , a long tracker period implies the suppression of all non - adiabatic modes @xcite .
moreover , in the case of a sub - dominant tracker the late - time evolution of the perturbations is even independent of the quintessence field initial conditions .
finally , during the domination regime , which is reached after the present epoch , the non - adiabatic modes are constant and decaying .
we have derived four first - order equations to describe large - scale perturbations in quintessence scenarios . during each of the five main regimes of quintessence behaviour ,
these equations have constant coefficients , enabling analytic solution of the perturbations by eigenvector decomposition .
we have seen that during the kinetic regime there is a growing isocurvature mode which then remains constant until tracking begins . however , if the quintessence field undergoes a long period of tracking , there remain only adiabatic perturbations which , in case of a sub - dominant tracker , are independent of its initial conditions .
a low initial quintessence energy density @xcite , or a long kinetic period , may prevent the non - adiabatic modes disappearing completely .
a. g. riess et al .
j. * 116 * , 1009 ( 1998 ) , ` astro - ph/9805201 ` ; p. garnavich et al .
, ap . j. * 509 * , 74 ( 1998 ) , ` astro - ph/9806396 ` ; s. perlmutter et al . , ap .
j. * 517 * , 565 ( 1998 ) , ` astro - ph/9812133 ` ; g. efstathiou et al .
, mon . not .
. soc . * 330 * , l29 ( 2002 ) , astro - ph/0109152 . b. ratra and p. j. e. peebles , phys .
rev . d**37 * * , 3406 ( 1988 ) . c. wetterich , nucl .
phys . b**302 * * , 668 ( 1988 ) ; e. j. copeland , a. r. liddle , and d. wands , ann . n. y. acad .
sci . * 688 * , 647 ( 1993 ) ; r. r. caldwell , r. dave , and p. j. steinhardt , phys .
lett . * 80 * , 1582 ( 1998 ) , ` astro - ph/9708069 ` ; p. g. ferreira and m. joyce , phys .
rev . d**58 * * , 023503 ( 1998 ) , ` astro - ph/9711102 ` ; e. j. copeland , a. r. liddle , and d. wands , phys .
rev . d*57 * , 4686 ( 1998 ) , gr - qc/9711068 ; i. zlatev , l. wang , and p. j. steinhardt , phys .
lett . * 82 * , 896 ( 1999 ) , ` astro - ph/9807002 ` ; a. r. liddle and r. j. scherrer , phys .
rev . d*59 * , 023509 ( 1999 ) , astro - ph/9809272 ; v. sahni and a. starobinsky , int .
* d9 * , 373 ( 2000 ) , ` astro - ph/9904398 ` .
r. r. caldwell , r. dave , and p. j. steinhardt , phys .
lett . * 80 * , 1582 ( 1998 ) , ` astro - ph/9708069 ` ; p. t. p. viana and a. r. liddle , phys .
rev . d**57 * * , 674 ( 1998 ) , ` astro - ph/9708247 ` ; f. perrotta and c. baccigalupi , phys . rev . d**59 * * , 123508 ( 1999 ) , ` astro - ph/9811156 ` .
p. brax , j. martin , and a. riazuelo , phys .
rev . d**62 * * , 103505 ( 2000 ) , ` astro - ph/0005428 ` .
l. r. abramo and f. finelli , phys .
rev . d**64 * * , 083513 ( 2001 ) , ` astro - ph/0101014 ` ; m. kawasaki , t. moroi , and t. takahashi , phys .
b * 533 * , 294 ( 2002 ) , ` astro - ph/0108081 ` ; r. dave , r. r. caldwell , and paul j. steinhardt , phys .
rev . d**66 * * , 023516 ( 2002 ) , ` astro - ph/0206372 ` .
v. mukhanov , h. feldman , and r. brandenberger , phys . rep .
* 215 * , 203 ( 1992 ) .
m. malquarti and a. r. liddle , phys .
d**66 * * , 023524 ( 2002 ) , ` astro - ph/0203232 ` . | we carry out a comprehensive study of the dynamics of large - scale perturbations in quintessence scenarios .
we model the contents of the universe by a perfect fluid with equation of state @xmath0 and a scalar field @xmath1 with potential @xmath2 .
we are able to reduce the perturbation equations to a system of four first - order equations . during each of the five main regimes of quintessence field behaviour ,
these equations have constant coefficients , enabling analytic solution of the perturbation evolution by eigenvector decomposition .
we determine these solutions and discuss their main properties . |
owing to its extraordinary properties @xcite , graphene has already been implemented in various kinds of applications / devices @xcite and led to the emergence of a new class of materials ; ultra - thin two - dimensional crystal structures .
nowadays , among the members of this new era , especially the ultra - thin transition metal dichalcogenides ( tmds ) have attracted considerable interest.@xcite even though , they are only few - atom - thick , mx@xmath3-type structures have remarkable chemical and mechanical stability@xcite and offer new physics as various quantum confinement effects amplified in quasi - two dimension @xcite . as a result of this confinement effect , the band gap increases and transforms to a direct band gap with decreasing number of layers which makes them promising candidates for nanoscale field - effect transistors and for solar cell applications.@xcite recently , possibility of vacancy creation in tmds under electron irradiation,@xcite bandgap transition in tungsten dichalcogenides@xcite , existence of tightly bound negative
trions@xcite and strain - engineered electronic properties have been reported.@xcite furthermore , we reported that with strain application mose@xmath3 , and possibly other tmds , show significant red shift in their raman spectrum and undergo a direct to indirect bandgap transition.@xcite synthesis and characterization of tungsten diselenide ( wse@xmath3 ) has been an active field of research with applications in photovoltaic and photoconductive devices and recently monolayer wse@xmath3 has become a popular choice for nanoscale devices@xcite . here , we present an experimental and theoretical investigation of the electronic properties and lattice dynamics of bulk , few layer and single layer wse@xmath3 .
we find that the @xmath1 and @xmath0 modes are almost degenerate for bulk wse@xmath2 whereas these modes are well separated for other members ( mos@xmath3 , mose@xmath3 and ws@xmath3 ) of the tmds .
interestingly , this degeneracy is lifted as the dimensionality is lowered from 3d ( bulk limit ) to 2d ( monolayer ) where the @xmath1 and @xmath0 modes are separated by @xmath412 @xmath5 .
calculated vibrational spectrum show that the lifting the degeneracy is closely related to the uniaxial strain induced on monolayer wse@xmath2 due to interaction with the substrate and/or sample preparation procedure .
lastly , we show that the band gap of wse@xmath2 goes through rather soft indirect to direct band gap crossover from bulk to monolayer and the band gap shows almost triple band degeneracy for bi and tri - layers as evidenced by our photoluminescence measurements .
monolayer and few - layer wse@xmath2 flakes were exfoliated from wse@xmath2 single crystals ( 2dsemiconductors.com ) onto 90 nm sio@xmath2/si ( mti inc . )
substrates using conventional mechanical exfoliation technique .
the thickness of the wse@xmath2 flakes was confirmed by three complementary methods , atomic force microscopy ( afm ) , raman spectroscopy , and photoluminescence ( pl ) .
non - contact afm line scans on the monolayer wse@xmath2 , typically resulted in @xmath40.7 nm step height corresponding to single unit cell lattice constant for wse@xmath2 fig.[fig1](d ) .
pl and raman measurements were performed using very low power intensity ( 10@xmath6w/@xmath6m@xmath7 ) on 2 - 3@xmath6m@xmath7 circle to avoid local heating or damaging effect .
the results presented in this manuscript were reproduced on more than fifty samples .
theoretical calculations for equilibrium and strained structures were carried out in the framework of density functional theory ( dft ) , using the projector augmented wave ( paw ) method@xcite as implemented in the vasp code.@xcite the generalized gradient approximation ( gga ) of perdew - burke - ernzerhof ( pbe ) was used for the exchange - correlation functional @xcite . to calculate the raman spectrum of single layer wse@xmath2 under biaxial ( uniaxial ) strain , hexagonal ( rectangular ) unitcell with one side parallel to the direction of stretch
is taken into account .
phonon frequencies and phonon eigenvectors were calculated by using both the density functional perturbation theory ( dfpt)@xcite and the small displacement method.@xcite
similar to graphite , wse@xmath3 crystals possess lamellar structure with bernal stacking where the individual layers are weakly coupled to the adjacent layers by van der waals ( vdw ) forces . as shown in fig .
[ fig1](a - b ) , each monolayer wse@xmath3 ( 1h - wse@xmath2 ) consists of se - w - se atoms wherein tungsten atoms are sandwiched between trigonally arranged sheets of selenium atoms .
our calculations reveal that bulk wse@xmath3 has an 1.21 ev indirect band gap where the valence band maximum ( vbm ) is located both at the k symmetry point while the conduction band minimum ( cbm ) is along the @xmath8-k direction ( fig .
[ fig1](c ) green - dashed curve ) .
we note that conduction bands located at k and @xmath8-k is only separated by 0.04 ev in energy and hence we expect that its electronic properties will be strain and dimensionality dependent as a result of significant changes in the hybridization .
since bulk wse@xmath2 is an indirect band gap semiconductor , the photoluminescence signal is expected to be rather weak for bulk wse@xmath3 as observed in fig .
[ fig1](f ) .
interestingly , 7 to 11 layers display two distinct pl peaks ( hot luminescence ) located at 1.39 ev and 1.59 ev where the former probes the indirect band gap ( @xmath8 to @xmath8-k ) and the latter is associated with the direct band gap transition ( k to k ) ( see inset of fig . [ fig1](f ) ) .
we note that the overall pl signal measured on few - layer flakes is orders of magnitude weaker in intensity as compared to bilayer and monolayer wse@xmath2 .
( color online ) ( a ) side and tilted view of single layer wse@xmath3 .
( b ) brillouin zone of wse@xmath3 and ( c ) gga+so band structures .
fermi level is set to zero .
blue , red and green colors are for single layer , bilayer and bulk structures , respectively .
( d ) atomic force microscopy ( afm ) images taken on wse@xmath2/sio@xmath2 and ( e ) bare sio@xmath2 substrate .
afm scan on sapphire substrate is given as comparison .
( f ) pl measurements on wse@xmath2 flakes with various thicknesses and ( g ) peak positions.,width=321 ] electronic band structure calculations for bilayer and monolayer wse@xmath2 show that the band gap increases from 1.21 ev in bulk to 1.23 ev ( direct - indirect almost degenerate ) for bilayer and 1.25 ev ( direct ) for monolayer wse@xmath2 fig .
[ fig1](c ) . experimentally observed pl signal for bulk , bilayer and monolayer wse@xmath2 peaks at 1.57 , 1.57 and 1.64 ev , respectively ( fig.[fig1](f ) ) .
we see that in contrast to the general case , for wse@xmath2 , lda and gga band dispersions differ significantly .
while gga finds single layer wse@xmath2 as a semiconductor with a direct bandgap of 1.56 ev , lda gives indirect bandgap of 1.68 ev .
here , both lda and gga finds vbm at the k symmetry point and cbm is located at k ( in between @xmath8 and k ) point in gga ( lda ) .
it appears that the gga exchange correlation better approximates the ground state characteristics of the wse@xmath2 crystal structure .
here it is also worth to note that spin - orbit interaction results in two significant effects on the electronic band dispersion : ( i ) shifting down the valence band energies at the @xmath8 point ( ii ) band splitting at the vicinity of k and m symmetry points . clearly , the existence of an intrinsic electric field breaks the inversion symmetry in the crystal structure .
therefore , it is seen that the degeneracy of the doubly degenerate valence and conduction bands of single layer wse@xmath2 are removed by spin - orbit interaction and a band splitting occurs .
furthermore , when the spin - orbit interaction is taken into account in our calculations , conduction band edges at k and k-@xmath8 points show band splitting that allows various direct and indirect transitions ( even in few - layered wse@xmath2 structures ) . in our study
, all the band structure calculations presented are performed by considering spin - orbit interaction together with gga .
previously , indirect to direct band crossover has been observed in other transition metal dichalcogenides @xcite and is consistent with our results .
however , we note that the indirect to direct transition is rather steep for mos@xmath2 where it s bilayer form is an indirect band gap ( 1.6 ev ) semiconductor while monolayer structure has a direct band gap ( 1.9 ev ) . in contrast
, bilayer wse@xmath2 possesses almost triple band gap degeneracy where the k@xmath9k and k@xmath9@xmath8-k gap values are almost degenerate and the difference between first and second cbms is @xmath10(k)=40 mev which is close to thermal broadening ( @xmath430mev ) . as a result of this band degeneracy , the integrated pl intensity ( fig .
[ fig1](e ) ) of bilayer wse@xmath2 is of the same order of magnitude as in monolayers , i.e i@xmath11/i@xmath12@xmath4 1 - 10 , which compares with 100 - 1000 for mos@xmath2 .
we also note that the pl signal for bilayer wse@xmath2 is rather broad and can be described by at least two lorentzian peaks .
as the number of layers increases , the band degeneracy is gradually lifted , wse@xmath2 becomes truly an indirect band gap semiconductor and the pl intensity decreases by orders of magnitude . .calculated direct and indirect transitions between vbm and cbms .
the difference between the first and second cbm at k point is also given as @xmath13@xmath14(k ) .
experimental value is given in parenthesis .
all energies are given in ev . [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] ( color online ) ( a ) raman shift for 1-layer , 2-layer and bulk wse@xmath3 .
( b ) phonon spectrum of bulk wse@xmath3 and ( c ) atomic displacements for optical modes between 100 - 250 @xmath5 .
( d ) phonon dispersions for bilayer and ( e ) single layer structure and ( f ) atomic displacements for optical phonon modes of single layer wse@xmath3.,width=321 ] next we turn our attention to the anomalous lattice vibrations of wse@xmath2 . for determination of lattice dynamics
we use raman spectroscopy which is one of the most useful non - destructive technique for the characterization of low - dimensional materials .
our measurements reveal that the most prominent raman peak for bulk wse@xmath2 is located at 252.2 @xmath5 ( fig .
[ fig2](a ) ) while other semiconducting tmds are characterized by well separated @xmath17 and @xmath18 raman peaks .
the stark difference between wse@xmath2 and other tmds already points towards an anomaly in the phonon dispersion of wse@xmath2 .
following from fig .
[ fig2](b ) , calculated phonon dispersion confirms that bulk wse@xmath2 displays only one strong vibrational mode around 250 @xmath5 consistent with our experiments .
interestingly , for flakes thinner than four layers , an additional peak appears at roughly 5 - 11 @xmath5 above the first order raman peak .
we find that the frequency difference between these two peaks is 5 , 6 and 11 @xmath5 for trilayer , bilayer , and monolayer wse@xmath2 flakes , respectively .
this peak is absent in few - layer flakes does not correspond to any new / additional raman modes in the calculated phonon dispersion curve , which describes the raman spectrum well .
bulk and single layer wse@xmath2 can be classified in the space group p63/mmc and p6m2 , respectively .
similar to the mos@xmath3 and mose@xmath3 counterparts,@xcite the unitcell of bulk wse@xmath2 has eighteen phonon branches corresponding to six acoustic and twelve optical phonon modes . while the modes at @xmath4170 ( @xmath4300 ) @xmath5 are 4-fold ( 2-fold ) degenerate , six modes become almost degenerate at @xmath4250 @xmath5 at the @xmath8 point . for a better understanding of the phonon spectrum and the prominent peak in the measurement
, we present phonon symmetry representations of related modes and corresponding atomic motions in figs .
[ fig2](b ) and ( c ) .
it is seen that the 4-fold branch at @xmath4170 cm @xmath19 is formed by the vibrational motions corresponding to the e@xmath20 and e@xmath15 modes . among these
only the e@xmath15 mode is raman - active .
however , in backscattering experiments on a surface perpendicular to the c - axis , the e@xmath15 mode is forbidden and is not observed in our experiments .
furthermore , the decomposition of the six - fold phonon branch at @xmath4250 @xmath5 ( at @xmath8 point ) can be described as @xmath21 . among these the @xmath0 and @xmath1 modes
are raman - active and are degenerate in energy . as a result of this degeneracy
, only one raman peak can be observed in our measurements at 250 @xmath5 .
next , we pay attention to the emergence of the second raman peak in few and single layer wse@xmath2 that appears @xmath411 @xmath5 above the main peak .
though , from bulk to single layer wse@xmath2 , there is no visible change in the calculated phonon dispersion ( figs .
[ fig2](b , d , e ) , our measurements show that a new peak develops ( fig .
[ fig2](a ) ) and this new peak appears to be very sensitive to the number of layers .
for single layer wse@xmath2 , decomposition of the vibration representation is calculated to be @xmath22 .
here , the @xmath23 , @xmath24 and @xmath25 modes correspond to the @xmath26 , @xmath0 and @xmath1 modes for bulk respectively .
since the @xmath24 and @xmath27 modes are almost degenerate , only one raman peak is expected to be observed in the experiment which contradicts our experimental results . before discussing more on the discrepancy
, we note that presence of contaminants at the wse2/interface and/or directly bonded to the tmd monolayer might cause small alterations to the raman spectrum .
however , if the contaminants are directly bonded ( chemisorption ) to the tmd , one would expect to observe drastic differences in the raman spectrum as a result of renormalization of phonon dispersion . since the overall monolayer raman spectrum is similar to few - layer
, we eliminate this possibility .
if the contaminants are chemically interacting locally , this would change the raman signal from those local regions ( bloch waves intermix ) but would be small comparing to overall raman signal due to the geometrical considerations and reduced raman sensitivity of the locally interacting region . to explain the emergence of the new peak , we consider external factors , such as compressive and tensile strain acting on few - layer and monolayer wse@xmath2 flakes , which is likely to be induced by the interaction with the substrate .
to test the possibility of strain effects on wse@xmath2 as the origin for such splitting of the raman peak , we calculate the phonon dispersion of monolayer wse@xmath2 after applying 1@xmath28 biaxial compressive and tensile strain . as seen from fig .
[ fig3](a ) , such biaxial deformation only results in a collective softening / hardening of the vibrational modes and does not lift the degeneracy of the @xmath0 and @xmath1 modes ( @xmath24 and @xmath25 in single layer ) implying that these modes remain degenerate as long as the hexagonal symmetry of the monolayer wse@xmath2 is retained .
next , we apply uniaxial strain that is likely to occur along with the biaxial strain on monolayer wse@xmath2 exfoliated on si / sio@xmath2 surface . for hexagonally ordered crystal structures uniaxial strain can be applied in two main directions : armchair and zigzag ( fig .
[ fig3](b ) ) .
interestingly , in the presence of uniaxial strain , degeneracy of the branches forming the most prominent raman peak are broken .
while most of the modes between 200 - 280 @xmath5 loose their symmetry that determines the raman activity , our symmetry analysis shows that two of the branches that correspond to the @xmath24 ( lower ) and @xmath25 ( upper ) modes are still raman active . for compressive ( tensile ) uniaxial strain along the armchair direction
the difference between the @xmath24 and @xmath25 modes is 16 ( 18 ) @xmath5 .
similarly , for compressive ( tensile ) strain along the zigzag direction , splitting is calculated to be 13 ( 17 ) @xmath5 . here
the @xmath24-@xmath25 splitting induced by uniaxial strain is in good agreement with our experimental data .
it is worth to note that independent from the direction of both tensile and compressive strains both have the same splitting effect on the raman peaks .
according to our phonon dispersion calculations performed on bulk wse@xmath2 , the out - of - plane ( a@xmath15 ) and in - plane ( e@xmath16 ) modes are degenerate in energy , consistent with the experimental measurements taken on bulk wse@xmath2 .
experimentally , as the dimension is lowered from 3d to 2d ( bulk to monolayers ) , raman mode located at around @xmath4 250 @xmath5 splits in two peaks at @xmath4 250 @xmath5 and 261 @xmath5 . to provide an explanation for the broken degeneracy , we recalculate the phonon dispersion of monolayer wse@xmath2 under uniaxial and biaxial strain / stress .
our theoretical results imply that the degeneracy of these two modes is lifted only if the crystal symmetry is broken , i.e. in the presence of uniaxial strain .
when the monolayer wse@xmath2 is under very little unaxial strain or stress , our theoretical calculations show a remarkable match with the experimental data . considering above arguments , we next argue about the possible origin of uniaxial strain on thin wse@xmath2 flakes . since our sio@xmath2 substrates typically display 4 - 8 surface roughness ( fig .
[ fig1](d - e ) ) that is of the same order of a single unit cell thickness , even perfect sio@xmath2/wse@xmath2 interface is likely to induce a mild strain to the wse@xmath2 monolayer .
the presence of the surface roughness on sio@xmath2 can be observed clearly when compared to sapphire which yields only 1 - 2 surface roughness .
typically , these rough features are asymmetric in shape and therefore strain induced on the few - layer wse@xmath2 is expected to have both biaxial and uniaxial component .
however , since the biaxial strain does not lift the degeneracy only the uniaxial component of the total strain results in raman peak splitting .
moreover , in the presence of residual contaminants at the wse@xmath2/sio@xmath2 interface and/or residues deposited after the exfoliation step , the effect of strain is likely to be amplified . from this perspective
, monolayers are most affected by the strain effects while thick flakes remain mostly unaltered .
another scenario might be associated with the unintentional strain induced during the exfoliation process . in such case ,
the wse@xmath2 monolayers are deposited on the sio@xmath2 substrates during the exfoliation and is not necessarily related to the surface residue .
however , we note that our results were confirmed on 50 different wse@xmath2 monolayers where the same results have been found making this scenario .
another indirect confirmation of presence of uniaxial strain comes from the changes in the pl peak position .
we also note that the band gap at the k - k symmetry point is expected to be independent from the dimensionality . on the contrary ,
pl measurements show that the k - k gap increases abruptly for the monolayer , likely due to the presence of small uniaxial strain ( fig .
[ fig1](g ) ) as is confirmed by our dft calculations .
we also point out that in the presence of large density of defects , the degeneracy can be lifted .
however , similar measurements taken on more than fifty independently prepared samples makes this case unlikely . in conclusion , phonon dispersion and the electronic properties of bulk to monolayer wse@xmath2
have been studied both experimentally and theoretically . unlike the other members of the tmds
, the @xmath0 and @xmath1 modes are degenerate in bulk wse@xmath2 and the degeneracy is lifted for as the dimension is lowered . on the contrary to our experimental results , calculated phonon dispersion show that these modes remain degenerate independent from the dimensionality .
however , theoretically the degeneracy is only lifted when the crystal symmetry is broken , i.e. in the presence of unaxial strain which might be induced by the interaction with the substrate , residues , and/or exfoliation process .
these results provide deeper understanding in the vibrational properties of tmds , especially on a material with unique phonon dispersion .
this work was supported by the flemish science foundation ( fwo - vl ) and the methusalem programme of the flemish government .
computational resources were partially provided by tubitak ulakbim , high performance and grid computing center ( tr - grid e - infrastructure ) .
h. s. is supported by a fwo pegasus marie curie long fellowship .
j. n. coleman , m. lotya , a. oneill , s. d. bergin , p. j. king , u. khan , k. young , a. gaucher , s. de , r. j. smith , i. v. shvets , s. k. arora , g. stanton , h. y. kim , k. lee , g. t. kim , g. s. duesberg , t. hallam , j. j. boland , j. j. wang , j. f. donegan , j. c. grunlan , g. moriarty , a. shmeliov , r. j. nicholls , j. m. perkins , e. m. grieveson , k. theuwissen , d. w. mccomb , p. d. nellist , and v. nicolosi , science * 331 * , 568 ( 2011 ) .
t. m. g. mohiuddin , a. lombardo , r. r. nair , a. bonetti , g. savini , r. jalil , n. bonini , d. m. basko , c. galiotis , n. marzari , k. s. novoselov , a. k. geim , and a. c. ferrari , phys .
b * 79 * , 205433 ( 2009 ) . | typical raman spectra of transition metal dichalcogenides ( tmds ) display two prominent peaks , @xmath0 and @xmath1 , that are well separated from each other .
we find that these modes are degenerate in bulk wse@xmath2 yielding one single raman peak . as the dimensionality
is lowered , the observed peak splits in two as a result of broken degeneracy .
in contrast to our experimental findings , our phonon dispersion calculations reveal that these modes remain degenerate independent of the number of layers .
interestingly , for minuscule biaxial strain the degeneracy is preserved but once the crystal symmetry is broken by uniaxial strain , the degeneracy is lifted .
our calculated phonon dispersion for uniaxially strained wse@xmath2 shows a perfect match to the measured raman spectrum which suggests that uniaxial strain exists in wse@xmath2 flakes possibly induced during the sample preparation and/or as a result of interaction between wse@xmath2 and the substrate .
furthermore , we find that wse@xmath2 undergoes an indirect to direct bandgap transition from bulk to monolayers which is ubiquitous for semiconducting tmds . these results not only allow us to understand the vibrational properties of wse@xmath2 but also provides detailed insight to their physical properties . |
in wireless communications , information is transmitted in the form of waves and space is considered as the physical medium for information transfer . hence , as a physical process , waves propagate in space via line of sight or multiple paths due to reflection , diffraction and scattering by objects present in the physical environment . like any other physical phenomenon
, wave propagation is governed by the laws of physics .
these laws determine the process itself as well as the amount of diversity waves carry along their path .
the spatial diversity of multipath influences the amount of information that can be communicated through wave propagation , thus , using spatial diversity of multipath we can ensure better system performances , including capacity improvement , high transmission rate , improved bit error rate etc . , @xcite .
in effect , the study of the spatial degrees of freedom of different multiple antenna systems ( i.e. , multi - user mimo systems , distributed mimo systems , mimo cognitive radio systems etc . ) has gained renewed attention and has more recently been addressed by @xcite .
this motivates to study the fundamental limits that space imposes on the degrees of freedom of band limited signals observed over finite spatial and temporal windows . in this paper ,
our aim is to determine the upper limit to the degrees of freedom of signals available in a band limited multipath wavefield when the wavefield is observed in , or coupled to a limited source - free region of space over a finite time window .
we may assume that multiple antennas or sensors are located in the region of space to sample the observable multipath field for signal processing or communication purposes .
we , however , aim to find an upper bound on the available degrees of freedom without explicitly considering a specific propagation condition , physical setup or application and thus , to show that the coupling of time and band limited multipath signals into a spatial region is fundamentally limited by a finite number of spatial modes . throughout the paper
, we will frequently refer to the radius/ size of the @xmath0 multipath observation region .
our derived result has great significance in a wide range of applications , including ( ) measuring the number of receive antennas required to sample a given region to maximize the performance gain , ( ) characterizing broadband beamforming techniques for next generation wireless communication to provide high quality video and audio , ( ) developing interference alignment scheme for mimo wireless networks , ( ) characterizing the degrees of freedom of distributed multi - antenna communications for broadband transmissions .
we review the degrees of freedom available in spatially diverse multipath fields in different contexts .
earlier works , @xcite focused on multipath fields that exhibit rich scattering and there are independent fading paths between transmitter and receiver antenna elements . according to these works , available degrees of freedom
is the minimum number of transmit and receive antenna elements and channel capacity can be improvement remarkably by increasing number of the antenna elements .
however , insufficient antenna spacing violates the assumption of independent fading and prevent channel capacity to increase linearly with degrees of freedom @xcite .
the impact of fading correlation on spatially diverse multipath fields was studied by a large number of research works ( e.g. , @xcite ) .
afterwards , independent works @xcite provided the characterization of the spatial degrees of freedom in multi - antenna systems as a function of the area , geometry of the antenna arrays and the angular spread of the physical environment .
in addition , @xcite estimated the degrees of freedom available in source - free narrowband multipath fields observed over a spatial window and showed that the available degrees of freedom scales with the spatial dimension in terms of wavelengths .
in contrast , poon et al .
@xcite and franceschetti @xcite applied antenna theory and slepian s theory of spectral concentration , respectively , to derive a fundamental limit on the degrees of freedom available in a wideband multi - antenna systems for a given constraint on the area of the spatial region and observation time and defined the degrees of freedom as a product of spatial degrees of freedom and degrees of freedom of the wideband channel itself .
since for wideband transmissions , space time , and frequency are strongly coupled , available bandwidth and observation time over space respectively differ from actual bandwidth and observation time depending on the available spatial information , the works of @xcite did not take this into account . in another approach , @xcite characterized multi - antenna systems in a wideband transmission regime and stated that in case of wideband frequency transmission , space and time are strongly coupled .
however , how information is conserved in space - time was left as an open and important problem .
the analysis in this paper considers a wideband multipath wavefield observed within a limited source - free region of space over a finite time window .
the signals observable within this wavefield are studied as solutions to the helmholtz wave equation @xcite and they are encoded in infinite but countable number of orthogonal waveforms or spatial modes .
this mathematical framework is similarly used in @xcite . however , in comparison , our derived result is more accurate , since we have considered the affect of available spatial information not only on the frequency bandwidth but also on the observation time .
further , the degrees of freedom result provided in @xcite is derived by using a complex geometrical argument to extend the narrowband degrees of freedom result of @xcite to a broadband scenario and resulted in a loose bound .
further , it is unclear , for different spatial modes , how the usable ( effective ) bandwidth varies from the given frequency bandwidth . in this work , on the contrary
, the degrees of freedom result is derived in a simple manner . moreover
, we clarify that at each spatial mode , how ( and why ) the observable signals are band limited within an effective frequency bandwidth rather than the given frequency bandwidth .
in addition , we illustrate that beyond a certain spatial mode , the effective bandwidth becomes zero which in turn , truncates the wavefield from its infinite representation to a finite number of spatial modes .
afterwards , by counting the number of spatial modes required to represent any signal within the given multipath field , we derive an analytical expression to determine the degrees of freedom of the signal wavefields and @xmath0 wavefields . ] .
we depict the strong coupling relation between space and time as well as space and frequency in spatially diverse wideband multipath fields .
we show that the effective observation time is fixed , independent of spatial modes , different from given observation time and essentially related to the spatial dimension of the observable region . whereas , for broadband transmissions , at each spatial mode , the observable signal is band limited within an effective frequency bandwidth , since even though the usable bandwidth at the lower spatial modes is equal to the given frequency bandwidth , for the higher modes , the usable bandwidth is less than the given frequency bandwidth .
the coupling relations also indicate that for spatially diverse wideband multipath fields , the classical degrees of freedom result of _ time - bandwidth product _ can not be extended directly to the product of spatial degrees of freedom and _ time - bandwidth product _ as shown in @xcite , rather the degrees of freedom result should portray how the time and band limited signals are coupled to a limited region of space .
our derived degrees of freedom result evidently portrays the impact of the coupling relations on the available degrees of freedom in spatially diverse wideband multipath fields .
we also show the affect of the acceptable signal to noise ratio ( snr ) on the available degrees of freedom of each spatial mode .
the reminder of the paper is organized as follows . in section [ problem ] , the problem statement together with background on shannon s time - frequency degrees of freedom and the eigenbasis expansion of the wavefield are discussed . in section
[ results ] , we present our main results , while , section [ plot ] provides graphical analysis of our derived results .
next , section [ insights ] elaborates the physical insights of the main results and briefly discusses the applications .
we summarize the main contributions of this paper in section [ conclusion ] .
in this paper , we consider a multipath field band limited to @xmath1 $ ] and observed over a time window @xmath2 $ ] within a @xmath0 spatial window enclosed by a spherical region of radius @xmath3 . here
, @xmath4 represents the mid band frequency .
any signal sampled or recorded within this spatial region can be expressed as a function of space and time whose spectra lies within the frequency range and whose time function lies within the time interval . since it is not possible to confine any waveform in both time and frequency , we consider that the spectrum is confined entirely within the frequency range and the time function is negligible outside the time interval .
+ we now express the space - time signal as rcl ( * _ x_*,t ) = ^_-(*_x_*,)e^j t d where @xmath5 is the fourier transform of @xmath6 with respect to @xmath7 , * _ x _ * represents a position in 3d space , such that @xmath8 denotes the euclidean distance of * _ x _ * from the origin , which is the center of the region of interest and @xmath9 . due to the band limitedness , @xmath5 is assumed to be zero outside the band @xmath10 $ ] .
thus , the space - time signal can be rewritten as rcl [ timespace ] ( * _ x_*,t ) = ^2(f_0+w)_2(f_0-w)(*_x_*,)e^jt d. in this work , we aim to answer the fundamental question : _ given constraints on the size of the observation region , frequency bandwidth and observation time window , what is the upper bound on the degrees of freedom of signals observable within multipath wireless fields ?
_ we will utilize shannon s result @xcite that provides the degrees of freedom of temporal signals for band limited communications over a single channel .
shannon s result states that if the transmission is band limited to @xmath11 $ ] and limited to the time interval @xmath12 $ ] , the available time - frequency degrees of freedom is limited to @xmath13 .
we provide more detailed reasoning of shannon s result in the next subsection .
let us now review the reasoning behind the time - bandwidth product result @xcite . in time domain
we have a wideband signal and in frequency domain this signal can be expressed as a spectrum .
the mapping between these two domains is the fourier transform .
the spectrum is then expanded over the frequency range with the help of fourier series expansion .
this expansion represents the time domain signal by a weighted sum of orthogonal basis functions .
given the signal is approximately time limited to @xmath12 $ ] and its spectrum is band limited to @xmath11 $ ] , the minimum number of terms required in the sum to satisfy both of these constraints provide the available degrees of freedom of the signal , @xmath13 .
note that there may be slight discrepancy as the time domain signal obtained by the fourier series expansion over the time interval will not be strictly limited within the frequency band , rather it may contain some frequency component outside the band .
however , in another approach , @xcite argued that roughly @xmath13 samples are enough to approximate any signal in energy for the best choice of a complete set of band limited functions which possess the property of being orthogonal over a given finite time interval .
afterwards , the time - bandwidth product result was formalized by several authors @xcite for various other configurations . however , shannon s result only accounts for broadband transmission over a single channel . in multipath wireless fields sampled over a region of space , spatial diversity
is exploited , for instance , in mimo communications , providing several independent channels over which information can be transmitted . to account for the spatial diversity of wireless fields , we start with the spherical harmonics analysis of the wavefield observed within a region of space .
we consider the space - frequency spectrum @xmath5 in as a scalar wavefield observed within a @xmath0 spherical region of finite radius @xmath3 generated by a source or distribution of sources and scatterers that exist outside the region of interest at some radius @xmath14 .
hence , @xmath5 satisfies the helmholtz wave equation ( in the region of interest ) @xcite rcl [ helmholtz ] ^2 ( * _ x _ * , ) + k^2 ( * _ x _ * , ) = 0 where @xmath15 is the laplacian , @xmath16 is the scalar wavenumber , @xmath17 is the wave velocity and @xmath18 is the angular frequency which can be expressed in terms of usual frequency @xmath19 as @xmath20 .
note that even though we only consider scalar waves , our derived results are equally valid for vector waves @xcite . in the spherical coordinate system , the wavefield @xmath5 in
can be decomposed into spherical harmonics which form an orthogonal basis set for the representation of the wavefield . using the jacobi - anger expansion @xcite and the addition theorem ( * ? ? ?
* theorem 2.8 ) , we can express @xmath5 as follows rcl [ expansion ] ( * _ x _ * , ) = _ n=0^ _ m =- n^n _ nm(r , ) y_nm ( ) where spatial mode @xmath21 and spatial order @xmath22 ( @xmath23 ) are integers , such that for any particular mode @xmath24 , there are @xmath25 orders , @xmath26 is the unit vector in the direction of nonzero vector * _ x _ * , @xmath27 are the spherical harmonics and @xmath28 are the harmonic coefficients . these harmonic coefficients can be expressed as the product of the frequency dependent coefficients @xmath29 and the spherical bessel functions of the first kind @xmath30 @xcite as rcl [ signal ] _
nm(r , ) _ nm ( ) j_n(r ) .
further , we can think of @xmath28 as the space - frequency spectrum encoded in @xmath24 modes and @xmath22 orders . in this work ,
we frequently refer to @xmath31 mode space - frequency spectrum which represents the spectrum at a particular mode @xmath24 and order @xmath22 . on the contrary
, @xmath32 mode space - frequency spectrum refers to the @xmath32 mode spectrum considering all of the @xmath25 orders .
+ note that @xmath28 depends only on frequency and radial coordinate @xmath33 of vector @xmath34 , not on angular information of vector @xmath34 .
also note that the spherical harmonics @xmath27 exhibit the following orthonormal property @xcite rcl [ orthogonal ] _
y_nm()y^ _ ( ) d= _ n _ m where the integration is taken over the unit sphere @xmath35 and
@xmath36 is the kronecker delta function which is defined as @xmath37 and @xmath38 if @xmath39 .
+ the expansion of the wavefield can be viewed as a weighted sum of orthogonal spherical waveforms encoded in an infinite but countable number of spatial modes .
this expansion was used in @xcite to represent general narrowband multipath fields observed over a region of space .
the work of @xcite truncated the spherical harmonic expansion of the wavefield to a finite number of spatial modes which contain most of the energy .
the number of spatial modes in the truncated expansion indicate the spatial degrees of freedom ( i.e. , the number of independent channels ) .
hence , in this work , we need to find a suitable way to truncate this expansion for broadband multipath fields observed over a region of space .
in this section , we derive an upper bound to the available degrees of freedom of any signal observed within a band limited multipath field over a spherical region of finite radius for a finite time interval . before proceeding to the main result presented in this work , we can reasonably ask , how the observable time and band limited signals are coupled to a limited region of space for each spatial mode ? in order to answer this question , we first show the coupling of time limited signals to a finite spatial window for each spatial mode @xmath24 .
lets consider the multipath wavefield is generated by a single farfield source transmitting a time domain signal .
the wavefield is enclosed within a spherical region of radius @xmath3 .
hence , time required for the time domain signal to travel across the diameter of the spherical region is @xmath40 .
further , the wavefield is observed over a time window @xmath2 $ ] . as a result
, the observable multipath field captures information content of the time domain signal over a time interval @xmath41 . in the following theorem , we formalize this statement for the @xmath32 mode space - time wavefield @xmath42 generated by the @xmath32 mode time domain signal @xmath43 .
[ observation time of the spatial modes ] given a multipath field observed over a spherical region of radius @xmath3 for a time interval @xmath44 that is encoded in a countable number of spatial modes @xmath24 , then it is possible to capture information about the underlying @xmath32 mode time domain signal @xmath43 over an effective time interval @xmath45further , this effective time interval @xmath46 is independent of the spatial mode index @xmath24 .
proof of the theorem is provided in appendix [ a1 ] .
observing the @xmath32 mode space - time wavefield @xmath47 over a time window @xmath2 $ ] within a spatial window @xmath48 is equivalent to observing the @xmath32 mode time domain signal @xmath43 over the time window @xmath49 $ ] .
hence , the effective time interval is essentially related to the spatial observation region .
this indicates the coupling relation between space and time .
further , the effective time interval is fixed and independent of spatial modes @xmath24 . in the next subsection
, we show the coupling relation between space and frequency .
this relation truncates the expansion in to a finite number of spatial modes .
the performance of wireless communication systems is highly determined by noise . ideally ,
if the wireless communication systems are noiseless , it would be possible to measure signals with infinite precision and each spatial mode @xmath24 would have an effective bandwidth equal to the given frequency bandwidth , i.e. , from @xmath50 to @xmath51 .
however , in practical systems , signals are perturbed by noise .
hence , it is not possible to detect signals within the band of frequencies where the signal to noise ratio ( snr ) drops below a certain threshold @xmath52 .
this threshold is dependent on the antenna/ sensor sensitivity or the robustness of the signal processing method to noise . to determine how noise affects the available bandwidth at each spatial mode @xmath24 ,
let us assume that @xmath53 is the white gaussian noise on the surface of the spherical region ( at radius @xmath3 ) associated with the antenna/ sensor at the angular position @xmath54 .
hence , from and , the space - frequency spectrum on the sphere is rcl [ field ] ( r , , ) = _ n=0^ _ m =- n^n _ nm ( ) j_n(r)y_nm()+ _ r ( , ) . in the following theorem and corollary
, we characterize the white gaussian noise at the different modes .
[ white gaussian noise in @xmath55 given a zero mean white gaussian noise with variance @xmath56 in @xmath57 represented by a random variable @xmath58 where @xmath59 , such that for any function @xmath60 the complex scalar rcl _ i _ ^2 _ r ( ) _ i^ ( ) d = _ r ( ) , _ i ( ) is also a zero mean gaussian random variable with variance @xmath61 .
* eqn 8.1.35 ) considering @xmath62 to be the orthonormal basis functions @xmath63 , the spatial fourier coefficients for the noise is rcl [ noise ] _ nm ( ) = _ ^2 _ r ( , ) y_nm^ ( ) d. applying theorem @xmath64 , @xmath65 are also zero mean gaussian random variables with variance rcl e\{|_nm()|^2 } = _ 0 ^ 2 ( ) _ ^2 |y_nm^()|^2 d= _ 0 ^ 2 ( ) where the noise power is independent of the mode .
further , since the noise is white gaussian , the noise power is the same at all frequencies @xmath18 , i.e. , rcl e\{|_nm()|^2 } = _ 0 ^ 2 .
based on corollary @xmath66 , we can define the @xmath31 mode space - frequency spectrum at radius @xmath3 as rcl [ coefficient ] _
nm(r , ) = _ nm ( ) j_n(r)+ _ nm ( ) and we assume that the noise and the signal are not dependent on each other . here , the @xmath31 mode space - frequency spectrum @xmath67 takes white gaussian noise @xmath65 into account .
this white gaussian noise has the property that each spatial mode is perturbed independently of all the others .
further , @xmath29 is the @xmath31 mode signal spectrum band limited over the range @xmath1 $ ] .
in contrast , for a fixed value of the radius , @xmath68 can be treated as a function of frequency .
however , it is evident from fig .
[ sbessel ] that except for the the @xmath69 order , the spherical bessel functions @xmath70 start small before increasing monotonically to their maximum .
therefore , for frequencies less than a critical frequency @xmath71 , the magnitude of the @xmath32 order spherical bessel function @xmath72 is negligible .
[ ht ] vs. argument @xmath73 for @xmath74.,title="fig : " ] + + at each spatial mode @xmath75 , for frequencies less than a critical frequency @xmath71 , the snr is less than the threshold @xmath52 . as a result , we can not detect the space - frequency spectrum for frequencies less than @xmath71 .
in addition , as depicted in fig .
[ sbessel ] , the spherical bessel functions @xmath76 start more slowly as @xmath24 increases .
thus , after a certain spatial mode @xmath24 , the critical frequency @xmath71 is larger than the lower bound of the given bandwidth @xmath50 and the useable ( effective ) bandwidth of this mode and the modes above varies from the given frequency bandwidth .
the following theorem provides the effective bandwidth available at each spatial mode @xmath24 .
any wireless multipath wavefield observed within a spherical region of space bounded by radius @xmath3 over a frequency range @xmath10 $ ] is encoded in a finite number of spatial modes @xmath77 where the effective frequency bandwidth of the @xmath32 mode space - frequency spectrum @xmath67 is rcl[bw ] w_n = 2w , & + f_0+w- f_n , & + 0 , & + here , @xmath78 is the lowest spatial mode beyond which the critical frequency @xmath79 , @xmath80 is the lowest spatial mode beyond which the critical frequency @xmath81 and rcl [ cf ] f_n = + ( ) with the threshold @xmath52 depicting the ability of the system to detect signals buried in noise .
further , we consider that the power of the @xmath32 mode signal spectrum @xmath29 is finite and bounded for all modes @xmath24 , orders @xmath22 and frequencies @xmath18 , i.e. , rcl [ emax ] e\{|_nm()|^2}^2 , hence , the maximum snr of the signal spectrum @xmath29 for any mode @xmath24 is rcl [ k ] ( snr ) _
proof of the theorem is given in appendix [ a2 ] .
the effective frequency bandwidth of each spatial mode is related to the spatial observation region and varies from the given frequency bandwidth depending on the critical frequency @xmath71 .
this portrays the strong coupling relation between space and frequency .
further , for @xmath82 , the critical frequency @xmath71 is greater than the upper bound of the given frequency range @xmath51 . therefore , we can truncate the expansion in to a finite number of spatial modes as rcl [ expansion1 ] ( * _ x _ * , ) = _ n=0^n _ _ m =- n^n
_ nm(r , ) y_nm ( ) . using , the upper bound for the spatial modes @xmath80 is rcl [ nmax ] n_= e ( f_0+w ) + ( ) where @xmath83 is the ceiling value , since by definition spatial modes are integers .
we are now in a position to provide an upper bound to the available degrees of freedom of wideband signals observed over finite spatial and temporal windows . in order to do so ,
it is useful to think of @xmath31 mode space - frequency spectrum @xmath84 in time domain , in which case we obtain rcl[s ] _
nm(r , t)= _ _ n _ nm(r , ) e^jt d where @xmath42 is the inverse fourier transform of @xmath84 with respect to @xmath18 and the integration is taken over @xmath85 with @xmath86 $ ] for @xmath87 and @xmath88 $ ] for @xmath89 where @xmath71 is defined in .
+ we expand @xmath90 over the frequency range using the fourier series expansion , similar to @xcite , as follows rcl [ s ] _
nm(r,)= _ = -^
c_nm(r ) e^-j where the fourier coefficients @xmath91 are proportional to the samples of @xmath42 and @xmath92 is the effective frequency of the @xmath32 mode defined in .
the result is obtained from when @xmath93 .
it illustrates that the samples of @xmath42 determine the coefficients @xmath94 in the fourier series expansion .
therefore , analogous to shannon s work @xcite , we can reconstruct the @xmath31 mode space - time signal @xmath42 from its samples as follows rcl [ ftruncation ] _
nm(r , t)&= & _
= -^_nm(r,)e^j2w_0n(t- ) where @xmath95 is the mid band frequency of the @xmath31 mode .
we obtain by substituting the fourier series in , applying and then exchanging integration and summation .
+ hence , it is possible to reconstruct the space - time signal @xmath6 by summing the @xmath31 mode space - time signals for all possible values of @xmath24 and @xmath22 over all propagation directions , i.e. , rcl ( * _ x_*,t ) & = & _ n=0^n _ _ m =- n^n _ nm(r , t ) y_nm ( ) and substituting yields rcl [ stsignal ] ( * _ x_*,t ) & = & _ n=0^n _ _ m =- n^n
_ = -^_nm(r,)e^j2w_0n(t- ) y_nm ( ) .
observe that the spherical harmonics @xmath63 are orthogonal over the spherical region as shown in .
further , considering rcl _ ( t ) = e^j2w_0n(t- ) , the functions @xmath96 are orthogonal over time .
proof of the orthogonality of the functions @xmath96 is provided in appendix [ a3 ] .
therefore , following the same reasoning as shannon @xcite , discussed in section [ shannon ] , the minimum numbers of terms required in the sum that satisfy the constraints on observation region size , bandwidth and observation time window provide the available signal degrees of freedom within the given multipath field . here
, @xmath97 can be truncated to @xmath98 $ ] .
we truncate @xmath97 based on the fact that the @xmath31 mode space - time signal @xmath42 is band limited to @xmath92 .
hence , we can determine @xmath42 by taking samples @xmath99 units apart .
now , in order to limit @xmath42 within the interval @xmath46 , @xmath100 is non - zero for only the appropriate values of @xmath97 , such that @xmath101 .
this means that the degrees of freedom of the @xmath31 mode space - time signal @xmath42 is @xmath102 .
hence , the total degrees of freedom of any signal available in the given multipath field considering all modes and orders is given by rcl [ eq : d ] d= _
n=0^n_max_m =- n^n ( w_nt_eff+1 ) where for different values of @xmath24 , @xmath92 is provided in and @xmath46 is defined in .
using our previous results , we now provide the following theorem for the degrees of freedom of any signal observed within a broadband multipath field .
given a multipath wireless field band limited to @xmath1 $ ] over the time interval @xmath2 $ ] within a @xmath0 spherical region of radius @xmath3 , then the degrees of freedom of any signal observable within this multipath field is upper bounded by rcl [ dspace ] d^2 + 2w ( t+ ) where @xmath80 is defined in and from theorem @xmath103 , @xmath78 is the lowest spatial mode beyond which the critical frequency @xmath79 . using , rcl [ nmin ] n_= e ( f_0-w ) + ( ) . since by definition spatial modes are integers , @xmath78 is defined as a ceiling value .
proof of the theorem is given in appendix [ a4 ] . in the next section ,
we graphically illustrate our derived results .
consider @xmath104 as the wavelength corresponding to the mid band frequency @xmath4 .
therefore , @xmath3 , @xmath105 , and @xmath44 can be represented in terms of @xmath104 or @xmath4 as follows @xmath106 $ } \nonumber \\ & t&= d /f_0 , ~~\mbox{$d\in[0,\infty)$}\end{aligned}\ ] ] where @xmath107 , @xmath108 and @xmath109 are real numbers .
furthermore , @xmath110 and @xmath111 represents the extreme broadband scenario . hence , can be rewritten as @xmath112\end{aligned}\ ] ] where considering @xmath113 , @xmath114 and @xmath115 in fig .
[ stf ] , degrees of freedom @xmath116 in is plotted as a function of radius of the spherical region and fractional bandwidth . it is evident from the figure that for a given observation time window , there is a sub - quadratic growth in available degrees of freedom with increasing region size and bandwidth .
note that we consider a scenario with a small value of @xmath117 .
increasing @xmath117 ( which is equivalent of minimizing the affect of noise ) we can achieve higher signal degrees of freedom .
[ ht ] vs. radius @xmath3 vs. fractional bw at a fixed observation time window ( @xmath118 ) for @xmath119 .
radius , fractional bw and observation time are defined in .,title="fig : " ] next , we portray the affect of snr on signal degrees of freedom considering different values of @xmath117 . it is evident from fig .
[ rho ] that for a given observation time window , increasing the value of @xmath117 , we can obtain a growth in degrees of freedom as a function of ( a ) radius of the observable region and ( b ) bandwidth , respectively .
[ ht ] at a fixed time window ( @xmath118 ) ( a ) for a fixed bandwidth ( @xmath120 ) , ( b ) for a fixed radius of the region ( @xmath121).,height=245 ] at a fixed time window ( @xmath118 ) ( a ) for a fixed bandwidth ( @xmath120 ) , ( b ) for a fixed radius of the region ( @xmath121).,height=245 ] we now present the signal degrees of freedom as a function of bandwidth at a fixed observation time window .
the parameter is the radius of the observable spherical region with @xmath122 .
the results in fig . [ space ] ( a ) demonstrate that the degrees of freedom increases sub - quadratically as the radius of the observable region increases .
+ on the contrary , it is clear from fig . [ space ] ( b ) that considering a fixed bandwidth , by increasing the radius of the observable region , it is possible to obtain a rapid non - linear growth in the degrees of freedom as a function of observation time .
[ ht ] considering @xmath123 ( a ) for a fixed time window ( @xmath118 ) , ( b ) for a fixed bandwidth ( @xmath120).,height=245 ] considering @xmath123 ( a ) for a fixed time window ( @xmath118 ) , ( b ) for a fixed bandwidth ( @xmath120).,height=245 ] + note that the two scenarios mentioned above clearly indicate that we can obtain significantly high signal degrees of freedom only by increasing the radius of the observable region irrespective of bandwidth or observation time .
this significant growth is also true if we increase the bandwidth but keep the radius of the region and observation time constant .
+ from the results , increasing the frequency or the radius leads to a sub - quadratic growth in the degrees of freedom . on the other hand , increasing the observation time window or the snr does not provide such a significant increase in the degrees of freedom .
in this section , we elucidate the physical insights of the main results and attempt to briefly discuss about their applications . we can make the following comments based on theorem @xmath124 : * from a spatial diversity perspective , in shannon s proposed communication model @xcite , wideband signals are encoded in only one spatial mode or one channel over which information is transmitted . on the contrary , our proposed model contains spatially diverse wideband signals that are encoded in a finite number of spatial modes @xmath24 . as a result
, intuitively the available degrees of freedom of any wideband signal observed over finite spatial and temporal windows should be shannon s degrees of freedom result extended form one spatial mode to @xmath24 modes .
our derived result clearly comply with this intuition .
the number of modes are the number of independent channels available to receive information due to the availability of measurements over a spatial region .
this means that each mode or channel has its own time - frequency degrees of freedom . spatial diversity , therefore ,
provides a number of independent channels over which time - frequency information can be transmitted .
* the degrees of freedom result does not agree with the well established result of evaluating degrees of freedom of spatially diverse wideband signals as a product of spatial degrees of freedom and time - frequency degrees of freedom @xcite .
however , in the propagation of waves even though space , time and frequency are separate entities , in spatially diverse wideband transmissions , space and time as well as space and frequency are strongly coupled , the results of @xcite fail to show those coupling relationships . on the contrary , our derived result takes the coupling relations in account . *
shannon s work considers broadband transmission over a single channel and shows that the channel has _ `
time - bandwidth product + 1 ' _ degrees of freedom . on the contrary , in addition to broadband transmission , we take spatial diversity in account . therefore , in this work , we consider broadband transmission over finite @xmath24 number of channels .
our analysis indicates that each of these channels has _ ` effective time - effective bandwidth product + 1 ' _ degrees of freedom .
this means that considering spatial diversity , we can capture more information from broadband transmission . for higher modes ,
the effective/ usable bandwidth is less than the measured bandwidth and so not all spatial modes can covey the same amount of information . therefore , from theorem @xmath103 , for modes
@xmath24 above @xmath78 , the _ ` effective time - effective bandwidth product + 1 ' _ decreases as the mode @xmath24 increases .
let us consider that the threshold is equivalent to the maximum snr of the @xmath32 mode signal spectrum @xmath29 , i.e. , @xmath125 .
this means that signals below the frequency @xmath126 are submerged in noise and can not be detected . as a result , from we obtain rcl[asymptotic ] d & & _ + _ + & & + _ which yields the following observations : * for non - spatially diverse multipath fields @xmath127 , in , @xmath128 reduces to 1 , @xmath129 reduces to @xmath130 and @xmath131 becomes zero .
thus , non - spatially diverse multipath fields provide @xmath13 degrees of freedom . * for narrowband wavefields @xmath132 , in , @xmath128 reduces to @xmath133 , whereas , both @xmath129 and @xmath131 become zero . hence , there are @xmath134 degrees of freedom available in narrowband wavefields operating at the mid band frequency @xmath4 . as a result , in terms of wavelengths ,
degrees of freedom available in narrowband wavefields is @xmath135 where @xmath136 is the wavelength corresponding to the mid band frequency @xmath4 . *
if any signal observable within a multipath field is representable with only one sample in time domain @xmath137 , then , by substituting @xmath138 in , we obtain + rcl [ t=0 ] d&&(e ( f_0 + w ) + 1)^2 + 4w^2 + & & + 2w . +
this equation indicates that even when there is only one sample available in time domain , spatial diversity influences the amount of information that can be captured within the observable region .
the derived result represents the degrees of freedom of signals observable within a broadband multipath field over finite spatial and time windows , assuming the signals are submerged in noise for frequencies less than the critical frequency @xmath126 and are not detectable .
this result is consistent with shannon s time - frequency degrees of freedom result @xcite when we take sample at a single spatial point @xmath127 .
further , is consistent with the degrees of result derived in ( * ? ? ?
44 ) at wavelength @xmath104 when we consider narrowband frequency transmissions @xmath132 . the degrees of freedom result obtained in this paper can be used to provide insights and bounds in the following areas : * in the context of spatial broadband communications such as wireless mimo communications , this work addresses the fundamental question of how the spatial degrees of freedom is interrelated to the time - frequency degrees of freedom .
the result provides insights into gains or losses of degrees of freedom in space and time - frequency analysis . * for broadband beamforming the degrees of freedom
characterises the maximum resolution that can be obtained @xcite .
the greater the degrees of freedom the higher the resolution can be obtained .
this has particular importance in this area since we have more variables to work with to perform broadband beamforming .
the performance of beamforming in wireless networks improves with the available degrees of freedom and has been shown in @xcite . in next generation of wireless communications capable of transmitting high quality video and audio ,
array gain is obtained by using broadband beamforming which exploits the spatial degrees of freedom and the effective bandwidth of each of these spatial degrees .
our work shows that as more spatial degrees are exploited for beamforming , for a receive antenna occupying a limited spatial region , the effective bandwidth of the higher spatial degrees or modes @xmath24 are less than the bandwidth and decreases with the mode index @xmath24 . * for broadband reception of signals by antennas placed within a given spatial region , initially linear growth in the degrees of freedom
is obtained with increasing number of antennas .
once the number of antennas is greater than @xmath139 , the increase in degrees of freedom reduces with each antenna added until number of antennas is equivalent to @xmath140 .
after that no gain can be obtained .
this is because the wavefield constraint results in correlations between channels when the number of antennas becomes too large . *
interference alignment is a promising new area introduced in the last two decades .
this seeks to solve the spectrum scarcity in wireless communications by utilizing the available degrees of freedom in space , time and frequency @xcite .
however , in mimo wireless networks , interference alignment uses the parallel channels in space offered by spatial degrees of freedom for alignment .
our derived results can be used to develop an interference alignment scheme for mimo wireless networks .
our results show that optimal signal alignment needs to consider that not all spatial channels are equal and can place interference on the spatial channels that have the lower time - frequency degrees of freedom . hence
, the interference channels should be placed in the spatial channels corresponding to the higher spatial modes .
further , the degrees of freedom analysis of this work provides the maximum degrees of freedom that can be utilized in these broadband communications with interference . *
recently , distributed mimo communications have seen an increase in importance due to the popularity of sensor and ad - hoc networks .
distributed mimo includes all muti - user communication configurations where the communications input and outputs are distributed over different users .
works of @xcite have studied the spatial degrees of freedom for these considering different channel conditions and showed performance gains .
these works looked at only narrowband transmissions , however , practical wireless transmissions are performed over a bandwidth .
considering this , our work shows the maximum degrees of freedom available over space , time and frequency for users in a limited spherical region cooperating to receive broadband information .
also , we show how the time - frequency degrees of freedom is distributed over the spatial modes .
this paper provides an upper bound to the degrees of freedom of any signal observed within a band limited multipath wireless field over finite spatial and temporal windows .
this upper bound is obtained characterizing the multipath field as solution to helmholtz wave equation encoded in a finite number of spatial modes .
the analysis of the work shows that the effective observation time is independent of spatial modes and related to the spatial dimension of the observable region .
further , for broadband transmissions , at each spatial mode , the observable signals are band limited within an effective frequency bandwidth and depending on the mode , the effective bandwidth varies from the given frequency bandwidth .
these findings imply that when both spatial diversity and broadband transmissions are taken in account , space and time as well as space and frequency can not be decoupled . + the degrees of freedom result derived in this work takes the coupling relations into account and portrays the interrelation between spatial degrees of freedom and time - frequency degrees of freedom . from a spatial diversity perspective ,
shannon s proposed communication model considers wideband signal encoded in only one spatial mode or one channel over which information is transmitted , the available degrees of freedom of spatially diverse wideband signal encoded in finite number of spatial modes @xmath24 is shannon s degrees of freedom result extended form one spatial mode to @xmath24 modes .
this means that each mode or channel has its own time - frequency degrees of freedom .
+ we also show that analogous to time , space can be treated as an information bearing object , since degrees of freedom increases sub - quadratically as the size of the observable spatial region increases irrespective of bandwidth or time window .
further , the derived result portrays how the degrees of freedom is affected by the acceptable snr at each spatial mode .
let @xmath47 be the inverse fourier transform of @xmath141 , then , the inverse fourier transform of is rcl[ot ] _
nm(r , t)= a_nm(t)p_n ( ) where the time domain coefficients @xmath142 are the inverse fourier transform of @xmath29 and the legendre polynomials @xmath143 are the inverse fourier transform of @xmath30 . in , we can consider @xmath144 as the @xmath32 mode time domain signal .
therefore , the @xmath32 mode space - time signal @xmath145 is a convolution between the @xmath32 mode time domain signal @xmath144 with the legendre polynomial @xmath146 . the convolution with the legendre polynomial @xmath146 represents the wavefield constraint and information content in the @xmath32 mode space - time signal @xmath145 is contained in @xmath144 .
further , from the definition in @xcite , legendre polynomials @xmath143 are defined only for @xmath147 .
this characteristic of legendre polynomials is also evident from fig .
[ lp ] . for @xmath148 . ]
+ we observe the @xmath32 mode space - time signal @xmath47 over a time window @xmath2 $ ] within a spatial window @xmath48 .
therefore , for the given wavefield constraint , at any particular radius @xmath33 , we can observe information in the the @xmath32 mode time domain signal @xmath144 over the time window @xmath149 $ ] . if we consider that the @xmath32 mode space - time signal is observed within a sphere of radius @xmath3 , it is possible to capture information about the @xmath32 mode time domain signal @xmath43 over a maximum of @xmath41 time interval .
hence , observing the @xmath32 mode space - time signal @xmath47 within a spherical region of radius @xmath3 over the time window @xmath2 $ ] is equivalent to observing the @xmath32 mode time domain signal @xmath43 over a maximum time window @xmath49 $ ] .
observable multipath field on the surface of the sphere ( at radius @xmath3 ) is rcl ( r , , ) = _ n=0^_m =- n^n _ nm(r , ) y_nm ( ) applying parserval s theorem with respect to the spherical harmonics , the average power of the observable multipath field from all propagation directions is a sum over the average power in the different modes , such that rcl _
e\{|(r , , ) |^2 } d= _
n=0^_m =- n^n e\{|_nm(r,)|^2}. since the noise is independent of the signal , using we obtain rcl [ parserval ] e\{|_nm(r,)|^2}= e\{|_nm()|^2 } |j_n(r)|^2 + e\{|_nm()|^2}. according to corollary @xmath66 , @xmath150 .
therefore , we can rewrite as rcl e\{|_nm(r,)|^2}= e\{|_nm()|^2 } |j_n(r)|^2 + _ 0 ^ 2 . from this , signal to noise ratio ( snr ) for the @xmath31 mode at frequency @xmath18
is rcl ( snr)_nm ( ) & = & [ snrn ] + & & ( snr ) _ , |j_n(r)|^2 [ snr ] where follows from and . + next
, based on the properties of the bessel functions @xcite and the spherical bessel functions @xcite and following a few intermediate steps , we can derive the following bound for the spherical bessel functions for large @xmath24 rcl [ j ] |j_n(z)| ^n , where @xmath151 is the gamma function .
therefore , rcl [ snr3 ] ( snr)_nm ( ) ^n+1 .
now , using the exponential inequality , @xmath152 for @xmath153 , we rewrite as rcl ( snr)_nm ( ) _ , e^(e r / c-2n ) , where @xmath154 . since @xmath155 , for @xmath75 , we have rcl ( snr)_nm ( ) < ( snr ) _ , e^(e r / c-2n ) . note that at the @xmath31 mode
, it is not possible to detect signals within the band of frequencies where the snr drops below a certain threshold @xmath52 .
hence , for a frequency to be usable to capture information at the @xmath31 mode , the snr must be larger than or equal to the threshold @xmath52 .
the frequency at which the @xmath156 is at least equal to the threshold @xmath52 is the critical frequency @xmath71 ( where @xmath157 ) .
therefore , rcl[eq : detec ] ( snr ) _ ,
e^(2e f_n r / c-2n ) = .
this result is easily derived based on the reasoning provided in section [ effectivebw ] . here
, we briefly discuss the reasoning : as depicted in fig .
[ sbessel ] , expect for the @xmath69 order , spherical bessel functions show a high pass characteristics . as a result , for frequencies less then a critical frequency @xmath71 ,
the magnitude of the @xmath32 @xmath158 order spherical bessel function is negligible .
therefore , at each spatial mode @xmath75 , for frequencies less than the critical frequency @xmath71 , it is not possible to maintain the snr at least equal to the threshold @xmath52 .
+ now , making @xmath71 the subject of the formula in yields .
this means that for spatial modes @xmath75 , signals below frequency @xmath71 are not detectable since will not be satisfied .
observe that for any particular mode @xmath159 , if @xmath160 , the effective bandwidth of that mode is @xmath161 .
in addition , if @xmath162 , the effective bandwidth of this mode and modes above this is zero .
it should also be noted that for a fixed value of radius , @xmath163 is active within the frequency range @xmath164 as depicted in fig [ sbessel ] , hence , effective bandwidth of the @xmath69 mode is @xmath165 .
these arguments are written mathematically as .
since rcl _ ( t ) = e^j2w_0n(t- ) , rcl [ timebasis ] _ -^ _ ( t )
^_(t ) dt & = & _ -^
e^j2w_0n(t- ) e^-j2w_0n(t- ) dt + & = & _ -^ dt + & = & \ { rl 0 & + & . where @xmath166 and for @xmath167 , @xmath168 . here ,
is derived using the fact that ( * ? ? ?
11 ) rcl & _ -^ & dt + & = & \ { rl 0 & + & . thus , the functions @xmath96 are orthogonal over time .
we start from and by expanding the sum , we obtain s. hua , c. geng , t. gou , and s. a. jafar , `` degrees of freedom of mimo x networks : spatial scale invariance , one - sided decomposability and linear feasibility , '' in _ eee international symposium on information theory proceedings ( isit ) _ , 2012 , pp .
20822086 .
a. ozgur , o. leveque , and d. tse , `` spatial degrees of freedom of large distributed mimo systems and wireless ad hoc networks , '' _ ieee journal on selected areas in communications _ ,
31 , no . 2 ,
pp . 202214 , february 2013 .
r. a. kennedy , p. sadeghi , t. d. abhayapala , and h. m. jones , `` intrinsic limits of dimensionality and richness in random multipath fields , '' _ ieee transactions on signal processing _ ,
55 , no . 6 , pp .
25422556 , june 2007 .
a. s. y. poon , r. w. brodersen , and d. n. c. tse , `` degrees of freedom in multiple - antenna channels : a signal space approach , '' _ ieee transactions on information theory _
51 , no . 2 ,
pp . 523536 , feb .
2005 .
f. bashar , s. a. salehin , and t. d. abhayapala , `` analysis of degrees of freedom of wideband random multipath fields observed over time and space windows , '' in _ ieee statistical signal processing workshop _ , gold coast , australia , 29 jun - 02 jul 2014 , accepted .
f. bashar and t. d. abhayapala , `` degrees of freedom of band limited signals measured over space , '' in _ the 12th international symposium on communications and information technologies ( iscit 2012 ) _ , 2012 , pp .
735 740 .
, `` prolate spheroidal wave functions , fourier analysis and uncertainty - iii : the dimension of the space of essentially time- and band - limited signals , '' _ bell system technical journal _ , vol . 41 , pp .
12951336 , july 1962 .
m. dollard , `` on the time - bandwidth concentration of signal functions forming given geometric vector configurations , '' _ ieee transactions on information theory _ , vol .
10 , no . 4 , pp . 328338 , oct .
1964 .
a. bostrom , g. kristensson , and s. strom , _ transformation properties of plane spherical and cylindrical scalar & vector wave functions_.1em plus 0.5em minus 0.4emelsevier science publishers , 1991 , ch . 4 , pp . 165210 .
j. krolik and d. swingler , `` multiple broad - band source location using steered covariance matrices , '' _ ieee transactions on acoustics , speech and signal processing _ , vol .
37 , no .
1481 1494 , oct
. 1989 . a. el - keyi and b. champagne , `` cooperative mimo - beamforming for multiuser relay networks , '' in _ ieee international conference on acoustics , speech and signal processing , 2008 , icassp 2008 _ , march 2008 , pp .
27492752 .
k. gomadam , v. r. cadambe , and s. a. jafar , `` a distributed numerical approach to interference alignment and applications to wireless interference networks , '' _ ieee transactions on information theory _ , vol .
57 , no . 6 , pp . 33093322 , june 2011 .
m. franceschetti , m. d. migliore , and p. minero , `` the capacity of wireless networks : information - theoretic and physical limits , '' _ ieee transactions on information theory _
55 , no . 8 , pp .
34133424 , august 2009 . | the study of degrees of freedom of signals observed within spatially diverse broadband multipath fields is an area of ongoing investigation and has a wide range of applications , including characterising broadband mimo and cooperative networks .
however , a fundamental question arises : given a size limitation on the observation region , what is the upper bound on the degrees of freedom of signals observed within a broadband multipath field over a finite time window ? in order to address this question , we characterize the multipath field as a sum of a finite number of orthogonal waveforms or spatial modes . we show that ( ) the effective observation time " is independent of spatial modes and different from actual observation time , ( ) in wideband transmission regimes , the effective bandwidth " is spatial mode dependent and varies from the given frequency bandwidth .
these findings clearly indicate the strong coupling between space and time as well as space and frequency in spatially diverse wideband multipath fields . as a result
, signal degrees of freedom does not agree with the well - established degrees of freedom result as a product of spatial degrees of freedom and time - frequency degrees of freedom .
instead , analogous to shannon s communication model where signals are encoded in only one spatial mode , the available signal degrees of freedom in spatially diverse wideband multipath fields is the _ time - bandwidth _ product result extended from one spatial mode to finite modes .
we also show that the degrees of freedom is affected by the acceptable signal to noise ratio ( snr ) in each spatial mode .
degrees of freedom , multipath propagation , spatial sampling , broadband mimo networks , distributed mimo . |
late time cosmic acceleration at the present epoch has almost been a de facto phenomenon since the late nineteens .
advances in cosmological observations during the past two decades reveal strong evidences in favour of this accelerated expansion of the universe .
these evidences have been brought forth _ la _ independent astrophysical observations like supernovae type ia ( sne ia ) luminosity distance modulus as a function of redshift @xcite , observational hubble data ( ohd ) @xcite , cosmic microwave background ( cmb ) shift parameter @xcite and baryon oscillation spectroscopic survey ( boss ) data @xcite .
a good deal of attempts have been taken to explain this accelerated expansion assuming the presence of some exotic fluid , namely _ dark energy _ , in huge abundances in the universe . though there exists a lot of dark energy models
( see for example @xcite and references therein ) with standard as well as exotic ideas ; the canonical and non - canonical scalar fields are the most promising candidates till date .
of late , robert r. caldwell and eric v. linder @xcite categorized these scalar field models in two broad classes namely freezing " and thawing " dark energy , based on the asymptotic behavior of the scalar field potential . in thawing models ,
dark energy equation of state @xmath0 initially remains at @xmath11 and deviates from @xmath11 near present epoch whereas just the opposite behavior of @xmath0 is witnessed in freezing models .
thawing models , in which we are interested in the present article , are broadly classified into two categories : ( i ) quintessence ( for which @xmath0 moves to @xmath12 ) , and ( ii ) phantom ( where @xmath0 is less than @xmath11 )
. a third possibility has also been explored in which lead to both quintessence and phantom behavior of @xmath0 . in these slow - rolling scalar field models with nearly flat potential , initially the kinetic energy of the field is much smaller than the potential energy .
this is because of the initial large hubble damping which keeps the field nearly frozen at @xmath13 at earlier era i.e. , in radiation and matter dominated eras . due to the expansion of the universe , energy density of the universe decreases .
after the radiation and matter dominated eras , the field energy density becomes comparable to the background energy density of the universe resulting in the deviation of the field from its frozen state , thereby leading to deviation of @xmath0 from @xmath11 .
slow - roll scalar field thawing models can be characterized by different relations between @xmath0 and the scale factor @xmath14 of the universe .
some typical examples of cpl parametrization ( eq .
( [ cpl ] ) ) @xcite , pngb models ( eq .
( [ pngb ] ) ) and algebraic thawing models ( eq . ( [ algethaw ] ) ) are included in the work by e. v. linder @xcite .
the corresponding equation of state parameterizations are respectively given by , @xmath15 @xmath16 @xmath17 where @xmath18 is a parameter which is inversely proportional to the symmetry breaking energy scale and @xmath19 and @xmath20 are two free parameters .
in the present work , we propose a two parameter generalization for this thawing dark energy models as @xmath21 where @xmath22 is an arbitrary function of scale factor @xmath14 . in this article , we have chosen @xmath22 as @xmath23 , where @xmath24 and @xmath25 are two arbitrary parameters . in this context
we would like to mention that choice of @xmath22 can be made otherwise and it would be interesting to see if there exists any observational constrain on the form of @xmath22 which is beyond the scope of this article . with the chosen form of @xmath23 for @xmath26 and
@xmath27 our proposal exactly overlaps with cpl thawing dark energy model @xcite . for @xmath26 and @xmath28 , our proposal leads to pngb thawing dark energy model @xcite which have been studied exclusively for scalar fields dark energy with pngb potential @xcite .
for suitable choice of the parameters @xmath25 and @xmath24 , our model can approximately reflect algebraic thawing @xcite as well .
as it turns out , all the existing ( and probably , upcoming ) thawing dark energy models fall in this broad minimal parametrization with different values of the parameters @xmath25 and @xmath24 .
so , rather than proposing individual models , it is quite reasonable to construct a minimal generic form of parametrization , analyze it and search for possible constraints on the parameters from present - day observations .
this is the primary objective of the present article .
along with this view , we also draw some comparisons among the results obtained for different values of @xmath25 ( i.e. , for @xmath26 , @xmath29 and @xmath30 ) with different fixed values of @xmath24 and vice - versa .
we further provide justification for this proposed generalized form of thawing dark energy model against the other existing thawing models by comparing them with ours .
moreover , we constrain our model by latest supernova type ia data from union2.1 compilation @xcite , newly released observational hubble parameter data @xcite , cosmic microwave background shift parameter data from wmap 9 year results @xcite and the latest baryon oscillation spectroscopic survey ( boss ) data from sdss - iii data release @xcite . for such analyses we have five parameters in total namely @xmath24 , @xmath25 , @xmath31 , @xmath3 and @xmath4 ( where @xmath3 and @xmath4 are matter density at present epoch normalized to critical density and hubble parameter at present epoch respectively ) . since the value of @xmath32 ( the normalized radiation density at present epoch ) is very low we do not treat it as a parameter and consider @xmath33 @xcite for numericals .
our analysis also helps in comparing the standard diagnostics with model independent ones , and reveals the pros and cons of each one .
the major conclusions of the paper are as follows : * existing thawing dark energy models @xcite can be generalized in the form of eq .
( [ ourmodel ] ) as we have presented in this article .
our minimal generalization of thawing dark energy models ( eq .
( [ ourmodel ] ) ) with two parameters @xmath25 and @xmath24 leads to different existing thawing models namely cpl ( @xmath34 ) , pngb ( @xmath35 ) and the algebraic thawing ( suitable choices of @xmath36 ) .
* results obtained for different @xmath25 values ( with different values of @xmath24 ) barely differ from the observational point of view . other way around , we can say that values of @xmath25 ( with different values of @xmath24 ) can hardly affect the best fit values as well as the @xmath37 c.l.s of matter density parameter @xmath3 and hubble parameter at the present epoch @xmath4 .
also the best fit plots for redshift evolution of average deceleration parameter @xmath6 and the growth of matter perturbations in terms of evolution of growth factor @xmath5 with redshift @xmath38 ( best fit plots ) remain unaffected when the values of @xmath25 and @xmath24 are altered accordingly .
therefore it is difficult to provide a unique dark energy eos @xmath0 for the thawing dark energy models as different values of @xmath25 with different values of @xmath24 lead to the same cosmological dynamics . * the best fit values and the @xmath37 c.l.s of eos at the present epoch @xmath31
does leave little trace on model discrimination for thawing dark energy . here
we discuss the fact that the values of @xmath25 and @xmath24 can be constrained by @xmath10 ( @xmath39 ) plots @xcite for thawing dark energy models .
more importantly , best fit @xmath10 plots can also serve as a model discriminator for the thawing dark energy models .
the non - linear @xmath10 plots can be realized for values of @xmath25 other than @xmath40 with different values of @xmath24 .
this is an important issue as pngb and cpl parameterizations can result only in linear @xmath10 plots and recent works on scalar field dark energy models point towards the non - linear @xmath10 plots @xcite . *
most importantly , the model - independent parameters like statefinder pair @xmath41 @xcite and the so called @xmath8 @xcite parameter do play a crucial role in discriminating among different dark energy models .
study of these parameters in the context of our generalized thawing model , therefore , reveals the fact that unlike the standard parameters mentioned in 2nd major conclusion above , these parameters indeed serve as model discriminators for different thawing dark energy models i.e. , these parameters can identify the different values of @xmath25 as well as @xmath24 in our generalized model . + the paper is organized as follows . in the next sec .
we propose the generalization for the thawing dark energy models and mention the standard as well as the model independent parameters .
3 briefly describes the various observational data we used . in the sec .
4 we present the results obtained by the analyses of the various observational data . in the sec .
5 we discuss our results and put forward the conclusions of the present work .
we propose a minimal two parameter generalization for thawing dark energy eos @xmath0 as , @xmath42 where @xmath22 is an arbitrary function of scale factor @xmath14 of the universe .
we study the dynamical universe with radiation , matter and thawing dark energy obeying the proposed eos @xmath0 with @xmath43 .
the proposed choice of @xmath22 here , for the generalized thawing model is motivated by the following findings : \i ) for @xmath26 and @xmath27 , our model is exactly same as cpl parametrization ( eq [ cpl ] ) .
\ii ) for @xmath26 and @xmath44 ( f being the parameter described in sec .
1 ) , our model is exactly same as pngb model ( eq [ pngb ] ) .
\iii ) algebraic thawing case ( eq [ algethaw ] ) can also approximated for certain choices of @xmath24 and @xmath25 in terms of @xmath20 and @xmath19 .
\iv ) for values of @xmath25 other than @xmath40 , generalized thawing dark energy eos takes the form @xmath48 in order to test the validity of our generalized model we show in fig .
[ w - wexample ] , the theoretically predicted @xmath10 @xmath49 plots for different thawing models that arise for different values of @xmath25 and @xmath24 ( we will put constrains on this @xmath10 plane with direct observational data later in this paper ) .
we find from fig .
[ w - wexample ] , theoretically obtained @xmath10 plane for different combinations of @xmath24 and @xmath25 in our model satisfy the allowed regions for the same @xcite . in fig .
[ w - wexample ] , the left plot is for quintessential thawing with @xmath50 and the right one is for the case of thawing originated in phantom scenarios with @xmath51 . for @xmath26 with @xmath27 (
dotted ) we get cpl thawing ( eq .
( [ cpl ] ) ) and for @xmath26 with @xmath52 ( solid and dot - dashed respectively ) we get pngb thawing ( eq . ( [ pngb ] ) ) .
the plots in black in fig .
[ w - wexample ] indicate these two models in the @xmath10 plane .
the orange plots are for the algebraic thawing model with @xmath53 ( dotted lines ) , @xmath54 ( solid lines ) and @xmath55 ( dot - dashed lines ) .
the results with higher values of @xmath25 are shown by the green ( @xmath56 ) and blue plots ( @xmath29 ) .
the dotted , solid and dot - dashed lines in these cases corresponds to @xmath57 respectively . in this section
we discuss the constraints on the model parameters of our genralized thawing dark energy eos as proposed in the work by caldwell @xcite . in fig .
[ nc ] red region shows the allowed region of the parameter space @xmath58 which is allowed for thawing dark energy with our generalized eos .
it is also necessary to point out that our generalized eos can represent dark energy models other than thawing .
the region of @xmath58 parameter space except the red zone represents these models .
this allowance of @xmath25 and @xmath24 values in fig .
[ nc ] is also reflected in fig .
[ w - wexample ] .
as is well - known , any dark energy model must at least probe three parameters directly from observations : \i ) the present value of equation of state ( eos ) for dark energy ( @xmath31 ) \ii ) the present value of matter density ( @xmath3 ) \iii ) the hubble parameter today ( @xmath4 ) .
nevertheless , dark energy model building today is tightly constrained by several observations , which , taken together , leave out a very narrow window through which the model should pass .
so , from today s perspectives , apart from the above three good old parameters , the supplementary parameters which one needs to address are the following : the statefinder pair @xmath41 @xcite serves as a geometrical diagnostic to probe the properties of dark energy in a model independent manner .
this pair @xmath41 has been studied extensively in the earlier works @xcite . for the late universe ( @xmath59 ) , which is well approximated by the presence of matter and dark energy
, the statefinder pair @xmath41 can be expressed as , @xmath60 where @xmath14 is the scale factor of the universe and @xmath61 is the dark energy density parameter . in the late universe we have @xmath62 , @xmath63 being the matter density parameter . for @xmath64 model , it can be checked that the statefinder pair @xmath41 takes the value @xmath65 and @xmath66 . any deviation in @xmath67 from @xmath40 and @xmath68 from @xmath69 , indicates the existence of varying dark energy in the universe .
the @xmath70 parameter proposed by sahni @xcite , is another tool to distinguish the dynamical dark energy from the cosmological constant .
the uncertainty in matter density parameter allows significant errors in cosmological reconstructions of dark energy .
@xmath70 parameter can in practice differentiate between the models , independent of the matter density parameter .
the @xmath70 diagnostic has been studied well in the earlier works @xcite .
@xmath70 parameter is defined in terms of hubble parameter which can directly be measured in cosmological observations .
the two - point @xmath70 @xcite diagnostic is given by , @xmath71 where @xmath72 .
it can be easily seen that for cosmological constant @xmath73 and when @xmath74 , @xmath75 ( @xmath76 ) represents the case of quintessence ( phantom ) @xcite .
this is how @xmath70 evaluated at two different redshifts ( @xmath77 and @xmath78 ) can help in distinguishing the dark energy model .
needless to mention that this procedure is independent of @xmath3 and @xmath4 .
the three - point diagnostic @xmath8 @xcite is defined by , @xmath79 for @xmath64 model @xmath80 .
another dimensionless parameter , which is useful for determining the beginning of cosmic acceleration in dark energy model , is the average deceleration parameter @xmath6 , defined as @xcite , @xmath81 where @xmath82 is the deceleration parameter .
we use eqs .
( [ r ] , [ s ] , [ om3 ] , [ qaveqn ] ) for evaluating the statefinder pair @xmath41 , @xmath8 and @xmath6 for the case of our generalization of thawing dark energy model .
further more , we investigate the growth factor @xmath5 in the context of this proposed generalized thawing eos . for this purpose
we assume the generalized thawing dark energy models proposed here , are decoupled from the cold matter sector .
this would lead to the effect that the galaxy cluster formation is not directly influenced by the existence of dark energy .
but the presence of dark energy alters the hubble expansion rate which affects the growth of inhomogeneities in the cold matter sector . in the linear regime of matter perturbations , the evolution of the inhomogeneities are governed by the relation @xcite @xmath83 where @xmath84 is the matter density contrast with @xmath85 being the matter density .
the growth factor @xmath5 is defined as @xcite , @xmath86 .
( [ ge ] ) can be written in terms of growth factor @xmath5 ( defined in eq .
( [ gf ] ) ) as , @xmath87 the growth equation can be expressed in terms of the redshift @xmath38 by the relation @xmath88 .
the growth factor is well approximated by the ansatz @xcite @xmath89 where @xmath90 is termed as `` growth index '' .
the growth factor @xmath5 is affected by dark energy models via @xmath91 .
for the purpose of putting constraints on the generalized thawing dark energy eos , we use the latest supernova type ia ( sne ia ) data from the union 2.1 compilation @xcite , observational hubble data ( ohd ) @xcite , cosmic microwave background data ( cmb ) from 9 year wmap results @xcite and boss data from sdss - iii @xcite .
there are a total of 607 data points ( 580 data point from sne ia , 25 from ohd , and 1 each from cmb and boss ) .
we make a combined @xmath92 analyses of the data sets comprising of all 607 data points to constrain our model parameters @xmath31 , @xmath3 and @xmath4 , as well as to confront with the model - independent parameters mentioned in section 2 .
this makes our analysis robust .
luminosity distance ( @xmath93 ) measurement of distant supernovae with redshifts @xmath38 is the first observational data to probe the current acceleration of the universe and the dark energy properties as well .
the most recent compilation of the supernova type ia data is given by union 2.1 dataset @xcite .
the data is tabulated in terms of distance modulus @xmath94 with redshift @xmath38 .
the distance modulus can be written as @xmath95 where @xmath96 ( speed of light in vacuum is normalized to unity ) and @xmath97 with @xmath98 given by @xmath99 .
@xmath92 of sne ia data is given by , @xmath100 ^ 2 \,\,,\ ] ] marginalizing over the nuisance parameter @xmath101 , one gets the @xmath92 as , @xmath102 where @xmath103 , @xmath104 and @xmath105 are given by , @xmath106 ^ 2 \nonumber \\ b&= & \sum_{i}\left [ \frac { \mu_{\rm obs}(z_i ) - \mu(z_i ; w_x^0 , \omega_m^0 , \mu_0=0)}{\sigma_i } \right ] \nonumber\\ c&= & \sum_{i}\frac { 1}{\sigma_i^2}\end{aligned}\ ] ] measurements of hubble parameters from differential ages of galaxies provide another way to probe the late time acceleration of the expanding universe .
jimenez @xcite first utilized this idea of measuring hubble parameter through the differential age method .
simon @xcite and later stern @xcite provides the values of the hubble parameter in the redshift range @xmath107 and @xmath108 respectively .
a total of 21 ohd data points are recorded at present in the literature @xcite . with the data release 7 ( dr7 ) from sloan digital sky survey ( sdss )
zhang @xcite provides 4 new values of hubble parameters at different redshifts .
all the 25 ohd data points , used in this work to constrain the model parameters , are listed in table [ tab - hubdata ] .
the @xmath92 function for the analysis of this observational hubble data can be defined as @xmath109 ^ 2 \,\,.\end{aligned}\ ] ] .hubble parameter ( @xmath110 ) versus redshift ( @xmath38 ) data from @xcite . here
@xmath110 and @xmath111 are in @xmath112 . [
cols="^,^,^ " , ] usually from statistical analysis , it is inferred that the models having @xmath113 in the range @xmath114 are strongly supported , models with @xmath115 are moderately supported , and those with @xmath116 are unsupported from perspective of a given data .
however , in cosmology , with the rapid increase of the number of data points @xmath117 ( we remind the reader that we have used combined dataset ) , eq ( [ bic ] ) shows that the @xmath113 value is always going to increase with introduction of new model parameter(s ) @xmath118 .
this does not essentially mean that the models with least number of parameters are always favored by observations , though it may appear to be so .
for example , we know @xmath64 model ( with the least number of parameters ) @xcite fits the sne ia data only in the low redshift region i.e , for @xmath119 , and in this vein , most of the models pay the price just because they have additional parameters , though they , in fact fair well with observations . on this note it should be mentioned that , as demonstrated in @xcite the above information criteria should better be replaced by bayesian evidence calculation , which gives a value after integrating over all probable states , and hence , does not suffer from any such limitations of aic or bic .
hence , nowadays , most of the cosmological models are relying more on bayesian evidence calculation , rather than @xmath120 or @xmath113 calculation .
we hope to address this issue in near future .
we are in the era of precision cosmology .
observational data are improving day by day .
but these are the error bars that the data come with makes the constraints on the models poor .
therefore it is very necessary to reduce the error bars which can improve the constraints on the model parameters further .
we used type ia supernova data , baryon oscillation spectroscopic survey data , observational hubble data and the cosmic microwave shift parameter data to constrain the models parameters . among all these data supernova data influences the analysis
the most i.e. , the constrained parameters space depend on supernova data to a great extent .
supernovae data are not that precise at the present moment as it comes with large error bars
. improving supernova data can probably give us the improved and satisfactory results in future .
we are thankful to supernova cosmology project ( union 2.1 ) http://supernova.lbl.gov/union/ ; wilkinson microwave anisotropy probe http://lambda.gsfc.nasa.gov/product/map/current/ and sloan digital sky survey http://www.sdss.org/ for providing us with the online data .
we also thank s. das and b. k. pal for some useful comments .
abraham r. g. , glazebrook k. , mccarthy p. j. , crampton d. , murowinski r. , jorgensen i. , roth k. , hook i. m. _ et al .
_ , 2004 , astron
. j. * 127 * 2455 [ astro - ph/0402436 ] .
akaike h. , 1974 , ieee trans .
automatic control * 19 * 716 .
ali a. , sami m. , sen a. a. , 2009 , phys .
d * 79 * 123501 [ arxiv:0904.1070 [ astro-ph.co ] ] .
ali a. , gannouji r. , sami m. , 2010 , phys .
d * 82 * 103015 [ arxiv:1008.1588 [ astro-ph.co ] ] .
ali a. , gannouji r. , sami m. , sen a. a. , 2010 , phys .
d * 81 * 104029 [ arxiv:1001.5384 [ astro-ph.co ] ] .
bennett c. l. , larson d. , weiland j. l. , jarosik n. , hinshaw g. , odegard n. , smith k. m. , hill r. s. _ et al . _ , 2012 , arxiv:1212.5225 [ astro-ph.co ] . bento m. d. c. , bertolami o. , sen a. a , 2003 , phys
. rev .
d * 67 * 063003 [ astro - ph/0210468 ] .
bento m. d. c. , bertolami o. , santos n. m. c. , sen a. a , 2006 , j. phys .
ser . * 33 * 197 [ astro - ph/0512076 ] .
beringer j. , 2012 , [ particle data group collaboration ] , phys .
d * 86 * 010001 blake c. , kazin e. , beutler f. , davis t. , parkinson d. , brough s. , colless m. and contreras c. _ et al .
_ , 2011 , mon . not .
soc . * 418 * 1707 [ arxiv:1108.2635 [ astro-ph.co ] ] .
caldwell r. r. , linder e. v. , 2005 , phys .
lett . * 95 * 141301 [ astro - ph/0505494 ] .
chen s. h. , dent j. b. , dutta s. , saridakis e. n. , 2011 , phys .
d * 83 * 023508 chen x. , wang b. , pan n. , gong y. , 2011 , phys .
b * 695 * 30 [ arxiv:1008.3455 [ astro-ph.co ] ] . clemson t. g. , liddle a. r. , 2009 , mon . not .
* 395 * ( 2009 ) 1585 [ arxiv:0811.4676 [ astro - ph ] ] .
davis t. m. , mortsell e. , sollerman j. , becker a. c. , blondin s. , challis p. , clocchiatti a. , filippenko a. v. _ et al .
_ , 2007 , astrophys . j. * 666 * 716 [ astro - ph/0701510 ] .
das s. , ghosh s. , holten j. -w .
van , pal s. , 2011 , int .
j. mod .
d * 20 * 1235 [ arxiv:0906.1044 [ astro-ph.co ] ] . dutta k. , sorbo l. , 2007 , phys . rev .
d * 75 * 063514 [ astro - ph/0612457 ] .
dutta s. , saridakis e. n. , scherrer r. j. , 2009 , phys .
d * 79 * 103005 [ arxiv:0903.3412 [ astro-ph.co ] ] .
dutta s. , scherrer r. j. , 2009 , phys .
b * 676 * 12 [ arxiv:0902.1004 [ astro-ph.co ] ] .
dutta s. , scherrer r. j. , 2010 , phys .
d * 82 * 043526 [ arxiv:1004.3295 [ astro-ph.co ] ] .
gaztanaga e. , cabre a. , hui l. , 2009 , mon . not .
soc . * 399 * 1663 [ arxiv:0807.3551 [ astro - ph ] ] .
gupta g. , saridakis e. n. , sen a. a. , 2009 , phys .
d * 79 * 123013 [ arxiv:0905.2348 [ astro-ph.co ] ] .
harko t. , lobo f. s. n. , 2010 , eur .
j. c * 70 * 373 [ arxiv:1008.4193 [ gr - qc ] ] .
hirano k. , komiya z. , 2011 , int .
j. mod .
d * 20 * 1 [ arxiv:1006.3641 [ astro-ph.co ] ] .
hossain m. w. , sen a. a. , phys .
b * 713 * ( 2012 ) 140 [ arxiv:1201.6192 [ astro-ph.co ] ] .
kessler r. , becker a. , cinabro d. , vanderplas j. , frieman j. a. , marriner j. , davis t. m. , dilday b. _ et al .
_ , 2009 , , astrophys .
j. suppl .
* 185 * 32 [ arxiv:0908.4274 [ astro-ph.co ] ] .
komatsu e. _ et al .
_ [ wmap collaboration ] , 2009 , astrophys .
j. suppl .
* 180 * 330 [ arxiv:0803.0547 [ astro - ph ] ] .
komatsu e._et al . _
[ wmap collaboration ] , 2011 , astrophys .
j. suppl . * 192 * 18 [ arxiv:1001.4538 [ astro-ph.co ] ] .
kowalski m. _ et al .
_ [ supernova cosmology project collaboration ] , 2008 , astrophys .
j. * 686 * 749 [ arxiv:0804.4142 [ astro - ph ] ] .
kujat j. , scherrer r. j. , sen a. a. , 2006 , phys .
d * 74 * 083501 [ astro - ph/0606735 ] .
linder e. v. , 2003 prl * 90 * 091301 [ astro - ph/0208512 ] .
linder e. v. , 2008 , gen .
grav . * 40 * 329 [ arxiv:0704.2064 [ astro - ph ] ] .
li x. z. , sun c. b. , xi p. , 2009
, jcap * 0904 * 015 [ arxiv:0903.4724 [ gr - qc ] ] .
moresco m. , cimatti a. , jimenez r. , pozzetti l. , zamorani g. , bolzonella m. , dunlop j. , lamareille f. _ et al .
_ , 2012 , jcap * 1208 * 006 [ arxiv:1201.3609 [ astro-ph.co ] ] .
nesseris s. , shafieloo a. , 2010 , mon . not .
soc . * 408 * 1879 [ arxiv:1004.0960 [ astro-ph.co ] ] .
riess a.g .
_ et al . _
[ supernova search team collaboration ] , 1998 , astron . j. * 116 * 1009 [ astro - ph/9805201 ] .
riess a. g. , strolger l. g. , casertano s. , ferguson h. c. , mobasher b. , gold b. , challis p. j. , filippenko a. v. _ et al .
_ , 2007 , astrophys .
j. * 659 * 98 [ astro - ph/0611572 ] .
riess a. g. , macri l. , casertano s. , sosey m. , lampeitl h. , ferguson h. c. , filippenko a. v. , jha s. w. _ et al . _ , 2009 , astrophys . j. * 699 * 539 [ arxiv:0905.0695 [ astro-ph.co ] ] .
rosenfeld r. , frieman j. a. , 2007 , phys .
d * 75 * 043513 [ astro - ph/0611241 ] .
sahni v. , saini t. d. , starobinsky a. a. , alam u. , 2003 , jetp lett .
* 77 * ( 2003 ) 201 [ pisma zh .
* 77 * 249 ] [ astro - ph/0201498 ] .
sahni v. , shafieloo a. , starobinsky a. a. , 2008 , prd * 78 * 103502 [ arxiv:0807.3548 [ astro - ph ] ] .
scherrer r. j. , sen a. a. , 2008 , phys .
d * 77 * ( 2008 ) 083515 [ arxiv:0712.3450 [ astro - ph ] ] .
scherrer r. j. , sen a. a. , 2008 , prd * 78 * 067303 [ arxiv:0808.1880 [ astro - ph ] ] .
schwarz g. , 1978,ann .
* 6 * 461 .
sen s. , sen a. a. , sami m. , 2010 , phys .
b * 686 * ( 2010 ) 1 [ arxiv:0907.2814 [ astro-ph.co ] ] .
simon j. , verde l. , jimenez r. , 2005 , phys .
d * 71 * 123001 [ astro - ph/0412269 ] .
stern d. , jimenez r. , verde l. , kamionkowski m. , stanford s. a. , 2010 , jcap * 1002 * 008 [ arxiv:0907.3149 [ astro-ph.co ] ] .
suzuki n. , rubin d. , lidman c. , aldering g. , amanullah r. , barbary k. , barrientos l. f. , botyanszki j. , 2012 , apj * 746 * 85 [ arxiv:1105.3470 [ astro-ph.co ] ] .
tsujikawa s. , arxiv:1004.1493 [ astro-ph.co ] . | we propose a two parameter generalization for the dark energy equation of state ( eos ) @xmath0 for thawing dark energy models which includes pngb , cpl and algebraic thawing models as limiting cases and confront our model with the latest observational data namely sne ia , ohd , cmb , boss data .
our analysis reveals that the phantom type of thawing dark energy is favoured upto @xmath1 confidence level .
these results also show that thawing dark energy eos is not unique from observational point of view . though different thawing dark energy models are not distinguishable from each other from best - fit values ( upto @xmath2 c.l.s ) of matter density parameter ( @xmath3 ) and hubble parameter ( @xmath4 ) at present epoch , best - fit plots of linear growth of matter perturbation ( @xmath5 ) and average deceleration parameter ( @xmath6 )
; the difference indeed reflects in best - fit variations of thawing dark energy eos , model - independent geometrical diagnostics like the statefinder pair @xmath7 and @xmath8 parameter .
we are thus led to the conclusion that unlike the standard observables ( @xmath3 , @xmath4 , @xmath5 , @xmath6 ) , the model - independent parameters ( @xmath9 ) and the variations of eos ( in terms of @xmath10 plots ) serve as model discriminators for thawing dark energy models .
cosmology : dark energy , thawing dark energy models , cosmological parameters , sne ia , ohd . |
supernova 1987a provides a uniquely detailed case for the study of the late evolution of massive stars with many surprising features .
it is one of the very few sne in which the progenitor was studied before the explosion .
the internal composition of the progenitor structure and that of the ejecta have been the subject of intense scrutiny . the sn environment is dominated by a mass - loss nebula ( or `` circumstellar envelope '' - cse ) with many different features capable of betraying various phases and events during the life of the progenitor . the cse can in turn be probed through its response to the ionizing flash from the sn explosion , followed by light echoes in the optical , followed by the collision with the cse of the ejecta from the explosion .
we are studying the spatial structure of this nebula , as well as its velocity structure seen in narrow emission lines , in order to construct as complete as possible a history of the mass loss phases of the star . with both the velocity structure and the full three - dimensional spatial structure that light echoes reveal
, one can interpret observations in terms of the kinematic ages and therefore the evolutionary phases involved in mass loss .
we are presenting these observations in a series of works detailing the spatial structure of the inner cse ( paper i : crotts , kunkel & heathcote 1995 ) , its velocity structure ( this paper ) and the spatial structure of the outer cse ( in preparation ) . while the velocity and distance from the sn of a feature yield a timescale characteristic of how long ago that material was ejected from the surface of the progenitor , these timescales can also be inferred more indirectly from compositional variations within the cse : structures arising at different times result from different layers of the star , and hence should have different compositions . detecting this level of differentiation across the cse
requires the spatial resolution of @xmath1 fos spectra to compare the rings ( panagia et al .
1996 ) , indicating much less nitrogen in the outer rings than the inner ring , suggesting that the outer rings were ejected some @xmath7 yr earlier than the inner rings whose kinematic age is @xmath8 yr ( crotts & heathcote 1991 ) .
we shall here test this compositional age measure against the kinematic age of the outer ring , and attempt to explain possible discrepancies .
this is a difficult observational problem best attacked through the combination of ground - based and @xmath1 measurements . over the entire central region where they are expected to show the largest variation in velocity ,
the ors are found within only @xmath50.3 arcsec of the er , yet are many times fainter than the er . this situation cries out for the high angular resolution of @xmath1 , but also requires high spectral resolution ( preferably at least @xmath9 ) which is available on @xmath1 ( ghrs and stis / mama ) only in the uv for transitions in which the ors are too faint .
such high spectral resolution is readily available in the optical from the ground , but seeing limits its usefulness to the portions of the ors separated from the er by more than about 1 arcsec . as we shall see
, the @xmath10 , stis g750 m setup for [ n ii]@xmath26583 can detect velocity offsets of some er and or features if sufficient care is taken . fortunately , for different segments of the ors , one or the other observational alternative offers some way of measuring their velocity , and for some these results can be intercompared . to look at signals at different spatial scales and spectral resolutions , we used a total of four different kinds of data : from the ground ( ctio 4m / echelle ) , single - slit spectra to study the er and or features 1 - 2 arcsec from the er and multi - slit spectra to study fainter features at larger radii , and on @xmath1 , stis spectra of the er and or compared to wfpc2 images , the latter used to fix the angular positions of different features in the stis spectrograph slit relative to one another . to appreciate the relevance of each data set to the problem , it is helpful to see ( figure 1 ) the placement of the stis and ctio / echelle slits relative to the rings and adjacent stars .
( `` nor '' and `` sor '' refer to the northern and southern or , respectively . ) the data used here , already described in sonneborn et al .
( 1998 ) ( their figure 6 ) , were taken on ut 1997 apr 26 , day 3715 after core collapse .
it was a 643s stis integration in the g750 m mode ( 6295 - 6867 , resolution @xmath11 1.02 , or 46.5 km s@xmath4 at [ n ii]@xmath12 ) .
the 2 arcsec - wide longslit centered on the sn was oriented at pa@xmath13 , nearly parallel to the er s projected major axis , and hence included the entire er as well as much of the ors , including the segments which might be expected to be seen expanding most rapidly from the sn along the line of sight . because of the wide spectrograph slit used the position of a feature along the dispersion axis depends on both its velocity and its position perpendicular to the slit .
thus in order to measure velocities within the rings , we must compare their dispersed images seen in the stis spectra to their undispersed locus measured in wfpc2 images .
for this purpose we have used an image through the narrow band f658n filter , which primarily admits flux in [ n ii]@xmath26583 obtained on day 3478 ( go program 6437 , kirshner et al . ) which had a total exposure time of 5400s .
the possible motion of the ring emission locus over this 237 day interval between the stis and wfpc2 observations should be negligible , less than 0.1 pixel ( c.f .
plait et al . 1995 ) .
these data were taken with the same setup used in crotts & heathcote ( 1991 ) , using the f/8 echelle spectrograph on the ctio 4 m , with the 31.6 l mm@xmath4 echelle grating and kpgl3 cross - disperser ( 527 l mm@xmath4 , blazed at 5500 in first order ) , which together deliver a resolving power @xmath14 ( corresponding to 7.5 km s@xmath4 ) with the 1 arcsec slit used .
we used the red , long - focus camera and red collimator , covering 5530 - 7050 .
we concentrate primarily on the narrow emission line at [ n ii]@xmath12 . for the slit decker used
, each order is 19 arcsec wide .
data such as these were taken on numerous occasions between 1989 and 1997 , and we chose the best of these , in terms of good seeing and signal - to - noise ratio ( @xmath15 ) of the or features .
those chosen were from ut 1996 dec 1 ( day 3569 after core collapse , shortly before the @xmath1 stis observation ) , and have seeing of about 0.8 arcsec fwhm .
these data are described in xu & crotts ( 1999 ) .
they employed the 79 l mm@xmath4 echelle grating on the ctio 4m / echelle , along with an order separation filter that selected h@xmath16 and [ n ii ] lines , projected through a set of five slits separated by 12.5 arcsec perpendicular to their length .
the free velocity range of 120 km s@xmath4 between successive slits was sufficient to prevent overlap of detectable lmc emission components .
the data considered here were part of a 6@xmath176 arcmin raster of north - south slits roughly centered on the sn ( with one actually passing through the sn ) , with additional single , long slits oriented east - west in preliminary exposures to check that there were no observable components lying beyond the 120 km s@xmath4 no - overlap range .
these spectra have a velocity resolution of 10 km s@xmath4 .
exposures were obtained for 1 hour in each position , and up to five velocity components were disentangled from each spectrum , analyzed at each point on a grid with 13 arcsec sampling north - south and 12.5 arcsec east - west .
these data were taken several years before the others , on days 2160 - 2162 .
these four datasets are combined to map the velocity field of sn 1987a s cse .
the ground - based data provides good measurements of the velocities of the ors north and south of the sn , which are missed by stis , but because of limited spatial resolution are less useful for the portions of the ors closer to the er , which stis covers well .
the emission from the portion of the sky containing the outer cse , however , has too low an emission measure for @xmath1 to study easily , but is seen more easily by a telescope with a faster f - ratio . we consider each of these measurements in turn .
the g750 m stis spectra of sonneborn et al .
( 1998 ) , their figure 6 , show excellent images of the er and a portion of the nor and sor in [ n ii]@xmath2@xmath26548 , 6583 and h@xmath16 ( as well as several other transitions ) taken on day 3715 , which can be compared to nearly contemporaneous images from wfpc2 . the [ n ii]@xmath26583 line has by far the greatest @xmath15 ( with the h@xmath16 signal being smeared by thermal broadening . )
the difference between the positions of features in the dispersed stis data compared to the direct wfpc2 image allows us to measure the relative velocities between different parts of the cse .
unfortunately , the region of the nor seen within the er is largely obscured by the sn ejecta and in part by the blue - shifted smear of the `` hot spot '' ( sonneborn et al .
1998 , michael et al . 1998 ) on the er overlapping the nor .
nonetheless , the entire er and a large portion of the northern sor and southern nor can be compared between the stis and wfpc2 data in this way .
figure 2 shows the results of the best bilinear fit of the g750 m stis locus of the er to the position of er from the day 3478 wfpc2 exposure with the f658n filter .
the loci in each image are defined as the set of points of maximum brightness defined along radial crosscuts through the sn .
( note that the scale in one dimension is expanded to make the er appear roughly circular . )
this shows a close correspondence between the two loci with the stis points following the wfpc2 track , matching most features in great detail , to within an r.m.s .
difference of 0.12 stis pixels .
the anamorphic magnification of this stis mode at this wavelength ( bowers et al .
1998 ) is 0.92786 ( compressed in the dispersion direction ) . in comparison ,
the stis - to - wfpc2 fit requires a 0.91058 compression along the dispersion , corresponding to a difference in velocity of @xmath18 stis pixels , or @xmath19 between the northern and southern side of the er , in the sense that the southern side is redshifted .
this implies a velocity shear of 14.3 km s@xmath4 across the slit , which corresponds ( since the slit is well - aligned with the er ) to a radial expansion velocity i.e. outward from the sn , of @xmath20 km s@xmath4 .
this is a new , independent measurement of the er expansion velocity , but it is in good agreement with some previous values ( 10.3 km s@xmath4 : crotts & heathcote 1991 , cumming 1994 ) and is not far from others ( 11 km s@xmath4 : panagia et al . 1996 , 8 km s@xmath4 : meaburn et al .
1995 ) .
figure 3a shows the position of the stis and wfpc2 or signals after alignment of the er signals as in figure 2 .
the residual offset of the sor and especially the nor is a clear sign of the velocity difference between the ors and the er centroid .
intercalibrating the data in this way eliminates any possible sources of systematic error due to uncertainties in the standard stis and wfpc2 platescales , at the cost of only slightly increasing the random error due to the uncertainty of aligning the two er signals .
the useful sample of the ors excludes those portions where they approach the er , and also where they run nearly parallel to the dispersion axis .
additionally , the sor signal is corrupted over a large extent by the continuum from star 3 .
the uncontaminated samples are shown in figure 3a , and with the sor signal magnified in figure 3b .
this remaining sample permits one useful measurement along the sor at a point 1.36 arcsec east of the sn ( labeled `` a '' in figure 1 ) , and three measurements of comparable accuracy along the nor , 1.04 , 1.40 and 1.68 arcsec west ( points `` b '' through `` d '' in figure 1 ) .
these yield velocity values of 21.5 km s@xmath4 ( blueshifted ) and 22.3 , 24.6 and 21.1 km s@xmath4 ( redshifted ) , respectively , relative to the 289 km s@xmath4 heliocentric redshift of sn 1987a , with @xmath21 errors of about 2 km s@xmath4 , except for the last point , with an error of about 3 km s@xmath4 .
as discussed by crotts & heathcote ( 1991 ) , the velocity field of the er , best traced by [ n ii]@xmath26583 , has a centroid at 289 km s@xmath4 , with little velocity shear across the er at pa@xmath22 , almost parallel to the er major axis , but with strong velocity shears on either side , at pa@xmath23 and @xmath24 .
there are additional , fainter emission patches at larger radius or velocity offsets from the centroid .
occasionally such faint sources correspond to transient spots ( crotts , kunkel & mccarthy 1989 , hanuschik 1990 , crotts & heathcote 1991 , cumming & meikle 1993 , cumming 1994 ) , but at the time of these observations no such spots were revealed in optical imaging .
these fainter , outrigger sources observed here are almost certainly due to the ors .
figure 4 shows the portion of the echelle order centered on the [ n ii]@xmath25 line , showing the broad emission lines from the sn and stellar continua extending horizontally , and the lmc interstellar [ n ii ] emission extending vertically .
the bright patch in the center is emission from the cse .
the underlying ism is fit linearly in the spatial dimension , interpolated across the cse , and subtracted off .
the spectral continuum and broad emission is fit quadratically , and likewise interpolated and subtracted .
this subtraction produces the residual distribution of flux shown in the contour plot of the echellogram in figure 5a ( with a vertical scale of 0.26 arcsec and horizontal scale of 3.50 km s@xmath4 per pixel , at [ n ii]@xmath26583 ) .
there are notable protrusions from the nearly elliptical flux contours , particularly to the south at roughly the centroid velocity and to the north slightly to the red of centroid .
furthermore , the flux contours at low levels become more rectangular , as if flux might be appearing at the corners . we know the flux distribution in [ n ii]@xmath26583 of both the er and ors from the hst images , and have good constraints on the global velocity structure of the er . furthermore , comparing the width of the er feature in the [ n ii ] and h@xmath16 echellograms , we see a significant difference most likely due to thermal broadening . accounting for the different atomic masses producing the lines we find the temperature of the gas at day 3569 to be roughly @xmath26k ( to within roughly 20% ) .
( this compares to other temperature measurements at late times : @xmath27k on day 1278 , from [ o ii ] and [ n ii ] recombination [ wang 1991 , sonneborn et al . 1997 ] ; 55000k on @xmath5day 300 from [ o iii ] [ wampler & richichi 1989 ] .
the lower temperature we measure on day 3569 may be due to the cooling of the gas , or due to h @xmath16 and [ n ii ] emission arising from different gas .
recombination model calculations predict @xmath28k throughout the er by day 2000 [ lundqvist & fransson 1996 ] . ) we can also measure the spatial and spectral line spread functions of the spectrograph from the width of the stellar continuum and night sky lines in the echellogram .
given these parameters we can model the image of the cse produced by the spectrograph on the detector and construct the synthetic echellogram shown in figure 5b . at high flux levels , it appears very similar to the actual data , but at low flux levels differs by the absence of the extensions mentioned above .
the width of the spatial and spectral line spread functions , temperature , and placement of the slit are allowed to vary over a grid of values spanning approximately @xmath2910% in each of these parameters , in order to produce a family of model echellograms , and the minimum @xmath30 ( in terms of the square - root of counts per pixel ) , within the region covered by non - zero flux from the model .
the resulting best fit model is shown in figure 5b , and the difference between the data and fit shown in figure 5c .
this corresponds to a best fit of 9.5 km s@xmath4 in radial expansion velocity of the ring away from the sn , with a probable error of about 5 - 10% ( which is difficult to estimate given systematic errors in other parameters ) .
note that the number of contours in the residual features is several times fewer than in the original , figure 5a , and that the contour levels are five times finer in figure 5c , so that the extreme residual features are only about 7% of the original peak flux , and cover regions on the scale of single pixels , so that residual features within the region of the model are at most about 1% of the total .
this compares to expected poisson errors from photon counting statistics at the 0.3% level , so a much better fit would be difficult to achieve . outside the region modeled in 3.2.1
, several significant residuals appear : to the north and south at roughly the centroid velocity , and to higher and lower velocities at roughly the centroid position along the slit .
these four residuals are centered at ( @xmath3121.7 km s@xmath4 , 0.39 arcsec north ) , ( @xmath322.8 km s@xmath4 , 2.18 arcsec south ) , ( @xmath324.2 km s@xmath4 , 1.98 arcsec north ) , and ( @xmath3320.3 km s@xmath4 , 0.05 arcsec south ) , and have flux amounting to 2.8% , 2.8% , 2.0% and 1.1% , of the total , respectively .
note that these radii along pa@xmath23 correspond closely to those measured from the wfpc2 image for the northern sor , southern sor , northern nor , and southern nor respectively : 0.71 arcsec north , 2.08 arcsec south , 1.96 arcsec north , and 0.08 arcsec south , with the first and fourth measurements being somewhat uncertain due to confusion with the er and sn , respectively .
these positions agree between the echelle and @xmath1 datasets to within about 0.2 arcsec r.m.s .
, less than an echelle pixel ( 0.26 arcsec ) .
we can use the relative strength of the er and ors in the wfpc2 images to predict relatively how much light is entering the echelle spectrograph slit . from the stis g750 m data , we know that the or flux in the vicinity of the er ( but beyond @xmath50.3 arcsec to avoid being overwhelmed by the er ) are 2.5 - 3.5 times stronger in the [ n ii]@xmath26583 line than in h@xmath16 . additionally , the throughput at the sn redshift in the wfpc2 f658n filter is 3.8 times greater at [ n ii]@xmath26583 than at h@xmath16 , hence the or signal in the wfpc2 f658n image is 92% due to [ n ii ] .
this implies that even if the h@xmath16 flux were to vary by 100% of its strength relative to the [ n ii]@xmath26583 flux at different points around the ors , the flux of the or [ n ii]@xmath26583 flux can still be predicted to within 9% .
the flux values inferred from wfpc2 for the south sor is 2.0% and for the north nor is 1.8% of the er signal within the slit , close to the values seen in the echellogram .
the brightness of the north sor and south nor is difficult to determine in the wfpc2 image , again because of the contributions by the er and sn . given that most of the signal in the residual echelle features corresponds to the actual or fluxes , it would appear that errors in the flux could only decenter the signal by at most about 1 pixel in velocity , or about 4 km s@xmath4 .
we will take this as the probable error on the echelle or measurements . to interpret these or velocity measurements , we compute a simple model for or loci in a homologous velocity flow , without imposing any constraint that the or and er velocities should compare , and then require that the projected shape of each or be consistent with its observed shape and location . for the purpose of computing homologous velocities
, we make the approximation that all three rings are coaxial . for the nor , which appears nearly elliptical , we adopt an inclination angle set by assuming that the ring is really circular and is only seen as an ellipse in projection .
this inclination angle is [email protected] ( burrows et al .
1995 ) with an error that we infer to be about 1@xmath34 .
in contrast , the er has an inclination measured variously as 43@xmath35 ( jakobsen et al .
1991 ) , [email protected] ( burrows et al .
1995 ) , or 44@xmath36 ( plait et al . 1995 ) .
the sor is not truly elliptical , but the best fit ellipse corresponds to an inclination of 31@xmath34 ( burrows et al .
for the sor we adopt two alternative geometries : a ) that it is inclined by 31@xmath34 , and b ) that it is inclined by 45@xmath34 , parallel to the er , but distorted from a circle .
perhaps neither alternative is correct , but these at least span a reasonable range of possibilities . under these assumptions ,
the nor is expanding outwards from the sn at 26.1 km s@xmath4 , while the sor expands at either a ) 25.5 km s@xmath4 , or b ) 26.3 km s@xmath4 ( along the minor axis ) , according to a best fit to the observed velocities from stis+wfpc2 and the ctio 4m / echelle .
these fits are shown in figures 6a , b and c ( for the nor and sor models a and b , respectively ) .
all three fits are reasonable , with reduced @xmath30 near unity .
sor model b ( 31@xmath34 inclination ) fits slightly better than model a. the 1@xmath37 errors on all of these derived expansion velocities are about 2 km s@xmath4 .
( note that the above results appear consistent with those quoted by cumming & lundqvist [ 1994 ] . )
these velocities should be compared the radial distance of the ors from the sn in order to derive a kinematic timescale . for the sake of discussion
, we adopt a semimajor axis length of 1.7 arcsec for both ors .
( we measure a slightly smaller value than the semimajor axis lengths found by burrows et al . [
1995 ] : 1.77 and 1.84 arcsec . ) for a distance to sn 1987a of 50 kpc , these models correspond to or distances from the sn of 0.58 pc for the nor , 0.52 pc for the sor ( model a ) , or 0.56 pc for the sor ( model b , along the minor axis ) .
compared to the corresponding expansion velocities , these correspond to kinematic timescales of 21700 , 19900 and 20800 yr , respectively , indicating the time in which the ors would have coasted at their current velocities to their current positions .
similarly , the er , with a semimajor axis of 0.86 arcsec ( plait et al . 1995 ) or 0.21 pc , would require 19500 years expanding at 10.5 km s@xmath4 .
clearly , these kinematic ages are consistent within the errors ( of @xmath38 11% ) .
the longslit survey by xu & crotts ( 1999 ) of the 6 arcmin square around sn 1987a was intended to observe structure on scales considerably larger than the spatial sampling of 13 arcsec , but nonetheless provides some information on the cse of the sn . as seen in light echoes ( crotts & kunkel 1991 , crotts 1999 ) , there is an echo feature apparently centered on the sn extending some 5 pc towards earth , corresponding to an angular scale of 20 arcsec .
a search for features on this scale in the longslit data set reveals two features ( figure 7 ) .
one , centered near 279 km s@xmath4 ( 10 km s@xmath4 blueshift with respect to the sn ) is found between 25 arcsec east and 13 arcsec west , and between 18 arcsec north and 4 arcsec south of the sn .
a second feature at 301 km s@xmath4 ( 12 km s@xmath4 redshift ) extends between 13 and 38 arcsec west , and 4 arcsec north and 15 arcsec south of the sn .
these appear to form a possibly bipolar , or double - lobed , nebula , with the near side to the northeast , assuming material is being ejected .
the axis of symmetry of this structure would lie along @xmath39 .
these features are not perfectly coherent in velocity , however . while the @xmath15 detection level of individual segments of these features is low ,
there appear to be significant gradients in velocity , so that the differences from the sn velocity centroid can extend up to 15 km s@xmath4 .
it is evident from figure 7 , however , that patches of roughly this spatial scale and velocity coherence exist in locations far away from the sn , such that there is a significant probability that one or both features are chance superpositions in projection against the sn .
one possible method of investigating the association of these two features with the sn is to determine if their elemental abundances are more similar to the anomalous values found in the rings ( e.g. lundqvist & fransson 1996 , sonneborn et al .
1998 ) , or those typical in the lmc .
we are currently pursuing this test of the origin of this gas .
why do the kinematic and compositional timescales ( panagia et al . 1996 ) disagree ?
of course , the simple kinematic timescale @xmath40 does not account for the acceleration of the rsg wind by the lower density ( and not directly observed ) bsg wind , so that at a given radius the winds interface i.e. the freshly shocked rsg material , is actually expanding faster than the rsg wind just exterior to the shock . since
this exterior wind has presumably coasted freely since leaving the progenitor , @xmath41 for this material is close to the actual time @xmath42 since the material left the progenitor . however
, this may not be so for the material actually observed . in comparing @xmath41 for different parts of the cse we wish to know whether the ratio of @xmath43 is constant throughout .
recent models of interacting winds within a wind - compressed disk outflow ( collins et al .
1999 ) can lead to velocity fields over the winds interface which produce a nearly constant @xmath43 throughout ( collins et al .
1999 , fig . 2 and a. frank , private communication ) , while simultaneously producing a shape for the cse consistent with that observed .
other models ( blondin & lundqvist 1993 , martin & arnett 1995 ) have a more complicated velocity flow , in which @xmath43 is nonuniform even on scales as small as @xmath44 cm .
thus while a constant @xmath43 is consistent with the known morphology of the cse , it is by no means assured .
in contrast , the compositional ratios computed by panagia et al .
( 1996 ) are highly time - dependent , since the system is far from equilibrium while still recovering from the pulse of euv photons emitted during the first hours of the sn explosion .
while the elemental ratios can be recovered from a time - dependent analysis of line ratios , the particular line ratios arising from the same gas can vary by large amounts over the course of time ( lundqvist and fransson 1991 ) .
the panagia et al .
study does not sufficiently explore these time - dependent effects .
it is unclear that a compositional difference between the er and ors can be disentangled from these effects given the single epoch of data and relatively simple models considered , as a unique solution to the elemental abundances requires knowledge of the temperature and ionization shortly after it was heated and ionized by the sn flash ( p. lundqvist , private communication ) .
if the extended two - lobed nebula apparently in figure 7 is indeed associated with sn 1987a , its expansion velocity is 10 - 15 km s@xmath4 slower than the or velocities , but roughly equal to the er expansion velocity . if this nebula is associated with the sn , what is its three - dimensional configuration ?
this is difficult to know certainly given the absence of light echoes from the outer regions of this double - lobed object .
this absence might simply be due to insufficient light travel time to cover the whole structure , which would probably require 30 yr or more . in two - dimensional projection ,
the correspondence of this structure to gas at smaller radii is not obvious .
its symmetry axis ( @xmath39 ) is far from the axis of the ring system ( @xmath45 ) or its equatorial plane .
since this nebula does extend significantly to the north and south of the sn , however , if it does surround the sn , it likely envelops the ors , too . in this case , the ors are expanding at @xmath525 km s@xmath4 into a medium expanding at only 10 - 15 km s@xmath4 , and are therefore being accelerated by the bsg wind into the slower rsg wind medium .
it is not clear that the er has been accelerated by a proportionally large fraction of its own velocity , implying that it might be younger than the ors .
this runs in the same sense as the inequality between er and or ages inferred by panagia et al .
regardless of the nature of the extended nebulosity , it is evident that the inner portions of the cse are expanding at different velocities .
this is different from at least the initial conditions assumed in all simulations of the sn s cse , with the exception of collins et al .
other simulations tend to produce an asymmetric boundary between the bsg and rsg winds with an isotropy in matter density , not velocity .
it is interesting that both can be probed using not only narrow - line emission , but also light echoes , which trace dust density .
furthermore , if the cse terminates in a contact discontinuity against an interstellar bubble of constant pressure ( chevalier & emmering 1989 ) , the shape of this contact discontinuity is a measure of the momentum flux as a function of angle around the sn , since pressure must be matched along the discontinuity s surface .
thus in directions from the sn of higher outward ram pressure @xmath46 at a given radius , the wind will be forced to a halt at a larger radius .
this will be investigated in paper iii .
regardless of the shape of the cse , its size scale combined with the outflow velocity , especially where gas is simply coasting and not being accelerated along a wind interface , provides an estimate of the minimum age of the cse .
it is only a lower limit because mass has accumulated in the contact discontinuity .
a cse radius of 5 pc and a velocity scale of @xmath515 km s@xmath4 implies a minimum cse lifetime of @xmath5350000 yr .
this compares to model predictions to the duration of the rsg phase of sn 1987a of about 190000 yr ( woosley , pinto & ensman 1988 ) minus a terminal bsg phase beginning 10000 - 20000 yr before explosion ( woosley et al .
1988 , saio , kato & nomoto 1988 ) .
while the terminal bsg phase lifetime is in agreement with observations , the rsg phase appears too short in the models .
this would imply that the sn progenitor depleted its core helium supply earlier than expected .
further work is needed to explain the production of these three rings .
lloyd , obrien and kahn ( 1995 ) interpret the rings in terms of interacting winds from the progenitor in the presence of a companion star .
also , chevalier and dwarkadas ( 1995 ) propose that the rings arose from the collision of the bsg and rsg winds , with the further action of ionization by the progenitor , but no complete interacting - wind model has yet to be formulated .
models ( meyer 1997 , soker 1999 ) that naturally explain the formation of a three - ring system fail in other important respects , either in not predicting ( meyer 1997 ) the cavity interior to the three rings , as evidenced by the rapid fall and only gradual rise of the radio signal from ejecta / cse interaction ( turtle et al .
1987 , staveley - smith 1993 ) , or in failing to incorporate ( soker 1999 ) the diffuse nebulosity seen in the outer cse ( crotts & kunkel 1991 , crotts 1998 ) and between the rings ( crotts et al . 1995 , see also lundqvist & fransson 1996 ) .
the eventual model , in order to be successful , must now account for the motions seen in the rings and beyond .
the authors would like to thank peter lundqvist and the anonymous referee for helpful comments .
this research was supported by grant ast 90 - 22586 from the nsf , stsci grant ar-05789.01 - 94a and nasa ltsa grant nag5 - 3502 to a.p.s.c .
figure 1 : a schematic layout of the three rings , traced from the day 3478 wfpc2 f658n image , showing the inner , equatorial ring ( er ) , southern outer ring ( sor ) and northern outer ring ( nor ) . also shown
are the positions of stars 2 and 3 , and the sn ( central dot ) .
overlaying these are the locations of the three ctio 4m / echelle spectrograph slits ( 1.0 arcsec wide at pa 20 , 60 and 130 ) , and the edges of the day 3715 stis g750
m slit ( 2.0 arcsec wide ) . the labels `` a '' through `` d '' rest immediately above the stis samples referred to in the text and used to measure or velocities .
the scale in terms of arcseconds and light - years ( for @xmath47 kpc ) are shown .
figure 2 : the loci of the day 3715 stis g750 m data of the er ( squares ) fit to overlay the day 3478 wfpc2 f658n image ( solid curve ) .
this bilinear fit is compared to the anamorphic distortion of the stis g750 m setup to yield the velocity shear across the er , which corresponds to radial expansion away from the sn of @xmath20 km s@xmath4 . note that the scales on the two axes are unequal , to account for the ring s inclination .
figure 3a shows the position of the stis and wfpc2 or signals in relation to the er loci once the er signals are aligned as in figure 2 .
the offset of the sor and especially the nor even at the dispersion position corresponding to the er centroid ( dashed line ) is a clear sign of the velocity offset of the ors versus the er centroid .
the nor can be split into three roughly equal segments , 1.04 , 1.40 and 1.68 arcsec west of the er centroid , with velocity offsets between the two data sets corresponding to 22.3 , 24.6 and 21.1 km s@xmath4 ( redshifted ) .
figure 3b shows the same data as in figure 3a , but with the sor signal magnified .
a single measurement of sor velocity is possible , at a point 1.36 arcsec east of the sn , yielding 21.5 km s@xmath4 ( blueshifted ) .
figure 4 shows the portion of the echelle order centered on the [ n ii]@xmath25 line , showing the broad emission lines from the sn and stellar continua extending horizontally , and the lmc interstellar emission in [ n ii ] extending vertically .
the bright patch in the center is emission from the cse .
this data is for pa@xmath23 , with north to the top .
wavelength ( vacuum , heliocentric ) is shown along the abscissa ( modified slightly by structure in the cse perpendicular to the slit ) .
these data have 8 km s@xmath4 velocity resolution .
figure 5a shows the same data as in figure 4 , once the underlying ism has been fit linearly in the spatial dimension , interpolated across the cse , and subtracted off .
the spectral continuum and broad emission have been fit quadratically , and likewise interpolated and subtracted .
echellogram pixels are denoted by the small tickmarks on both sides of both axes .
the vertical pixel scale is 0.26 arcsec and the horizontal scale is 3.50 km s@xmath4 at [ n ii]@xmath26583 ) .
there are notable protrusions from the nearly elliptical flux contours , particularly to the south at roughly the centroid velocity and to the north slightly to the red of centroid , and flux contours at low levels become more rectangular , as if flux might be appearing at the corners .
figure 5b shows a fit to brighter flux levels by adjusting for the temperature of the gas , velocity gradient , seeing , spectral resolution , and placement of the slit .
it reproduces most features seen in the actual data ( fig .
5b ) except for the extensions at large separations and velocity differences .
figure 5c shows the resulting residual between the actual data and the best fit .
note that the number of contours in the residual features is several times fewer than in the original , figure 5a , and that the contour levels are five times finer in figure 5c , so that the extreme residual features are only about 7% of the original peak flux , and cover regions on the scale of single pixels , so that residual features within the region of the model are at most about 1% of the total .
this compares to count poisson errors expected at the 0.3% level , so is close to the smallest possible residual .
figure 6a shows the best fit to the measured nor velocities from both stis+wfpc2 and ctio 4m / echelle data , assuming that the ring is intrinsically round . in figures 6a - c , the velocity flow is assumed to be homologous . under these assumptions ,
the nor is expanding radially outwards from the sn at 26.1 km s@xmath4 .
figure 6b shows the best fit to the measured sor velocities from both stis+wfpc2 and ctio 4m / echelle data , assuming that the ring is inclined by @xmath48 , the angle required to force the semimajor and deprojected semiminor axis to the same lengths . under these assumptions ,
the sor is expanding radially at 25.5 km s@xmath4 .
figure 6c shows the best fit to the measured sor velocities from both stis+wfpc2 and ctio 4m / echelle data , assuming that the ring is inclined by same value as the er , at @xmath49 , but distorted from a circle . under these assumptions , the sor is expanding radially at 26.3 km s@xmath4 ( along the observed minor axis ) .
figure 7 shows a 70 arcsec by 40 arcsec array of longslit echelle spectrograms from a field centered on sn 1987a ( the bright spot in the center ) , in the [ n ii]@xmath26583 line .
the data were collected five slits at a time as part of a larger ( 6 arcmin @xmath17 6 arcmin ) survey , and four such exposures are shown here , separated by bold black lines .
the free velocity range between successive slits is 120 km s@xmath4 , with 10 km s@xmath4 velocity resolution , and the slits are separated by 12.5 arcsec on the sky .
north is up , east is to the left ( from slit to slit ) , and higher velocities ( between slits ) is to the right , as indicated by the larger set of arrows .
two features , centered @xmath3110 km s@xmath4 and @xmath3312 km s@xmath4 with respect to the sn , are found between two slits east and one slit west , and between one to three slits west of the sn , respectively .
the first feature is found between the vertical pairs of white arrows , and the second between pairs of black arrows .
note that other features of similar size are also seen in different locations in the field , casting some doubt on the association of the two features with the sn . | we present several different measurements of the velocities of structures within the circumstellar envelope of sn 1987a , including the inner , equatorial ring ( er ) , the outer rings ( ors ) , and the diffuse nebulosity at radii @xmath0 pc , based on ctio 4 m and @xmath1 data .
a comparison of stis and wfpc2 [ n ii]@xmath26583 loci for the rings show that the er is expanding in radius at ( 10.5@xmath3 ) km s@xmath4 ( with the northern or expanding along the line of sight at @xmath526 km s@xmath4 , and for the southern or ,
@xmath523 km s@xmath4 ) .
the best fit to ctio 4m / echelle spectra of the [ n ii]@xmath26583 line show the ors expanding at @xmath523 km s@xmath4 along the line of sight .
accounting for inclination , the best fit to all data for the expansion in radius of both ors is 26 km s@xmath4 .
the ratio of the er to the or velocity is nearly equal to the ratio of the er to the or radius , so the rings are roughly homologous , all having been created @xmath520,000 yr before the sn explosion .
this makes the previously reported , large compositional differences between the er and ors difficult to understand .
additionally , a grid of longslit 4m / echelle spectra centered on the sn shows two velocity components over a region roughly coextensive with the outer circumstellar envelope extending @xmath55 pc ( 20 arcsec ) from the sn .
one component is blueshifted @xmath510 km s@xmath4 relative to the systemic velocity of the sn , while the other is redshifted by a similar amount .
these features may represent a bipolar flow expanding from the sn , in which the ors are propelled 10 - 15 km s@xmath4 faster than that of the surrounding envelope into which they propagate .
the kinetic timescale for the entire nebula is @xmath6350,000 yr ( and probably more since material may be accumulating in an outer contact discontinuity ) . the kinematics of these different structures constrain possible models for the evolution of the progenitor and its formation of a mass loss nebula . |
only about twenty years ago it was put forward by bonioff @xcite and feynmann @xcite the idea of building up of computers based on the principles of quantum mechanics which was developed concretely by david deutsch @xcite in 1985 . since then it has been discussed in the literature the possible implications of quantum entanglement schrodinger and non - locality in the broad new and promising filed of research embracing quantum information and quantum computation nielsen , mathqc , pqi .
entanglement and non - locality is fascinating , and although at the same time it is conterintuitive , it can have revolutionizing impact on our thinking about information processing and computing .
einstein podolsky and rosen , noticed the implications of quantum entanglement and proposed a _
experiment , the epr experiment @xcite in order to show that quantum mechanics was not a complete theory which could explain nature correctly .
probably one of the most notable and dramatic among various concepts developed through the application of quantum mechanics to the information science and which is a consequence of entanglement and non - locality , is teleportation put forward by bennett _
et al _ @xcite , given rise to a new field of research .
quantum teleportation is an experimental reality and it holds tremendous potential for applications in the fields of quantum communication and computing @xcite . for instance , it can be used to build quantum gates which are resistant to noise and is intimately connected with quantum error - correcting codes @xcite .
as feymann put it , `` the double - slit experiment has in it the heart of quantum mechanics '' .
the double - slit experiment has been performed with many different kinds of particles ranging from photons @xcite , electrons @xcite , neutrons @xcite and atoms dlatoms . in this article
we propose a scheme of teleportation of the states of motion of atoms interacting with two - slit screens and the electromagnetic field in cavities .
we assume that the atoms are rydberg atoms with relatively long radiative lifetime @xcite .
we also assume that the cavities are microwave superconducting cavities of relatively large quality factors @xcite
let us assume two three - level lambda atoms @xmath0 prepared in the state @xmath1 and @xmath2 prepared in the state @xmath3 to pass through a double slit screen @xmath4 where behind slit @xmath5 ( at @xmath6 there is a cavity @xmath7 prepared in a coherent state @xmath8 and behind slit @xmath9 ( at @xmath10 there is a cavity @xmath11 prepared in a coherent state @xmath12 @xcite@xmath13 now , for a three - level lambda atom interacting with the electromagnetic field inside a cavity where the upper and the two degenerated lower states are @xmath14 @xmath15 and @xmath16 respectively , and for which the @xmath17 and @xmath18 transitions are in the far from resonance interaction limit , the time evolution operator @xmath19 for the atom - field interaction in a cavity @xmath20 is given by @xcite@xmath21where @xmath22 @xmath23 is the annihilation ( creation ) operator for the field in cavity @xmath20 , @xmath24 @xmath25 , @xmath26 is the coupling constant , @xmath27 is the detuning where @xmath28 , @xmath29 and @xmath30 are the frequencies of the upper and of the two degenerate lower levels respectively and @xmath31 is the cavity field frequency and @xmath32 is the atom - field interaction time . for @xmath33 , we get @xmath34where @xmath35and we have @xmath36where the even coherent state @xmath37 and odd coherent state @xmath38 are given by @xmath39@xcite , and we have used @xmath40 @xcite . taking into account ( [ ulambdapi ] ) , after atom @xmath0 has passed through the slits of screen @xmath4 and before it has passed through the cavities we have@xmath41after the atom @xmath0 passes through the cavities we get@xmath42then , after atom @xmath43 passes through the two slits @xmath5 and @xmath9 and
before it has interacted with the cavities we have@xmath44and after @xmath2 has passed through @xmath7 and @xmath45 we have@xmath46now if we detect @xmath47 and @xmath48 we get@xmath49 prepared in the state@xmath50to fly through the slits and cavities of @xmath4 .
after @xmath51 has passed through the slits we have@xmath52and after @xmath51 has passed through the cavities we have@xmath53if we detect @xmath54 we have@xmath55 now , we let atom @xmath51 evolve toward a screen @xmath56 near screen @xmath4 with one slit and a detector behind @xmath57 and detect @xmath51 at @xmath58 . if the time evolution operator for the evolution of @xmath51 toward the screen @xmath56 is @xmath59 we have , writing @xmath60 @xmath61 and @xmath60 @xmath62 , @xmath63for @xmath64 and setting @xmath65 since we assume that @xmath4 and @xmath56 are near enough such that this condition be true , we get@xmath66which are the states of motion of atoms @xmath0 and @xmath2 entangled with the cavity field states .
now we let atom @xmath0 to pass through a double slit screen @xmath67 with @xmath68 at @xmath69 and @xmath70 at @xmath71 .
before @xmath0 passes through these slits , making use of @xmath72 , we writelatexmath:[\[|\psi \rangle _ { a1-a2}=\frac{1}{n}\{|\gamma _ { 1}\rangle \langle \gamma _ { 1}|+|\gamma _ { 2}\rangle \langle \gamma _ { 2}|\}\{c_{b}|\zeta _ { 21}\rangle the evolution toward @xmath67 the operator @xmath59 and @xmath74 @xmath75 @xmath74 @xmath76 @xmath77 and @xmath78 @xmath79 we have@xmath80if we detect atom @xmath0 using a screen @xmath81 with one slit at @xmath82 and a detector behind this slit , we get@xmath83now , we assume that atom @xmath0 is far enough from the screen @xmath67 so that the spherical waves @xmath84 and @xmath85 have evolved in such a way that we can consider @xmath86 @xmath87 @xmath88and therefore we can write@xmath89this is the state of motion of @xmath2 which is far apart from the screen @xmath4 to be teleported to another atom @xmath90 . in order to do so we pass an atom @xmath90 through @xmath5 and @xmath91 in screen @xmath4 and @xmath7 and @xmath45
. then we have , after @xmath90 has passed through the slits@xmath92and after @xmath90 has passed through the cavities@xmath93now , we let atom @xmath2 to pass through a screen @xmath94 with two slits @xmath95 ( at @xmath96 ) and @xmath97 ( at @xmath98 ) and before @xmath2 passes through the slits , using @xmath99 , we have @xmath100and writing @xmath101 @xmath102 , @xmath101 @xmath103 , @xmath104 @xmath105 and @xmath104 @xmath106 we havelatexmath:[\]]and if we detect atom @xmath120 in state @xmath123 finally we get @xmath130and the teleportation is completed successfully . in the above disentanglement process
we can choose coherent field states with a photon - number distribution with a sharp peak at average photon number @xmath131 so that , to a good approximation , @xmath132 and @xmath133 , where @xmath134 is the integer nearest @xmath135 , and we could choose , for instance @xmath136 , so that we would have @xmath137 and @xmath138 . in this case atom @xmath120 would be detected in state @xmath123 with almost @xmath139 of probability .
therefore , proceeding this way , we can guarantee that the atomic and field states will be disentangled successfully as we would like . in fig .
1 we show the set - up of the teleportation experiment . concluding , we have presented a scheme of teleportation of the spacial state ( a state of the continuum ) of atoms making use of cavities and screens with slits .
teleportation shows dramatically the strange and fascinating behavior we can find in quantum mechanics due to entanglement .
the experiments involving screens with slits also shows the strange and fascinating aspects of quantum mechanics . in this work
we have combined teleportation with slit experiments and shown in one strike the beautiful and fascinating aspects of quantum mechanics . *
fig . 1 - * experimental set - up of the teleportation .
atom @xmath0 pass through the screen @xmath4 , cavities @xmath7 and @xmath45 , detector @xmath140 , screens @xmath141 and @xmath81 and finally its position is detected in detector @xmath142 .
atom @xmath2 pass through the screen @xmath4 , cavities @xmath7 and @xmath45 , detector @xmath143 , screens @xmath94 and @xmath108 and finally its position is detected in detector @xmath144 .
atom @xmath51 pass through the screen @xmath4 , cavities @xmath7 and @xmath45 , detector @xmath145 , screens @xmath56 and finally its position is detected in detector @xmath146 .
atom @xmath90 pass through the screen @xmath4 , cavities @xmath7 and @xmath45 , and detector @xmath147 .
atoms @xmath148 and @xmath149 are not shown in the figure .
e. schrdinger , naturwissenschaften , * 23 * , 807 ( 1935 ) ; _ ibden _ 823 ; _ ibden _ 844 ; proc . of the am .
* 124 * , 323 ( 1980 ) ; reprinted in _
quantum theory of measuremente _
( eddited by j. a. wheeler and w. h. zurek , princeton , 152 , 1983 ) . | we present a scheme of teleportation of the state of motion of atoms making use of screens with slits and cavities .
the fascinating aspects of quantum mechanics are highlighted making use of entanglement and the interaciton of atoms with slits .
pacs : 03.65.ud ; 03.67.mn ; 32.80.-t ; 42.50.-p keywords : double - slit experimant ; entanglement ; non - locality ; teleportation ; cavity qed . |
in this paper , which is the second part of work begun in @xcite , we present applications of the theoretical regularity results in the first part of this paper to finite element method approximation schemes .
the first application is to approximation of eigenvalues , @xmath19 , and eigenfunctions , @xmath20 , of the bloch operator , @xmath16 , associated to a periodic hamiltonian operator with inverse square potential at isolated points .
for example , one of our main results , theorem [ athm1.eig ] yields optimal orders of convergence for the finite element approximations of the eigenvalues of @xmath16 using graded meshes .
these rates are higher than those that can be obtained using standard meshes .
the second application is to the finite element method , again using graded meshes , applied to equations of the form @xmath21 .
the final section of this paper presents numerical tests showing good agreement with our theoretical results for this second problem .
hamiltonian operators with inverse square potentials arise in a variety of interesting contexts .
the standard example of a schrdinger operator with @xmath22 potential is a special case of the inverse square potentials we consider , where the function @xmath23 vanishes to order 1 at the singularity , and the results of this work apply to such operators .
but in addition , hamiltonians with true inverse square potentials arise in relativistic quantum mechanics from the square of the dirac operator coupled with an interaction potential , and they arise in the interaction of a polar molecule with an electron .
see @xcite for further applications of inverse square potentials to physics .
see also @xcite for related results on operators with singular coefficients .
thus it is interesting in several areas of physics to understand how to approximate solutions to equations involving such operators . before we can state our approximation results
, we must fix some notation and state the assumptions we make about our hamiltonian operators . consider a hamiltonian operator @xmath24 that is periodic on @xmath25 with triclinic periodicity lattice @xmath11 .
its fundamental domain is a parallelopiped whose faces can be identified under the symmetries of @xmath17 to form the torus @xmath26 , which is how we will denote this fundamental domain in the remainder of this paper .
let @xmath27 be a continuous function on @xmath28 that is given by @xmath29 for @xmath5 close to @xmath6 , is smooth except at the points of @xmath2 , and may be assumed to be equal to one outside a neighbourhood of @xmath2 .
we need two assumptions about the potentials @xmath0 that we will consider in this paper .
first , we assume that @xmath0 is smooth except at a set of points @xmath30 , near which it has singularities of the form @xmath3 , where @xmath7 is continuous on @xmath28 and smooth in polar coordinates around @xmath6 .
we denote this as follows .
@xmath31 assumption 1 , more precisely the continuity of @xmath7 at @xmath2 , allows us to formulate our second assumption .
namely , @xmath32 in particular , we assume that for all @xmath9 , @xmath8
. these assumptions are sharp in the sense that the analysis yields fundamentally different results if either one fails .
in particular , the value @xmath33 corresponds to the critical coupling for an isolated inverse square potential in @xmath25 where the system undergoes a transition between the conformal and non - conformal regimes @xcite .
if the first assumption fails , then the available analytic techniques are much weaker , see for instance @xcite . in either case ,
the approximation theorems in this paper fail if either assumption is violated .
more details of this are included in the first part , @xcite , and a study of the analysis when these assumptions are relaxed will be examined in a forthcoming paper .
we are interested in understanding the spectrum and generalised eigenfunctions of the operator @xmath17 . as usual , we do this by studying bloch waves . recall that if @xmath15 is an element of the first brillion zone of @xmath11 , that is , is an element of the fundamental domain of the dual lattice of @xmath11 , then a bloch wave with wave vector @xmath15 is a function in @xmath34 that satisfies the semi - periodicity condition @xmath35 it is well known that such a bloch wave can be written as @xmath36 for a function @xmath37 that is truly periodic with respect to @xmath11 and thus can be considered as living on the three - torus @xmath38 .
we define the @xmath15hamiltonian @xmath16 on @xmath14 by @xmath39 then we have further that if a bloch wave @xmath40 is a generalized eigenfunction of @xmath17 with generalized eigenvalue @xmath19 , then the function @xmath41 is a standard @xmath42-eigenfunction of @xmath16 with eigenvalue @xmath19 .
let @xmath43 , @xmath44 , be the eigenvalues of @xmath16 , arranged in increasing order , @xmath45 , and repeated according to their multiplicities .
that is , if @xmath46 denotes the eigenspace of @xmath16 corresponding to @xmath19 , then @xmath19 is repeated @xmath47 times . as usual , for our finite element approximation results
, we consider a sequence @xmath48 of finite dimensional subspaces of the domain of @xmath16 and let @xmath49 denote the riesz projection onto @xmath50 , that is , the projection in the bilinear form @xmath51 , ( for a suitable @xmath52 ) , associated to @xmath16 .
let @xmath53 be the associated finite element approximation of @xmath16 , acting on @xmath48 .
denote by @xmath54 the eigenvalues of the approximation @xmath55 , again arranged in increasing order , @xmath56 , and repeated according to their multiplicities and let @xmath57 be a choice of corresponding eigenfunctions ( linearly independent ) .
the spaces @xmath48 we use for our theorems are defined in terms of a sequence of graded tetrahedral meshes @xmath58 on @xmath38 ( sometimes called triangulations ) , given by sequential refinements , associated to a scaling parameter @xmath59 , of an original tetrahedral mesh @xmath60 .
we describe the meshing refinement procedure in detail in section [ sec3 ] .
we will take @xmath61 the finite element spaces associated to these meshes ( _ i.e. , _ using continuous , piecewise polynomials of degree @xmath62 ) . our first theorem , which is a theoretical result for the finite element method approximation of eigenvalues and eigenfunctions of @xmath16 using tetrahedralisations with graded meshes , is as follows .
[ athm1.eig ] let @xmath63 be an eigenvalue of @xmath16 and fix @xmath64 , @xmath65 .
let @xmath54 be the finite element approximations of @xmath63 associated to the nested sequence @xmath66 of meshes on @xmath38 defined by the scaling parameter @xmath67 and piecewise polynomials of degree @xmath62 .
also , let @xmath68 be an eigenbasis corresponding to @xmath54 .
then there exists a constant @xmath69 independent of @xmath70 such that the following inequalities hold for a suitable eigenvector @xmath71 : @xmath72 @xmath73 where the space @xmath74 is a weighted sobolev space defined below in equation [ eq.def.ws ] . for our second theorem
, we consider the finite element approximations of the equation @xmath75 where @xmath76 is the constant from theorem [ theorem1 ] below .
we then define the form @xmath77 and let @xmath78 be the solution of equation above .
we then define the usual galerkin finite element approximation @xmath79 of @xmath78 as the unique @xmath80 such that @xmath81 then theorem [ athm1.eig]together with the lax - milgram lemma and cea s lemma imply that we have the following @xmath82 quasi - optimal rate of convergence .
[ athm1.fem ] the sequence @xmath83 of meshes on @xmath84 defined using the @xmath59-refinement , for @xmath85 , @xmath64 , @xmath86 , and piecewise polynomials of degree @xmath62 , has the following property .
the sequence @xmath87 of finite element ( galerkin ) approximations of @xmath78 from equation satisfies @xmath88 where @xmath89 is independent of @xmath70 and @xmath90 .
these theorems are interesting because it is known that the convergence rate of a standard finite element method ( _ i.e. , _ based on quasi - uniform meshes ) is limited .
however , under the assumptions on our potentials , if we use graded meshes instead , we can obtain an approximation rate as fast as we like by using polynomials of sufficiently high degree in the elements .
this is due to the fact that although regularity of the associated bloch waves is limited in terms of standard sobolev spaces on @xmath38 , it is arbitrarily good with respect to weighted sobolev spaces .
we will recall the definition of these spaces and the relevant regularity results from @xcite along with additional background in section [ sec : background ] .
the remainder of the paper is organised as follows . in section [ sec3 ] , we first describe the @xmath59-refinement algorithm for the three dimensional tetrahedral meshes , which results in a sequence of meshes @xmath66 .
we then prove a general interpolation approximation result for the sequence of finite element spaces associated to this sequence of meshes . in section [ sec.fem ] we use this general approximation result to prove our main approximation results this section includes in particular the proofs of theorem [ athm1.eig ] and theorem [ athm1.fem ] , as well as an additional result about the condition number of the stiffness matrix associated to the finite element spaces , @xmath48 . in the last section , section [ sec.tests ] ,
we discuss results of numerical tests of the method for solving equations of the form @xmath91 and compare them to the theoretical results .
we would like to thank bernd ammann , douglas arnold , and catarina carvalho for useful discussions .
we also thank the leverhulme trust whose funding supported the fourth author during this project .
this project was started while hunsicker and nistor were visiting the max planck institute for mathematics in bonn , germany , and we are greatful for its support .
in this section we recall some definitions and results from @xcite , as well as the classical approximation result for lagrange interpolants ( see @xcite ) , that will be used in the proofs of the approximation theorems above .
first , these results are given in terms of weighted sobolev spaces which are defined as follows : @xmath92 these spaces have been considered in many other papers , most notably in kondratiev s groundbreaking paper @xcite .
the first result that we recall guarantees the existence of solutions of equations of the form @xmath93 for @xmath94 greater than some constant @xmath76 , and identifies the natural domain of @xmath16 .
let us fix smooth functions @xmath95 supported near points of @xmath2 such that the functions @xmath95 have disjoint supports and @xmath96 in a small neighbourhood of @xmath9 .
then theorem 1.1 and proposition 3.6 from @xcite combine to give right away the following result .
[ theorem1 ] let @xmath0 be a potential satisfying both assumptions 1 and 2
. then there exists @xmath97 such that @xmath98 is an isomorphism for all @xmath99 , all @xmath100 , and all @xmath101 .
moreover , for any @xmath102 satisfying @xmath103 , we can find constants @xmath104 such that @xmath105 we obtain , in particular , that @xmath16 has a natural self - adjoint extension , the friedrichs extension .
therefore , from now on , we shall extend @xmath16 to the domain of the friedrichs extension of @xmath106 , as in the above theorem .
let us denote by @xmath107 its domain .
then theorem [ theorem1 ] gives that @xmath108 for @xmath109 , and , in general , @xmath110 so that @xmath111 , since we assumed that @xmath112 .
we can now state a regularity theorem for the eigenfunctions of @xmath16 near a point @xmath9 , or equivalently , for bloch waves associated to the wavevector @xmath15 .
[ theorem2 ] assume that @xmath0 satisfies assumptions 1 and 2 and let @xmath113 satisfy @xmath114 , for some @xmath115
. then we can find constants @xmath104 such that @xmath116 in particular , @xmath117 where @xmath118 and @xmath119 is arbitrary .
see also @xcite for some related classical results in this area .
theorems [ theorem1 ] and [ theorem2 ] lead to an estimate for the distance from an element in the domain of @xmath16 to the approximation spaces that we construct using graded meshes . next , recall the definition of lagrange interpolants associated to a mesh .
let us choose @xmath84 to be a parallelopiped that is a fundamental domain of the lattice @xmath11 .
that is , @xmath120 and all @xmath121 disjoint .
let @xmath122 be a _ mesh _ on @xmath84 , that is a mesh of @xmath84 with tetrahedra @xmath123 .
we can identify this @xmath124 with a mesh @xmath125 of the fundamental region of the lattice @xmath126 ( that is , to the brillouin zone of @xmath126 ) .
fix an integer @xmath127 that will play the role of the order of approximation .
we denote by @xmath128 the finite element space associated to the degree @xmath62 lagrange tetrahedron .
that is , @xmath128 consists of all continuous functions @xmath129 such that @xmath130 coincides with a polynomial of degree @xmath131 on each tetrahedron @xmath132 and _ @xmath130 is periodic_. this means the values of @xmath130 on corresponding faces coincide , so @xmath130 will have a continuous , periodic extension to the whole space , or alternatively , can be thought of as a continuous function on @xmath124 .
we shall denote by @xmath133 the lagrange interpolant of @xmath134 .
let us recall the definition of @xmath135 .
first , given a tetrahedron @xmath136 , let @xmath137 $ ] be the barycentric coordinates on @xmath136 .
the nodes of the degree @xmath62 lagrange tetrahedron @xmath136 are the points of @xmath136 whose barycentric coordinates @xmath138 $ ] satisfy @xmath139 .
the _ degree @xmath62 lagrange interpolant _
@xmath135 of @xmath20 is the unique function @xmath140 such that @xmath141 at the nodes of each tetrahedron @xmath132 .
the shorter notation @xmath142 will be used when only one mesh is understood in the discussion . the classical approximation result for lagrange interpolants ( @xcite )
can now be stated .
[ athm1.classical ] let @xmath124 be a mesh of a polyhedral domain @xmath143 with the property that all tetrahedra comprising @xmath124 have angles @xmath144 and edges @xmath145 .
then there exists a constant @xmath146 such that , for any @xmath147 , @xmath148 finally , we recall two properties of functions in the weighted sobolev spaces @xmath149 that are useful for the analysis of the approximation scheme we use with graded meshes . the proofs of these lemmas are contained in @xcite and are based on the definitions and straightforward calculations .
[ claim2 ] let @xmath150 be a small neighborhood of a point @xmath151 such that on @xmath150 , @xmath10 is given by distance to @xmath6 .
let @xmath152 and denote by @xmath153 the region obtained by radially shrinking around @xmath6 by a factor of @xmath154
. then @xmath155 [ claim3 ] if @xmath156 , @xmath157 and @xmath158 on @xmath150 , then @xmath159 we can now continue to the definition of the mesh refinement technique and the proof of the general approximation theorem underlying our two main theorems .
our two main theorems follow from standard results , such as cea s lemma ( for the proof of theorem [ athm1.fem ] ) and the results used in @xcite ( for the proof of theorem [ athm1.eig ] ) , together with the following underlying approximation theorem : [ athm1.gen ] there exists a sequence @xmath66 of meshes of @xmath38 that depends only on the choice of a parameter @xmath160 , @xmath161 and @xmath162 , with the following property .
if @xmath102 , then the modified lagrange interpolant @xmath163 of @xmath20 satisfies @xmath164 where @xmath89 depends only on @xmath62 and @xmath165 ( so it is independent of @xmath70 and @xmath20 ) . in this section
we will define the mesh refinement process and prove theorem [ athm1.gen ] .
the first step is to describe the refinement procedure that results in our sequence of meshes ( or triangulations ) .
this is based on the construction in @xcite and in @xcite , thus we refer the reader to those papers for details , and here give only an outline and state the critical properties .
the second step is to prove a sequence of simple lemmas used in the estimates .
the third step is to prove the estimate separately on smaller regions .
this uses the scaling properties of the meshes in lemmas [ claim2 ] and [ claim3 ] together with theorem [ athm1.classical ] .
we continue to keep the approximation degree @xmath62 fixed throughout this section .
fix a parameter @xmath165 and let @xmath67 . in our estimates
, we will chose @xmath165 such that @xmath166 and @xmath162 .
let @xmath167 denote the smallest distance between the points in @xmath2 .
choose an initial mesh @xmath60 of @xmath84 with tetrahedra such that all singular points of @xmath0 ( _ i.e. , _ all points of @xmath2 ) are among the vertices of @xmath60 and no tetrahedron has more than one vertex in @xmath2 .
we assume that this mesh is such that if @xmath168 and @xmath169 are two opposite faces of @xmath84 , which hence correspond to each other through periodicity , then the resulting triangulations of @xmath168 and @xmath169 will also correspond to each other , that is , they are congruent in an obvious sense .
we start with a special refinement of an arbitrary tetrahedron @xmath136 that has one of the vertices in the set @xmath2 .
our assumptions then guarantee that all the other vertices of @xmath136 will not be in @xmath2 . motivated by the refinement in @xcite , we define our _ @xmath59-refinement algorithm for a single tetrahedron _ that divides @xmath136 into eight sub - tetrahedra as follows .
[ def.n1 ] * @xmath59-refinement for a single tetrahedron * : let @xmath170 be the four vertices of @xmath136 .
we denote @xmath136 by its vertex set @xmath171 .
suppose that @xmath172 , so that @xmath173 is the one and only vertex that will be refined with a ratio @xmath174 $ ] . we first generate new nodes @xmath175 , @xmath176 , on each edge of @xmath136 , such that @xmath177 for @xmath178 and @xmath179 for @xmath180 . note that the node @xmath175 is on the edge connecting @xmath181 and @xmath182 .
connecting these nodes @xmath175 on all the faces , we obtain 4 sub - tetrahedra and one octahedron .
the octahedron then is cut into four tetrahedra using @xmath183 as the common vertex .
therefore , after one refinement , we obtain eight sub - tetrahedra ( figure [ fig.tetrarefine ] ) , namely , @xmath184 ( left ) ; eight sub - tetrahedra after one @xmath59-refinement ( right ) , @xmath185.,title="fig : " ] ( left ) ; eight sub - tetrahedra after one @xmath59-refinement ( right ) , @xmath185.,title="fig : " ] [ def.n2 ] * @xmath59-refinement for a mesh*:let @xmath124 be a triangulation of the domain @xmath84 such that all points in @xmath2 are among the vertices of @xmath124 and no tetrahedron contains more than one point in @xmath2 among its vertices . then we divide each tetrahedron @xmath136 of @xmath124 that has a vertex in @xmath2 using the @xmath59-refinement and we divide each tetrahedron in @xmath136 that has no vertices in @xmath2 using the @xmath186-refinement .
the resulting mesh will be denoted @xmath187 .
we then define @xmath188 , where @xmath60 is the intial mesh of @xmath84 . according to @xcite ,
when @xmath189 , which is the case when the tetrahedron under consideration is away from @xmath2 , the recursive application of algorithm [ def.n1 ] on the tetrahedron generates tetrahedra within at most three similarity classes . on the other hand ,
if @xmath190 , the eight sub - tetrahedra of @xmath136 are not necessarily similar .
thus , with one @xmath59-refinement , the sub - tetrahedra of @xmath136 may belong to at most eight similarity classes .
note that the first sub - tetrahedra in algorithm [ def.n1 ] is similar to the original tetrahedron @xmath136 with the vertex @xmath191 and therefore , a further @xmath59-refinement on this sub - tetrahedron will generate eight children tetrahedra within the same eight similarity classes as sub - tetrahedra of @xmath136 .
hence , successive @xmath59-refinements of a tetrahedron @xmath136 in the initial triangulation @xmath60 will generate tetrahedra within at most three similarity classes if @xmath136 has no vertex in @xmath2 .
on the other hand , successive @xmath59-refinements of a tetrahedron @xmath136 in the initial triangulation will generate tetrahedra within at most @xmath192 similarity classes if @xmath136 has a point in @xmath2 as a vertex .
thus , our @xmath59-refinement is conforming and yields only non - degenerate tetrahedra , all of which will belong to only finitely many similarity classes .
recall that our initial mesh @xmath60 has matching restrictions to corresponding faces .
since the singular points in @xmath2 are not on the boundary of @xmath84 , the refinement on opposite boundary faces of @xmath84 is obtained by the usual mid - point decomposition .
therefore , the same matching property will be inherited by @xmath66 . in particular
, we can extend @xmath66 to a mesh in the whole space by periodicity . we will , however , not make use of this periodic mesh on the whole space . for each point @xmath9 and each @xmath193 ,
we denote by @xmath194 the union of all tetrahedra of @xmath195 that have @xmath6 as a vertex .
thus @xmath194 is obtained by scaling the tetrahedra in @xmath196 by a factor of @xmath197 with center @xmath6 . in particular ,
the level @xmath198 refinements of @xmath60 give rise to a mesh on @xmath199 .
define @xmath200 according to definition [ def.n2 ] , both @xmath201 and @xmath202 are triangulated using the @xmath59-refinement . for each tetrahedron in @xmath196
we use the @xmath59-refinement for a single tetrahedron , while for @xmath201 we use the @xmath186refinement for meshes , which is , of course , a uniform refinement . then
, we can decompose @xmath84 as the union @xmath203 where each set in the union is a union of tetrahedra in @xmath66 .
[ rk.45 ] note that the size of each simplex of @xmath66 contained in @xmath201 is @xmath204 , the size of each simplex of @xmath66 contained in @xmath205 is @xmath206 , and the size of @xmath207 is @xmath208 .
in addition , the number of tetrahedra in @xmath66 is @xmath209 ( see algorithm [ def.n2 ] ) .
we now define the finite element approximation @xmath210 to the equation @xmath93 , where @xmath66 is obtained by applying @xmath70 times the @xmath59-refinements to @xmath60 , where @xmath211 , @xmath212 , @xmath162 , and @xmath213 satisfies theorem [ theorem1 ] .
then @xmath214 is defined for any @xmath215 by @xmath216 note that theorem [ theorem1 ] gives that the finite element solution @xmath217 is well defined by .
the approximation properties of @xmath214 are discussed in theorem [ athm1.eig ] .
note that the singular expansion of theorem [ theorem2 ] shows that the value of an eigenvalue @xmath20 of @xmath16 at a singular point in @xmath2 may not be defined .
therefore , we must define the _ modified _ degree @xmath62 lagrange `` interpolant '' @xmath218 associated to the mesh @xmath66 , such that @xmath219 alternatively , we can take the modified lagrange interpolant to be zero on the whole tetrahedron that contains a singular point . by construction ,
the restriction of @xmath66 to @xmath205 scales to the restriction of @xmath220 to @xmath221 . from now on ,
we refer to @xmath222 as the _ modified _ interpolation defined in .
the following lemma is based on the definition of the @xmath59-refinement and the discussion in remark [ rk.45 ] .
[ claim1 ] for all @xmath223 , @xmath224 where @xmath225 is the dilation with ratio @xmath226 and center @xmath6 .
recall that @xmath227 and @xmath112 .
that is , @xmath0 satisfies assumptions 1 and 2
. we can now give the proof of theorem [ athm1.gen ] .
recall that @xmath196 consists of the tetrahedra of the initial mesh @xmath60 that have @xmath6 as a vertex and that all the regions @xmath228 are away from each other ( they are closed and disjoint ) .
we used this to define @xmath229 .
the region @xmath194 is obtained by dilating @xmath230 with the ratio @xmath231 and center @xmath6 .
finally , recall that @xmath232 .
let @xmath233 be any of the regions @xmath201 , @xmath205 , or @xmath207 .
since the union of these regions is @xmath84 , it is enough to prove that @xmath234 for a constant @xmath89 independent of @xmath233 and @xmath70 .
the result will follow by squaring all these inequalities and adding them up .
in fact , since @xmath235 , it is enough to prove @xmath236 again for a constant @xmath89 independent of @xmath233 and @xmath70 . if @xmath237 , the estimate in follows right away from theorem [ athm1.classical ] . for the other estimates , recall that @xmath238 , where @xmath239 and @xmath162 .
we next establish the desired interpolation estimate on the region @xmath240 , for any fixed @xmath9 and @xmath241 .
let @xmath242 .
from lemmas [ claim2 ] and [ claim1 ] , we have @xmath243 since @xmath244 is equivalent to @xmath245 , we can apply theorem [ athm1.classical ] with @xmath246 to get @xmath247 now applying lemma [ claim2 ] to scale back again and using also @xmath67 , we get that the right hand side in @xmath248 this proves the estimate in for @xmath240 .
it remains to prove this estimate for @xmath249 . for any function @xmath250 on @xmath207 , we let @xmath251 be a function on @xmath230 . therefore , by lemma [ claim2 ] @xmath252 and by [ claim1 ] ( which follows from the definition of the meshes @xmath253 and from the fact that interpolation commutes with changes of variables ) , @xmath254 now let @xmath130 be a smooth cutoff function on @xmath230 such that @xmath255 in a neighborhood of @xmath6 and @xmath256 at every other node of @xmath230 .
define @xmath257 .
then , by , @xmath258 since @xmath130 vanishes in the neighborhood of @xmath6 we can consider multiplication by @xmath130 as @xmath259 times a degree 0 b - operator . thus it is a bounded operator on any weighted sobolev space .
thus @xmath260 where @xmath89 depends on @xmath62 and , through @xmath130 , the nodes in the triangulation . using , , , , lemma [ claim3 ] , and theorem [ athm1.classical ]
, we have @xmath261 this proves the estimate of equation for @xmath262 and completes the proof of theorem [ athm1.gen ] .
we can now turn to the proofs of the theorems stated in the introduction .
first , theorem [ athm1.eig ] follows from our general approximation result , theorem [ athm1.gen ] , and the standard results on approximations of eigenvalues and eigenvectors ( eigenfunctions in our case ) discussed , for instance , in @xcite .
more precisely , using the notation introduced in the introduction , we have the following .
let us denote by @xmath46 the eigenspace of @xmath16 corresponding to the eigenvalue @xmath19 and by @xmath263 , the subspace consisting of vectors of length one .
then the following result is well known ( see for instance equations ( 1.1 ) and ( 1.2 ) in @xcite ) .
we state it only for our operator @xmath16 , although it is valid for more general self - adjoint operators with compact resolvent .
[ thm.eig.a ] there exists a constant @xmath264 with the following property .
let @xmath265 be a finite dimensional subspace and @xmath266 the projection in the energy norm .
let @xmath267 be an eigenbasis of @xmath268 , namely @xmath269 , with the @xmath54 arranged in increasing order in @xmath193 .
then @xmath270 and , furthermore @xmath271 for a suitable eigenvector @xmath71 .
the proof of theorem [ athm1.eig ] will then be obtained from theorem [ thm.eig.a ] as follows .
( of theorem [ athm1.eig ] ) .
we need to estimate @xmath272 . to this end , let us notice that any @xmath273 satisfies @xmath274 .
theorem [ theorem1 ] then gives @xmath275 for a suitably large @xmath276 that depends on @xmath19 and @xmath277 .
a bootstrap argument then gives for any @xmath278 that @xmath279 .
theorem [ athm1.gen ] then gives for @xmath280 ( thus @xmath281 ) , the following .
@xmath282 the proof of theorem [ athm1.eig ] is now complete .
next , the proof of theorem [ athm1.fem ] follows from theorem [ athm1.gen ] , the lax - milgram lemma and cea s lemma .
we note some consequences of this theorem .
[ rk.opt ] first , in the case @xmath283 , by the estimate in equation , we have @xmath284 as long as the index in theorem [ athm1.fem ] is chosen such that @xmath285 .
as in the classical finite element method , a duality argument yields the following @xmath42-convergence result .
in addition to the assumptions and notation in theorem [ athm1.fem ] , assume that @xmath286
. then the following @xmath42 estimate holds @xmath287 we sketch the proof by using the duality argument in weighted sobolev spaces .
consider the equation @xmath288 ( so we use periodic boundary conditions on @xmath84 . )
the definition of the galerkin projection @xmath79 of @xmath78 , equation , then gives @xmath289 where @xmath290 is the finite element solution of equation on @xmath66 .
we also have @xmath291 by theorem [ theorem1 ] , since @xmath292 . therefore ,
applying theorem [ athm1.fem ] to @xmath293 and @xmath294 , we have @xmath295 this completes the proof .
it is important that the discrete system @xmath48 that we use is well - conditioned for us to be able to realise the theoretical approximation bounds in practice .
thus we need to additionally obtain upper and lower bounds on the eigenvalues of the stiffness matrix that arises in calculation .
recall the standard nodal basis function @xmath296 of the space @xmath297 .
it consists of functions that are equal to @xmath298 at one node and equal to zero at all the other nodes . for convenience ,
we now instead consider the rescaled bases @xmath299 , where @xmath300 is the diameter of the support patch for @xmath296 .
then , we consider the scaled stiffness matrix @xmath301 from our graded finite element discretization . in practice , @xmath302 can be obtained from the usual stiffness matrix @xmath303 by a diagonal preconditioning process .
we point out that similar scaled matrices were considered in @xcite for condition numbers of other galerkin - based methods . for a symmetric matrix @xmath304
, we shall denote by @xmath305 the largest eigenvalue of @xmath304 and by @xmath306 the smallest eigenvalue of @xmath304 .
thus the spectrum of @xmath304 is contained in @xmath307 $ ] , but is not contained in any smaller interval .
we first have the following estimates regarding properties of functions . [ thm.inverse ]
let @xmath123 be a tetrahedron in the mesh @xmath66 and let @xmath308 denote the diameter of @xmath123 .
then , for any @xmath309 and @xmath310 , there exists a constant @xmath264 independent of @xmath70 , @xmath311 and @xmath276 , such that @xmath312 furthermore , writing @xmath313 , where @xmath314 are rescaled basis functions , then @xmath315 we shall show and since is a standard result in @xcite .
recall that all the tetrahedra @xmath123 belong to a finite class of shapes ( or similarity classes ) in our graded triangulation .
thus , the bounded constant @xmath89 in follows from the inverse estimates in @xcite . as for , note @xmath316 . based on the definition of the basis function @xmath317 and of the graded mesh , @xmath318 on the reference tetrahedron @xmath319 ,
both @xmath320 and @xmath321 are norms for the finite element function @xmath322 , where @xmath323 is obtained by the usual scaling process and the summation on @xmath324 is for all the nodes in @xmath319 . based on equivalence of all norms for a finite dimensional space
, we have @xmath325 this , together with , implies @xmath326 which completes the proof
. therefore , we have the following estimates on the eigenvalues of the stiffness matrix . [ lem.upperbound ]
let @xmath302 be the stiffness matrix from the finite element discretization corresponding to the rescaled nodal basis @xmath327 of the space @xmath297 in equation . then , @xmath328 where the constant @xmath329 is independent of the mesh level @xmath70 .
let us fix the mesh level @xmath70 .
all the constants below will be independent of @xmath70 .
let @xmath330 be the tetrahedra forming our mesh @xmath66 .
let @xmath331 be arbitrary and write @xmath332 and @xmath333 .
then , by lemma 3.4 in @xcite we have @xmath334 by the inverse inequality and the estimate , we further have @xmath335 where @xmath308 is the diameter of the tetrahedron @xmath123 .
this completes the proof .
[ lem.n3 ] we use the same notation as the one for lemma [ lem.upperbound ] .
the smallest eigenvalue of the stiffness matrix @xmath302 , @xmath336 for any @xmath337 , we use the notation @xmath338 , @xmath339 , and @xmath308 denotes the diameter of @xmath123 , as before . in view of ,
the inverse estimate , hlder s inequality , and the sobolev embedding estimate , we then have @xmath340 then , we have the estimate on the condition number .
[ thm.main11]let @xmath341 be the stiffness matrix .
then the condition number @xmath342 satisfies @xmath343 the constant @xmath89 depends on the finite element space , but not on @xmath344 .
using @xmath345 , we obtain the estimate by lemmas [ lem.upperbound ] and [ lem.n3 ] .
we now present the numerical tests for the finite element solution defined in approximating possibly singular solutions to equation [ eq.pde.eq ] . to be more precise ,
suppose that our periodicity lattice is @xmath346 and we choose our fundamental domain @xmath347 ^ 3 $ ] to be a cube of side length 2 .
we impose periodic boundary condition on the following model problem @xmath348 where @xmath349 , @xmath350 , @xmath351 , and the cut - off function @xmath352 for @xmath353 and @xmath354 for @xmath355 ; in the tests , we chose @xmath356 . note that if @xmath357 , it is clear that the operator @xmath106 is positive on @xmath358 ( see theorem [ theorem1 ] ) .
we use the @xmath359 linear finite element method on triangulations graded toward the origin with grading ratio @xmath360 ( recall that @xmath361 corresponds to the quasi - uniform refinement . ) to enforce the periodic boundary condition for the finite element functions , we use meshes where all the boundary nodal points are symmetric about the mid - plane between opposite faces of the cube .
any set of the symmetric nodes will be associated to the same shape function in the discretization .
for example , nodes on edges of the cube generally have three mirror images over two mid - planes ( two direct mirror images and the third is symmetric over the line of intersection of these two mid - planes ) , and these four points are associated to the same shape function .
consequently , the eight vertices of the cube are associated to the same shape function through symmetry
. see figure [ cube ] for example .
our first tests are for equation with @xmath363 and @xmath364 . according to theorem [ athm1.fem ]
, the optimal rate of convergence for the finite element solution should be obtained on triangulations with any @xmath365 , since @xmath366 .
the convergence rates @xmath367 associated to triangulations with different values of @xmath59 are listed in table [ tab.5.3 ] .
starting from an initial triangulation , we compute the rates based on the comparison of the numerical errors on triangulations with consecutive @xmath59-refinements , @xmath368 where @xmath369 is the finite element solution on the mesh after @xmath193 @xmath59-refinements .
recall the dimension of the finite element space grows by a factor of 8 with one @xmath59-refinement . by theorem [ athm1.fem ] , for a sequence of optimal meshes , the error @xmath370 is reduced by a factor of 2 for linear finite element approximations with each @xmath59-refinement.thus , @xmath371 implies that the optimal rate of convergence in theorem [ athm1.fem ] is achieved .
table [ tab.5.3 ] clearly shows that the convergence rates @xmath367 approach @xmath298 for all values of the grading parameter @xmath59 .
this is in agreement with our theory that the optimal rates of convergence are obtained for any triangulations with @xmath365 , since the singularity in the solution is not strong enough to be detectable for linear finite elements .
all our numerical tests ( tables [ tab.5.3],[tab.5.4 ] [ tab.5.5 ] , and corresponding eigenvalue computations ) convincingly verify theorem [ athm1.eig ] by comparing the rates of convergence for different singular potentials on different graded triangulations for the model operator in .
the theoretical upper bounds @xmath372 of the optimal range for the grading parameter @xmath59 are also clearly demonstrated in these numerical results . in these tests ,
the initial triangulation of the unit cube consists of 12 tetrahedra and we consecutively refine the mesh using the @xmath59-refinements up to level 7 that includes @xmath373 tetrahedra and roughly 4.2 million unknowns .
numerical experiments show that the condition numbers of our discrete systems grow by a factor of 4 for consecutive refinements , regardless of the value of @xmath59 , which resembles the estimates given in @xcite for the laplace operator
. however , the values of @xmath59 affect the magnitude of the condition numbers .
in general , smaller @xmath59 leads to bad shapes for the tetrahedra and therefore results in larger condition numbers .
the preconditioned conjugate gradient ( pcg ) method was used as the numerical solver for the discrete systems .
i. babuka and a. k. aziz .
survey lectures on the mathematical foundations of the finite element method . in _ the mathematical foundations of the finite element method with applications to partial differential equations ( proc .
maryland , baltimore , md . , 1972 )
_ , pages 1359 . academic press , new york , 1972 . with the collaboration of g. fix and r. b. kellogg .
i. babuka and j. e. osborn .
estimates for the errors in eigenvalue and eigenvector approximation by galerkin methods , with particular attention to the case of multiple eigenvalues .
, 24(6):12491276 , 1987 . | let @xmath0 be a _ periodic _
potential on @xmath1 that is smooth everywhere except at a discrete set @xmath2 of points , where it has singularities of the form @xmath3 , with @xmath4 for @xmath5 close to @xmath6 and @xmath7 is continuous , @xmath8 for @xmath9 .
we also assume that @xmath10 and @xmath7 are smooth outside @xmath2 and @xmath7 is smooth in polar coordinates around each singular point .
let us denote by @xmath11 the periodicity lattice and set @xmath12 .
in the first paper of this series @xcite , we obtained regularity results in weighted sobolev space for the eigenfunctions of the schrdinger - type operator @xmath13 acting on @xmath14 , as well as for the induced @xmath15hamiltonians @xmath16 obtained by resticting the action of @xmath17 to bloch waves . in this paper
we present two related applications : one to the finite element approximation of the solution of @xmath18 and one to the numerical approximation of the eigenvalues , @xmath19 , and eigenfunctions , @xmath20 , of @xmath16 .
we give optimal , higher order convergence results for approximation spaces defined piecewise polynomials .
our numerical tests are in good agreement with the theoretical results . |
for a knot @xmath10 , the 4genus of @xmath11 , @xmath12 , is the minimum genus of a smoothly embedded surface in @xmath13 bounded by @xmath11 .
although the study of this invariant has been a focus of knot theoretic research for over 50 years , it remains an intractable invariant to compute ; for instance , the determination of the 4genus for knots with 10 or fewer crossings has just been recently completed , with even the computation for individual knots being the subject of papers , for instance @xcite .
the depth of continuing interest in the 4genus is indicated by the application of the deepest tools now available in low - dimensional topology : kronheimer - mrowka s study of 4-dimensional gauge theory @xcite , ozsvth - szab s development of heegaard - floer theory @xcite , and rasmussen s work on khovanov homology @xcite , have each been used to establish milnor s conjecture that for torus knots @xmath14 , @xmath15 .
work in the 1960s identified the central role of algebraically defined witt groups to understanding the 4genus . as we will review in a brief appendix , to each knot
@xmath11 there is naturally associated a witt class , @xmath16 , the witt group of hermitian forms over the rational function field , having involution induced by @xmath17 .
( here @xmath18 is represented by the matrix @xmath19 , where @xmath20 is an integer matrix associated to @xmath11 , the _ seifert matrix_. ) a fundamental result states @xmath21 , where @xmath7 is defined as follows : for a class @xmath22 , the rank of @xmath1 , @xmath0 , is the minimum dimension of a square hermitian matrix representing @xmath1 .
for a given class @xmath1 , determining @xmath0 can be very difficult and the most effective tools for bounding @xmath0 are based on bounds on signature functions associated to the class @xmath1 .
a few of the early papers that applied signatures to the 4genus are @xcite .
the goal of this paper is to more closely examine the function @xmath0 .
we define an easily computed invariant @xmath5 which provides stronger bounds on @xmath0 than were previously known .
we then prove that @xmath5 completely determines @xmath0 , modulo torsion in the witt group .
more precisely , our main result states .15 in * theorem * _ for all @xmath9 , @xmath23 and @xmath8 . _ .15 in * outline
* let @xmath24 denote the witt group of nonsingular hermitian bilinear forms over a field @xmath3 with ( possibly trivial ) involution . in the case of @xmath25 or @xmath26 ( with involution given by conjugation )
, a relatively simple exercise shows that @xmath27 , where @xmath28 is the signature .
for @xmath29 the situation is more complicated .
the diagonal form @xmath1 with diagonal @xmath30 $ ] is not witt trivial , and thus @xmath31 , but @xmath32 ; note however that since @xmath32 , @xmath1 represents an element of order four in @xmath33 , and thus @xmath34 . more generally , for @xmath35 , @xmath36 .
details are presented in section [ secwq ] .
the arguments in section [ secwq ] are fairly basic , but they illustrate the structure of the proof of our main theorem regarding @xmath37 . in section [ sectiondefs ]
we will set the notation to be used throughout the paper and define the function @xmath6 .
we will also discuss explicit means for computing @xmath6 . in section [ secbounds ] we will prove the first part of the main theorem : for @xmath9 , @xmath38 . following this
we have a realization result , showing in section [ secrealize ] that for any form @xmath1 , there is a form @xmath39 having an identical signature function and for which @xmath40 .
finally , in section [ sector ] it is shown that a class with trivial signature function represents 4torsion in @xmath37 .
the main theorem is an immediate consequence of this result .
the paper concludes with section [ sectionnorm ] which describes how @xmath7 leads naturally to a norm on @xmath41 and then section [ sectionapps ] presenting examples of the computation of this norm , with specific applications to determining the 4genus of low - crossing number knots .
.1 in _ acknowledgments _ thanks are due to pat gilmer and neal stoltzfus for early discussions regarding this work .
thanks are also due to jim davis and andrew ranicki for their insights regarding the structure of the witt group @xmath37 .
special thanks go to stefan friedl for his thoughtful commentary on an earlier version of this paper .
the proof that @xmath8 for @xmath43 is in structure the same as the proof that @xmath44 for @xmath35 . the proof we give here
is broken up into three mains steps corresponding to sections [ secbounds ] , [ secrealize ] and [ sector ] . a fourth concluding step is identical in both settings .
[ thmmain ] if @xmath35 then @xmath44 .
@xmath45 1 .
@xmath46 : for @xmath35 this is immediate from the definition of the signature . in the case of @xmath37
the corresponding proof will reduce to a careful algebraic calculation , based on the details of the definition of @xmath5 given in the next section .
the argument occupies section [ secbounds ] .
if @xmath35 , there is a class @xmath47 with @xmath48 and @xmath49 : the form @xmath39 is simply the form represented by the identity matrix of dimension @xmath50 . in the case of @xmath37 ,
the construction of the appropriate form @xmath39 for which @xmath51 and @xmath40 is more delicate .
3 . if @xmath32 then @xmath52 : this depends on the structure of @xmath33 . as described for instance in @xcite , there is a split short exact sequence @xmath53 where @xmath54 is the finite field with @xmath55 elements , @xmath55 a prime integer .
the groups @xmath56 are all 4torsion , and @xmath57 , with the isomorphism given by the signature . in the case of @xmath37 ,
the exact sequence is replaced with the sequence @xmath58 ) \to w(\qq(t ) ) \to \oplus_{\alpha } w(\qq(\alpha)),\ ] ] where the @xmath59 are all unit complex roots of symmetric irreducible rational polynomials . for the analysis of these witt groups
, we turn to the references @xcite .
conclusion : since the signatures are the same , we have @xmath60 .
certainly @xmath61 , but by construction , @xmath62 has a representative of rank exactly @xmath63 .
the desired equality follows .
the argument in the case of @xmath37 is identical .
for a nonsingular matrix @xmath64 with entries in @xmath4 that is hermitian with respect to the involution induced by @xmath17 , the signature function @xmath65 is defined by @xmath66 .
this is well - defined function on @xmath67 except at the finite set of points that correspond to poles among the entries @xmath64 .
since @xmath64 is hermitian , @xmath68 is symmetric about @xmath69 ; this justifies the restriction to the interval @xmath70 . as given in the next definition ,
taking averages and differences gives the levine @xcite and milnor @xcite signature functions .
( according to matumoto @xcite the jumps in the levine signature function are determined by the milnor signatures .
notice that the factor of @xmath69 in the definition of @xmath71 implies that @xmath72 represents half the jump in the signature function at @xmath73 . )
if @xmath9 , and @xmath64 is a matrix representing the class @xmath1 , then:.05 in 1 .
@xmath74 .. 1 in 2 .
@xmath75 .. 1 in 3 . for @xmath76
we define @xmath77 and @xmath78 .
an elementary argument shows that these functions are well defined ; that is , @xmath79 and @xmath80 depend only on the class in @xmath37 represented by the matrix @xmath64 .. 1 in * example .
* figure [ siggraph ] illustrates a possible signature function on the interval @xmath70 . we will construct a class with signature function having such a graph later , being more specific about the points @xmath81 . for the specific matrix used ,
the values of the signatures at the discontinuities will not be known , but upon averaging , the values will be as shown in the figure .
in particular , the values of the jumps at the five discontinuities are @xmath82 $ ] . to define the function @xmath83 we need to focus on the set of discontinuities of the signature function and include in that set the value @xmath84 for technical reasons .
that set has a natural decomposition , indexed by symmetric irreducible rational polynomials . throughout
this paper polynomials will be laurent polynomials @xmath85 $ ] and symmetric means @xmath86 .
we will view these polynomials as defined on the unit complex circle . for @xmath9 ,
let @xmath87 denote the ( finite ) set of discontinuities of @xmath79 .
for each @xmath88 , @xmath89 is an algebraic number with rational irreducible polynomial . for
each symmetric irreducible polynomial @xmath90 let @xmath91 be the set @xmath92
. given this we can define @xmath6 .
let @xmath93 . 1 . for each @xmath90 set @xmath94.05 in 2 .
@xmath95 consider the graph of the signature function illustrated in figure [ siggraph ] .
we assume that the five jumps occur at the roots of the same irreducible symmetric polynomial .
( we will see that as part of a general realization result , theorem [ thmrealize ] , such an example does occur .
for example , the @xmath59 can be primitive @xmath96roots of units , the zeroes of the cyclotomic polynomial @xmath97 .
this is the alexander polynomial of the torus knot @xmath98 , although this torus knot does not have this signature function . ) in this example , the values of the signatures ( at the @xmath81 along with the value at 0 ) are @xmath99 $ ] and the values of the jumps are @xmath82 $ ] . the value of @xmath6 for this example will be @xmath100 notice that this is greater than the maximum absolute value of the signature function , which is 6 .
.1 in * computations * for any hermitian matrix @xmath64 representing a class in @xmath37 , standard mathematical computer packages can be used to diagonalize @xmath64 and to arrange that the diagonal entries are laurent polynomials .
factoring these diagonal entries and removing factors of the form @xmath101 ensures that these diagonal entries have factorizations as @xmath102 , so that the @xmath103 are distinct symmetric polynomials ( @xmath104 ) with exponent one . by symmetry , the values of the @xmath103 at points on the unit circle are real , and thus the signs can be determined ( that is , the numerical approximation of @xmath105 will be given as @xmath106 for some small @xmath107 , which can be ignored in the determination of the sign . ) with this , the signature function can be approximated with necessary accuracy . at this same time
, the roots of the @xmath103 on the unit circle will be identified .
these computation are sufficient to completely determine the value of @xmath108)$ ] .
for any matrix representative @xmath64 of a class @xmath9 , dim(@xmath64 ) @xmath109 . to simplify notation , we will let @xmath110 , @xmath111 , and @xmath112 .
( for instance , for the signature function in figure [ siggraph ] ; @xmath113 and @xmath114 . )
since @xmath64 is a hermitian matrix with entries in @xmath4 , we can diagonalize @xmath64 , clear denominators , and remove square factors and factors of the form @xmath101 in the diagonal entries .
thus , there is a diagonalization where each diagonal entry factors as the product of distinct symmetric irreducible rational polynomials .
let @xmath115 be one such diagonalization .
note that the discontinuities of the signature function can occur only at roots of the diagonal factors .
let @xmath116 be the set of discontinuity points for the signature function on @xmath67 .
let @xmath90 be an irreducible symmetric polynomial such that @xmath79 has a discontinuity at @xmath117 and @xmath118 . by reordering
, we can assume that @xmath115 has diagonal @xmath119,\ ] ] where each @xmath120 and @xmath121 is a product of distinct irreducible symmetric polynomials , none of which are @xmath90 .
let @xmath122 . evaluating the diagonal at @xmath123 , where @xmath124 is a number close to but smaller than @xmath59
, we denote the count of positive entries in @xmath125 $ ] to be @xmath126 and the number of negative entries to be @xmath127 .
similarly , we denote the number of positive entries in @xmath128 $ ] by @xmath129 and the number of negative entries by @xmath130 .
here are some elementary calculations : * @xmath131 * @xmath132 * @xmath133 if we switch from @xmath134 to @xmath135 the only change in signs occurs because of the change in the sign of @xmath90 at @xmath136 , and thus the signs of all the diagonal entries with @xmath90 factors change , so that we have @xmath127 positive entries and @xmath126 negative entries .
it follows , using a little arithmetic for the second calculation , that * @xmath137 * @xmath138 note that @xmath139 .
the following inequalities are verified by simply substituting for @xmath140 and @xmath141 in each : * @xmath142 * @xmath143 given that @xmath144 and @xmath145 are nonnegative , we have * @xmath146 * @xmath147 these equations hold at each @xmath148 , so , multiplying by 2 and taking the maximums we find * @xmath149 * @xmath150 this proves the theorem , except we have not dealt yet with the signature at 0 . but clearly @xmath151 , the signature of @xmath64 evaluated near @xmath152 ( that is , ( @xmath153 ) , so the proof is complete .
we now want to show that every step function @xmath154 satisfying certain criteria occurs as the signature function for some class @xmath9 and for that class , @xmath155 . for a step function
@xmath154 , let @xmath156 .. 05 in let @xmath157 denote the set of integer valued step functions defined on @xmath67 satisfying the following list of conditions .
if @xmath158 , then:.05 in 1 .
the set of discontinuities of @xmath159 is finite and @xmath159 is continuous at @xmath84 .. 05 in 2 . for all @xmath73 , @xmath160 .. 05 in 3 . for all @xmath73 , @xmath161 . .05 in 4
if @xmath162 then @xmath89 is the root of an irreducible symmetric rational polynomial . 5 . if @xmath163 and @xmath164 satisfy @xmath165 for some symmetric irreducible rational polynomial @xmath90 , then @xmath166 .
the definitions in the previous section were given purely in terms of the signature function of a class @xmath167 , so the definition extends to @xmath157 as now described .
let @xmath168:.05 in * @xmath169 .
* for an irreducible symmetric polynomial @xmath90 , @xmath170 .. 05 in * for each @xmath90 , @xmath171.05 in * @xmath172 for all @xmath158 , @xmath173 . since @xmath174 and @xmath175 is the jump in the signature function at each discontinuity @xmath59 , @xmath176 if @xmath73 is not a point of discontinuity
. thus , at each discontinuity @xmath59 , @xmath177 . for each @xmath90 , for all
@xmath178 we have @xmath166 .
it follows that @xmath179 . for a fixed @xmath90 , @xmath180 and
we have now seen that mod 2 , all the terms in the set of values over which the maxima are being taken are equal .
thus , if @xmath181 , @xmath182 . we have already seen that this sum equals @xmath183 , modulo 2 .
finally , since @xmath184 and each of these elements equal @xmath183 modulo 2 , the maximum also equals @xmath183 modulo 2 .
[ thmrealize ] suppose that @xmath158 .
there exists a hermitian matrix @xmath64 of rank @xmath185 having signature function @xmath159 .
to construct @xmath64 we begin with the diagonal matrix @xmath186 of rank @xmath187 in which @xmath103 appears as a factor of exponent one of the first @xmath188 entries .
if this condition taken over all @xmath189 does not specify all the entries of @xmath190 , we make the remaining entries all 1 .
next , we change the sign of some of the diagonal entries to form @xmath191 so that the signature near 1 is @xmath183 .
this is possible , since by the previous lemma , @xmath192 .
to continue the modification , we must introduce a family of polynomials , @xmath193 , @xmath194 . for a dense set of @xmath195
, @xmath196 is a rational polynomial having its only root on the upper half circle at @xmath197 . for @xmath198
close to 0 , @xmath199 is positive and for @xmath73 close to @xmath69 , @xmath199 is negative .
if some of the diagonal entries of @xmath191 are multiplied by @xmath196 , the value of the signature is unchanged for @xmath200 .
the signature can change for values of @xmath201 .
however , jumps can continue to appear only at the roots of the @xmath103 , as well as , possibly , @xmath195 .
the construction of the desired form consists of making such modifications to create a form with signature function @xmath154 .
suppose that @xmath202 is a root of @xmath90 , one of the @xmath103 , corresponding to a nontrivial jump .
suppose also that the form @xmath203 has been constructed so that its signature function agrees with @xmath159 for all @xmath204 .
we want to alter @xmath203 , building @xmath205 , so that its signature function is unchanged for @xmath204 and has the same jump at @xmath59 as @xmath154 ; that is , @xmath206 .
pick a @xmath207 with @xmath208 small .
suppose that the first @xmath209 entries of @xmath203 are the ones divisible by @xmath90 , and that the number of remaining entries is @xmath210 .
if the desired form @xmath205 is to have a jump of @xmath211 at @xmath59 , then ( when evaluated at a point @xmath124 close to but less than @xmath59 ) the number of positive and negative entries in @xmath205 among the first @xmath209 diagonal entries must be @xmath212 and @xmath213 .
similarly , if the signature to the left of @xmath59 is to be @xmath214 , we must have the number of positive and negative entries among the last @xmath210 diagonal entries of @xmath215 be @xmath216 and @xmath217 .
( recall that the jump is determined by the first @xmath209 entries , since only those change sign as @xmath73 increases near @xmath59 . )
the desired sign distribution of the diagonal can be achieved by multiplying some of the diagonal entries by @xmath196 .
the only concern is that each of the numbers @xmath212 , @xmath213 , @xmath216 , and @xmath217 , must be nonnegative
. this will be the case as long as @xmath218 and @xmath219 , which is ensured by our initial choice of the dimension of @xmath186 to be @xmath187 .
( note that in the definition of @xmath187 one of the two maximum is over the numbers @xmath220 , and @xmath221 . )
observe that since @xmath214 is unchanged , no jump has been introduced at @xmath195 .
here we prove the main theorem : for @xmath9 , @xmath8 . for the given form @xmath1 , we apply theorem [ thmrealize ] to the function @xmath222 to find find a hermitian matrix @xmath64 of dimension @xmath5 .
if we denote the class represented by @xmath64 in @xmath37 by @xmath39 , then @xmath223 .
it follows that @xmath224 .
lemma [ lemmasig ] below then shows that @xmath225 represents and element of order 1 , 2 , or 4 , in @xmath37 .
thus , @xmath226 . since @xmath39 is constructed to have a representative of rank @xmath227 , clearly @xmath228 . on the other hand ,
if follows immediately from the definition of @xmath6 that @xmath229 for any @xmath9 .
thus we have @xmath230 .
the proof of the theorem is complete , given the next lemma .
[ lemmasig ] for a class @xmath9 , if @xmath231 , then @xmath232 .
background for the structure of witt groups is contained @xcite for symmetric bilinear forms . the specifics in the case of hermitian forms
are contained in @xcite .
a more complete description is in ranicki s book @xcite , with the details of the structure of hermitian forms over number rings presented in @xcite . for each symmetric irreducible @xmath233 $ ]
there is a homomorphism @xmath234/\left<\delta(t)\right>),\ ] ] defined as follows .
if @xmath9 is represented by a diagonal matrix @xmath64 with diagonal entries @xmath235 $ ] where each @xmath120 and @xmath121 is a symmetric irreducible polynomial prime to @xmath90 , then @xmath236 $ ] .
this induces split exact sequence @xmath237 ) \to w(\qq(t ) ) \to \oplus_\delta w(\qq[t , t^{-1}]/\left<\delta(t)\right>).\ ] ] according to @xcite ( see also @xcite for an elementary argument ) the inclusion @xmath238)$ ] is an isomorphism , so the previous sequence can be rewritten as @xmath239/\left<\delta(t)\right>).\ ] ] the field @xmath240/\delta(t)$ ] is an algebraic extension of @xmath241 , @xmath242 , a field with involution given by @xmath243 . denoting by @xmath244 the fixed field of the involution , an element in the witt group of hermitian forms over this @xmath242 is of finite order ( actually 4torsion ) if and only for all complex embeddings of @xmath242 that restrict to a real embedding of @xmath244 , the signature of the corresponding complex hermitian form is 0 .
this condition on the embedding implies that @xmath59 maps to a unit complex root of @xmath90 . since the jump function for @xmath1 is 0 at all unit roots of @xmath90 , it is then clear that the signature of @xmath245 is also 0 .
it follows that @xmath1 maps to an element of finite order in @xmath246/\left<\delta(t)\right>).$ ] in particular , @xmath247 ) \cong w(\qq)$ ] .
but any element in the image of @xmath33 has constant signature function , and in our case this implies that @xmath248 is represented by a class in @xmath33 with 0 signature .
but @xmath249 , where @xmath250 is the torsion subgroup of @xmath33 and satisfies @xmath251 .
it now follows that as desired @xmath232 , using the fact the exact sequence is split exact .
in order to compare @xmath6 as a bound on @xmath7 with bounds based on the maximum of the signature function , we want to view these functions as norms on a vector space . the function @xmath6 on @xmath37 is multiplicative : @xmath252 for @xmath253 .
thus @xmath6 induces a well - defined rational valued function on @xmath254 .
the same is not true for @xmath7 since it can be nonzero on torsion elements in @xmath37 , but we can define a stable version of @xmath7 by @xmath255 .
it follows then by theorem [ thmmain ] that @xmath256 , so @xmath257 also determines a well - defined rational valued function on @xmath258 .
if we define @xmath259 , then @xmath159 also defines a function on @xmath258 . for all @xmath260 , @xmath261 .
suppose the the maximum value of @xmath262 occurs at @xmath117 and @xmath263 is the largest value of a discontinuity that is less than @xmath117
. then @xmath264 .
the conclusion now follows from the definition of @xmath265 ( which equals @xmath266 ) .
recall that a norm on a vector space @xmath267 is a function @xmath268 satisfying @xmath269 for all @xmath270 , @xmath271 if and only if @xmath272 , and @xmath273 for all @xmath274 and @xmath1 .
an immediate consequence of lemma [ lemmasig ] is the following . both @xmath257 and @xmath159
are norms on @xmath258 .
if @xmath268 is a norm on a vector space @xmath267 , the unit ball of @xmath268 is defined by @xmath275 .
here we illustrate the strength of @xmath6 over basic signature bounds in determining the rank of a witt class .
we begin with a specific class , @xmath276 defined below .
we then expand on this to consider all linear combinations @xmath277 .
let @xmath278 and let @xmath279 .
these are the sixth and tenth cyclotomic polynomials , having roots at @xmath280 for @xmath281 and @xmath282 , respectively , on @xmath67 .
we let @xmath283 be the class in @xmath37 with diagonal representative @xmath284 $ ] and let @xmath285 be the class with diagonal @xmath286 $ ] .
the graphs of the signature functions of @xmath287 , and @xmath288 are illustrated in figure [ fig25 ] .
these signature functions occur for the knots @xmath289 , @xmath290 , and @xmath291 .
( the choice of signs simplifies some of the calculations that follow . ) , @xmath285 , and @xmath288.,title="fig : " ] .2 in , @xmath285 , and @xmath288.,title="fig : " ] .2 in , @xmath285 , and @xmath288.,title="fig : " ] the maximum absolute values of the signature for these three forms are seen to be @xmath292 , @xmath293 , and @xmath294 . in the first two cases we have the same result for @xmath6 : @xmath295 and @xmath296 .
however , for @xmath276 we have the set of jumps at the tenth roots of unity are @xmath297 and the signatures at the tenth roots of unity are @xmath298 .
thus , the sum of the two maximum is @xmath299 .
these calculations lead to the following theorem , where the knots @xmath300 and @xmath290 are as found in the tables at @xcite .
the witt rank of @xmath276 is @xmath301 .
in particular , the knot @xmath302 has 4genus 3 .
the algebraic statements are demonstrated in the discussion preceding the statement of the theorem .
for the geometric result it follows from the algebra that @xmath303 .
but it is known ( for example , see @xcite ) that @xmath304 and @xmath305 , so @xmath306 .
.1 in * comment * this topological result can be obtained by using ozsvth - szab invariants @xcite or khovanov - rasmussen invariants @xcite , which apply only in the smooth category . in the topological category , neither the murasugi nor the tristram - levine signatures @xcite can give this genus bound .
we now compute and compare the values of @xmath307 and @xmath159 on the span of @xmath283 and @xmath285 in @xmath258 .
both are determined by their unit balls .
the value of @xmath308 for @xmath309 is given by @xmath310 for the value of @xmath6 we sum the the maximum absolute value of the signature at the points @xmath311 and @xmath312 , and maximum for the jump function at those two points .
the result is : @xmath313 the unit balls for these norms are drawn in figure [ balls ] ; the larger region represents the @xmath159 ball , and the smaller hatched region is the @xmath257 ball . and @xmath257.,title="fig : " ] in this figure , we see that the point @xmath314 is in the unit @xmath159 ball ; as seen earlier , the @xmath294 . also , as we computed , @xmath314 is not in the unit @xmath257 ball , but @xmath315 is , since @xmath316 another interesting point in the diagram is @xmath317 .
the graph of the signature function of @xmath318 is illustrated in figure [ signgraph2 ] .
the maximum absolute value of the signature is @xmath319 , but the value of @xmath257 is @xmath320 . .,title="fig : " ] .1 in * knot theoretic comment * it follows from this calculation that @xmath321 .
a straightforward knot theoretic exercise in fact shows shows that @xmath322 .
this presents another example for which @xmath6 detects the 4genus of a knot , but signatures do not .
more interesting , in this case the ozsvth - szab and rasmussen - khovanov invariants are both insufficient to determine the 4genus ; both turn out to give a lower bound of 1 .
here we summarize the geometric background related to @xmath37 invariants of knots .
details can be found in such references as @xcite .
every smooth oriented knot @xmath323 bounds a smoothly embedded oriented surface @xmath324 .
there is a _ seifert _ pairing @xmath325 given by @xmath326 , where @xmath327 is the map @xmath328 given by pushing off in the positive direction and @xmath329 is the linking number .
a simple observation is that the intersection number of classes @xmath330 is given by @xmath331 . in particular ,
any matrix representation of @xmath267 has determinant @xmath332 .
suppose the genus of @xmath244 is @xmath333 , and @xmath334 , where @xmath335 is properly embedded in @xmath13 and is of genus @xmath336 .
then @xmath337 is a closed surface in @xmath13 , and it bounds an embedded 3manifold @xmath338 .
an argument using poincar duality shows that the kernel @xmath339 of the inclusion @xmath340 is of dimension @xmath341 .
since @xmath342 is a @xmath343 dimensional subspace of @xmath344 , which is of dimension @xmath345 , a simple linear algebra argument shows that @xmath346 is of dimension at least @xmath347 .
another simple geometric argument implies that @xmath267 vanishes on the subspace of @xmath348 .
if we now write @xmath267 for a matrix representation of the seifert pairing , the form @xmath349 defines a hermitian pairing on the rational function field . as above
, if @xmath11 bounds a surface of genus @xmath336 in @xmath13 , then this forms vanishes on a subspace in @xmath350 of dimension @xmath351 .
thus the form splits as a direct sum of forms , one of which is metabolic and of dimension @xmath352 ; the other summand is of dimension @xmath353 . | the witt rank @xmath0 of a class @xmath1 in the witt group @xmath2 of a field with involution @xmath3 is the minimal rank of a representative of the class . in the case of the witt group of hermitian forms over the rational function field @xmath4 , we define an easily computed invariant @xmath5 and prove that modulo torsion in the witt group
, @xmath6 determines @xmath7 ; more specifically , @xmath8 for all @xmath9 .
the need to determine the witt rank arises naturally in the study of the 4genus of knots ; we illustrate the application of our algebraic results to knot theoretic problems , providing examples for which @xmath6 provides stronger bounds on the 4genus of a knot than do classical signature bounds or ozsvth - szab and rasmussen - khovanov bounds . |
solutions of polymers with attractive groups ( associating polymers ) exhibit a wide range of rheological properties that can be controlled through variation of temperature and concentration .
these materials enjoy applications ranging from viscosity modifiers in food or oil recovery to adhesives and coatings ( see ref .
@xcite and references therein ) . under certain conditions ,
the attractive groups of the polymers associate to form a network and the system undergoes a sol
gel transition .
the physical bonds between the attractive groups are reversible and , depending on their strength , can break and reform frequently on experimental time scales . the properties of these so - called _ weak _ or physical gels differ markedly from chemical gels in which the polymers are interconnected through covalent bonds . compared to chemical gels , the current understanding of physical gelation is still limited and even controversial @xcite .
scenarios for thermoreversible gelation include the possibility of _ discontinuous _ gelation , in which the gelation is accompanied by sol gel phase separation , and _ continuous _ gelation .
the theoretical treatments of tanaka and stockmayer @xcite predict that continuous gelation is a thermodynamic phase transition , whereas semenov and rubinstein arrive at the opposite conclusion @xcite .
simulations can provide specific information that is not easily obtained otherwise .
on the one hand , the microscopic structure of physical gels is difficult to determine experimentally , whereas it can be directly accessed ( within the limitations of the model used ) by means of simulations . on the other hand ,
the numerical calculations permit testing of theoretical hypotheses and approximations .
two models are widely used to study associating polymers . in the first category of models , used in the above - mentioned theories ,
there are many association sites ( `` stickers '' ) distributed along the polymer chain .
this model was studied extensively in early ( off - lattice ) monte carlo simulations by groot and agterof @xcite .
kumar and panagiotopoulos @xcite have investigated the thermodynamic properties of a lattice - based version of this model by monte carlo simulations and did not find any indication that gelation is a thermodynamic phase transition .
more recently , it has been observed that the dynamical properties of this model are similar to those of weak glass formers , in which the diffusion coefficient is described by an expression with a vogel fulcher form @xcite .
the second category of models consists of _ telechelic _ chains , in which the associative sites are located at both chain ends and typically represented by a single monomer ( see ref .
@xcite for a concise overview ) . at low concentrations , these chains
have been predicted to form flower - like micelles @xcite .
simulations have indeed confirmed this @xcite and found that the associative groups are located in the core of the micelles and the non - associative groups in the corona . at higher concentrations ,
micelles can be connected by `` bridging '' polymers for which both associative endgroups belong to different micelles , leading to the formation of a micellar gel .
the dynamics in such a system are governed by the hopping rate of the associative groups between different micelles @xcite .
simulations have indeed found that the diffusion properties of such solutions can be described by an arrhenius law @xcite , as predicted by tanaka and edwards @xcite . if the polymer chains are less flexible , qualitative structural changes occur , as intra - chain pairing is suppressed and the formation of flower - like micelles becomes energetically unfavorable .
this promotes the formation of a network structure at low polymer concentrations @xcite .
however , the dynamical properties of such solutions of semiflexible telechelic chains seem not to have been investigated . recently , petka _ et al . _
@xcite have used genetic engineering techniques to create artificial proteins consisting of a hydrophilic group flanked by two stiff hydrophobic blocks .
this triblock copolymer was found to exhibit gelation in response to variation of ph or temperature .
its significance lies in the possibility to independently tune the strength of the endgroup attractions that are responsible for gelation and the solvent retention capability of the chains , which is essential for the formation of a swollen gel .
however , the actual structure of the hydrogel , which is formed at low polymer concentrations , could only be conjectured . motivated by the experimental findings
, we have employed molecular dynamics simulations to investigate the dynamic and structural properties of a solution of triblock copolymers that can be viewed as a greatly simplified , coarse - grained model of the artificial proteins .
this model evidently does not capture all relevant properties of the experimental systems , but rather should be viewed as a first attempt to determine the generic properties of a solution of triblock copolymers with two stiff endgroups .
in order to study the gelation of triblock copolymers we employ molecular dynamics simulations , using the dl_poly_2 code @xcite .
the polymers have an a - b - a structure , where the a - blocks are rigid hydrophobic rods and the b - block is hydrophilic and semiflexible . in our coarse - grained model , the solvent is modeled implicitly and each copolymer block is composed of spherical units ( `` monomers '' ) that represent an effective segment .
the total length of each chain is set to 15 units , consisting of three a - monomers per hydrophobic block and nine b - monomers in the hydrophilic block .
this choice is mostly based upon practical considerations .
a minimum of three units is required to represent a rod - like endblock , whereas a longer central block would pose equilibration problems , given the computationally accessible time scales .
monomers of type a interact via an attractive lennard - jones potential , @xmath0 \ ; , \label{eq : lj}\ ] ] whereas the interactions between monomers of type b and the interactions between unlike pairs are purely repulsive , @xmath1 we set @xmath2 and @xmath3 and cut off all interactions at @xmath4 . in order to express our results in reduced units we use @xmath5 and @xmath6 as units of energy and length , respectively . the reduced coupling , or inverse reduced temperature , @xmath7 is varied between 1 and 2 in the simulations .
the semiflexible character of the hydrophilic block is controlled by a harmonic angle - dependent potential , @xmath8 where @xmath9 and @xmath10 .
the value of @xmath11 was chosen in accordance with the model proposed in ref .
@xcite . in combination with the large value for @xmath12 it causes the chains to adopt an extended structure ( but without having the rod - like structure that would be obtained for @xmath13 ) , and gelation
is anticipated to occur at relatively low polymer concentrations .
the @xmath14-helical structure in the artificial proteins @xcite is mimicked by making the endblocks fully rigid .
all monomer units within a chain are connected via a harmonic bond potential @xmath15 where @xmath16 and @xmath17 .
the simulations are performed in the canonical ( @xmath18 ) ensemble , in a cubic box of linear dimension @xmath19 with periodic boundary conditions .
the total number of chains equals @xmath20 , corresponding to a monomeric packing fraction of only @xmath21 , i.e. , roughly twice the overlap threshold .
the temperature is controlled by means of the nos - hoover thermostat @xcite .
the equations of motion are integrated using a `` leap - frog '' verlet scheme @xcite , with a time step ( in reduced units ) @xmath22 . in all runs ,
the system is first equilibrated for four million steps ; for some low temperatures , even longer equilibration periods are used .
subsequently , 40 million time steps are carried out for high temperatures and 200 million time steps for low temperatures .
after the equilibration period , the configuration of the system is recorded every 1000 time steps for analysis of structural ( e.g. , chain conformations and percolation of the system ) and dynamical ( e.g. , single - chain diffusion ) properties .
in addition , the energies and the stress tensors are calculated and recorded every 100 time steps for the calculation of the specific heat and the stress autocorrelation function . despite the simplifications made in this coarse - grained model , the required simulation effort is still appreciable .
the total amount of cpu time corresponds to approximately 2.5 years on a single 2.0ghz intel xeon processor .
in order to characterize structural changes that take place in this system upon variation of the temperature , we employ an approach used in the study of micelle formation . for each configuration ,
bundles of endblocks are identified .
an endblock is considered part of a bundle if its center monomer lies within a distance @xmath23 from the center monomer of an endblock that is already part of the bundle .
our results turn out to be insensitive to the precise value of @xmath24 $ ] and we have chosen @xmath25 .
a configuration contains @xmath26 bundles of @xmath27 endblocks and the bundle - size distribution is defined as the thermal average @xcite @xmath28 figure [ fig : bundle_dist ] shows @xmath29 for five values of the inverse temperature @xmath30 , illustrating the formation of bundles of hydrophobic blocks as the temperature is lowered . at high temperature ( @xmath31 ) , the distribution function decays monotonically , with a single peak at @xmath32 ( isolated endblocks ) .
this corresponds to a regular solution of chains that are not associated . for lower temperatures ( @xmath33 )
, however , an additional `` shoulder '' appears in the distribution function , which develops into a peak that increases in height and shifts to larger bundle sizes if the temperature is further decreased .
this signals the formation of bundles in which large numbers of hydrophobic endblocks participate , resulting in important structural changes in the solution .
the occurrence of an inflection point in @xmath29 has been taken as a criterion for the critical micelle point @xcite . here , we associate the appearance of a shoulder in fig . [
fig : bundle_dist ] with the onset of the bundling process .
the corresponding characteristic inverse temperature @xmath34 lies between @xmath35 and @xmath36 .
a similar criterion was employed in ref .
@xcite [ but note that the quantity @xmath37 employed in this reference differs from @xmath29 ] , and a comparable temperature dependence was observed for the bundle - size distribution .
there is , however , a marked difference in the morphology of the bundles .
the flexible telechelic polymers studied in ref .
@xcite form flower - like micelles in which the chains take the shape of a loop .
both endgroups of each chain lie in the core of the same micelle and the central block lies in the corona @xcite . for more rigid chains ,
two effects occur .
the stiff endblocks have a tendency to align inside the bundle , giving it the appearance of a microcrystalline domain .
in addition , the semiflexible character of the central block prevents the chain from adopting a loop - like conformation .
this second effect is illustrated in fig .
[ fig : ring ] , in which we compare the average number of endblocks per bundle to the average number of endblocks in a bundle that belong to the same chain .
as shown , single - chain loops are essentially absent at all temperatures , confirming the prohibitively large energy penalty incurred by ring formation .
instead , each chain takes an extended conformation , with both hydrophobic endgroups participating in different bundles .
thus , even the formation of a continuous network becomes possible , in which the bundles of endblocks act as network junction points @xcite .
however , following the experimental observations in ref .
@xcite , we have chosen a markedly lower concentration than in earlier simulation studies . before investigating whether the formation of a network is nevertheless possible ,
we consider the energetic aspects of bundling . as illustrated in fig .
[ fig : cv ] , the specific heat @xmath39 exhibits a pronounced but relatively broad maximum around @xmath40 , corresponding to the creation of bundles of attractive endblocks .
the specific - heat maximum was found to occur at a temperature _ below _ the onset of micelle formation ( at @xmath41 ) in ref .
our data do not permit us to conclude this unambiguously .
indeed , for @xmath42 the simulations become almost prohibitively expensive , owing to the slow dynamic evolution of the system .
thus , the present system does not lend itself well to the application of finite - size scaling techniques for the determination of the nature and precise location of the bundling transition .
for example , in case of a continuous phase transition , the height of the specific - heat maximum will increase ( up to corrections to scaling ) as @xmath43 .
the exponent @xmath44 is typically rather small ( e.g. , @xmath45 for ising - type criticality @xcite ) , so that even doubling the number of chains would only increase the peak height by an amount comparable to the statistical accuracy of the data . in order to determine whether bundle formation indeed leads to the emergence of a connected network structure , we consider the percolation probability .
geometric percolation of the polymer chains in the solution is a necessary condition for gelation .
however , whereas _ chemical _ gelation coincides with the occurrence of geometric percolation @xcite , _ physical _ gelation has been suggested to take place only far below the percolation point @xcite .
we consider our polymer solution to be percolating if a connected path ( composed of chains that bridge the bundles of endblocks ) exists between any pair of opposite sides of the simulation cell .
the percolation probability , which is defined as the probability that a configuration is percolating , is plotted as a function of inverse temperature in fig .
[ fig : percolation ] .
the system always percolates for @xmath46 , i.e. , near the characteristic inverse temperature @xmath34 for bundle formation .
as the percolation probability certainly can exhibit strong finite - size effects , this determination must only be viewed as an estimate for the percolation threshold in the thermodynamic limit .
the observation that @xmath34 and the percolation threshold coincide reaffirms our interpretation that the telechelic chains become interconnected through bundle formation and form an spanning network . this behavior appears to differ from what has been observed for the solution of flexible telechelic chains studied in ref .
@xcite , which exhibits a comparable temperature dependence in the bundle - size distribution but is reported to exhibit geometric percolation at all temperatures .
figure [ fig : snapshot ] shows a typical configuration , obtained in a simulation performed at @xmath47 .
a network of interconnected hydrophobic junction points is indeed clearly discernable .
the onset of percolation affects the single - chain conformations as well .
this is illustrated by means of the temperature dependence of the end - to - end distance @xmath48 , see fig .
[ fig : re2 ] . for comparison ,
the figure also includes the end - to - end distance for an identical system in which the monomers in the endblocks experience a purely repulsive interaction [ see eq .
( [ eq : repulsion ] ) ] . whereas @xmath48 increases for both systems as the temperature
is lowered , the end - to - end distance increases more rapidly for the chains with attractive endblocks than for the purely repulsive chains . because of the semiflexible character of the chains , the _ relative _ change in @xmath48 is only several percent , but nevertheless the effect is clearly most pronounced for @xmath49 , i.e. , near the percolation threshold @xmath34 .
we ascribe it to the conformational changes induced by the network formation .
owing to the low polymer concentration , the bundles are relatively widely separated , forcing the connecting chains to adopt an extended conformation .
this observation is reinforced by considering the spatial correlations between endblocks .
indeed , the rod - like structure of the hydrophobic endblocks leads to an internal structure in the bundles that is absent in the models studied in refs .
figure [ fig : rdf ] shows the hydrophobe hydrophobe radial distribution function @xmath50 ( calculated from their center - of - mass separation ) at different values of @xmath30 . as the temperature
is decreased , two distinct features can be identified in this distribution function . the increasing maxima at short separations , which all lie at distances within the bundle size ( cf .
[ fig : bundle_dist ] ) , correspond to intra - bundle alignment of endblocks .
the emergence of this microcrystalline morphology can be understood from the fact that in an aligned bundle each endblock experiences a large number of monomer monomer interactions with surrounding endblocks . the cutoff distance employed in the lennard - jones potential ( [ eq : lj ] ) is larger than the maximum distance between monomers on fully - aligned ( close - packed ) endblocks , so that even a single pair of rods can have @xmath51 pair interactions . as shown in fig .
[ fig : ring ] , the average number of rods per bundle increases rapidly from approximately @xmath52 at @xmath53 to almost @xmath54 at @xmath55 , leading to tightly - bonded bundles .
it is this bonding that makes the resulting network resistant to external stress .
a second feature arises in fig .
[ fig : rdf ] at lower temperatures . as shown in the inset , an additional peak appears at a position that roughly coincides with the calculated average bundle separation , which varies from @xmath56 at @xmath53 to @xmath57 at @xmath47 .
thus , this peak characterizes the ordered arrangement of the bundles at low temperatures , and we conclude that the radial distribution function reflects the simultaneous emergence of both short - range and medium - range order upon cooling . in order to determine whether the structural changes observed in the triblock copolymer solution indeed correspond to gelation , we consider the dynamical properties as a function of temperature . evidently , bundle formation and the formation of a percolating network
are anticipated to have a strong influence on the diffusion properties of the polymers .
figure [ fig : msd ] shows the mean - square displacement of the center - of - mass of polymers at different values for @xmath30 . at high temperatures ,
we observe the standard behavior in which the dynamics cross over from ballistic motion at short times to diffusive motion at long times . at low temperatures ,
an intermediate regime appears where the dynamics are slowed down , indicative of the arrested dynamics resulting from network formation .
comparable observations were reported by kumar and douglas @xcite in a monte carlo study of a lattice model of an associating polymer solution and by bedrov _
et al . _
@xcite for micellar solutions , although it should be noted that in both studies the polymer concentration was considerably higher than in the current system ( which has @xmath58 ) and that in the micellar system the change in dynamic behavior was not associated with the formation of a network structure .
the dynamic behavior seen in the low - density gel is similar to that found in glass - forming materials , but the underlying mechanism is different .
the temporary localization of the triblock copolymers is caused by the strong intra - bundle interactions experienced by the endblocks , rather than by caging or jamming effects .
we also note that the width of the ballistic regime depends on the simulation model .
if water molecules are included explicitly or implicitly via a friction coefficient ( brownian dynamics ) , the ballistic regime may be rather narrow or even not observable at all .
for all investigated temperatures , the polymers eventually diffuse and we extract the diffusion coefficient @xmath59 from this long - time behavior , see fig . [
fig : diffusion ] . at low temperatures , a second ( narrow )
diffusive regime arises for times between the ballistic and the intermediate regime .
this corresponds to intra - domain motion of the endblocks .
the diffusion coefficient exhibits an exponential dependence on inverse temperature over the entire temperature range that we investigated , but two regimes can be discerned , separated near @xmath60 . since both percolation and micelle formation occur near this temperature , it is not possible to uniquely attribute the strong decrease of @xmath59 to either of these two phenomena . for the `` sticker '' model of ref .
@xcite , in which association sites are distributed along the polymer chain , the diffusion coefficient was found to be well described by a vogel fulcher law . in our system , the diffusive dynamics follow an arrhenius law , @xmath61 , where @xmath62 is the effective activation energy , similar to strong glass - formers ( cf . ref .
@xcite ) and micelle - forming telechelic polymers @xcite .
the arrhenius - type behavior suggests that dynamic relaxation , which takes place through the exchange of an endblock between two bundles ( network junctions ) , is controlled by an energetic barrier .
this barrier has a clear thermodynamic origin , namely the strong attraction between endblocks . from fig .
[ fig : diffusion ] , @xmath62 is estimated to be approximately @xmath63 in the low - temperature regime , consistent with our earlier estimate of the number of interacting monomers in a pair of hydrophobic rods .
interestingly , the similarity between the dynamic properties observed in this system and those of the micellar system studied in ref .
@xcite suggest that while the semiflexible character of the center blocks and the presence of rod - like hydrophobic endblocks change the structural properties of the solution , these differences do not qualitatively affect the dynamic behavior .
since gelation will be accompanied by a dramatic increase in viscosity @xmath64 , we compute this quantity by integrating the stress autocorrelation function @xmath65 @xcite , @xmath66 here , @xmath65 is defined as @xmath67 where @xmath68 is the volume of the system and @xmath69 assumes the values @xmath70 , @xmath71 , @xmath72 .
the average @xmath73 is taken over all time origins @xmath74 .
@xmath75 is the stress tensor of the system @xcite @xmath76 where @xmath27 is the monomer mass , @xmath77 is the @xmath14-component of the velocity of atom @xmath78 , @xmath79 is the @xmath80-component of the vector @xmath81 separating monomers @xmath78 and @xmath82 , and @xmath83 is the @xmath14-component of the force exerted on monomer @xmath78 by monomer @xmath82 .
the sum runs over all @xmath84 monomers . for the calculation of @xmath65
we employ a fast fourier transform @xcite , which accelerates the calculation by several orders of magnitude compared to the direct calculation method .
figure [ fig : viscosity ] shows @xmath64 as a function of @xmath30 . above the percolation threshold ( @xmath85 ) , the viscosity increases gradually with decreasing temperature .
however , in accordance with the behavior of the diffusion coefficient , @xmath64 starts to increase rapidly at the onset of percolation and micelle formation , and is described by an arrhenius law .
experimentally , gelation is characterized by the appearance of a plateau in the stress autocorrelation function @xmath65 . in our simulations
, we observe such a plateau for all temperatures below the percolation threshold .
the plateau extends to longer times upon decreasing temperature , but eventually @xmath65 decays to zero .
since we use the `` atomistic '' ( i.e. , monomer - based ) representation of the stress tensor @xmath86 in eq .
( [ eq : stress - corrfn ] ) , the results exhibit relatively large fluctuations .
figure [ fig : autocorr ] shows a representative example .
the plateau value @xmath87 is small , but clearly nonzero , as shown in the inset .
@xmath87 is found to be only weakly dependent on temperature , and no clear trend can be identified , implying that the rapid increase in @xmath64 ( fig .
[ fig : viscosity ] ) arises from an increase in relaxation time rather than from a variation in @xmath87 .
the integration ( [ eq : visc ] ) partially suppresses the statistical fluctuations present in @xmath65 , but the uncertainties in @xmath64 still reflect the computational challenges , in particular at low temperatures . while the strong increase in viscosity follows unambiguously from fig .
[ fig : viscosity ] , we emphasize that the largest relaxation times ( to be discussed below ) are still much smaller than the experimentally observed relaxation times for physical gels , which range from microseconds to seconds @xcite .
thus , only for ( computationally inaccessible ) temperatures far below the percolation threshold would the system investigated here undergo a sol
gel transition . following ref .
@xcite , we employ the non - gaussian parameter @xmath88 for the single - chain center - of - mass displacement @xcite , @xmath89 to estimate a characteristic time .
this parameter equals zero for both the ballistic and the diffusive regime . as shown in fig .
[ fig : nongaussian ] , @xmath88 increasingly deviates from zero as the temperature is decreased , reflecting the heterogeneous dynamics resulting from the hopping of endblocks between bundles .
also the maximum in @xmath88 , from which we extract a characteristic time @xmath90 , shifts to larger times upon cooling . at short and long times
, @xmath88 goes to zero , confirming the expected behavior for the ballistic and diffusive regimes , respectively . at intermediate times
, however , @xmath88 deviates from zero and the deviation increases as the temperature is lowered .
the positive deviation indicates the presence of anomalously fast chains , very similar to the heterogeneous dynamics observed in glass - forming liquids @xcite .
figure [ fig : tau ] displays @xmath90 as a function of @xmath30 , on a log linear scale .
we see that the values of @xmath90 , which extend up to approximately @xmath91 for the lowest temperatures , also follow an arrhenius law , with a similar change in slope as observed for the diffusion coefficient and the viscosity . the mean - square displacement corresponding to @xmath90 ( cf .
[ fig : msd ] ) is much smaller than the typical `` hopping distance '' or bundle separation ( cf .
[ fig : rdf ] ) , comparable to what is observed for glassy systems .
likewise , the observed values of @xmath90 are much smaller than the relaxation times that follow from the extent of the plateau in @xmath65 , which are approximately given by @xmath92 .
finally , we demonstrate the similarity of our dilute polymer solution to a glass - forming liquid by considering how well the stokes
einstein relation is obeyed . for a newtonian fluid
, we expect the product of the diffusion coefficient and the viscosity to obey @xmath93 , where @xmath94 is the hydrodynamic radius .
the product of @xmath59 and @xmath64 is plotted as a function of @xmath30 in fig .
[ fig : se ] .
the behavior of the end - to - end distance ( fig .
[ fig : re2 ] ) suggests that @xmath94 will change only weakly with temperature , and even may increase as the temperature is lowered . in combination with the linear temperature dependence of the numerator of the stokes
einstein relation ( i.e. , inverse dependence on @xmath30 ) , the steep _ increase _ in @xmath95 for @xmath96 indicates a clear breakdown of this relation once micellization and network formation set in .
the diffusion coefficient is _ larger _ than would be predicted on the basis of the viscosity shown in fig .
[ fig : viscosity ] . the precise microscopic origin of this behavior , however , remains to be determined .
as shown in the inset of fig .
[ fig : autocorr ] , this breakdown does _ not _ result from a temperature dependence in the modulus @xmath87 .
we have studied the structural and dynamical properties of a solution of associative a - b - a triblock copolymers .
the semiflexible character of the center ( b ) blocks , in combination with the attraction between endgroups , allows these polymers to form a gel at remarkably low concentrations .
the molecular dynamics simulations presented here form a natural extension of earlier work on semiflexible chains @xcite , which however only addressed structural properties .
in addition , we observe dynamic effects that bear close resemblance to those reported in ref .
@xcite for a lattice - based model studied by monte carlo simulations and to those reported in ref .
@xcite for micelle formation .
however , the solutions in either of these studies had a significantly higher polymer concentration .
furthermore , the micelle solution was found to exhibit geometric percolation at all temperatures and the dynamic changes were linked to the thermodynamic micelle transition .
we observe that all dynamic changes are correlated with micellization and the simultaneous emergence of a percolating network of polymer chains , in which bundles of rigid endblocks act as network junctions . upon a further decrease in temperature ,
the hydrophobic blocks tend to align within a bundle , forming a microcrystalline - like structure .
the resulting strong binding of the chains is responsible for the mechanical stability of the gel - like network . on a larger scale ,
the bundles distribute more regularly at lower temperature , as indicated by the appearance of peak at the average bundle separation in the hydrophobe
hydrophobe radial distribution function .
the change in the dynamical behavior of the solution that occurs upon micellization and network formation is reflected in the diffusion constant , the viscosity and the maximum in the non - gaussian parameter .
the temperature dependence of all these properties changes near the percolation threshold and is well described by an arrhenius law , similar to what is observed for strong glass formers .
the activation barrier in our system has a clear _ thermodynamic _ origin , namely the strong attraction between endblocks that are part of the same microdomain or `` bundle . ''
thus , there are similarities with the diffusion of diblock copolymers in a lamellar phase , which also exhibits an exponential decay with temperature @xcite .
recently , this type of dynamics has attracted attention in the context of slow dynamics in systems with frustration - limited domains ( see ref .
@xcite and references therein ) .
finally , we note that , while we observe a finite plateau in the stress autocorrelation function , as would be expected for a gel - forming material , the dynamics in our systems are still faster than in actual experimental gels .
thus , the sol gel transition only occurs at a temperature far below the percolation threshold .
helpful comments by ken schweizer are gratefully acknowledged .
this work is supported by the american chemical society petroleum research fund under grant no .
38543-g7 and by the national science foundation through an itr grant ( dmr-03 - 25939 ) via the materials computation center at the university of illinois at urbana - champaign .
access to computational facilities at oak ridge national laboratory ( via grant no .
cnms2003 - 005 at the center for nanophase materials sciences ) was supported by the u.s .
department of energy , under contract de - ac05 - 00or22725 with ut - battelle , llc . | molecular dynamics simulations have been employed to study the formation of a physical ( thermoreversible ) gel by amphiphilic a - b - a triblock copolymers in aqueous solution . in order to mimic the structure of hydrogel - forming polypeptides employed in experiments [ w.a .
petka _
et al . _ ,
science * 281 * , 389 ( 1998 ) ] , the endblocks of the polymer chains are modeled as hydrophobic rods representing the alpha - helical part of the polypeptides whereas the central b - block is hydrophilic and semi - flexible .
we have determined structural properties , such as the hydrophobic cluster - size distribution function , the geometric percolation point and pair correlation functions , and related these to the dynamical properties of the system . upon decrease of the temperature , a network structure
is formed in which bundles of endblocks act as network junctions .
both at short and medium distances an increased ordering is observed , as characterized by the pair correlation function .
micelle formation and the corresponding onset of geometric percolation induce a strong change in dynamical quantities , e.g. , in the diffusion constant and the viscosity , and causes the system to deviate from the stokes
einstein relation .
the dynamical properties show a temperature dependence that is strongly reminiscent of the behavior of glass - forming liquids .
the appearance of a plateau in the stress autocorrelation function suggests that the system starts to exhibit a solid - like response to applied stress once the network structure has been formed , although the actual sol gel transition occurs only at a considerably lower temperature . |
the observations of isomers in the interstellar medium is becoming increasingly more common as new and larger molecules are being detected .
currently , 135 molecules have been detected in interstellar and circumstellar environments .
excluding diatomic molecules , @xmath1030% of all interstellar molecules have observed isomeric counterparts . in addition , as the number of atoms in an interstellar molecule increases , so do the number of isomers . for example , of the 9 interstellar molecules containing 8 atoms , 5 are isomers . with an empirical formula c@xmath1h@xmath2o@xmath1 , acetic acid ( ch@xmath0cooh ) , glycolaldehyde
( ch@xmath1ohcho ) and methyl formate ( ch@xmath0ocho ) are isomers ; and for h@xmath1c@xmath11 , hexapentaenylidene and triacetylene ( hc@xmath11h ) are isomers .
isomerism may provide important clues to the formation of interstellar molecules .
as such , an important class of isomers are cyanide ( -cn ) and isocyanide ( -nc ) species .
hydrogen cyanide ( hcn ) , methyl cyanide ( ch@xmath0cn ) , vinyl cyanide ( ch@xmath1chcn ) , and ethyl cyanide ( ch@xmath0ch@xmath1cn ) all have large abundances and are easily detectable in a number of interstellar environments and are readily seen in the spectra of several molecular line surveys ( e.g. friedel et al.2004 , nummelin et al .
however , while there have been detections of hnc in interstellar and cometary environments , there have been no confirmed detection of the isocyanides ch@xmath0nc , ch@xmath1chnc , and ch@xmath0ch@xmath1nc .
the first tentative detection of ch@xmath0nc was toward sgr b2(oh ) by cernicharo et al .
( 1988 ) from observations of the @xmath12 , @xmath13 , and @xmath14 transitions . though the lines were highly contaminated from other known interstellar species , they calculated an ch@xmath0nc / ch@xmath0cn abundance ratio of @xmath100.03 - 0.05 . since this initial claim
, there have been no further observations to confirm the existence of interstellar ch@xmath0nc , in particular toward the lmh hot core of sgr b2(n ) , the best known source of large molecules .
thus , in this work , we confirmed the existence of ch@xmath0nc and attempted to observe other large isocyanides by observing the same ( or similar ) low energy transitions of the cyanides ch@xmath0cn , ch@xmath1chcn , ch@xmath0ch@xmath1cn , and hc@xmath2cn and their corresponding isocyanide isomers toward the the high - mass star forming region , sgr b2(n ) .
the observations using the nrao 100-m robert c. byrd green bank telescope ( gbt ) were made in 2004 february 24 , february 29 and march 29 .
table 1 lists the molecular line parameters of the transitions of the cyanides and isocyanides sought .
the species , calculated rest frequencies , transition quantum numbers , line strengths , upper level energies , and continuum levels are listed in the first six columns of table 1 .
the gbt half - power beamwidth ( @xmath15 ) is given as 740@xmath16/@xmath17 [ ghz ] .
the k - band receiver set used has a frequency range covering 18 ghz to 22.5 ghz . for the observations reported herein
, the average beamwidth @xmath18=39@xmath16 corresponds to an average gaussian beamsize of @xmath19=@xmath20/[4 ln(2 ) ] = 1723 arcsec@xmath9 and an average beam efficiency @[email protected] .
the gbt spectrometer was configured in its 8 intermediate frequency ( if ) , 200 mhz , 3-level mode which permits observing four 200 mhz frequency bands at a time in two polarizations through the use of offset oscillators in the if .
this mode affords 24.4 khz channel separation .
antenna temperatures are on the @xmath22 scale ( ulich & haas 1976 ) with estimated 20% uncertainties .
the sgr b2(n - lmh ) j2000 pointing position employed was @xmath23= 17@xmath2447@[email protected] , @xmath27 = -28@xmath2822@xmath2917@xmath16 and the lmh systemic lsr source velocity of + 64 km s@xmath30 was assumed .
automatically updated dynamic pointing and focusing corrections were employed based on realtime temperature measurements of the structure input to a thermal model of the gbt ; zero points were adjusted typically every two hours or less using the calibrators 1626 - 298 and/or 1733 - 130 .
data were taken in the off - on position - switching mode , with the off position 60@xmath29 east in azimuth with respect to the on source position . a single observation consisted of two minutes in the off source position followed by two minutes in the on source position .
the two polarization outputs from the spectrometer were averaged in the final data reduction process to improve the signal - to - noise ratio .
figure 1 shows our attempt to detect cyanide and isocyanide species toward sgr b2(n ) with the gbt .
the panels in the left column show the detections of ch@xmath0cn , ch@xmath1chcn , ch@xmath0ch@xmath1cn and hc@xmath2cn .
we have successfully detected all species in the left column and for the first time the @xmath31 transition of interstellar ch@xmath0ch@xmath1cn is reported . in the panels containing the transitions of ch@xmath0cn , ch@xmath1chcn and ch@xmath0ch@xmath1cn , the intensity scale ( ordinate ) decreases , showing that the emission from the larger species is in general weaker . the abscissa is the radial velocity with respect to the local standard of rest ( lsr ) calculated for the rest frequency of the transition shown at the top left of each panel assuming a source velocity of + 64 km s@xmath30 .
dashed lines show lsr velocities of + 64 , + 73 and + 82 km s@xmath30 .
molecules with no hyperfine ( hf ) structure previously observed with the gbt such as propenal ( ch@xmath1chcho ) , propanal ( ch@xmath0ch@xmath1cho ) , and glycolaldehyde ( ch@xmath1ohcho ) show absorption components at + 64 and + 82 km s@xmath30 and an emission component at + 73 km s@xmath30 ( hollis et al .
2004a , b ) . in general ,
the cyanide spectra shown in the left panel of figure 1 are consistent with this source velocity structure . in the cases of ch@xmath0cn and ch@xmath1chcn , however , the source velocity structure is confused by molecular hf splitting .
the fiducial on each hf component denotes the position corresponding to + 64 km s@xmath30 .
thus , it can be seen that only the f=0 - 1 components are clear of any complication .
the panels in the right column of figure 1 show the passbands containing the transitions of ch@xmath0nc , ch@xmath1chnc , ch@xmath0ch@xmath1nc and hc@xmath2nc .
we clearly detect the absorption and emission components at + 64 , + 73 and + 82 km s@xmath30 from ch@xmath0nc and there is possible narrow emission from ch@xmath1chnc near + 64 km s@xmath30 .
the primary absorption component of ch@xmath0nc at + 64 km s@xmath30 is more than an order of magnitude weaker in intensity than the corresponding feature of ch@xmath0cn , however this is the first interstellar report of this line .
no emission or absorption is seen from either ch@xmath0ch@xmath1nc or hc@xmath2nc beyond a 1 @xmath32 detection limit ( @xmath10 2 mk ) .
the @xmath4 ( @xmath5 ) ch@xmath0nc observations using the berkeley - illinois - maryland association ( bima ) millimeter array were made in 2001 april 20 in its c configuration toward sgr b2(n - lmh ) .
the pointing position was the same as used for the gbt .
table 2 lists the molecular line parameters of each species in the ch@xmath0nc passband .
the species , the calculated rest frequencies , the transition quantum numbers , the product of the line strength and the square of the dipole moment ( @xmath33 ) , and the upper level energies are listed in the first five columns of table 2 .
the synthesized beamwidth for these observations is @xmath34@xmath35@xmath36 = 23.@xmath169@xmath357.@xmath161 corresponding to a gaussian beamsize of @xmath8=192 arcsec@xmath9 .
the quasar 1733 - 130 was used to calibrate the antenna based gains and uranus was used as the flux density calibrator .
the absolute amplitude calibration of these sources is accurate to within @xmath1020% .
the passbands were calibrated automatically during data acquisition ; a technical description of auto - calibration can be found at http://astron.berkeley.edu/@xmath10plambeck/technical.html .
the spectral window containing the ch@xmath0nc transitions has a bandwidth of 50 mhz and is divided into 128 channels for a spectral resolution of 0.39 mhz per channel .
all data were combined , imaged , and self - calibrated using the miriad software package ( sault , teuben & wright 1995 ) . figure 2 shows fiducials for the @xmath4 ( @xmath5 ) transitions of ch@xmath0ncin a passband dominated by strong emission from ch@xmath0ch@xmath1cn which is largely confined to the lmh hot core ( liu & snyder 1999 ) . the rest frequency located in the upper left of
the panel corresponds to the @xmath37 transition of ch@xmath0ch@xmath1cn for an assumed lsr velocity of + 64 km s@xmath30 .
dashed lines denote lsr velocities of + 64 , + 73 and + 82 km s@xmath30 .
the transition of ch@xmath0ch@xmath1cn clearly shows strong emission components at the lmh systemic velocity of + 64 km s@xmath30 and also at + 73 km s@xmath30 .
the @xmath38 ( @xmath39 ) ch@xmath0nc transitions are labeled in this passband assuming an lsr velocity of + 64 km s@xmath30 and , except for the @xmath40 transition , all lie in a region that is clear from any other molecular line emission . however , no transition of ch@xmath0nc is detected above a 1 @xmath32 detection limit ( @xmath100.3 jy beam@xmath30 ) .
the 1 @xmath32 rms noise level fiducial is shown at the left of the panel .
we have presented two disparate sets of observations : gbt single antenna results with a gaussian beamsize of @xmath8=1723 arcsec@xmath9 , and bima array results obtained with a gaussian beamsize of @xmath8=192 arcsec@xmath9 . given the gbt beamsize is a factor of 9 greater than the bima beamsize , the spatial scale of the gbt will couple better to the extended emission toward sgr b2(n ) .
conversely , the bima array will tend to couple better to the 28 arcsec@xmath9 ( @xmath105@xmath16 diameter ) of the lmh hot core .
hence , we expect that the gbt observations will be characterized by a lower excitation temperature ( t@xmath41 ) than the excitation temperature associated with the bima observations .
first , we will derive the cyanide and isocyanide column densities from the gbt observations .
then , we will show from bima observations that ch@xmath0nc must be extended . for the gbt observations , following the formalism outlined in hollis et al .
( 2004b ) , the total beam - averaged column density of a molecular species obtained with a single element radio telescope is given as : in both equations , the line shapes are assumed to be gaussian ; @xmath44 is the telescope beam efficiency ; t@xmath41 is the excitation temperature ; @xmath45 is the product of the fitted line intensity ( mk ) and line width ( km s@xmath30 ) ; @xmath46 is the rotational partition function ; @xmath33 is the product of the transition line strength and the square of the dipole moment ( debye@xmath9 ) ; and @xmath47 is the upper rotational energy level ( k ) . in equation
( 1 ) , @xmath17 is the transition frequency ( mhz ) and @xmath482.7 k is the cosmic background temperature . in equation ( 2 ) , @xmath49 is the source continuum temperature ( k ) , and @xmath50 is the lower rotational energy level ( k ) . using the gbt at k - band , hollis et al .
( 2004b ) found a glycolaldehyde excitation temperature of @xmath51=8 k that satisfied the detected integrated line intensities of the absorption components at + 64 and + 82 km s@xmath30 and the emission component at + 73 km s@xmath30 . since the lsr velocity components of the cyanide and isocyanide molecules in figure 1 appear to be similar , we used @xmath51=8 k to find the total column density of each of the detected species .
columns ( 7 ) to ( 12 ) of table 1 summarize the gaussian fitting intensities and line widths for each detected transition or 3 @xmath32 upper limits for the intensities of non - detections at lsr velocities of + 64 , + 73 and + 82 km s@xmath30 . in order to determine the intensity and line width of the individual hf and velocity components of the ch@xmath0cn and ch@xmath1chcn lines , multiple gaussian fits
were used . by fixing the frequency separation between hf components
, we fit the + 64 , + 73 and + 82 km s@xmath30 velocity components for each hf component .
if the fitting routine found a solution for the intensity of a velocity component less than the 3 @xmath32 rms noise level , it was omitted , and the fit recalculated for one less velocity component .
the results are listed in columns 7 - 12 of table 1 . in figure 1a ,
the solid line is the result of gaussian fitting the 3 hf structure components to the + 64 , + 73 and + 82 km s@xmath30 lsr velocity structure ( see table 1 ) .
furthermore , as a residual to this methodology , the small absorption dip near 50 km s@xmath30 in figure 1a is not due to a hf component nor is it a velocity component of ch@xmath0cn .
the column densities listed in column ( 13 ) of table 1 are the sum of all detected velocity components , or , in the cases of ch@xmath0cn and ch@xmath1chcn , of all velocity and hf components .
the column densities of the cyanide molecules are within a factor of 3 or 4 of the ch@xmath0cn column density of ( @xmath1010@xmath52 @xmath53 ) .
furthermore , the column density measured from the + 64 km s@xmath30 velocity component accounts for @xmath1075% of the total ch@xmath0cn column density .
the + 64 km s@xmath30 velocity component is often the only one resolved at higher frequencies and is the basis for the majority of ch@xmath0cn column density measurements at 1 and 3 mm wavelengths .
the similarity in column densities between cyanides is also seen at higher resolution toward other sources ( remijan et al .
the column density ratio ch@xmath0nc
/ ch@xmath0cn is @xmath100.02 and is relevant if and only if the two species are co - spatial .
the possible detection of ch@xmath1chnc gives a column density ratio ch@xmath1chnc
/ ch@xmath1chcn of @xmath100.005 . from remijan
et al . ( 2003 ) , the total beam averaged column density obtained with an interferometer is : in equation ( 3 ) , @xmath55 is the product of the fitted line intensity ( jy beam@xmath30 ) and line width ( km s@xmath30 ) , @xmath34 and @xmath36 are the fwhm beam widths ( arcsec ) and @xmath56 $ ] is the beam filling factor where @xmath8 is the solid angle subtended by the synthesized beam of the interferometer and @xmath57 is the solid angle subtended by the source emission . for the bima array data , @xmath34@xmath35@xmath36 = 23.@xmath169@xmath357.@xmath161
. columns ( 7 ) to ( 12 ) of table 2 summarize the gaussian fitting intensities and line widths for ch@xmath0ch@xmath1cnand the 3 @xmath32 upper limits for the intensities of the non - detections at lsr velocities of + 64 , + 73 and + 82 km s@xmath30 .
the column density of ch@xmath0ch@xmath1cn in column ( 13 ) is based on the excitation temperature of methanol ( ch@xmath0oh ) of 170 k ( pei , liu & snyder 2000 ) .
no emission for the @xmath12 @xmath58 transition of ch@xmath0nc was seen in the bima array sgrb2(n ) data .
however , if ch@xmath0nc and ch@xmath0cn are co - spatial ( i.e. , our gbt ch@xmath0nc / ch@xmath0cn ratio of @xmath100.02 is valid ) , then observations of ch@xmath0cn by other investigators which yield large ch@xmath0cn column densities predict that we should have detected ch@xmath0nc with the bima array if the source is compact . in an extensive survey encompassing the @xmath59 , @xmath60 and @xmath61 transitions of ch@xmath0cn at 1 mm wavelengths , nummelin et al .
( 2000 ) obtained a ch@xmath0cn beam averaged column density of @xmath104.2@xmath3510@xmath62 @xmath53 for a 352 k lmh hot core temperature .
these observations were performed using the 15 m sest telescope with a @xmath10600 arcsec@xmath9 beamsize .
owing to the lower spatial resolution ( @xmath102 km s@xmath30 channel@xmath30 ) at 1 mm wavelengths and the large line widths ( 15 - 20 km s@xmath30 ) of ch@xmath0cn lines toward sgr b2(n - lmh ) ( nummelin et al .
2000 ) , it is not possible to resolve the individual velocity components of the ch@xmath0cn lines .
thus , the calculated column density is based solely on the + 64 km s@xmath30 lsr velocity component .
in addition to the the column density and temperature , nummelin et al . also found a beam filling factor of the ch@xmath0cn emission toward the sgr b2(n ) hot core of 0.014 . in their data analysis ,
the column density , rotational temperature and beam filling factor were given as free parameters .
from these values , an integrated line intensity was calculated for each transition and was compared to the observed integrated line intensity until a `` best - fit '' model was obtained .
the best - fit model was found by finding the minimum of the reduced @xmath63 function ( nummelin et al .
2000 ) . assuming the model explains the physics accurately , the measured values of the rotational temperature , column density and beam filling factor by nummelin at al.toward the sgr b2(n ) hot core are reported at a 68.3% confidence interval ( 1 @xmath32 ) . given the @xmath10600 arcsec@xmath9 beamsize of the 15 m sest telescope , and a beam filling factor of 0.014 , the measured source size of the ch@xmath0cn emitting region toward the sgr b2(n ) hot core is @xmath108.5 arcsec@xmath9 .
this size is very similar to the ch@xmath0cn cores toward w51e1 ( 7.9(6)arcsec@xmath9 ) and w51e2 ( 6.0(9)arcsec@xmath9 ) ( remijan et al .
furthermore , the size is also similar to the emitting regions of ch@xmath1chcn and ch@xmath0ch@xmath1cn towards the sgr b2(n ) hot core measured by liu & snyder ( 1999 ) at subarcsecond resolution .
thus , @xmath108.5 arcsec@xmath9 is a reasonable measure of the ch@xmath0cn emitting region toward the sgr b2(n ) hot core . using the measured rotational temperature of 352 k , the scaled nummelin et al .
( 2000 ) ch@xmath0cn column density within the 192 arcsec@xmath9 beamsize of the bima array is @xmath109.9@xmath3510@xmath64
using the column density ratio of 0.02 , the expected ch@xmath0nc column density of the + 64 km s@xmath30 lsr velocity component should be @xmath102@xmath3510@xmath62 @xmath53 for co - spatial isomers .
solving equation ( 3 ) for @xmath55 assuming a column density of @xmath102@xmath3510@xmath62 @xmath53 , we should have detected a ch@xmath0nc integrated line intensity of @xmath1057 jy beam@xmath30 km s@xmath30 for the @xmath65 transition .
assuming a typical line width of 15 km s@xmath30 for the lmh hot core of sgr b2(n ) , the expected peak intensity of the @xmath66 transition is @xmath103.8 jy beam@xmath30 .
however , the bima array data shows no ch@xmath0nc emission greater than the 1 @xmath32 rms upper limit of @xmath100.3 jy beam@xmath30 .
the most likely explanation for the lack of a ch@xmath0nc detection is that the emission is extended and the array can not couple to it . in the case of extended emission ,
the ratio of the measured line intensity , or in our case the 1 @xmath32 rms upper limit of @xmath100.3 jy beam@xmath30 , to the expected line intensity of @xmath103.8 jy beam@xmath30 is equal to the ratio of the solid angle subtended by the synthesized beam of the interferometer ( @xmath8 ) to the solid angle subtended by the source emission ( @xmath57 ) .
this is because the synthesized beam of the interferometer is only collecting a fraction of the flux from the extended emission region which we assume to be uniformly distributed .
therefore our observations indicate that @xmath8/@xmath57 @xmath67 0.3/3.8 . solving for the source size assuming a gaussian beamsize of 192 arcsec@xmath9
, we find @xmath57@xmath62432 arcsec@xmath9 or equivalently , a linear size of @xmath6100@xmath16(1.7@xmath29 ) , which is much larger than the typical lmh hot core diameter of @xmath105@xmath16 ( hollis et al .
this strongly suggests that the spatial distribution of ch@xmath0nc is extended and would not show any significant concentration toward the lmh hot core .
defrees , mclean , & herbst ( 1985 ) theoretically calculated a ch@xmath0nc / ch@xmath0cn abundance ratio for dense interstellar clouds in the range of 0.1 - 0.4 by assuming that both species are formed from ion precursors and destroyed by gas phase , ion molecule reactions .
the defrees et al .
calculations show that when ch@xmath0@xmath68 collides with hcn , a collision complex is formed that rapidly equilibrates to protonated methyl isocyanide ( ch@xmath0nch@xmath68 ) and protonated methyl cyanide ( ch@xmath0cnh@xmath68 ) due to the large amount of energy produced from the initial collision .
as the ensemble of ions continues to relax , the internal energy of ch@xmath0cnh@xmath68 becomes insufficient to overcome the barrier to isomerize to ch@xmath0nch@xmath68 .
the end result finds ch@xmath0cn formed preferrentially over ch@xmath0nc .
we performed ab initio calculations in order to quantify the energy differences between cyanide and isocyanide isomers in this work .
table 3 lists each molecular species ( col . 1 ) , the zero - point corrected bonding energy ( zpcbe in col . 2 ) and the relative zpcbe ( col . 3 ) .
column ( 2 ) zpcbe represents the complete dissociation of the molecule into its constituent atoms . note that the relative bonding energies between isomers are very large ( @xmath68500 @xmath7 ) .
thus , the cyanide molecules are much more stable than their isocyanide isomers , and therefore it is highly unlikely that any thermal process can satisfactorily explain the formation of isocyanide molecules .
this indicates that a non - thermal process is necessary to convert a cyanide into an isocyanide species . in recent laboratory experiments performed by hudson & moore ( 2004 ) ,
cyanide species were bombarded by both protons and uv - photons to determine what would happen to the structure of the molecule . bombarding a pure ice sample of ch@xmath0cn at 24 k , the proton experiment yielded ketenimine ( ch@xmath1cnh ) , another isomer of ch@xmath0cn , as well as ch@xmath0nc in a slightly lower abundance .
similarly , the uv - photon experiment at 12 k yielded ch@xmath0nc as the primary product .
these products were also seen when ch@xmath0cn was embedded in n@xmath1 ice and then irradiated .
hudson & moore ( 2004 ) also note that the temperature in both experiments did not affect the formation of the end products . in an uv - photon experiment performed at 100 k ,
the destruction of ch@xmath0cn was faster than at 12 k , but the same end products were seen .
this suggests that the reactions were driven more by the radiation dose than the sample temperature ( hudson & moore 2004 ) , reinforcing our suggestion that non - thermal processes may be the primary route to the formation of interstellar isocyanides .
laboratory experiments involving ch@xmath1chcn , ch@xmath0ch@xmath1cn , and cyanoacetylene ( hc@xmath1cn ) , all showed similar results for isocyanide isomers .
it is likely that a similar reaction involving hc@xmath2cn will produce hc@xmath2nc .
furthermore , these laboratory experiments also suggest that all isocyanides are likely to be formed under the conditions which interstellar ch@xmath0nc was detected .
based on the column density ratios and upper limits in table 1 , we did not detect any emission from ch@xmath0ch@xmath1nc and hc@xmath2nc because of the possibility that we did not integrate long enough to detect these species .
on the other hand , hudson & moore ( 2004 ) also conducted experiments with cyanides embedded in h@xmath1o ice and found that hydrogen addition and subtraction reactions were taking place in both radiation bombardment reactions in which _ no _ isocyanides were formed . in these experiments
hc@xmath1cn hydrogenated to ch@xmath1chcn and then to ch@xmath0ch@xmath1cn .
based on the aforementioned laboratory results , we will suggest how large cyanides and their corresponding isocyanide isomers may be produced in the extended regions surrounding sgr b2(n ) .
the sgr b2(n ) region is known to contain widespread shocks ( chengular & kanekar 2003 ) .
such shocks have been presumed to be responsible for the formation and distribution of large aldehyde molecules toward this region ( hollis et al .
2004a , b ) .
furthermore , it is also well known that there is ambient uv flux and cosmic ray flux in these regions . on the surfaces and mantles of dust grains , larger species such as ch@xmath0cn and hc@xmath1cn
may be present .
ch@xmath1chcn and ch@xmath0ch@xmath1cn are formed as a result of h - addition reactions , thus depleting hc@xmath1cn . as a shock passes ,
these large molecules are expelled from the grain surface into the gas phase . at this point
uv and cosmic rays may interact with the cyanide species and form isocyanides . as a result
, the isocyanides show an extended distribution . in hot cores ,
there is a lower uv and cosmic ray flux due to the large column density of gas and dust .
furthermore , thermal processes are the primary mechanism to liberate molecules off grain surfaces ( tielens & hagen 1982 ; hasegawa , herbst , & lueng 1992 ) .
indeed , this is initially why these regions were believed to be the ideal places to form large molecules ( ehrenfreund & charnley 2000 ) .
also , in these compact regions , it is believed warm gas phase chemistry in addition to grain surface chemistry can form large molecules which may account for the enhanced abundances of ch@xmath1chcn and ch@xmath0ch@xmath1cn in hot core regions ( liu & snyder 1999 , see also ch@xmath0ch@xmath1cn column density in table 2 with respect to table 1 ) .
however , thermal processes are not enough to form isocyanide species after the cyanides have been expelled from the grain and hence there should be a depletion of isocyanides towards hot cores .
our observations support this hypothesis . in summary
, we have calculated the chemical bonding energies for cyanide molecules studied herein ( ch@xmath0cn , ch@xmath1chcn , ch@xmath0ch@xmath1cn , & hc@xmath2cn ) and found they are much greater ( @xmath68500 @xmath7 ) than that of the corresponding isocyanide isomers . as a consequence ,
cyanide interstellar molecules are most likely preferentially formed at the expense of isocyanide isomers . moreover , in the case of ch@xmath0cn , it is seen in the lmh hot core and in the surrounding medium while ch@xmath0nc appears to be deficient in the lmh hot core .
this suggests that large - scale , non - thermal processes such as shocks or enhanced uv flux in the surrounding medium account for the conversion of ch@xmath0cn to ch@xmath0nc while the lmh hot core is dominated by thermal processes which can not produce a significant amount of ch@xmath0nc .
hudson & moore ( 2004 ) have shown that radiation bombardment of ch@xmath0cn in ice analog experiments can produce ch@xmath0nc , thus tending to support what is observed .
additionally , since ch@xmath0nc is deficient in the lmh and ch@xmath0cn is prevalent , this suggests that the relative abundance ratio between these two isomers is not meaningful because the molecules are not co - spatial .
this may well apply to other cyanide and isocyanide isomers .
j.m.h . gratefully acknowledges research support from h.a .
thronson , assistant associate administrator for technology , nasa science mission directorate .
we also greatly thank an anonymous referee whose comments and suggestions provided additional insight to this manuscript .
lccccccccccccc ch@xmath0cn & 18396.7252(7 ) & 1@xmath70 - 0@xmath70 & 1.332 & 0.88 & 10.6 & -275(25 ) & 5.8(7 ) & 160(20 ) & 7.9(9 ) & @xmath677 & ... & & + & & @xmath71 & & & & & & & & & & & + & 18397.9965(6 ) & 1@xmath70 - 0@xmath70 & 2.220 & 0.88 & 10.6 & -400(15 ) & 7.0(3 ) & 310(19 ) & 5.9(4 ) & @xmath677 & ... & 10.8(20 ) & + & & @xmath72 & & & & & & & & & & & + & 18399.8924(3 ) & 1@xmath70 - 0@xmath70 & 0.444 & 0.88 & 10.6 & -180(25 ) & 8.3(9 ) & @xmath677 & ... & @xmath677 & ... & & 0.02(2 ) + & & @xmath73 & & & & & & & & & & & + ch@xmath0nc & 20105.754(1 ) & 1@xmath70 - 0@xmath70 & 4.000 & 0.88 & 10.2 & -48(2 ) & 8.4(4 ) & 11(3 ) & 6.4(9 ) & @xmath677 & ... & 0.2(1 ) & + & & & & & & & & & & & & & + & & & & & & & & & & & & & + ch@xmath1chcn &
18512.158(13 ) & 2@xmath74 - 1@xmath75 & 0.375 & 3.51 & 11.1 & -62(2 ) & 4.9(1 ) & @xmath677 & ... & @xmath677 & ... & & + & & @xmath72 & & & & & & & & & & & + & 18513.311(7 ) & 2@xmath74 - 1@xmath75 & 0.701 & 3.51 & 11.1 & -82(2 ) & 4.9(1 ) & @xmath677 & ... & @xmath677 & ... & 37.2(30 ) & + & & @xmath76 & & & & & & & & & & & + & 18514.400(27 ) & 2@xmath74 - 1@xmath75 & 0.167 & 3.51 & 11.1 & -19(3 ) & 4.1(5 ) & @xmath677 & ... & @xmath677 & ... & & 0.005(2 ) + & & @xmath77 & & & & & & & & & & & + ch@xmath1chnc & 19993.916(1 ) & 2@xmath74 - 1@xmath75 & 1.500 & 3.67 & 10.5 & 8(2 ) & 1.5(4 ) & @xmath677 & ... & @xmath677 & ... & 0.2(5 ) & + & & & & & & & & & & & & & + & & & & & & & & & & & & & + ch@xmath0ch@xmath1cn & 18377.721(3 ) & 2@xmath75 - 1@xmath78 & 1.500 & 1.56 & 10.0 & -24(2 ) & 2.8(2 ) & 20(2 ) & 2.8(2 ) & -20(2 ) & 2.8(2 ) & 3.5(4 ) & + & & & & & & & & & & & & & @xmath670.3 + ch@xmath0ch@xmath1nc & 18803.209(10 ) & 2@xmath74 - 1@xmath75 & 1.500 & 1.56 & 9.8 & @xmath677 & ... & @xmath677 & ... & @xmath677 & ... & @xmath671.1 & + & & & & & & & & & & & & & + & & & & & & & & & & & & & + hc@xmath2cn & 18638.617(2 ) & 7 - 6 & 7.000 & 2.70 & 10.9 & -63(2 ) & 5.8(2 ) & 97(1 ) & 16.0(5 ) & -59(2 ) & 8.0(3 ) & 6.1(2 ) & + & & & & & & & & & & & & & @xmath670.03 + hc@xmath2nc & 19616.504(2 ) & 7 - 6 & 7.000 & 2.84 & 9.7 & @xmath675 & ... & @xmath675 & ... & @xmath675 & ... & @xmath670.2 & + lccccccccccc h@xmath1cccc & 80383.96(2 ) & 9@xmath79 - 8@xmath80 & 153.8 & 19.3 & @xmath670.3 & ... & @xmath670.3 & ... & @xmath670.3 & ... & @xmath676.0@xmath3510@xmath52 + hcooch@xmath0 & 80395.135(126 ) & 9@xmath81 - 9@xmath79 e & 0.7 & 28.8 & @xmath670.3 & ... & @xmath670.3 & ... & @xmath670.3 & ... & @xmath674.7@xmath3510@xmath82 + ch@xmath0ch@xmath1cn & 80404.898(12 ) & 9@xmath81 - 8@xmath83 & 126.1 & 23.8 & 4.6(1 ) & 16.0(8 ) & 1.7(3 ) & 5.1(6 ) & 0.6(2 ) & 13.6(4.0 ) & 3.1@xmath3510@xmath64 + ch@xmath0nc & 80405.504(120 ) & 4@xmath84 - 3@xmath84 & 29.6 & 73.0 & @xmath670.3 & ... & @xmath670.3 & ... & @xmath670.3 & ... & + ch@xmath0nc & 80414.660(120 ) & 4@xmath85 - 3@xmath85 & 50.7 & 37.4 & @xmath670.3 & ... & @xmath670.3 & ... & @xmath670.3 & ... & + ch@xmath0nc & 80420.062(120 ) & 4@xmath86 - 3@xmath86 & 63.3 & 16.0 & @xmath670.3 & ... & @xmath670.3 & ... & @xmath670.3 & ... & + ch@xmath0nc & 80421.910(120 ) & 4@xmath87 - 3@xmath87 & 67.6 & 8.9 & @xmath670.3 & ... & @xmath670.3 & ... & @xmath670.3 & ... & @xmath674.0@xmath3510@xmath88 + ccc ch@xmath0cn & -29,070,027 & 0 + ch@xmath0nc & -29,060,541 & 9,486 + ch@xmath1chcn & -37,408,030 & 0 + ch@xmath1chnc & -37,399,372 & 8,658 + ch@xmath0ch@xmath1cn & -37,672,368 & 0 + ch@xmath0ch@xmath1nc & -37,663,671 & 8,697 + hc@xmath2cn & -53,818,766 & 0 + hc@xmath2nc & -53,807,547 & 11,219 + | we present strong detections of methyl cyanide ( ch@xmath0cn ) , vinyl cyanide ( ch@xmath1chcn ) , ethyl cyanide ( ch@xmath0ch@xmath1cn ) and cyanodiacetylene ( hc@xmath2cn ) molecules with the green bank telescope ( gbt ) toward the sgr b2(n ) molecular cloud .
attempts to detect the corresponding isocyanide isomers were only successful in the case of methyl isocyanide ( ch@xmath0nc ) for its @xmath3 transition , which is the first interstellar report of this line . to determine the spatial distribution of ch@xmath0nc , we used archival berkeley - illinois - maryland association ( bima ) array data for the @xmath4 ( @xmath5 ) transitions but no emission was detected . from ab initio calculations ,
the bonding energy difference between the cyanide and isocyanide molecules is @xmath68500 @xmath7 ( @xmath612,000 k ) .
thus , cyanides are the more stable isomers and would likely be formed more preferentially over their isocyanide counterparts .
that we detect ch@xmath0nc emission with a single antenna ( gaussian beamsize(@xmath8)=1723 arcsec@xmath9 ) but not with an interferometer ( @xmath8=192 arcsec@xmath9 ) , strongly suggests that ch@xmath0nc has a widespread spatial distribution toward the sgr b2(n ) region .
other investigators have shown that ch@xmath0cn is present both in the lmh hot core of sgr b2(n ) and in the surrounding medium , while we have shown that ch@xmath0nc appears to be deficient in the lmh hot core .
thus , large - scale , non - thermal processes in the surrounding medium may account for the conversion of ch@xmath0cn to ch@xmath0nc while the lmh hot core , which is dominated by thermal processes , does not produce a significant amount of ch@xmath0nc .
ice analog experiments by other investigators have shown that radiation bombardment of ch@xmath0cn can produce ch@xmath0nc , thus supporting our observations . we conclude that isomers separated by such large bonding energy differences are distributed in different interstellar environments , making the evaluation of column density ratios between such isomers irrelevant unless it can be independently shown that these species are co - spatial . |
since the minimal supersymmetric standard model ( mssm ) @xcite suffers from the @xmath3-problem @xcite and the little hierarchy problem , some non - minimal supersymmetric models have recently attracted much attention , among which the most intensively studied is the next - to - minimal supersymmetric standard model ( nmssm ) @xcite . in the nmssm
there is no dimensionful parameters in the supersymmetry - conserving sector and the @xmath3 term is dynamically generated through the coupling between the two higgs doublets and a newly introduced singlet higgs field which develops a vacuum expectation value of the order of the susy breaking scale .
the nmssm provides two ways to alleviate the little hierarchy problem .
one is to relax the lep ii lower bound on the mass of the sm - like higgs boson , @xmath6 , by diluting @xmath7 coupling through the singlet component of @xmath6 and/or by suppressing the visible decay @xmath8 through introducing new decay of @xmath6 @xcite .
the other is to push up the higgs boson mass with a large @xmath0 , which can be seen from the tree level upper bound of the higgs boson mass @xcite @xmath9 where @xmath10 , @xmath11 and @xmath0 is the mixing parameter between the singlet and doublet higgs fields defined in eq.([superpotential ] ) .
note that the choice of a large @xmath0 to solve the little hierarchy may be limited by the perturbativity of the theory at the scale @xmath12 since the value of @xmath0 is increasing with the energy scale@xcite . if this scale @xmath12 is the grand unification ( gut ) scale , @xmath0 should be less than about 0.7 at weak scale , leading to an upper bound on the higgs boson mass of about 150 gev @xcite . however , the bound on @xmath0 from the perturbativity consideration can be relaxed by embedding the nmssm in some more complex frameworks .
for example , in the fat higgs model @xcite , by completing the nmssm ( or nmssm - like models ) with an appropriate strong dynamics at an intermediate scale ( much lower than the gut scale ) , @xmath0 can be as large as 2 at weak scale and the higgs boson mass can be pushed up to about 400 gev . in this work ,
regardless the detailed forms of the ultraviolet physics , we treat the nmssm as an effective theory and examine the current experimental constraints on its parameter space .
such phenomenological studies on the higgs boson and supersymmetry are pressing since the mystery of the higgs sector will be unveiled at the lhc in the near future .
if the sm - like higgs boson is discovered with a mass above the mssm upper bound , the nmssm ( or other nmssm - like models ) with a large @xmath0 , generally called @xmath0susy @xcite , will be immediately favored since it not only inherits all the advantages of the mssm , such as unifying gauge couplings and providing a dark matter candidate , but also is free from the @xmath3-problem and the little hierarchy problem .
for the phenomenological studies of these models , a primary work is to examine the current experimental constraints on their parameter space .
we note that various constraints on the nmssm have been studied in the literature , but different constraints were considered in different papers .
for example , in @xcite the authors mainly considered the lep ii constraints and put emphasize on small @xmath0 case .
the package nmssmtools @xcite encoded various constraints ( like the lep ii searches for the higgs boson , the cosmic dark matter and the stability of the higgs potential ) , but it is still not complete since it does not include the indirect constraints from precision electroweak measurements and the muon anomalous magnetic moment . in this work
we consider all these constraints and especially focus on the case with a large @xmath0 .
as will be shown from our study , with the increase of @xmath0 , the parameter space is getting more stringently constrained . to figure out the allowed parameter space is helpful for exploring such low energy supersymmetry at the lhc and
also may shed some light on constructing the ultraviolet physics from the bottom - up view .
this paper is organized as follows . in sec.ii
we briefly describe the structure of the nmssm with emphasis on its difference from the mssm . in sec.iii
we summarize the constraints considered in this work and briefly discuss their characters . in sec .
iv we scan over the nmssm parameter space and display the region allowed by all these constraints . in sec .
v we give our conclusions .
the nmssm extends the matter fields of the mssm by adding one gauge singlet superfield @xmath13 , and its superpotential takes the form @xcite @xmath14 where @xmath15 , @xmath16 and @xmath17 are squark superfields , @xmath18 and @xmath19 are slepton superfields , @xmath20 and @xmath21 are higgs doublet superfields . the soft susy breaking terms are given by @xmath22 note that just like the mssm , the nmssm has the feature that susy breaking induces the electroweak symmetry breaking . before susy breaking ( i.e. without the soft breaking terms ) ,
the higgs scalars have zero vevs in the supersymmetric vacuum of the scalar potential and thus the electroweak symmetry is not broken .
after susy breaking ( i.e. with the soft breaking terms ) , the higgs scalars develop non - zero vevs in the physical ( non - supersymmetric ) vacuum of the scalar potential and hence the electroweak symmetry is spontaneously broken and the @xmath3 parameter is generated @xmath23 .
since both the electroweak symmetry breaking and the @xmath3 parameter generation are induced by susy breaking , their scales should be naturally at the susy breaking scale ( the scale of soft breaking mass parameters ) .
the differences of the nmssm and mssm come from the higgs sector and the neutralino sector@xcite . in the higgs sector of the nmssm
there are three cp - even and two cp - odd higgs bosons . in the basis
@xmath24 $ ] , the mass - squared matrix elements for cp - even higgs bosons are @xmath25 in the basis @xmath26 $ ] with @xmath27 , the mass - squared matrix elements for the cp - odd higgs bosons are @xmath28 as shown in eq.([ma ] ) , we can choose @xmath29 instead of @xmath30 as a free parameter .
so compared with the mssm , the nmssm has three additional parameters : @xmath0 , @xmath31 and @xmath32 .
conventionally , @xmath0 is chosen to be positive while @xmath31 and @xmath32 can be either positive or negative .
note that eqs.([cp - even ] ) and ( [ cp - odd ] ) indicate that the parameters @xmath33 and @xmath34 affect the mixings between doublet and singlet higgs fields , while @xmath32 only affects the squared - mass of the singlet higgs field . in the neutralino sector ,
the nmssm predicts one extra neutralino . in the basis
@xmath35 the neutralino mass matrix is given by @xcite @xmath36 this mass matrix is independent of @xmath32 , and the role of @xmath0 is to introduce the mixings of @xmath37 with @xmath38 and @xmath39 , and @xmath40 is to affect the mass of @xmath41 . from eq.([cp - even],[cp - odd],[mass matrix ] ) one can learn that in the limit @xmath42 , the singlet field have no mixing with the doublet field and thus is decoupled . in this case , the nmssm can recover the mssm .
before we proceed to discuss experimental constraints on the parameters of the nmssm , we take a look at the bounds on @xmath0 and @xmath31 from the requirement that the theory should keep perturbative under a certain scale @xmath12 .
the renormalization group equations ( rges ) for @xmath0 and @xmath31 under the scale @xmath12 take the following form @xcite @xmath43 where @xmath44 and @xmath45 are the @xmath46 and @xmath47 gauge couplings .
these rges indicate that the values of @xmath0 and @xmath31 increase with the energy scale .
the requirement of perturbativity till the cut - off scale @xmath12 , i.e. , @xmath48 and @xmath49 , will set upper bounds on @xmath0 and @xmath31 at weak scale ( throughout this paper , without specification all input parameters are defined at weak scale ) .
for example , if we assume that new dynamics appears at @xmath50tev , we get @xmath51 and for @xmath52 , @xmath31 must be less than 1.2 ; while if @xmath12 is chosen to be the gut scale , a stringent bound @xmath53 is obtained @xcite . in our following numerical study
we let @xmath0 and @xmath31 to vary below 2 and 1 , respectively , and this corresponds to set @xmath54 tev . in our study
we consider the following constraints on the parameters of the nmssm : * constraints on the neutralino and chargino sector , which include : the bound from invisible @xmath55 decay @xmath56 mev ; the upper bounds on neutralino pair productions at lep ii @xmath57 ( @xmath58 ) and @xmath59 ; and the lep ii bound on the lightest chargino mass @xmath60 gev .
these bounds will mainly constrain the parameters @xmath61 , @xmath4 and @xmath3 . *
lower bounds on sparticle masses from lep and tevatron experiments @xcite @xmath62 where @xmath63 denotes the masses for the first two generation squarks .
these constraints will put lower bounds on the soft breaking masses for sleptons and squarks .
* the lep ii lower bound on the charged higgs boson mass , @xmath64 gev , which gives a lower bound on @xmath29 through the relation @xmath65 .
* constraints from the direct search for higgs boson at lep ii @xcite , which include various channels of higgs boson productions @xcite .
they will constrain the parameters @xmath29 , @xmath66 , @xmath0 as well as the masses and the chiral mixing of top squarks in a complex way . *
constraint from the relic density of cosmic dark matter , i.e. @xmath67 @xcite , assuming the lightest neutralino is the dark matter particle .
the relic density will constrain the parameters @xmath68 , @xmath4 , @xmath3 , @xmath29 , @xmath66 and @xmath0 in a complex way @xcite . *
constraint from the stability of the higgs potential , which requires that the physical vacuum of the higgs potential with non - vanishing vevs of higgs scalars should be lower than any local minima .
also , the scale of the higgs soft breaking parameters should not be much higher than the electroweak scale to avoid the fine - tuning problem .
here we set 1 tev as the upper bound of the soft breaking parameters in the higgs sector .
this will constrain the parameters @xmath29 , @xmath3 , @xmath32 , @xmath0 and @xmath69 . *
constraints from precision electroweak observables such as @xmath70 , @xmath71 and @xmath72 , or their combinations @xmath73 @xcite .
we require the predicted @xmath74 in the nmssm to be compatible with the lep / sld data at @xmath75 confidence level or equivalently @xmath76 .
we take the correlation coefficient of @xmath74 from @xcite in calculating @xmath77 .
this requirement will constrain the parameters @xmath78 , @xmath29 as well as the soft breaking parameters in the third generation squark sector . *
constraint from @xmath79 , whose measured value is @xmath80 and the sm prediction is @xmath81 for @xmath82 gev @xcite . in our analysis
we require @xmath83 is within the @xmath84 range of its experimental value .
it has been shown that the susy contribution to @xmath85 might be sizeable for large @xmath86 @xcite .
* constraint from the muon anomalous magnetic momentum @xmath87 . now both the theoretical prediction and the experimental measurement of @xmath87
have reached a remarkable precision , but they show a significant deviation @xmath88 @xcite . in our analysis
we require the susy effects to account for such deviation at @xmath84 level .
the character of the susy contribution to @xmath87 is that it is suppressed by smuon masses but enhanced by @xmath66 . among the above constraints ,
( 1 - 6 ) and ( 9 ) have been encoded in the package nmssmtools @xcite . in our calculations
we extend it by including the constraints ( 7 ) and ( 8) .
the analytic expressions of @xmath74 and @xmath85 in the nmssm were given in our recent work @xcite .
in @xcite we also calculated the nmssm contribution to @xmath87 ( when we started that work , the results in @xcite had not yet published ) , where we extended the neutralino- and chargino - mediated mssm contributions @xcite to the nmssm and also considered the contributions from the higgs - mediated diagrams @xcite and from the barr - zee diagrams @xcite .
we checked that our @xmath87 results in @xcite agree with those in @xcite .
note that in our analysis we did not include the constraints from various @xmath89-decays @xcite because they are dependent on squark flavor mixings and thus involve additional parameters .
in this section , we scan over the nmssm parameter space to look for the region allowed by the constraints in the preceding section . since we are interested in the parameters sensitive to the constraints , we make some assumptions ( as conservative as possible ) for the other parameters such as soft breaking parameters in squark , slepton and gaugino sectors . for the parameters in squark sector , we assume the so - called @xmath90 scenario , which can maximize the lightest higgs boson mass @xcite .
this scenario assumes all the soft breaking masses in the squark sector to be degenerate @xmath91 with @xmath92 being the generation index .
it also assumes the trilinear couplings to be degenerate @xmath93 with @xmath94 .
we fix @xmath95 in our analysis since large @xmath96 can not only enhance the lightest higgs boson mass , but also decrease the contribution of the third generation squarks to the electroweak parameters , which has the same sign with the higgs contributions @xcite . for the parameters in slepton sector , we note that the slepton masses affect little on the constraints except the muon anomalous magnetic momentum . in our calculation
we assume all the soft breaking parameters in the slepton sector are degenerate and take a value of @xmath97 gev ( we will discuss the effects of its variation ) .
for the gaugino mass parameters , we assume the grand unification relation @xmath98 . with the above assumptions , the free parameters are reduced to seven ( @xmath0 , @xmath31 , @xmath32 , @xmath66 , @xmath29 , @xmath3 , @xmath4 ) and within the capability of our computer to perform a scan . during our scan , we first divide the varying range of @xmath0 into bins with each bin width being 0.1 and then we vary the values of other parameters in the following ranges @xmath99 with two hundred million samples in each bin and keeping the points satisfying the constraints , we finally get the allowed regions of these parameters .
our scan results indicate that the number of the survived samples for @xmath100 is much larger than that for @xmath101 , which means that the parameters for small @xmath0 are much less constrained than the case with large @xmath0 .
since we are interested in large @xmath0 , here we only show our scan results for @xmath101 .
in fig.[lam - tanb ] we display the parameters ( scatter plots ) satisfying all the constraints ( 1 - 9 ) in the plane of @xmath0 versus @xmath66 . also , we present a curve which is the upper bound on @xmath66 without considering the muon g-2 constraints . to get this curve , we fix @xmath0 and scan over the parameters in eq.([scan - region ] ) .
we adopt the important sampling method @xcite to optimize the varying range of @xmath66 .
fig.[lam - tanb ] shows that the upper bound on @xmath66 gets stronger as @xmath0 gets large , and when all the constraints are considered , @xmath0 is upper bounded by about 1.5 .
the underlying reason for this is that the constraints ( 1 - 8 ) , especially the constraint ( 7 ) , have limited the maximal value of @xmath86 , which decreases with the increase of @xmath0 .
since a large @xmath66 is needed to explain the deviation of the muon g-2 , @xmath0 must terminate at a certain value where the corresponding @xmath66 value is too small to explain the muon g-2 .
we have checked that the maximal value of @xmath0 is dependent on slepton mass .
for example , for slepton mass of @xmath102 gev , @xmath103 gev and @xmath104 gev , the bounds on @xmath0 are @xmath105 , @xmath106 and @xmath107 , respectively . in fig.[lambda - all ]
we display the nmssm parameters satisfying all the constraints in different planes .
we see that for a large @xmath0 the parameters @xmath29 , @xmath3 , @xmath4 and @xmath32 are also bounded in a certain region . for @xmath108 ,
these bounded regions are @xmath109 gev , @xmath110 gev , @xmath111 gev and @xmath112 gev . from the figure of @xmath2 versus @xmath0 in fig.[lambda - all ] one can see that the lower bound of @xmath2 increases as @xmath0 becomes large .
the reason is that the lep ii direct search for higgs boson mainly limits the mass and the couplings of the light cp - even higgs boson whose component is dominated by the doublet higgs field @xmath113 or @xmath114 . for @xmath115 , this higgs boson should be dominantly composed by @xmath113 field since @xmath116 is smaller than @xmath117 , and its mass is to be reduced by the off - diagonal elements @xmath118 and @xmath119 . as @xmath0 gets larger , these off - diagonal elements get larger and hence reduce the mass of the light cp - even higgs boson , which then requires a larger @xmath2 to compensate in order to satisfy the lep ii lower bound .
the figure of @xmath3 versus @xmath0 in fig.[lambda - all ] indicates that with the increase of @xmath0 , the upper bound of @xmath3 decreases .
this is because in the off - diagonal elements @xmath120 and @xmath121 ( which reduce the light cp - even higgs boson mass ) , @xmath33 is always associated with @xmath3 , and to meet the lep ii bound a large @xmath0 must be accompanied by a small @xmath3 .
the figure of @xmath4 versus @xmath0 in fig.[lambda - all ] shows that @xmath4 is also bounded in a narrow region .
this is because the relic density of the dark matter correlates the parameters @xmath29 , @xmath3 , @xmath4 , @xmath0 and @xmath69 in a complex way , and a large value for any of these parameters will limit severely the region of other parameters .
the figure of @xmath32 versus @xmath0 in fig.[lambda - all ] shows that the trilinear soft breaking parameter @xmath32 for the singlet field is also limited . this can be understood from the expressions of @xmath122 and @xmath123 .
the stability of the higgs potential requires both of them to be positive , which sets an double - sided bound on @xmath32 .
we also studied the relationship between the yukawa couplings @xmath0 and @xmath31 , and we found no correlation between them . even for @xmath124 , the value of @xmath31 can still vary from 0.3 to 1 .
next , we take a look at the higgs boson masses allowed by the constraints .
since a large @xmath0 can enhance the lightest cp - even higgs boson mass and thus avoid the little hierarchy problem , it is interesting to look at the dependence of the higgs boson masses on the parameter @xmath0 . in figs.[lambda - mh ] and
[ lambda - ma ] we show our scan results in @xmath0 versus @xmath125 plane and @xmath0 versus @xmath126 plane with @xmath125 being the lightest cp - even higgs boson mass and @xmath126 the lighter cp - odd higgs boson mass . from fig.[lambda - mh
] one can learn that the upper bound of @xmath125 increases with @xmath0 , which is expected from eq.([limit ] ) , and for @xmath124 the value of @xmath125 can reach 210 gev . from fig.[lambda - ma ] one can learn that with the increase of @xmath0 , a super light cp - odd higgs boson is gradually ruled out , and for @xmath127 it is bounded in the range @xmath128 gev .
the properties of these higgs bosons can be quite different from those in the mssm , and their phenomenology at the lhc was discussed in @xcite . finally , in order to understand the mechanism used to reproduce the correct dark matter abundance , we consider the properties in the neutralino sector . in the nmssm with large @xmath66 ,
the component of the lightest neutralino is either higgsino - dominant or bino - dominant for a light mass below 80 gev , but for a heavier mass it is bino - dominant . in fig.[lambda - chi ] we show our scan results in the plane of @xmath129 versus @xmath0 .
we see that with the increase of @xmath0 , the upper bound on @xmath129 becomes stringent and eventually it is constrained in the range of @xmath130 gev . about the next lightest neutralino @xmath131 we found that its mass is constrained in the range of @xmath132 gev for @xmath133 . in order to figure out the annihilation mechanism of @xmath134 in providing for the dark matter relic density , we compare the masses of @xmath131 and @xmath135 with @xmath134 in fig.[spectrum ] .
this figure indicates that @xmath131 is significantly heavier than @xmath134 . since in our scan the slepton masses
are fixed to 200 gev , also significantly heavier than @xmath134 , we conclude that the coannihilation of @xmath134 with @xmath131 or with a slepton is generally boltzmann - suppressed and plays an unimportant role in accounting for the dark matter relic density .
note that , as shown in fig.[spectrum ] , there are some samples around the funnel region @xmath136 and in this case the annihilation of @xmath134 through the s - channel exchange of a light @xmath135 becomes dominant @xcite .
the nmssm with a large @xmath0 is an attractive scenario since it can push up the upper bound on the sm - like higgs boson mass to solve the little hierarchy problem .
we examined the current experimental constraints on this scenario , which include the direct experimental bounds , the indirect constraints from precision electroweak measurements , the cosmic dark matter relic density , the muon anomalous magnetic moment , as well as the stability of the higgs potential .
our results showed that for a large @xmath0 the parameter space is severely constrained .
for example , for a smuon mass of 200 ( 500 ) gev the parameter space with @xmath5 is excluded , and for @xmath137 the allowed ranges are @xmath138 for @xmath78 , @xmath139 gev for @xmath2 , @xmath140 gev for @xmath3 , @xmath141 gev for @xmath4 and @xmath142 gev for @xmath32 . finally , we would like to point out that our conclusion may be qualitatively applicable to other nmssm - like models such as the minimal nonminimal supersymmetric standard model ( mnmssm ) @xcite , which has similar structure with the nmssm and can be viewed as the low energy realization of the fat higgs model @xcite . for example , it has been pointed out that for any singlet extensions of the mssm , regardless the form of its superpotential , a large @xmath0 is always accompanied by a small @xmath66 @xcite .
this property , as shown in our paper , can either limit the smuon mass or limit @xmath0 if we require the theory to explain the deviation of the muon anomalous magnetic momentum .
another example is about the constraint from dark matter . in the mnmssm
we expect that the constraint can limit the relevant parameters in a more stringent way than in the nmssm since the neutralino sector in the mnmssm is exactly same as in the nmssm but with fixed @xmath143 @xcite .
this work was supported in part by the national sciences and engineering research council of canada , and by the national natural science foundation of china ( nnsfc ) under grant no .
10505007 , 10725526 and 10635030 .
99 h. e. haber and g. l. kane , phys .
rept . * 117 * , 75 ( 1985 ) .
j. f. gunion and h. e. haber , nucl .
b * 272 * , 1 ( 1986 ) [ erratum - ibid .
b * 402 * , 567 ( 1993 ) ] .
j. e. kim and h. p. nilles , phys .
b * 138 * , 150 ( 1984 ) .
j. r. ellis , j. f. gunion , h. e. haber , l. roszkowski and f. zwirner , phys .
d * 39 * ( 1989 ) 844 .
m. drees , int .
j. mod .
a * 4 * ( 1989 ) 3635 .
u. ellwanger , m. rausch de traubenberg and c. a. savoy , phys .
b * 315 * ( 1993 ) 331 ; nucl .
b * 492 * ( 1997 ) 21 ; s. f. king and p. l. white , phys .
d * 52 * ( 1995 ) 4183 ; f. franke and h. fraas , int .
j. mod .
a * 12 * ( 1997 ) 479 ; b. a. dobrescu , k. t. matchev , jhep 0009 ( 2000 ) 031 .
r. harnik , g. d. kribs , d. t. larson and h. murayama , phys . rev .
d * 70 * , 015002 ( 2004 ) ; s. chang , c. kilic and r. mahbubani , phys .
d * 71 * , 015003 ( 2005 ) ; a. birkedal , z. chacko and y. nomura , phys .
d * 71 * , 015006 ( 2005 ) ; a. delgado and t. m. p. tait , jhep * 0507 * , 023 ( 2005 ) .
r. barbieri , l. j. hall , y. nomura and v. s. rychkov , phys .
d * 75 * , 035007 ( 2007 ) .
v. barger , p. langacker , h. s. lee and g. shaughnessy , phys .
d * 73 * , 115010 ( 2006 ) ; v. barger , p. langacker and g. shaughnessy , phys .
d * 75 * , 055013 ( 2007 ) ; phys . lett .
b * 644 * , 361 ( 2007 ) .
see u. ellwanger , j. f. gunion and c. hugonie , jhep * 0502 * , 066 ( 2005 ) ; u. ellwanger and c. hugonie , comput .
commun .
* 175 * , 290 ( 2006 ) .
j. p. derendinger and c. a. savoy , 237 , 307 ( 1984 ) ; n. k. falck , z. phys .
c30 , 247 ( 1986 ) ; r. b. nevzorov , m.a .
trusov , hep - ph/0112301 .
w. m. yao _ et al .
_ , particle data group , j. phys .
g * 33 * ( 2006 ) 1 .
s. schael , _ et al .
_ , eur .
j. c * 47 * , 547 ( 2006 ) . c. l. bennett _ et al .
_ , astrophys .
j. suppl .
* 148 * ( 2003 ) 1 ; d. n. spergel _ et al . _ ,
astrophys .
j. suppl .
* 148 * ( 2003 ) 175 .
g. belanger , f. boudjema , c. hugonie , a. pukhov and a. semenov , jcap * 0509 * , 001 ( 2005 ) ; v. barger , p. langacker and h. s. lee , phys . lett .
b * 630 * , 85 ( 2005 ) .
g. altarelli and r. barbieri , phys .
b * 253 * , 161 ( 1991 ) ; g. altarelli , r. barbieri and s. jadach , nucl .
b * 369 * , 3 ( 1992 ) [ erratum - ibid .
b * 376 * , 444 ( 1992 ) ] ; g. altarelli , r. barbieri and f. caravaglios , nucl .
b * 405 * , 3 ( 1993 ) ; phys . lett .
b * 314 * , 357 ( 1993 ) .
lep and sld collaborations , phys .
rept . * 427 * , 257 ( 2006 ) .
j. j. cao , j. m. yang , arxiv:0810.0751 [ hep - ph ] .
for a recent review , see j. p. miller , e. de rafael and b. l. roberts , rept .
phys . * 70 * , 795 ( 2007 ) ; d. stockinger , arxiv:0710.2429 [ hep - ph ] .
see , for example , t. ibrahim and p. nath , phys .
d * 62 * , 015004 ( 2000 ) ; s. p. martin and j. d. wells , phys .
d * 64 * , 035003 ( 2001 ) .
w. a. bardeen , r. gastmans , and b. lautrup , 46 , 319 ( 1972 ) ; j. p. leveille , 137 , 63 ( 1978 ) ; h. e. haber , g. l. kane , and t. sterling , 161 , 493 ( 1979 ) ; e. d. carlson , s. l. glashow and u. sarid , 309 , 597 ( 1988 ) ; j. r. primack and h. r. quinn , 6 , 3171 ( 1972 ) .
d. chang , w. -f .
chang , c .- h .
chou , and w .- y .
keung , phys .
d * 63 * , 091301 ( 2001 ) ; k. cheung , c. h. chou and o. c. w. kong , phys . rev .
d * 64 * , 111301 ( 2001 ) ; a. arhrib and s. baek , phys .
d * 65 * , 075002 ( 2002 ) .
f. domingo and u. ellwanger , arxiv:0806.0733 [ hep - ph ] .
j. f. gunion , arxiv:0808.2509 [ hep - ph ] .
g. hiller , phys .
d * 70 * , 034018 ( 2004 ) ; f. domingo and u. ellwanger , jhep * 0712 * , 090 ( 2007 ) ; z. heng , _ et al .
_ , phys .
d * 77 * , 095012 ( 2008 ) ; r. n. hodgkinson , phys .
b * 665 * , 219 ( 2008 ) .
s. heinemeyer , w. hollik and g. weiglein , jhep * 0006 * , 009 ( 2000 ) .
g. p. lepage , j. comput .
* 27 * , 192 ( 1978 ) .
l. cavicchia , r. franceschini and v. s. rychkov , phys . rev .
d * 77 * , 055006 ( 2008 ) .
j. f. gunion , d. hooper and b. mcelrath , phys .
d * 73 * , 015011 ( 2006 ) . c. panagiotakopoulos and k. tamvakis , phys .
b * 469 * , 145 ( 1999 ) .
a. menon , d. e. morrissey and c. e. m. wagner , phys .
d * 70 * , 035005 ( 2004 ) . | the next - to - minimal supersymmetric model ( nmssm ) with a large @xmath0 ( the mixing parameter between the singlet and doublet higgs fields ) is well motivated since it can significantly push up the upper bound on the sm - like higgs boson mass to solve the little hierarchy problem . in this work we examine the current experimental constraints on the nmssm with a large @xmath0 , which include the direct search for higgs boson and sparticles at colliders , the indirect constraints from precision electroweak measurements , the cosmic dark matter relic density , the muon anomalous magnetic moment , as well as the stability of the higgs potential .
we find that , with the increase of @xmath0 , parameters like @xmath1 , @xmath2 , @xmath3 and @xmath4 are becoming more stringently constrained .
it turns out that the maximal reach of @xmath0 is limited by the muon anomalous magnetic moment , and for smuon masses of 200 gev ( 500 gev ) the parameter space with @xmath5 is excluded . |
in random spin systems with competing ferromagnetic and antiferromagnetic interactions , the spin frustration effect occurs , leading to a spin - glass ( sg ) phase at low temperatures .
this situation may change when there is a majority of the ferromagnetic interactions and a minority of the antiferromagnetic interactions to create substantial spin frustration effect .
the system ( so - called reentrant ferromagnet ) exhibits two phase transitions at @xmath6 and @xmath4 ( @xmath21 ) : the reentrant spin glass ( rsg ) phase below @xmath6 and the ferromagnetic ( fm ) phase between @xmath6 and @xmath4 .
experimental studies on the dynamic magnetic properties have been carried out for reentrant ferromagnets such as ( fe@xmath22ni@xmath23)@xmath24p@xmath25b@xmath26al@xmath3 , @xcite cr@xmath27fe@xmath28,@xcite ( fe@xmath29ni@xmath30)@xmath31mn@xmath32,@xcite cdcr@xmath33in@xmath34s@xmath35 ( @xmath36 = 0.90 , 0.95 , and 1.00),@xcite and fe@xmath37al@xmath38.@xcite for cr@xmath27fe@xmath28,@xcite ( fe@xmath29ni@xmath30)@xmath31mn@xmath32,@xcite and fe@xmath37al@xmath38,@xcite the rsg phase exhibits aging phenomena , which are very similar to those observed in the sg phase of sg systems .
no aging phenomenon has been observed in the fm phase . for ( fe@xmath22ni@xmath23)@xmath24p@xmath25b@xmath26al@xmath3 , @xcite in contrast , not only the rsg phase but also the fm phase exhibit aging phenomena . a strikingly increased fragility to the magnitude of the magnetic field
is observed when passing from the low - temperature rsg region into the fm phase , implying a chaotic nature of the fm phase .
the effect of the probing field @xmath18 on the relaxation rate has to be carefully considered .
the dramatic decrease of the limiting field around @xmath6 may explain why other experiments on reentrant ferromagnets have not resolved an aging behavior in the fm phase .
for cdcr@xmath33in@xmath34s@xmath35 with @xmath36 = 0.90 , 0.95 , and 1.0,@xcite the aging behavior of the low frequency ac susceptibility ( absorption @xmath39 ) is observed both in the fm and rsg phases .
experimental studies on the aging dynamics have been limited to a macroscopic measurement such as the time evolution of zero - field cooled susceptibility , thermoremnant susceptibility , and the absorption of the ac magnetic susceptibility .
recently motoya et al.@xcite have studied time - resolved small angle neutron scattering on fe@xmath40al@xmath41 in order to probe the microscopic mechanism of slow dynamics .
the lorentzian form of the scattering pattern and the temperature variation of the inverse correlation length below @xmath6 show that the system is composed of only finite - sized clusters .
the size of the clusters gradually decreases with decreasing temperature
. cu@xmath0co@xmath1cl@xmath2-fecl@xmath3 graphite bi - intercalation compound ( gbic ) is one of typical 3d ising reentrant ferromagnets .
it has a unique layered structure where the cu@xmath0co@xmath1cl@xmath2 intercalate layer (= @xmath42 ) and fecl@xmath3 intercalate layers ( = @xmath43 ) alternate with a single graphite layer ( @xmath44 ) , forming a stacking sequence ( -@xmath44-@xmath42-@xmath44-@xmath43-@xmath44-@xmath42-@xmath44-@xmath43-@xmath44-@xmath45 ) along the @xmath46 axis . in the cu@xmath0co@xmath1cl@xmath2 intercalate layer ,
two kinds of magnetic ions ( cu@xmath47 and co@xmath47 ) are randomly distributed on the triangular lattice .
the static and dynamic magnetic properties have been reported in a previous paper.@xcite this compound undergoes successive transitions at the transition temperatures @xmath4 (= 9.7 k ) and @xmath6 (= 3.5 k ) .
a prominent nonlinear susceptibility is observed between @xmath6 and @xmath4 , suggesting a chaotic nature of the fm phase . in this paper
we report our experimental study on the aging dynamics of the rsg and fm phases of cu@xmath0co@xmath1cl@xmath2-fecl@xmath3 gbic using dc and ac magnetic susceptibility measurements .
our system is cooled from 50 k to @xmath12 ( @xmath48 ) in the absence of an external magnetic field .
this zfc aging protocol process is completed at @xmath49 , where @xmath50 is defined as an age ( the total time after the zfc aging protocol process ) .
then the system is aged at @xmath12 under @xmath51 until @xmath52 , where @xmath10 is a wait time .
the aging behavior of the zfc magnetic susceptibility @xmath14 has been measured under the various aging processes : ( i ) a wait time @xmath10 ( @xmath53 sec ) , @xmath12 ( @xmath54 k ) , and @xmath18 ( @xmath55 oe ) as parameters , and ( ii ) the @xmath12-shift and @xmath18-shift perturbations . the relaxation rate defined by @xmath56 ( see sec .
[ backa ] for the definition ) exhibits a peak at a characteristic time @xmath9 close to @xmath10 below @xmath4 , indicating the occurrence of the aging phenomena both in the rsg and fm phases .
we will also show that the @xmath13 dependence of @xmath14 around @xmath57 is well described by a stretched exponential relaxation [ @xmath58 ( see sec .
[ backb ] ) , where @xmath17 is an exponent and @xmath19 is a relaxation time nearly equal to @xmath9 .
we will show that @xmath17 , @xmath19 , and @xmath9 depend on @xmath12 , @xmath10 , and @xmath18 .
the local maximum of @xmath19 and @xmath9 around 5 k is observed , reflecting the chaotic nature of the fm phase .
a partial rejuvenation of the system occurs in @xmath8 under the positive shift in the fm phase .
after the sg system is cooled to @xmath12 ( @xmath61 ) through the zfc aging protocol at @xmath49 , the size of domain defined by @xmath62 grows with the age of @xmath50 and reaches @xmath63 just before the field is turned on at @xmath64 or @xmath52.@xcite the aging behavior in @xmath65 is observed as a function of the observation time @xmath13 .
after @xmath64 , a probing length @xmath66 corresponding to the maximum size of excitation grows with @xmath13 , in a similar way as @xmath62 .
the quasi - equilibrium relaxation occurs first through local spin arrangements in length @xmath67 , followed by non - equilibrium relaxation due to domain growth , when @xmath68 , so that a crossover between the short - time quasi - equilibrium decay and the non - equilibrium decay at longer observation times is expected to occur near @xmath69 .
the droplet model@xcite predicts the following two .
( i ) the equilibrium sg states at two temperatures with the difference @xmath70 are uncorrelated when the overlap length @xmath71 is smaller than @xmath63 , i.e. , so - called temperature ( @xmath12)-chaos nature of the sg phase .
( ii ) the equilibrium sg states at two fields with the difference @xmath72 are uncorrelated when the overlap length @xmath73 is smaller than @xmath63 .
the absorption @xmath39 of the ac magnetic susceptibility is evaluated from the spin auto - correlation function @xmath74 using the fluctuation - dissipation theorem ( fdt ) as@xcite @xmath75 where @xmath76 , @xmath77 ( typically @xmath78 sec ) , @xmath79 is the angular frequency , and @xmath13 is much larger than @xmath80 . in the auto - correlation function ,
the over - line denotes the average over sites and over different realizations of bond disorder , and the bracket denotes the average over thermal noises . for slow processes , the dispersion @xmath81 is approximated by @xmath82/t .
\label{eq02}\ ] ] in the quasi - equilibrium regime where the fdt holds , the zfc susceptibility @xmath59 is described by @xmath83/t , \label{eq03}\ ] ] where @xmath84 and @xmath10 is a wait time
. then the relaxation rate @xmath8 is described by @xmath85 which corresponds to @xmath86 . in spin glasses the aging manifests by the fact that @xmath87 shows a strong dependence on the value of @xmath10 .
it is predicted that @xmath88 can be described by an addition of the short - time and long - time contributions@xcite @xmath89 whereas the aging at the sg transition temperature @xmath90 leads to a multiplicative representation , @xmath91 here the equilibrium part can be described by power - law form @xmath92 with a temperature - dependent exponent @xmath93 ( @xmath94 ) .
these two forms ( additive and multiplicative representations ) are actually not so different for short times , since @xmath95 is approximately constant for @xmath96 , in the regime where @xmath97 varies most . the zfc susceptibility is described by either an additive form @xmath98 , \label{eq08}\ ] ] or a multiplicative form @xmath99 .
\label{eq09}\end{aligned}\ ] ] when the additive form of @xmath100 is used , the absorption @xmath60 can be rewritten as @xmath101 where @xmath102 and @xmath103 here we assume that the aging contribution @xmath95 is approximated by a scaling function of @xmath104 as@xcite @xmath105 in eq.([eq10 ] ) the first term @xmath106 [ @xmath107 is independent of @xmath79 , while the second term @xmath108 [ @xmath109 is a function of @xmath110 ( see sec . [ resultf ] for detail ) . here
@xmath93 ( @xmath94 ) and @xmath111 ( @xmath112 ) are exponents of quasi - equilibrium part and aging parts . in all cases
we observe a slow time decay ( aging ) of @xmath60 towards an asymptotic frequency dependent value @xmath113 ( stationary susceptibility ) : @xmath114 in the limit of @xmath115 , where @xmath108 is assumed to be zero .
( color online)(a ) contour plot of @xmath116 ( @xmath117 ) in the ( @xmath17 , @xmath118 ) plane , where @xmath119 , and the points with the same @xmath116 are connected by the same solid line .
the definition of @xmath9 and @xmath19 is given in the text .
( b ) plot of @xmath116 vs @xmath118 at various @xmath17 .
the expression for @xmath116 is given by eqs.([eq20 ] ) and ( [ eq21]).,width=283 ] here we present a simple review on the stretched exponential relaxation of @xmath65 of sg phase after the zfc aging protocol .
theoretically@xcite and experimentally @xcite it has been accepted that the time variation of @xmath14 may be described by a product of a power - law form and a stretched exponential function @xmath120 , \label{eq16}\ ] ] where the exponent @xmath118 may be positive and is very close to zero , @xmath17 is between 0 and 1 , @xmath19 is a characteristic relaxation time , and @xmath121 and @xmath122 are constants . in general , these parameters are dependent on @xmath10 .
this form of @xmath14 incorporates both the nonequilibrium aging effect through the stretched exponential factor @xmath123 $ ] in the crossover region ( @xmath69 and @xmath124 ) between the quasi equilibrium state and nonequilibrium state , and an equilibrium relaxation response at @xmath96 through a pure power - law relaxation ( @xmath125 ) .
note that ogielski@xcite fits his data by a stretched exponential multiplied by a power function .
for @xmath126 , ogielski@xcite fits it by a power law with a different temperature dependence of exponent @xmath118 . when @xmath127 , @xmath14 is well described by a power law form given by @xmath128 .
however , in the regime of @xmath129 , the stretched exponential relaxation is a very good approximation in spite of finite @xmath118 that is very small .
for all temperatures , @xmath14 increases with increasing @xmath13 and the relaxation rate @xmath8 , which is defined by @xmath130 exhibits a maximum at @xmath9 that is close to @xmath10 . using eq.([eq16 ] ) for @xmath14
, the relaxation rate @xmath8 can be derived as @xmath131\nonumber\\ & & ( t/\tau)^{-(m+n)}[(1-n)+m(t/\tau)^{n-1 } ] .
\label{eq18}\end{aligned}\ ] ] the condition that @xmath8 may have a peak at @xmath132 [ @xmath133 leads to the ratio @xmath119 satisfying the following equation @xmath134 the solution of eq.([eq19 ] ) can be exactly obtained as @xmath135 with @xmath136/(1-n ) , \label{eq21}\ ] ] where @xmath137 .
note that the value of @xmath116 is uniquely determined only by the values of @xmath17 and @xmath118 .
when @xmath138 , @xmath139 ( or @xmath140 ) , which is independent of @xmath17 .
figure [ fig01](a ) shows the contour plot of @xmath116 in the ( @xmath17,@xmath118 ) plane with @xmath141 and @xmath142 , where the points having the same @xmath116 are connected by each solid line . the value of @xmath116 is lower than 1 for @xmath143 ,
is equal to 1 for @xmath138 irrespective of @xmath17 , and is larger than 1 for @xmath144 .
figure [ fig01](b ) shows a plot of @xmath116 as a function of @xmath118 at various fixed @xmath17 .
the maximum value of @xmath8 at @xmath145 is given by @xmath146 when @xmath138 , @xmath147 is equal to @xmath148 [ @xmath149 with @xmath150 .
we used the same sample of cu@xmath0co@xmath1cl@xmath2-fecl@xmath3 gbic that was used in the previous paper.@xcite the detail of the sample characterization and synthesis were presented in the references of the previous paper.@xcite the dc magnetization and ac susceptibility of cu@xmath0co@xmath1cl@xmath2-fecl@xmath3 gbic were measured using a squid magnetometer ( quantum design , mpms xl-5 ) with an ultra low field capability option .
the remnant magnetic field was reduced to zero field ( exactly less than 3 moe ) at 298 k for both dc magnetization and ac susceptibility measurements . the ac magnetic field used in the present experiment
has a peak magnitude of the ac field ( @xmath151 ) and a frequency @xmath152 .
each experimental procedure for measurements is presented in the text and figure captions .
in our measurement of the time ( @xmath13 ) dependence of the zero - field cooled ( zfc ) magnetization ( @xmath153 ) , the time required for the zfc aging protocol and subsequent wait time were precisely controlled .
typically it takes @xmath154 sec to cool the system from 10 k to @xmath12 ( @xmath48 ) .
it takes another @xmath155 sec until @xmath12 becomes stable within the uncertainty ( @xmath156 k ) .
the system is kept at @xmath12 and @xmath51 for a wait time @xmath10 ( typically @xmath157 or @xmath158 sec ) . at time @xmath64
, external magnetic field @xmath18 is applied along any direction perpendicular to the @xmath46 axis .
the time for setting up a magnetic field from @xmath18 = 0 to @xmath159 oe is @xmath160 sec . in the zfc measurement ,
the sample is slowly moved through the pick - up coils over the scan length ( 4 cm ) .
the magnetic moment of the sample induces a magnetic flux change in the pick - up coils .
it takes 12 sec for each scan .
the data at @xmath13 is regarded as the average of @xmath153 measured over the scanning time @xmath161 between the times @xmath162 and @xmath162 .
thus the time window @xmath163 is a scanning time ( @xmath161 ) .
the measurement was carried out at every interval of @xmath164 , where @xmath165 is a pause between consecutive measurements .
typically we used ( i ) the time window @xmath166 sec for the measurements with three scans and @xmath167 sec or 30 sec , and ( ii ) the time window @xmath168 sec for the measurements with one scan and @xmath169 or 2 sec .
( color online)(a ) @xmath113 vs @xmath12 at various @xmath171 ( @xmath172 hz ) .
@xmath173 moe .
( b ) @xmath12 dependence of @xmath65 ( @xmath174 ) and @xmath175 ( @xmath176 ) at @xmath159 oe for cu@xmath0co@xmath1cl@xmath2-fecl@xmath3 gbic .
the measurement was carried out after the zfc aging protocol : annealing of the system at 50 k for 1200 sec at @xmath51 and quenching from 50 to 1.9 k. during the zfc measurement with increasing @xmath12 from 1.9 to 20 k , the system was aging at @xmath177 k ( zfc-1 measurement ) and @xmath178 k ( zfc-2 measurement ) for wait time @xmath179 sec in the presence of @xmath159 oe .
after the zfc measurement , the system was annealed at @xmath180 k for @xmath181 sec in the presence of @xmath18 (= 1 oe ) . the fc measurement ( denoted as fc-1 and fc-2 )
was carried with decreasing @xmath12 from 20 to 1.9 k. the @xmath12 dependence of fc-1 susceptibility is the same as that of fc-2 susceptibility.,width=283 ] in the previous paper@xcite the magnetic properties of cu@xmath0co@xmath1cl@xmath2-fecl@xmath3 gbic have been extensively studied from our results on dc susceptibility ( @xmath65 , @xmath175 ) , and ac magnetic susceptibility ( @xmath182 and @xmath39 ) . in fig .
[ fig02](a ) , as a typical example we show the @xmath12 dependence of the absorption @xmath113 at various @xmath171 , where @xmath173 moe .
it is concluded from this figure that that our system undergoes two phase transitions at @xmath183 k and @xmath184 k. there are a rsg phase below @xmath6 and an fm phase between @xmath6 and @xmath4 . in order to demonstrate the evidence of aging behavior in @xmath65 , we measured the @xmath12 dependence of @xmath65 (= @xmath185 ) in the following two zfc protocols .
the system was quenched from 50 to 1.9 k in the absence of @xmath18 before the measurement .
the susceptibility @xmath65 was measured with increasing @xmath12 from 1.9 k to @xmath186 k in the presence of @xmath18 (= 1 oe ) , where @xmath187 is in the fm phase .
the system was kept at @xmath187 for a wait time @xmath188 sec .
subsequently @xmath65 was measured with further increasing @xmath12 from @xmath187 to 20 k. the system was annealed at 50 k for @xmath189 sec in the presence of @xmath18 (= 1 oe ) .
the susceptibility @xmath175 was continuously measured from 20 to 1.9 k. in fig .
[ fig02](b ) we show the @xmath12 dependence of @xmath65 and @xmath175 ( @xmath159 oe ) thus obtained ( denoted as the zfc-1 and fc-1 processes ) .
we find that there is a remarkable increase of @xmath65 with increasing @xmath13 during the one stop at @xmath186 k. as @xmath12 again increases from @xmath187 , @xmath65 starts to decreases with increasing @xmath12 and merges with @xmath190 without the stop as a reference at @xmath191 k ( see the inset of fig .
[ fig02](b ) ) .
the state at @xmath192 k is uncorrelated with that at @xmath187 since the temperature difference @xmath70 is large so that the domain size @xmath193 is much larger than the overlap length @xmath71 . in another measurement ( denoted as the zfc-2 and fc-2 processes ) ,
the system was kept at @xmath194 k in the rsg phase for @xmath195 sec during the zfc process . in the inset of fig.2(b ) we show the @xmath12 dependence of @xmath65 in the zfc-2 process .
there is a remarkable increase of @xmath65 during the one stop at @xmath196 .
as @xmath12 again increases from @xmath196 , @xmath65 starts to increase with a derivative d@xmath65/d@xmath12 that is positive and much smaller than that of @xmath190 .
this susceptibility @xmath65 merges with @xmath190 at @xmath12 @xmath197 k. the state at @xmath198 k is uncorrelated with that at @xmath196 since the temperature difference @xmath199 is large so that the domain size @xmath200 is much larger than @xmath71 .
the @xmath12 dependence of @xmath175 in the fc-2 process is the same as that in the fc-1 process .
note that the gradual increase of @xmath65 with increasing @xmath12 from the stop temperature is also observed in the sg phase of the 3d ising spin glasses fe@xmath201mn@xmath201tio@xmath3@xcite and cu@xmath201co@xmath201cl@xmath2-fecl@xmath3 gbic.@xcite such an aging behavior in @xmath65 may be common to the rsg and sg phases .
this is in contrast to the gradual decrease of @xmath65 with increasing @xmath12 from the stop temperature in the fm phase .
( color online)(a ) and ( b ) relaxation rate @xmath202 [ @xmath203 vs @xmath13 at various @xmath12 . @xmath159
@xmath157 sec .
@xmath18 is applied along a direction perpendicular to the @xmath46 axis .
the zfc aging protocol : annealing of the system at 50 k for @xmath189 sec at @xmath51 , quenching from 50 k to @xmath12 , and then isothermal aging at @xmath12 and @xmath51 for a wait time @xmath10 .
the measurement was started at @xmath64 when the field @xmath18 is turned on.,width=283 ] ( color online)(a ) and ( b ) @xmath202 vs @xmath13 at various @xmath12 . @xmath159
@xmath204 sec.,width=283 ] ( color online)@xmath13 dependence of @xmath14 at various @xmath12 .
@xmath159 oe . @xmath204 sec .
the solid curves are fits to eq.([eq16 ] ) with @xmath138 for stretched exponential relaxation.,width=283 ] ( color online)@xmath9 vs @xmath12 and @xmath19 vs @xmath12 .
@xmath9 is a characteristic time at which @xmath8 has a maximum @xmath147 .
@xmath19 is the relaxation time for the stretched exponential relaxation . @xmath205
( a ) @xmath157 sec .
( b ) @xmath206 sec .
the solid lines are guide to the eyes .
the inset of ( a ) shows the plot of @xmath207 vs @xmath208 ( @xmath209 ) , where the solid line is a least - squares fit to eq.([eq23 ] ) .
the fitting parameters are given in the text.,width=283 ] we have measure the @xmath13 dependence of @xmath65 under the condition with various combinations of @xmath12 , @xmath18 , and @xmath10 .
figures [ fig03 ] and [ fig04 ] show the @xmath13 dependence of the relaxation rate @xmath8 at various @xmath12 ( @xmath210 k ) for @xmath157 sec and @xmath211 sec , respectively , where @xmath159 oe .
the relaxation rate @xmath8 exhibits a broad peak at a characteristic time @xmath9 in the fm phase as well as the rsg phase , indicating the aging behaviors . in fig .
[ fig05 ] we show the @xmath13 dependence of @xmath65 at @xmath159 oe for @xmath204 sec .
we find that @xmath14 is well described by a stretched exponential relaxation form given by eq.([eq16 ] ) with @xmath138 for @xmath212 sec .
the least squares fits of these data to eq.([eq16 ] ) with @xmath138 yields the parameters @xmath19 , @xmath17 and @xmath122 .
figure [ fig06](a ) shows the @xmath12 dependence of @xmath9 and @xmath19 for @xmath213 sec , where @xmath159 oe . the @xmath12 dependence of @xmath9 almost agrees well with that of @xmath19 for @xmath214 k. this indicates that the relaxation is dominated by a stretched exponential relaxation with @xmath138 ( see sec .
[ backb ] ) .
the relaxation time @xmath19 ( @xmath20 ) shows a broad peak centered around 5.5 k between @xmath4 and @xmath6 , and a local minimum around @xmath6 .
it decreases with further increasing @xmath12 below @xmath6 .
as far as we know , there has been no report on the broad peak of @xmath9 ( or @xmath19 ) in the fm phase of reentrant ferromagnts .
the existence of the broad peak around 5.5 k suggests the chaotic nature of the fm phase in our system .
the drastic increase of @xmath9 ( or @xmath19 ) below @xmath6 with decreasing @xmath12 is a feature common to the sg phases of typical sg systems .
note that at @xmath215 k no peak in @xmath8 is observed for @xmath216 sec .
the value of @xmath19 is estimated as @xmath217 sec , which is much larger than @xmath10 .
in contrast , as shown in fig .
[ fig06](b ) for @xmath204 sec , the value of @xmath9 is larger than that of @xmath19 in the fm phase .
both @xmath9 and @xmath19 show a broad peak at @xmath12 between 4 and 5 k. this peak temperature is slightly lower than that for @xmath157 sec , suggesting that the relaxation mechanism is dependent on @xmath10 . in the inset of fig .
[ fig06](a ) we show the plot of @xmath207 as a function of @xmath208 . in the limited temperature range ( @xmath218 k )
, @xmath207 can be approximated by an exponential @xmath12 dependence @xmath219 with @xmath220 sec@xmath221 and @xmath222 .
our value of @xmath223 is relatively larger than those derived by hoogerbeets et al .
( @xmath224)@xcite from an analysis of thermoremanent magnetization ( trm ) relaxation measurements on four sg systems : ag : mn ( 2.6 at .
% ) , ag : mn ( 4.1 at .
% ) , ag:[mn ( 2.6 at .
% ) + sb ( 0.46 at .
% ) ] , and cu : mn ( 4.0 at .
note that in their work the stretched exponential is taken as representative of the short time ( @xmath225 ) relaxation .
one must thus be careful in comparing the results .
( color online)@xmath147 vs @xmath12 for the three cases : ( i ) @xmath159 oe and @xmath204 sec ( @xmath174 ) , ( ii ) @xmath18 = 1 oe and @xmath157 sec ( @xmath176 ) , and ( iii ) ) @xmath18 = 10 oe and @xmath226 sec ( @xmath227 ) .
the soild lines are guides to the eyes.,width=283 ] figure [ fig07 ] shows the @xmath12 dependence of @xmath147 ( the peak height of @xmath8 at @xmath132 ) for the following three cases : ( i ) @xmath159 oe and @xmath204 sec , ( ii ) @xmath159 oe and @xmath228 sec , and ( iii ) @xmath229 oe and @xmath226 sec .
the peak height @xmath147 at @xmath159 oe exhibits two peaks around @xmath230 and at 7.0 k just below @xmath4 , independent of @xmath10 ( @xmath231 sec or @xmath211 sec ) .
note that the peak of @xmath147 at 7.0 k is strongly dependent of @xmath18 : it disappears at @xmath229 oe for @xmath226 sec .
( color online)(a ) @xmath17 vs @xmath12 and ( b ) @xmath122 vs @xmath12 , where @xmath204 sec ( @xmath174 ) and @xmath232 sec ( @xmath176 ) . @xmath159
the soild lines are guides to the eyes.,width=283 ] in fig .
[ fig08](a ) we show the plot of the exponent @xmath17 as a function of @xmath12 for @xmath157 and @xmath211 sec , where @xmath159 oe . for @xmath204 sec the exponent @xmath17 increases with increasing @xmath12 and exhibits a peak at @xmath233 k just above @xmath6 .
the exponent @xmath17 decreases with further increasing @xmath12 , showing a local minimum around 7.5 k. for @xmath157 sec , the exponent @xmath17 exhibits a local minimum ( @xmath234 ) around 2.5 k ( @xmath235 ) and a local maximum ( @xmath236 ) around 4.0 k. the local minimum of @xmath17 below @xmath6 is a feature common to that of sg systems , where @xmath17 has a local minimum at @xmath237 given by @xmath238.@xcite the exponent @xmath17 increases as @xmath12 approaches @xmath90 from the low-@xmath12 side .
the overall variation of @xmath17 below @xmath90 is qualitatively the same as that calculated for the sherrington - kirkpatrick model.@xcite figure [ fig08](b ) shows the @xmath12 dependence of the amplitude @xmath122 for @xmath157 and @xmath211 sec .
we find that @xmath122 is strongly dependent on @xmath12 .
irrespective of @xmath10 , the amplitude @xmath122 shows two local maxima at 3 and 7.0 k. for comparison , we evaluate the @xmath12 dependence of @xmath148 [ @xmath239 , which corresponds to eq.([eq22 ] ) with @xmath138 . as the values of @xmath17 and @xmath122 ,
here we use the experimental values at each @xmath12 shown in figs .
[ fig08](a ) and ( b ) .
we find that the @xmath12 dependence of @xmath148 thus calculated is in excellent agreement with that of @xmath147 ( see fig .
[ fig07 ] ) experimentally determined from the data of @xmath8 vs @xmath13 .
( color online)@xmath202 vs @xmath13 at various @xmath18 . ( a )
@xmath240 k and @xmath157 sec .
( b ) @xmath241 k and @xmath213 sec.,width=283 ] ( color online)(a ) @xmath9 vs @xmath18 and @xmath19 vs @xmath18 at @xmath242 and 7.0 k. the soild lines are guides to the eyes .
( a ) @xmath9 vs @xmath18 at @xmath12 3.3 and 7.0 k. the solid lines are least - squares fits to eq.([eq25 ] ) for data at low @xmath18 .
the fitting parameters are given in the text.,width=283 ] figures [ fig09](a ) and ( b ) show the @xmath13 dependence of @xmath8 at various @xmath18 , where @xmath12 = 3.3 and 7.0 k , and @xmath157 sec .
the relaxation rate @xmath8 shows a peak at @xmath9 .
this peak greatly shifts to the short-@xmath13 side with increasing @xmath18 . in fig .
[ fig10](a ) we make a plot of @xmath9 and @xmath19 as a function of @xmath18 at @xmath12 = 3.3 and 7.0 k , where @xmath19 is determined from the least squares fit of the data of @xmath65 vs @xmath13 to the stretched exponential relaxation with @xmath138 .
the value of @xmath9 is almost the same as that of @xmath19 at each @xmath12 and @xmath18 , indicating that the exponent @xmath118 is equal to zero ( see sec .
[ backb ] ) .
the time @xmath9 undergoes a dramatic decrease at low fields : @xmath243 2 - 3 oe at @xmath241 k and @xmath229 oe at @xmath242 k. note that the decrease of @xmath9 with @xmath18 has been reported for the zfc relaxation in the rsg phase of the reentrant ferromagnet fe@xmath38al@xmath37.@xcite the increase of @xmath17 and @xmath207 with increasing @xmath18 has been reported for the trm decay in the sg systems such as ag:[mn ( 2.6 at .
% ) + sb ( 0.46 at .
% ) ] @xcite and cu : mn ( 6 at .
% ) .@xcite the shift of @xmath9 for @xmath8 under the @xmath18-shift process is governed by the mean domain - size @xmath244 and overlap lengths @xmath245 defined by@xcite @xmath246 respectively , where @xmath247 and @xmath248 are the corresponding exponents for @xmath244 and @xmath245 , and @xmath249 is the magnetic field corresponding to a wall stiffness @xmath250 ( a typical energy setting the scale of free energy barriers between conformations ) . using the scaling concept , the shift of @xmath9 for @xmath8 under the @xmath18-shift can be approximated by@xcite @xmath251 in the limit of @xmath252 , where @xmath253 is constant proportional to @xmath247 @xmath254 . in fig .
[ fig10](b ) we make a plot of @xmath9 vs @xmath18 at @xmath242 and 7.0 k. we find that the data of @xmath9 vs @xmath18 is well described by eq.([eq25 ] ) with @xmath255 and @xmath256 sec for @xmath257 oe at @xmath240 k and @xmath258 and @xmath259 sec for @xmath260 oe at @xmath241 k. ( color online)(a ) @xmath147 vs @xmath18 , ( b ) @xmath17 vs @xmath18 , and ( c ) @xmath122 vs @xmath18 at @xmath242 and 7.0 k. @xmath157 sec . @xmath159
the soild lines are guides to the eyes.,width=283 ] in fig .
[ fig11](a ) we show the @xmath18 dependence of @xmath147 at @xmath242 and 7.0 k , where @xmath157 sec .
the value of @xmath147 is almost independent of @xmath18 for @xmath261 oe .
in contrast , the value of @xmath147 at 7.0 k greatly reduces with increasing @xmath18 .
this result agrees well with that reported by jonason and nordblad@xcite with the gradual decrease of @xmath147 at 16 k with @xmath18 and the dramatic decrease in @xmath147 at 30 k with @xmath18 at weak fields less than 0.2 oe . figure [ fig11](b ) shows the @xmath18 dependence of the exponent @xmath17 at @xmath240 and 7.0 k , where @xmath157 sec .
the @xmath18 dependence of @xmath17 at 7.0 k is rather different from that at 3.3 k at low @xmath18 .
the value of @xmath17 at @xmath242 k has a local minimum at @xmath262 oe , while the value of @xmath17 at @xmath241 k has a local maximum around 10 oe .
the value of @xmath17 tends to saturate to 0.82 - 0.84 at @xmath18 higher than 30 oe , independent of @xmath12 .
figure [ fig11](c ) shows the @xmath18 dependence of @xmath122 at @xmath242 and 7.0 k , where @xmath157 sec .
the amplitude @xmath122 at 7.0 k deviates from that at 3.3 k above 5 oe , indicating the fragility of the aging state in the fm phase against a weak magnetic field - perturbation . for comparison
, we make a plot of @xmath148 [ @xmath149 as a function of @xmath18 , where @xmath122 and @xmath17 experimentally determined are used .
there is a very good agreement between @xmath147 ( see fig .
[ fig11](a ) ) and @xmath148 .
( color online)@xmath202 vs @xmath13 at various @xmath10 ( @xmath263 sec ) . @xmath159 oe .
( a ) @xmath264 k. ( b ) @xmath241 k.,width=283 ] ( color online)(a ) @xmath9 vs @xmath10 and @xmath19 vs @xmath10 .
the solid line is a least - squares fit to @xmath265 for @xmath19 at 3.5 k ( @xmath266 sec ) .
the fitting parameters are given in the text .
( b ) @xmath17 vs @xmath10 , and ( c ) @xmath122 vs @xmath10 at @xmath267 and 7.0 k. the soild lines are guides to the eyes.,width=283 ] figure [ fig12 ] shows the @xmath13 dependence of @xmath8 at various @xmath10 for @xmath267 and 7.0 k , where @xmath159 oe . the relaxation rate @xmath8 shows a peak at @xmath9 .
this peak shifts to long-@xmath13 side with increasing @xmath10 .
the least squares fit of the data of @xmath14 vs @xmath13 for @xmath268 sec to eq.([eq16 ] ) yields the exponent @xmath17 , the relaxation time @xmath19 , and amplitude @xmath122 for the stretched exponential relaxation .
figure [ fig13](a ) shows the @xmath10 dependence of @xmath9 and @xmath19 at @xmath267 and 7.0 k. at @xmath269 k , @xmath9 is nearly equal to @xmath19 for @xmath270 sec , while at @xmath267 k @xmath9 is much longer than @xmath19 for @xmath271 sec .
both @xmath9 and @xmath19 increases with increasing @xmath272 it has been reported that @xmath19 varies exponentially with @xmath10 as @xmath273 for ag : mn ( 2.6 at .
% ) .@xcite our data of @xmath19 vs @xmath10 at 3.5 k are well fitted to this equation in spite of the limited @xmath10 region ( @xmath274 sec ) ; @xmath275 sec and @xmath276 sec .
figure [ fig13](b ) shows the @xmath10 dependence of @xmath17 at @xmath277 and 7.0 k. the exponent @xmath17 at @xmath267 k shows a local minimum at @xmath278 sec and increases with further increasing @xmath10 .
in contrast , the exponent @xmath17 at 7.0 k is smaller than that at 3.5 k for any @xmath10 .
it exhibits a local maximum around @xmath279 sec .
similar local minimum in @xmath17 vs @xmath10 is observed in the sg phase of the spin glass cr@xmath280fe@xmath281 [ @xmath282 k]:@xcite @xmath17 at @xmath283 k shows a local minimum around @xmath284 sec and increases with further increasing @xmath10 .
in contrast , there is no local minimum in @xmath17 vs @xmath10 in the rsg phase of ( fe@xmath29ni@xmath30)@xmath31mn@xmath32@xcite ( @xmath17 increases with increasing @xmath10 ) and in the sg phase of ag : mn ( 2.6 and 4.1 at .
% ) @xcite ( @xmath17 is independent of @xmath10 ) .
figure [ fig13](c ) shows the @xmath10 dependence of the amplitude @xmath122 at @xmath277 and 7.0 k. the amplitude @xmath122 at 3.5 k shows a local minimum around @xmath278 sec , while @xmath122 at 7.0 k is almost independent of @xmath10 . in summary , so far there has been no theory to explain the @xmath10 dependence of @xmath17 , @xmath19 , @xmath9 and @xmath122 in both the rsg and fm phases .
however , our results suggest that the rsg and fm phases are essentially nonequilibrium phases . the nature of the aging behavior in the fm phase is rather different from that in the rsg phase .
( color online)(a ) @xmath13 dependence of @xmath202 at @xmath285 k under the @xmath12 shift from @xmath286 to @xmath287 , where @xmath286 = 4.9 , 5.6 , and 6.3 k. @xmath159 oe .
the zfc aging protocol is as follows : quenching of the system from 50 k to @xmath286 , and isothermal aging at @xmath288 and @xmath51 for @xmath204 sec .
the measurement was started at @xmath64 , just after @xmath12 was shifted from @xmath286 to @xmath287 and subsequently @xmath18 (= 1 oe ) was turned on .
( b ) @xmath9 vs @xmath70 .
the solid line is a fitting curve to eq.([eq27 ] ) .
the fitting parameters are given in the text.,width=283 ] figure [ fig14](a ) show the relaxation rate @xmath8 measured at @xmath289 k for different initial temperature @xmath286 (= 4.9 , 5.6 , and 6.3 k ) , where @xmath204 sec and @xmath159 oe .
the measurement was carried out as follows .
after the zfc protocol from 50 k to @xmath286 below @xmath4 , the system was kept at @xmath286 for a wait time @xmath10 . immediately after @xmath12 was changed from @xmath286 to @xmath287 ( the positive @xmath12 shift ; @xmath290 where @xmath291 ) , the magnetic field was applied and the @xmath13 dependence of @xmath65 was measured . with increasing @xmath70 ,
the time @xmath9 shifts to the low-@xmath13 side .
the relaxation rate @xmath8 has a peak at @xmath292 - 460 sec at @xmath293 k and 1.4 k , suggesting a rejuvenation of the system during the positive @xmath12-shift . for @xmath294
k two peaks appear at 570 sec and 1210 sec , respectively , suggesting a coexistence of two characteristic ages in the system due to the partial rejuvenation and original aging at @xmath286 for @xmath204 sec .
this behavior is explained in terms of the overlap length @xmath71 , which becomes smaller as @xmath70 becomes large .
when @xmath295 corresponding to the case of @xmath296 , no initialization occurs in the system .
there is only one type of domain .
thus @xmath8 exhibits a peak around @xmath10 . for @xmath297 corresponding to the case of @xmath298 , some domains fractures into smaller domains of dimension @xmath71 .
thus @xmath8 has two peaks at @xmath9 ( @xmath299 ) and a time much shorter than @xmath9 .
the shift of @xmath9 for @xmath8 under the @xmath12-shift aging process is governed by the mean domain - size @xmath244 and overlap length @xmath300 defined by@xcite @xmath301 where @xmath302 is the corresponding exponent and @xmath303 is the temperature corresponding to the wall stiffness @xmath250 . using the scaling concepts ,
the shift of @xmath9 for @xmath8 under the @xmath12-shift can be approximated by@xcite @xmath304 in the limit of @xmath252 , where @xmath305 is a constant proportional to @xmath306 . in fig .
[ fig14](b ) we make a plot of @xmath9 vs @xmath70 at @xmath289 k. we find that the data of @xmath9 vs @xmath70 is described by eq.([eq27 ] ) with @xmath307 and @xmath308 sec for @xmath309 k. similar behaviors in @xmath8 under the @xmath12 shift have been observed in other reentrant ferromagnets . in the fm phase of ( fe@xmath22ni@xmath23)@xmath24p@xmath25b@xmath26al@xmath3,@xcite the amplitude of the maximum in @xmath8 at @xmath13 @xmath299 decreases with increasing @xmath310 ( both the positive and negative @xmath12-shift ) and a second maximum gradually develops at a shorter time , indicating a partial rejuvenation . in the rsg phase of ( fe@xmath29ni@xmath30)@xmath31mn@xmath32 double peaks in @xmath8 vs @xmath13
are observed under a temperature - cycling experiment:@xcite an initial wait time @xmath10 ( @xmath311 sec ) at @xmath12 , followed by a temperature cycle @xmath312 of the duration @xmath313 (= 300 sec ) .
( color online)(a ) and ( b ) @xmath171 dependence of @xmath113 at various @xmath12 . @xmath173 moe .
the solid lines are least - squares fits to eq.([eq15 ] ) for @xmath314 and 6.7 k. the fitting parameters are given in the text.,width=283 ] in previous paper@xcite we have reported the @xmath171 dependence of the equilibrium absorption @xmath113 at various @xmath12 in the vicinity of @xmath6 , respectively .
the absorption @xmath113 curves exhibit different characteristics depending on @xmath12 . above @xmath6
, @xmath113 shows a peak at a characteristic frequency , shifting to the low @xmath171-side as @xmath12 decreases .
below @xmath6 , @xmath113 shows no peak for @xmath315 hz . in figs .
[ fig15 ] ( a ) and ( b ) , for comparison we show the @xmath171 dependence of @xmath113 at fixed @xmath12 ( @xmath316 k and @xmath317 k ) .
the absorption @xmath113 decreases with increasing @xmath171 in the rsg phase , while it increases with increasing @xmath171 in the fm phase .
the @xmath171 dependence of @xmath318 can be expressed by a power law form ( @xmath319 ) , where the magnitude of the exponent @xmath93 is very small .
the sign of @xmath93 may be negative for the rsg phase and is positive for the fm phase . according to the fluctuation - dissipation theorem ,
the magnetic fluctuation spectrum @xmath320 is related to @xmath318 by @xmath321 .
then @xmath320 is proportional to @xmath322 , which is similar to @xmath323 character of typical sg systems .
note that for the spin glass cdcr@xmath324in@xmath38s@xmath35 ( @xmath325 k),@xcite @xmath93 ( @xmath326 ) is positive at @xmath90 and decreases with decreasing @xmath12 .
the sign of @xmath93 changes from positive to negative below 5 k : @xmath327 .
in contrast , @xmath113 for @xmath328 k increases with increasing @xmath171 .
the least squares fit of the data of @xmath113 vs @xmath171 for @xmath329 hz to the power law form ( @xmath319 ) yields the value of @xmath93 .
the exponent @xmath93 increase with increasing @xmath12 for @xmath330 k : @xmath331 at 6.1 k , @xmath332 at 6.7 k and @xmath333 at 7.2 k. ( color online)@xmath13 dependence of @xmath60 at ( a ) 3.75 k , ( b ) @xmath241 k , and ( c ) @xmath334 k. @xmath171 = 0.05 , 0.1 , 0.5 and 1 hz . the time @xmath13 is the time taken after the sample is quenched from 50 k to @xmath12 .
the solid lines are the least - squares fits to eq.([eq28 ] ) .
the fitting parameters are listed in table [ table01].,width=283 ] .[table01]exponents @xmath93 ( or @xmath111 ) determined from the least squares fits of @xmath60 at @xmath171 to the power law form by given by eq.([eq28 ] ) .
@xmath12 = 3.5 , 3.75 , 7 and 8.5 k. the parameters with the suffix `` c '' denote those determined from the least squares fits of the data to a power - law form @xmath335 instead of @xmath336 , where @xmath337 is regarded as a fitting parameter . [
cols="<,<,<,<,<",options="header " , ] we have also measured the @xmath13 dependence of @xmath60 at @xmath12 = 3.3 , 3.5 , 3.75 , 7 and 8.5 k , where @xmath51 , @xmath338 oe , and @xmath171 = 0.05 - 5 hz . the system was quenched from 100 k to @xmath12 at time ( age ) zero .
the origin @xmath64 is a time at which the temperature becomes stable at @xmath12 within @xmath156 k. the change of @xmath60 with @xmath13 below @xmath6 is not so prominent compared to that above @xmath6 , partly because of relatively small magnitude of @xmath39 below @xmath6 . in figs .
[ fig16](a)-(c ) , we shows the @xmath13 dependence of @xmath60 at various @xmath171 , where @xmath12 = 3.75 , 7.0 , and 8.5 k. we find that the decrease of @xmath39 with increasing @xmath13 is well described by a power - law decay of the quasi - equilibrium part @xmath339 where @xmath340 and @xmath341 are @xmath13-independent constants .
it is predicted that @xmath106 exhibits no @xmath110-scaling law [ @xmath106 @xmath342 . in the limit of @xmath343
, @xmath60 tends to @xmath340 which is assumed to be equal to @xmath113 .
the least squares fit of the data of @xmath60 at each @xmath12 to eq.([eq28 ] ) yields parameters listed in table [ table01 ] at @xmath12 = 3.5 , 3.75 , 7 , and 8.5 k. the exponent @xmath93 below @xmath6 is nearly equal to 0.07 - 0.084 . at 7.0 k ,
the exponent @xmath93 is dependent on @xmath171 : @xmath344 at @xmath345 hz and @xmath346 at @xmath347 hz .
the value of @xmath348 at @xmath241 k tends to decrease with increasing @xmath171 .
this seems to be inconsistent with our result that @xmath349 increases with increasing @xmath171 .
since the value of @xmath122 tends to increase with increasing @xmath171 , it follows that the second term of eq.([eq28 ] ) does not obey the @xmath110-scaling law as is predicted . in general , @xmath60 consists of the quasi - equilibrium part @xmath106 and the aging relaxation part @xmath108 .
the aging part @xmath108 is predicted to be described by a @xmath110-scaling law [ @xmath350.@xcite note that the @xmath110-scaling of @xmath60 is observed in the rsg phase ( @xmath283 k ) and the fm phase ( @xmath351 k ) of the reentrant ferromagnet cdcr@xmath352in@xmath353s@xmath35 ( @xmath354 k and @xmath355 k).@xcite the absorption @xmath60 mainly arises from @xmath108 but not from @xmath106 .
the aging relaxation part @xmath356 is assumed to be described by the power - law form [ @xmath357^{-b}$ ] , instead of @xmath358 , where @xmath359 for both temperatures and @xmath337 is an off - set time which takes into account the fact that the cooling procedure is not instantaneous .
the least squares fit of our data of @xmath60 vs @xmath13 at @xmath241 k to the power - law form @xmath360 yields the parameters listed in table [ table01 ] , where @xmath337 is one of the fitting parameters .
we find that the value of @xmath111 for @xmath345 hz ( @xmath361 ) is close to that for cdcr@xmath352in@xmath353s@xmath35.@xcite however , the value of @xmath111 for higher @xmath171 is still on the order of 0.047 - 0.07 , which is much smaller than 0.2 .
this result indicates that @xmath362 in our system .
the nature of the rsg and fm phases for cu@xmath0co@xmath1cl@xmath2-fecl@xmath3 gbic is summarized as follows .
the static and dynamic behavior of the rsg phase is characterized by that of the normal sg phase : a critical exponent @xmath363 for the sg parameter and a dynamic critical exponent @xmath364 for the characteristic relaxation time.@xcite the aging phenomena are clearly seen in the rsg phase , although no appreciable nonlinear magnetic susceptibility is observed .
the exponent @xmath17 for the stretched exponential relaxation exhibits a local minimum just below @xmath6 , and increases when @xmath12 approaches @xmath6 from the low-@xmath12 side .
the relaxation time @xmath19 for the stretched exponential relaxation drastically increases with decreasing @xmath12 below @xmath6 . in this sense
, it follows that the rsg phase below @xmath6 is a normal sg phase .
in contrast , the dynamic nature of the fm phase is rather different from that of an ordinary ferromagnet .
a prominent nonlinear susceptibility is observed between @xmath6 and @xmath4.@xcite aging phenomena and a partial rejuvenation effect under the @xmath12-shift are also seen in the fm phase .
this aging state disappears even in a weak magnetic field ( @xmath365 ; @xmath366 - 5 oe ) .
the relaxation time @xmath19 ( or @xmath367 ) as a function of @xmath12 exhibits a local maximum between @xmath6 and @xmath4 .
these results suggest that no long range ferromagnetic correlation exist in the rsg phase .
the chaotic behavior of the fm phase is rather similar to that observed in the rsg phase .
similar behaviors have been observed in the reentrant ferromagnets . in
( fe@xmath22ni@xmath23)@xmath24p@xmath25b@xmath26al@xmath3 @xcite the relaxation time diverges at a finite temperature ( @xmath368 ) with a dynamic critical exponent similar to that observed for normal sg transitions .
the fm phase just above @xmath6 shows a dynamic behavior characterized by an aging effect and chaotic nature similar to that of rsg phase . in cdcr@xmath33in@xmath34s@xmath35 with @xmath36 = 0.90 , 0.95 , and 1.0,@xcite the aging behavior of the low frequency ac susceptibility
is observed both in the fm and rsg phases , with the same qualitative features as in normal sg systems . in the mean - field picture,@xcite a true reentrance from the fm phase to the normal sg phase
is not predicted .
there exists a normal fm long range order in the fm phase .
this picture , which assumes infinite - range interactions , is not always appropriate for real reentrant magnets where the short - range interactions are large and random in sign and the spin symmetry is rather heisenberg - like than ising - like . in a picture proposed by aeppli
et al.@xcite , the system in the fm phase consists of regions which by themselves would order ferromgnetically and other regions forming paramagnetic ( pm ) clusters
. the frustrated spins in the pm clusters can generate random molecular fields which act on the unfrustrated spins in the infinite fm network . in the fm phase well above @xmath6 ,
the fluctuations of the spins in the pm clusters are so rapid that the fm network is less influenced by them and their effect is only to reduce the net fm moment .
on decreasing the temperature toward @xmath6 , the thermal fluctuations of the spins in the pm clusters become slower .
the coupling between the pm clusters and the fm network becomes important and the molecular field from the slow pm spins acts as a random magnetic field .
this causes a breakup of the fm network into finite - sized clusters . below @xmath6
, the ferromagnetism completely disappears , leading to a rsg phase .
we have shown that the @xmath13 dependence of @xmath8 is strongly dependent on @xmath12 , @xmath18 , and @xmath10 .
here we compare our results on @xmath8 with those observed in other reentrant ferromagnets . in order to facilitate the comparison ,
we have determined the time @xmath9 and the peak height @xmath147 from their original data of @xmath202 vs @xmath13 .
the first case is the results of @xmath8 for ( fe@xmath22ni@xmath23)@xmath24p@xmath25b@xmath26al@xmath3 ( @xmath369 k and @xmath370 k ) reported by jonason and nordblad.@xcite they have shown that @xmath8 exhibits a peak at @xmath9 in both rsg and fm phases , indicating that the aging phenomena occur in both phases .
these results are in good agreement with our results [ see figs . [ fig03 ] and [ fig04 ] ] .
the aging state in the fm phase is fragile against a very weak magnetic - field perturbation .
there is a threshold magnetic field @xmath371 , below which there is a linear response of the relaxation .
the field @xmath371 undergoes a dramatic decrease with increasing @xmath12 above @xmath6 .
the value of @xmath371 ( @xmath372 oe ) for @xmath209 is much lower than that for our system ( @xmath373 2 - 5 oe ) .
the time @xmath9 decreases with increasing @xmath12 at low @xmath12 , showing a local minimum around 23 k , and increases with increasing @xmath12 between 25 and 30 k. although there have been no data on @xmath8 above 30 k , it is assumed that @xmath9 shows a local maximum between 30 k and @xmath4 , since @xmath9 should reduce to zero above @xmath4 .
such a possible local maximum is similar to a local maximum of @xmath9 in the rsg phase observed in our system [ see figs . [
fig06](a ) and ( b ) ] . here
we discuss the @xmath12 dependence of the peak height @xmath147
. the peak height @xmath147 ( @xmath374 oe and @xmath375 sec ) exhibits a peak at 13 k just below @xmath6 , having a local minimum at 25 k , and tends to increase with further increasing @xmath12 .
although there have been no data on @xmath8 above 30 k , it is assumed that @xmath147 shows a local maximum between 30 k and @xmath4 , since @xmath147 should reduce to zero above @xmath4 .
these two peaks of @xmath147 vs @xmath12 are similar to two local maxima around @xmath6 and between @xmath6 and @xmath4 in our system .
next we discuss the @xmath18 dependence of the peak height @xmath147 . the peak height @xmath147 ( @xmath376 sec ) decreases with increasing @xmath18 at both @xmath377 and 30 k. the drastic decrease of @xmath147 at @xmath378 k occurs even at @xmath379 oe , while the decrease of @xmath147 at 16 k with @xmath18 is much weaker at low @xmath18 below 1 oe .
the @xmath18-dependence of @xmath147 at 16 and 30 k is similar to that at @xmath242 and 7.0 k for our system ( see fig .
[ fig11](a ) ) .
the second case is the results of @xmath8 for fe@xmath40al@xmath41 ( @xmath380 k and @xmath381 k ) .
motoya et al.@xcite have reported the @xmath13 dependence of @xmath8 at various @xmath12 only in the rsg phase , where @xmath18 = 10 oe and @xmath382 sec .
the relaxation rate @xmath8 shows a peak at @xmath9 .
the time @xmath9 decreases with increasing @xmath12 below @xmath6 , while the peak height @xmath147 exhibits a peak at 50 k well below @xmath6 .
these results are in good agreement with our results : see figs . [
fig06](a ) and ( b ) for @xmath383 and fig .
[ fig07 ] for @xmath384 having a local maximum in @xmath147 at 3.2 k below @xmath6 .
motoya et al.@xcite have also measured the @xmath13 dependence of @xmath8 at 10 k at various @xmath18 , where @xmath385 k and @xmath382 sec . both @xmath9 and @xmath147 decrease with increasing @xmath18 .
this result is in good agreement with our results : see fig .
[ fig10](a ) for @xmath386 and fig .
[ fig11](a ) for @xmath387 . as is described in sec .
[ backa ] , the @xmath13 dependence of @xmath14 and @xmath60 arises from that of the spin auto - correlation function @xmath88 and @xmath388 , respectively , where @xmath389 . here
we assume that @xmath88 is defined by either eqs.([eq05 ] ) or ( [ eq06 ] ) .
the function @xmath95 is approximated by a scaling function @xmath390 , where @xmath391 has a power - law form @xmath392 for @xmath393 . in the present work , @xmath10 takes various values as @xmath394 - 100 sec ( @xmath395 hz ) for @xmath60 and as a wait time @xmath396 sec for @xmath14 .
the function @xmath97 is independent of @xmath10 and is expressed by a power - law form ( @xmath397 ) given by eq.([eq07 ] ) . by the appropriate choice of @xmath13 and @xmath10 ( or @xmath80 ) , it is experimentally possible to separate the quasi - equilibrium part @xmath97 and the aging part @xmath95 .
our experimental results are as follows .
( i ) the absorption @xmath60 mainly comes from the quasi - equilibrium contribution @xmath97 .
this is also supported by the fact that no @xmath110-scaling is observed for @xmath60 .
the exponent @xmath93 is positive and very small ; typically @xmath398 at @xmath399 = 3.5 k in the rsg phase ( see table [ table01 ] ) .
this value of @xmath93 is considered to coincide with the exponent @xmath118 for the stretched exponential relaxation .
( ii ) the susceptibility @xmath14 for @xmath69 and @xmath400 mainly comes from the aging contribution . as shown in fig .
[ fig08 ] , the exponent @xmath17 depends on @xmath12 and is 0.78 at @xmath401 .
ozeki and ito@xcite have studied the nonequilibrium relaxation of the @xmath402 ising model in three dimensions .
they have shown that @xmath88 obeys the power - law form : @xmath403 for @xmath404 and @xmath405 and for @xmath406 , where @xmath407 and @xmath408 at @xmath409 .
the value of @xmath410 is consistent with that obtained by ogielski:@xcite @xmath411 at @xmath412 .
the exponent @xmath413 characterizes the nonequilibrium relaxation .
the exponent @xmath410 at @xmath90 is described by @xmath414 , where @xmath36 is the dynamic critical exponent and @xmath415 is the exponent of the sg order parameter . in the previous paper@xcite
we have reported the values of @xmath415 and @xmath36 as @xmath416 and @xmath417 , leading to @xmath418 .
the agreement between the theory and experiment is very good .
it is predicted that the @xmath13 dependence of @xmath60 for @xmath419 is dominated by that of the aging part @xmath356 : @xmath420 with @xmath421 .
experimentally it has been confirmed that @xmath60 in the sg phase obeys the @xmath110-scaling law with @xmath422 for fe@xmath201mn@xmath201tio@xmath3 ( @xmath423 k)@xcite and @xmath424 for cu@xmath201co@xmath201cl@xmath2-fecl@xmath3 gbic.@xcite we discuss the validity of the stretched exponential relaxation .
we show that @xmath14 can be well described by the stretched exponential relaxation with @xmath138 . however , we find that the least squares fit of the data of @xmath65 vs @xmath13 to the stretched exponential relaxation with @xmath425 does not work well in determining the values of @xmath17 and @xmath118 .
the value of @xmath17 is very sensitive to the small change in @xmath118 . in sec .
[ backb ] we assume that @xmath14 is given by eq.([eq16 ] ) .
then @xmath8 has a peak at @xmath426 described by eq.([eq20 ] ) . as shown in the contour plot of @xmath116 in the ( @xmath17,@xmath118 ) plane of fig .
[ fig01](a ) , the value of @xmath116 strongly depends on the values of @xmath17 and @xmath118 . in fig .
[ fig01](b ) we show the plot of @xmath116 as a function of @xmath118 for a fixed @xmath17 .
the value of @xmath116 takes 1 at @xmath138 and drastically decreases with increasing @xmath118 .
experimentally the value of @xmath116 is estimated from the ratio of @xmath9 to @xmath19 , where @xmath9 is the time at which @xmath8 takes a peak and @xmath19 is from the least squares fit of the data of @xmath65 vs @xmath13 to the stretched exponential relaxation with @xmath138 .
the ratio @xmath116 provides a good measure to determine whether the stretched exponential relaxation is valid for our system .
experimentally we find that the ratio @xmath116 is dependent on @xmath10 and @xmath12 ( @xmath427 k ) : @xmath428 for @xmath157 sec and @xmath429 for @xmath204 sec . in our simple model
, it follows that @xmath118 becomes negative when @xmath430 . furthermore , in sec .
[ disc ] , we show that @xmath431 and @xmath432 at @xmath6 . however , these values do not satisfy the inequality ( @xmath137 ) .
this inequality is required for the peak of @xmath8 to appear in the case of the stretched exponential relaxation . when @xmath433 , @xmath434 , which is independent of @xmath118 .
the value of @xmath116 becomes zero as @xmath17 tends to unity . in our system ,
@xmath17 = 0.73 - 0.81 . using the inequality [ @xmath435 ,
the upper limit of @xmath118 can be estimated as 0.0475 for @xmath436 ( or @xmath437 ) .
here we note that the exponent @xmath111 for @xmath356 [ @xmath109 is on the same order as the value of @xmath438 or @xmath439 .
in general the relationship between @xmath17 and @xmath111 may be derived from the relaxation rate @xmath8 as a function @xmath104 , where @xmath69 is replaced by @xmath440 .
cu@xmath1co@xmath0cl@xmath2-fecl@xmath3 gbic undergoes successive transitions at the transition temperatures @xmath4 ( @xmath441 k ) and @xmath6 ( @xmath442 k ) .
the fm phase of our system is characterized by the aging phenomena and nonlinear magnetic susceptibility .
the relaxation rate @xmath8 exhibits a characteristic peak at @xmath9 close to a wait time @xmath10 below @xmath4 , indicating the occurrence of aging phenomena in both the rsg and fm phases .
the aging behavior in the fm phase is fragile against a weak magnetic - field perturbation . in the fm phase there
occurs a partial rejuvenation effect in @xmath8 under the @xmath12-shift perturbation .
the time ( @xmath13 ) dependence of @xmath14 around @xmath13 @xmath20 is well approximated by a stretched exponential relaxation .
the relaxation time @xmath19 ( @xmath20 ) exhibits a local maximum around 5 k , reflecting a chaotic nature of the fm phase .
it drastically increases with decreasing temperature below @xmath6 , as is usually seen in the sg phase of sg systems .
we would like to thank h. suematsu for providing us with single crystal kish graphite , and t. shima and b. olson for their assistance in sample preparation and x - ray characterization .
early work , in particular for the sample preparation , was supported by nsf dmr 9201656 .
k. jonason , j. mattsson , and p. nordblad , phys .
* 77 * , 2562 ( 1996 ) .
k. jonason , j. mattsson , and p. nordblad , phys .
b * 53 * , 6507 ( 1996 ) . k. jonason and p. nordblad , j. mag . mag . mater . * 177 - 181 * , 95 ( 1998 ) .
k. jonason and p. nordblad , eur .
j. b * 10 * , 23 ( 1999 ) .
mitchler , r.m .
roshko , and w. ruan , phil . mag .
b * 68 * , 539 ( 1993 ) .
d. li , r.m .
roshko , and g. yang , phys . rev .
b * 49 * , 9601 ( 1994 ) .
e. vincent , v. dupuis , m. alba , j. hammann , and j .-
bouchaud , europhys .
lett . * 50 * , 674 ( 2000 ) .
e. vincent , f. alet , m. alba , j. hammann , m. ocio , and j.p .
bouchaud , physica b * 280 * , 260 ( 2000 ) .
v. dupuis , e. vincent , m. alba , and j. hammann , eur .
j. b * 29 * , 19 ( 2002 ) .
k. motoya , h. hioki , and j. suzuki , j. phys .
soc . jpn .
* 72 * , 3212 ( 2003 ) .
m. suzuki and i.s .
suzuki , phys .
b * 69*,144424 ( 2004 ) . l. lundgren , in _ relaxation in complex systems and related topics _ , edited by i.a .
campbell and c. giovannella ( plenum press , new york , 1990 ) p.3 .
d.s . fisher and d.a .
huse , phys .
b * 38 * , 373 ( 1988 ) ; * 38 * , 386 ( 1988 ) .
e. vincent , j. hammann , m. ocio , l .- p .
bouchaud , and l. f. cugliandolo , in _ complex behavior of glassy systems _ , edited by m. rub and c. prez - vicente ( springer verlag , berlin , 1997 ) p.184 .
t. komori , h. yoshino , and h. takayama , j. phys .
soc . jpn . *
68 * , 3387 ( 1999 ) ; * 69 * , 1192 ( 2000 ) ; * 69 * , suppl .
a , 335 ( 2000 ) . m. picco , f. ricci - tersenghi , and f. ritort , eur .
. j. b * 21 * , 211 ( 2001 ) .
l. berthier and j .- p .
bouchaud , phys .
b * 66 * , 054404 ( 2002 ) .
ogielski , phys .
b * 32 * , 7384 ( 1985 ) .
koper and h.j .
hilhorst , j. phys .
france * 49 * , 429 ( 1988 ) .
chamberlin , g. mozurkewich , and r. orbach , phys .
* 52 * , 867 ( 1984 ) .
r. hoogerbeets , w .- l
. luo , and r. orbach , phys .
lett . * 55 * , 111 ( 1985 ) .
r. hoogerbeets , w .- l .
luo , and r. orbach , phys .
b * 34 * , 1719 ( 1985 ) . m. alba , m. ocio , and j. hammann , europhys . lett . * 2 * , 45 ( 1986 ) .
l. lundgren , p. nordblad , and p. svedlindh , phys .
b * 34 * , 8164 ( 1986 )
. m. alba , j. hammann , m. ocio , ph .
refregier , and h. bouchiat , j. appl . phys .
* 61 * , 3683 ( 1987 ) .
p. granberg , p. svedlindh , p. nordblad , and l. lundgren , phys .
b * 35 * , 2075 ( 1987 ) .
p. nordblad , l. lundgren , p. svedlindh , l. sandlund , and p. granberg , phys .
b * 35 * , 7181 ( 1987 ) .
r. hoogerbeets , w .- l .
luo , and r. orbach , phys .
b * 35 * , 7185 ( 1987 ) .
d. chu , g.g .
kenning , and r. orbach , phys .
* 72 * , 3270 ( 1994 ) .
d. chu , g.g .
kenning , and r. orbach , phil . mag .
b * 71 * , 479 ( 1995 ) .
bernardi , h. yoshino , k. hukushima , h. takayama , a. tobo , and a. ito , phys .
lett . * 86 * , 720 ( 2001 ) .
m. suzuki and i.s .
suzuki , cond - mat/0308285 ; eur .
j.b , submitted ( 2003 ) .
d. sherrington and s. kirkpatrick , phys .
32 * , 1792 ( 1975 ) .
h. takayama , j. magn .
magn . mater . *
272 - 276 * , 256 ( 2004 ) .
jnsson , h. yoshino , and p. nordblad , phys .
lett . * 89 * , 097201 ( 2002 ) .
jnsson , h. yoshino , and p. nordblad , phys .
* 90 * , 059702 ( 2003 ) .
j. hammann , m. ocio , and e. vincent , in _ relaxation in complex systems and related topics _ , edited by i.a .
campbell and c. giovannella ( plenum press , new york , 1990 ) p.11 .
s.f . edwards and p.w .
anderson , j. phys .
f**5 * * , 965 ( 1975 ) .
g. parisi , phys .
lett . * 43 * , 1754 ( 1979 ) .
g. toulouse , j. phys .
( france ) lett . * 41 * , l447 ( 1980 ) .
g. aeppli , s.m .
shapiro , r.j .
birgeneau , and h.s .
chen , phys .
b * 28 * , 5160 ( 1983 ) .
y. ozeki and n. ito , phys .
b * 64 * , 024416 ( 2001 ) . v. dupuis , e. vincent , j .-
bouchaud , j. hammann , a. ito , and h. aruga katori , phys .
b * 64 * , 174204 ( 2001 ) .
suzuki and m. suzuki , phys .
b * 68 * , 094424 ( 2003 ) . | aging dynamics of a reentrant ferromagnet cu@xmath0co@xmath1cl@xmath2-fecl@xmath3 graphite bi - intercalation compound has been studied using ac and dc magnetic susceptibility .
this compound undergoes successive transitions at the transition temperatures @xmath4 ( @xmath5 k ) and @xmath6 ( @xmath7 k ) . the relaxation rate @xmath8 exhibits a characteristic peak at @xmath9 close to a wait time @xmath10 below @xmath4 , indicating that the aging phenomena occur in both the reentrant spin glass ( rsg ) phase below @xmath6 and the ferromagnetic ( fm ) phase between @xmath6 and @xmath4 .
the relaxation rate @xmath8 ( @xmath11 ) in the fm phase exhibits two peaks around @xmath10 and a time much shorter than @xmath10 under the positive @xmath12-shift aging , indicating a partial rejuvenation of domains .
the aging state in the fm phase is fragile against a weak magnetic - field perturbation . the time ( @xmath13 )
dependence of @xmath14 around @xmath15 is well approximated by a stretched exponential relaxation : @xmath16 $ ] .
the exponent @xmath17 depends on @xmath10 , @xmath12 , and @xmath18 .
the relaxation time @xmath19 ( @xmath20 ) exhibits a local maximum around 5 k , reflecting a chaotic nature of the fm phase .
it drastically increases with decreasing temperature below @xmath6 . |
the weyl formalism for static axisymmetric vacuum solutions in 3 + 1 dimensions allows one to write down a solution that represents an arbitrary number of collinear neutral black holes held in unstable equilibrium by a system of struts and strings @xcite .
the metric is constructed using the newtonian potential of a number of collinear rods of differing lengths and the same mass per unit length .
the thermodynamics of such configurations has been studied recently in @xcite .
the weyl construction was used by myers to produce an infinite periodic array of neutral black holes , in which case the strings and struts disappear @xcite .
it has also been long known that the c - metric for an accelerating black hole has an interesting weyl form @xcite .
it is constructed from the newtonian potential of a semi - infinite - line - mass ( silm ) and a second , finite , rod .
the silm is the position of the acceleration horizon and the finite rod is the position of the black hole horizon .
there is necessarily a cosmic string or strut in the spacetime which provides the force accelerating the black hole .
it is now easy to guess the form of the metric which describes a number , @xmath3 , of collinear black holes all accelerating in the same direction .
it is constructed from the newtonian potential of a silm and @xmath3 finite non - overlapping collinear rods of arbitrary lengths .
we examine the @xmath4 case in detail .
we show that the interpretation as two accelerating black holes is valid . in the `` newtonian limit '' where the accelerations are small compared to the inverse masses of the black holes ,
newton s law holds for each black hole with the acceleration produced by a combination of the forces due to any deficit angle cosmic strings or struts present and the gravitational attraction of the other black hole .
we show that the parameters may be chosen so that the outer black hole is pulled along by a cosmic string and the inner black hole follows behind accelerated purely by its gravitational attraction to the outer one .
a similar solution can be found for the @xmath5 case , in which the only deficit angle along the axis is a cosmic string attached to the outermost black hole and the other two are accelerated by gravitational forces alone .
we conjecture that similar solutions exist for any number of black holes .
any static axisymmetric vacuum solution of the einstein equations may be written in the following form [ weylone.eq ] ds^2 = ^2 ( dr^2 + dz^2 ) + r^2 ^-2 d^2 - ^2 dt^2 where @xmath6,@xmath7 are functions of @xmath8,@xmath9 . by using the substitutions @xmath10
the above metric is transformed into the weyl canonical form @xmath11 @xmath12 must satisfy laplace s equation @xmath13 and @xmath14 can be found by using @xmath15 { \nonumber}\\ { d\nu\over dz } & = & 2r { d\lambda\over dr}{d\lambda\over dz}\ .\end{aligned}\ ] ] in @xcite the case of a collinear array of arbitrarily spaced black holes of arbitrary masses was studied . in that case
the newtonian potential used for @xmath12 is the sum of the potentials of a number of finite nonoverlapping rods placed along the @xmath9-axis and the function @xmath14 is found by quadrature .
explicitly , let there be @xmath3 rods each centred at @xmath16 , @xmath17 , of length @xmath18 and with mass / unit length @xmath19 .
the distances to the two ends of rod @xmath20 are given by @xmath21 then = _ i = 1^n _ i(r , z ) ; = _ i , j = 1^n _ ij ( r , z ) , where _ i & = & 12 ( _ i - z_i_i -z_i ) + _ ij&= & 14(e(i,j)e(i , j)e(i , j)e(i,j ) ) where , for instance @xmath22 as stated in the introduction , the c - metric has the weyl form built from the potential of a semi - infinite line mass and a finite rod along the @xmath9-axis . consider the silm to stretch from a point @xmath23 to @xmath24 and the finite rod to be between @xmath25 and @xmath26 so that @xmath27 .
they both have mass / unit length @xmath28 as before .
we are clearly free to choose the origin of @xmath9 coordinates to be wherever we like and so we fix it by requiring that the following three equations have a solution for @xmath29 and @xmath30 [ aandm.eq ] 2 a c_i^3 - c_i^2 + m^2 = 0 , i = 0,1,2 .
in particular this means that @xmath31 .
then the following metric is the weyl metric constructed from the corresponding newtonian potential [ weylc.eq ] ds^2 = & - & a^-1 x_0 x_1 ^ -1 x_2 dt^2 + a^-1 r^2 [ x_0 x_1 ^ -1 x_2]^-1 d^2 + & + & m^24 a^3(c_0-c_1)^-2(c_1-c_2)^-2 y_01y_21r_0 r_1 r_2 y_02 ( dz^2 + dr^2 ) where _ i & = & z - c_i + r_i & = & ( r^2 + _ i^2)^12 + x_i & = & r_i - _ i + y_ij & = & r_ir_j + _ i_j + r^2 . the overall scale and normalisations of the @xmath32 and @xmath33 coordinates are chosen with hindsight so that the following coordinate transformation gives the c - metric in its most familiar form .
consider the coordinates @xmath34 related to @xmath35 by @xcite z & = & a^-1 ( x - y)^-2 ( 1 - maxy(x+y ) - xy ) + r & = & a^-1 ( x - y)^-2 ( -g(y))^12 ( g(x))^12 where the function @xmath36 is the cubic g ( ) & = & 1 - ^2 - 2ma ^3 + & = & - 2 m a ( -_2)(-_3)(-_4 ) . the roots of @xmath36 , @xmath37 are so labelled for historical reasons and there is the following relationship between them and the @xmath38 c_0= -m_2 , c_1 = -m _ 3 , c_2 = -m _ 4 . in terms of @xmath34 coordinates the metric ( [ weylc.eq ] ) becomes the familiar [ c.eq ] ds^2 = 1a^2(x - y)^2 .
in the maximally extended form of the ordinary c - metric above there are actually two accelerating black holes , one covered by the coordinates shown above and the other on the opposite side of the acceleration horizon , accelerating in the opposite direction and out of causal contact with the first .
it is possible however to construct metrics in which there are two ( or more ) black holes on the same side of a common acceleration horizon , accelerating in the same direction .
( these then each have their `` mirror black hole '' on the other side of the acceleration horizon but we will be ignoring these in what follows . ) these are formed by taking the weyl form with a potential constructed from a silm and two ( or more ) finite rods along the z - axis .
consider the silm and first rod as in the previous section and add a second rod between @xmath39 and @xmath40 where @xmath41 as illustrated in figure [ fig1.fig ] .
the corresponding weyl metric is [ weyl2.eq ] ds^2 = & -&^-1 x_0 x_1 ^ -1 x_2 x_3 ^ -1 x_4 dt^2 + ^-1 r^2 [ x_0 x_1 ^ -1 x_2 x_3 ^ -1 x_4]^-1 d^2 + & + & ^24 y_01y_21y_03y_34y_14y_232 r_0 r_1 r_2 r_3 r_4 y_02 y_04 y_31 y_42 ( dz^2 + dr^2 ) where @xmath6 is a constant with units @xmath42 and @xmath43 is a dimensionless constant which we discuss further below .
recall that we chose the origin of @xmath9 so that the equations ( [ aandm.eq ] ) have a solution for @xmath29 and @xmath30 .
now let @xmath44 be such that [ aandmhat.eq ] 2(c_i -)^3 - ( c_i -)^2 + ^2 = 0 , i = 0,3,4 have solutions for @xmath45 and @xmath46 .
we want this metric to become the c - metric in three different limits ( i ) c_1 & & c_2 + ( ii ) c_3 & & c_4 + ( iii ) c_2 & & c_3 in limit ( ii ) we want to obtain the c - metric exactly as it appears in ( [ weylc.eq ] ) . in limit ( @xmath20 ) ( ( @xmath47 ) ) the c - metric
would be expressed in terms of parameters @xmath48 ( @xmath49 ) instead of @xmath50 . in order for the limits to work we require @xmath6 and @xmath43 to behave in the following ways ( i ) ( , ) & & ( , ( c_0-c_3)^-1(c_3-c_4)^-1 ) as c_1 c_2 + ( ii ) ( , ) & & ( a , ma(c_0-c_1)^-1(c_1-c_2)^-1 ) as c_3 c_4 + ( iii ) ( , ) & & ( |a , |m|a(c_0-c_1)^-1(c_1-c_4)^-1 ) as c_2 c_3 where @xmath51 and @xmath52 depend on @xmath53 via 2|a(c_i - |k)^3 - ( c_i - |k)^2 + |m^2 = 0 , i = 0,1,4 for some @xmath54
. these conditions are not enough to fix the dependence of @xmath6 and @xmath43 on @xmath55 but there are solutions .
let black hole 1 ( bh1 ) coordinates @xmath34 be defined as before z & = & a^-1 ( x - y)^-2 ( 1 - maxy(x+y ) - xy ) + r & = & a^-1 ( x - y)^-2 ( -g(y))^12 ( g(x))^12 . let black hole 2 ( bh2 ) coordinates , @xmath56 be given by z - & = & ^-1 ( -)^-2 ( 1 - ( + ) - ) + r & = & ^-1 ( -)^-2 ( -())^12 ( ( ) ) ^12 . where the function @xmath57 is the cubic ( ) & = & 1 - ^2 - 2 ^ 3 + & = & - 2 ( -_2)(-_3)(-_4 ) and c_0 -= -_2 , c_3 -= -_3 , c_4 -= -_4
. then we can transform the metric ( [ weyl2.eq ] ) into bh1 and bh2 coordinates and it becomes , respectively [ bh1.eq ] ds^2 = & & 1a^2 ( x - y)^2and [ bh2.eq ] ds^2 = & & 1 ^ 2 ( -)^2 .
let us examine the solution close up to bh1 in bh1 coordinates .
we consider the limit @xmath58 and @xmath59 . in that limit
, metric ( [ bh1.eq ] ) tends to [ nearbh1.eq ] a^-2(x-_2)^2 where f(x ) & = & a^-1(x-_2)^-2(1 - max_2(x+_2 ) - _ 2 x ) - c_3 a^-1(x-_2)^-2(1 - max_2(x+_2 ) - _ 2 x ) -c_4 + p & = & a ^-1 + q & = & ^2a^3m^2(c_1-c_4)^2(c_0-c_1)^2 ( c_1-c_2)^2(c_0-c_3)^2 ( c_1-c_3)^-2(c_0-c_4)^-2 . from this
we can not read off the surface gravity of the black hole as we can in the asymptotically flat case because the timelike coordinate @xmath32 is not minkowski time at spatial infinity but , rather , a rindler - like time . to see this we consider the limit of ( [ weyl2.eq ] ) as @xmath60 .
d^2 + d^2 + d^2 + ^2 d^2 where ^2&= & ^2 ( - z ) + ^2&= & ^2 r^2 ( - z)^-1 + & = & 1 t + & = & 1 and @xmath61 .
this tells us that the period , @xmath62 , of the euclidean time coordinate @xmath63 that would be required for the euclideanized solution to have no conical singularity at the acceleration horizon is @xmath64 .
we can , however , read off from ( [ nearbh1.eq ] ) the period , @xmath65 , that the euclidean time coordinate @xmath66 would have to have in order for the euclideanized solution to be regular on the horizon of bh1 .
bh1 = 2 we can also examine the smoothness conditions along the axis in either direction from bh1 .
let the period of @xmath33 necessary for smoothness on the axis pointing towards the acceleration horizon , _
i.e. _ @xmath67 and @xmath68 , be @xmath69 and the period of @xmath33 necessary for smoothness on the axis pointing towards the outer black hole bh2 , _ i.e. _ @xmath67 , @xmath70 be @xmath71 .
when @xmath8 is small , the limit @xmath72 corresponds to @xmath73 and examining ( [ nearbh1.eq ] ) in that limit we can read off _
1r = 2 on the other side , @xmath74 when @xmath75 and _ 1l
= 2 we can repeat this all for bh2 , looking at ( [ bh2.eq ] ) near the horizon of bh2 . in this limit
the metric becomes [ nearbh2.eq ] ^-2(-_2)^2 where ( ) & = & ^-1(-_2)^-2(1 - _ 2(+_2 ) - _ 2 ) - c_2 ^-1(-_2)^-2(1 - _ 2(+_2 ) - _ 2 ) -c_1 + & = & ^-1 + & = & ^2 ^ 3 ^ 2(c_1-c_4)^2(c_0-c_3)^2 ( c_3-c_4)^2(c_2-c_4)^-2 the period of @xmath76 required for regularity on the bh2 horizon is _
bh2 = 2 the period of @xmath33 required for smoothness on the axis pointing towards bh1 and the acceleration horizon is _
2r = 2 notice that this is equal to @xmath71 as it should be because these refer to the same part of the axis , that between the black holes .
the period of @xmath33 required for smoothness on the outer axis is _
2l = 2 .
it is possible to choose the parameters @xmath38 , @xmath77 so that @xmath78 .
indeed this condition implies ( c_0-c_4)(c_1-c_2 ) = ( c_0-c_2)(c_3-c_4 ) which can be rewritten as ( c_1-c_2)= ( c_3-c_4 ) and finally as ( c_3 - c_4 ) = ( c_1-c_2)(c_0-c_3 ) ( c_0-c_1)^-1 .
any choice of @xmath79 and @xmath80 can be made and @xmath81 is then fixed .
but it is not possible to have either @xmath82 or @xmath83 , which is not a surprise given the same is true for the c - metric .
thus , on euclideanizing the solution we may choose the period of imaginary time to make both black hole horizons regular but if we do so there will be a conical singularity at the acceleration horizon .
we check that in the weak field limit , _
i.e. _ @xmath84 and @xmath85 and when the distance between the black holes is large compared to their masses , the acceleration of bh2 is given by newton s laws .
we may realise this limit by taking @xmath86 and keeping @xmath87 and @xmath88 fixed . in this limit
we have the following c_0-c_1 & & 12a - m + o(ma ) + c_1-c_2 & & 2 m + o((ma)^2 ) + c_0-c_3 & & 12 - + o ( ) + c_3-c_4 & & 2 + o(()^2 ) .
the net force outwards ( to the left ) on bh2 due to the combination of possible strings and struts is given by t_2 = 14expanding this out and setting @xmath89 we find [ newtontwo.eq ] t_2 = + mr^2 + correction the first terms of the correction are cubic in @xmath90 and @xmath91 and quadratic in @xmath92 and @xmath93 . in order for the above to be a consistent approximation
we require that @xmath94 not be fixed but tends to infinity in such a way that @xmath90 and @xmath91 are larger than @xmath92 and @xmath93 in the limit .
we might choose , for example , @xmath95 , @xmath87 and @xmath88 fixed and @xmath96 . similarly the net force to the left ( towards bh2 ) on bh1 due to the strings and/or struts is t_1 = 14 and in the same limit as above this gives [ newtonone.eq ] t_1 = m a - mr^2 + correction then ( [ newtontwo.eq ] ) and ( [ newtonone.eq ] ) show that the acceleration of each black hole is due to the net tension on it due to the strings and/or struts together with the gravitational attraction of the other black hole .
we can ask , is it possible to choose the period of @xmath33 so that bh1 is free of strings or struts so that its acceleration is due entirely to its gravitational attraction towards bh2 ?
the condition for this is @xmath97 .
this reduces to the condition [ nostrings.eq ] ( c_0-c_1)(c_3-c_4 ) = ( c_1-c_4)(c_2-c_3 ) .
this can be satisfied by choosing @xmath98 , @xmath99 and @xmath100 arbitrarily .
@xmath101 is independent and can be chosen to satisfy ( [ nostrings.eq ] ) .
note that in the weak field limit as defined above this condition is approximately @xmath102 so indeed the acceleration of the black hole is due to its gravitational attraction towards the outer one .
it is clear how to generalize the above results to the case of three or more black holes . in the case of three ,
we add a third finite rod on the @xmath9-axis between @xmath103 and @xmath104 to the left of the existing two .
the metric will be constructed from ( [ weyl2.eq ] ) by multiplying @xmath105 by @xmath106 and @xmath107 by the inverse of this .
the rest of the metric is multiplied by a factor . @xmath43 and @xmath6 will now be functions of the parameters @xmath108 such that the metric reduces to the known c - metric in the appropriate limits . adding more black holes proceeds similarly .
we can again calculate the various periods required for regularity .
we can also ask whether the @xmath38 may be chosen so that there is only one cosmic string pulling the outermost black hole with the others being accelerated purely because of the gravitational attraction of the others . this is a more difficult question to answer in the general case . in the case of three black holes
there are two conditions , & c_14 c_23 c_16 c_25 c_02 = c_01 c_13 c_24 c_15 c_26 + & c_16 c_25 c_36 c_45 c_02 c_04 = c_01 c_03 c_15 c_26 c_35 c_46 where @xmath109 . these do have solutions with @xmath110 for @xmath111 , for example : c_6 = 0,c_5= a , c_4=a+1 , c_3=a+2 , c_2=a+3 , c_1=a+4 , c_0 = a + 4 + f where @xmath112 , and @xmath113 is the positive root of @xmath114 .
we conjecture that such solutions do exist for any number of black holes .
then one could conceive of an accelerating version of the myers black hole array in which there is an infinite collinear array of accelerating black holes with a common acceleration horizon and no cosmic strings or struts in sight .
we also conjecture that there are solutions with any number of accelerating black holes for which it would be possible to choose the period of imaginary time to be such that the euclidean solution is regular on all the black hole horizons at once .
these multi accelerating black hole solutions are classically unstable , assuming that the deficit angles of the nodal singularities are fixed as they would be if the solutions are approximations to spacetimes with field theory cosmic strings of fixed tension
. a little displacement of one of the black holes will disturb the balance of forces required to maintain the solution and presumably cause some black holes to collide and/or some to be left behind .
this instability prevents us from asking whether the euclideanized solutions can be interpreted as instantons for multi - pair production of black holes or the production of some pairs in the background of other accelerating pairs .
this is even before we confront the problem that the uncharged solutions do not give rise to regular instantons even if we allow cosmic strings in space because of the impossibility of matching the acceleration and black hole horizon temperatures .
( nevertheless the euclidean action may straightforwardly be calculated as in @xcite since the singularities are conical and integrable .
using the work in @xcite , calculating the action is just a matter or calculating the difference in the acceleration horizon areas between the instanton and background and the areas of the horizons of the black holes produced . )
these problems might be resolved were we able to give charge to the black holes .
the obvious method to try is that of the charging transformation given in @xcite .
however this only works when the time coordinate is asymptotically minkowki time ( and in any case gives all the black holes the same charge to mass ratio and so might not be able to produce a regular instanton ) . here
, our asymptotic time coordinate is rindler time and the charging transformation gives a new metric of unknown physical interpretation which has zero net charge at infinity .
finding charged versions of these solutions requires more ingenuity .
we would like to thank jerome gauntlett , david kastor , rob myers and harvey reall for useful discussions and gary gibbons for telling us about the weyl form of the c - metric .
fd was supported in part by an epsrc advanced fellowship .
m. s. costa and m. j. perry , `` interacting black holes , '' _ nucl .
* b591 * ( 2000 ) 469487 , http://www.arxiv.org/abs/hep-th/0008106[hep-th/0008106 ] .
r. c. myers , `` higher dimensional black holes in compactified space- times , '' _ phys .
* d35 * ( 1987 ) 455 .
b. godfrey , `` horizons in weyl metrics exhibiting extra symmetries , '' _ gen .
grav . _ * 3 * ( 1972 ) 316 .
d. m. eardley , g. t. horowitz , d. a. kastor , and j. traschen , `` breaking cosmic strings without monopoles , '' _ phys .
lett . _ * 75 * ( 1995 ) 33903393 , http://www.arxiv.org/abs/gr-qc/9506041 [ gr - qc/9506041 ] . | we show how the weyl formalism allows metrics to be written down which correspond to arbitrary numbers of collinear accelerating neutral black holes in 3 + 1 dimensions .
the black holes have arbitrary masses and different accelerations and share a common acceleration horizon . in the general case ,
the black holes are joined by cosmic strings or struts that provide the necessary forces that , together with the inter black hole gravitational attractions , produce the acceleration . in the cases of two and three black holes , the parameters
may be chosen so that the outermost black hole is pulled along by a cosmic string and the inner black holes follow behind accelerated purely by gravitational forces .
we conjecture that similar solutions exist for any number of black holes .
@[email protected] , @[email protected] .
qmul - ph-01 - 05 , hutp-00/a039 |
green s functions @xcite are encountered as response functions , time - ordered expectation values , certain solutions of boundary - value problems or resolvent kernels .
this introduction to green s functions is based on their role as kernels of differential equations . the procedures to construct solutions to a differential equation with an external source or with an inhomogeneity term
are put together to _ derive _ the dyson equation for the green s function of the inhomogeneous system .
very different areas of physics such as , for example , electrodynamics ( see section ii and @xcite ) and quantum transport ( see section iii and @xcite ) , can profit from such green s function formalisms .
green s function formalisms do not present a cure - all for solving differential equations , because essentially the problem of finding the solution of the differential equation is shifted to that of finding the corresponding green s function .
this can , however , be a simplification and even give access to the solution of a more general class of problems .
starting from ordinary differential equations , this short review is meant to introduce how the corresponding green s functions are defined and how they are involved in constructing the solutions for different types of differential equations .
although bearing similarities , the green s function formalism can go beyond perturbation theory .
special emphasize is laid here upon the parallels between homogeneous and inhomogeneous ordinary differential and green s functions equations .
the starting point is a homogeneous differential equation @xmath0 which we suppose is exactly solvable , although @xmath1 will not explicitly be needed .
@xmath2 is some differential operator which may include multiplication with a constant or even another function of @xmath3 .
@xmath3 is either a space or time variable .
what we need to know is the solution @xmath4 of the corresponding green s functions equation @xmath5 @xmath4 is a tensor - like function of two arguments , @xmath2 only acting on the first of them .
like @xmath4 replaces @xmath6 on the left , the zero on the right side of ( 1 ) is replaced by a @xmath7-distribution in ( 2 ) .
there is no general recipe , but knowing @xmath1 can help to get @xmath8 .
having @xmath4 , the construction of a solution @xmath9 of the differential equation with a source term @xmath10 on the right side @xmath11 is straight forward . of course , a solution @xmath12 of ( 1 )
can be added independently of @xmath10 , so we only need to proove that the integral term from ( 3 ) satisfies ( 4 ) : @xmath13 \ ; q(\xi ' ) \nonumber \\ & = & \int d\xi ' \ ; \delta(\xi-\xi ' ) \ ; q(\xi ' ) = q(\xi ) \nonumber \end{aligned}\ ] ] ( all integrals are understood to range over the entire @xmath3-space . ) instead of a source term , the differential equation can contain a potential term which we shall call an inhomogeneity . @xmath14
that is a @xmath3-dependent function @xmath15 multiplied with @xmath6 .
@xmath16 is not included in @xmath2 , because we assume that it so much complicates the equation that a standard solution is no longer known .
the minus sign is a useful convention .
we put the inhomogeneity term on the right @xmath17 to make ( 5 ) formally look like ( 4 ) . with @xmath18 playing the role of @xmath10 a formal solution is constructed analogously to ( 3 ) : @xmath19 this presents an implicit equation for @xmath20 , the so - called _ lippmann - schwinger _ equation .
we have replaced the differential equation by an integral equation . inserting the solution of the homogeneous equation @xmath12 also for @xmath20 in the integral on the right side of ( 6 ) would give the _ born _ approximation @xmath21 , which is appropriate if @xmath18 is a small perturbation compared to @xmath22 .
however , it is the virtue of the green s functions method that in contrast to perturbation theory the inhomogeneity need _ not _ be a small deviation . although not yet providing an explicit solution for @xmath20 , one can make use of ( 6 ) in numerical calculations ( see section ii and @xcite ) . as ( 2 )
is the corresponding green s function equation with a point source to the homogeneous equation ( 1 ) , the equation defining the green s function @xmath23 for the inhomogeneous case derived from ( 5 ) reads : @xmath24 \ ; g(\xi,\xi ' ) = \delta(\xi-\xi')\ ] ] however , while supposing that we know or can easily guess @xmath8 , there is no hint yet how to calculate @xmath25 . finally we have to treat the most general case with source and inhomogeneity term : @xmath26 in analogy to ( 3 ) and ( 4 ) a yet formal solution can be written down using ( 7 ) : @xmath27 here we have the freedom to add any solution of ( 6 ) to the integral . but let us start again solving the problem directly from ( 8) which we rewrite as @xmath28 we compare this with ( 5a ) and construct a solution of the same form as ( 6 ) .
@xmath29 takes the role of @xmath12 , because it would satisfy ( 8a ) if the right side were zero .
@xmath20 has to be replaced by @xmath30 . using ( 3 ) for @xmath29
we obtain @xmath31 ( 10 ) could also have been established by a different line of thought . in ( 8) putting everything that differs from the homogeneous equation ( 1 ) on the right side gives @xmath32
if it were not for the inhomegeneity term @xmath18 , the solution would be ( 3 ) . and
if it were not for the source term @xmath10 , we could use ( 6 ) .
although this is , of course , not a correct way to solve non - homogeneous differential equations , we can understand ( 10 ) as an ansatz adding these two contributions .
the @xmath12-part does not have to be written twice .
and @xmath30 has to appear instead of @xmath20 in the integral taken from ( 6 ) .
this still unknown @xmath30 leaves the neccessary freedom to somehow counterbalance the @xmath10-term not present in ( 5 ) and ( 6 ) , which justifies ( 10 ) as an ansatz .
whichever way it was obtained , ( 10 ) is an implicit integral equation for @xmath30 as ( 6 ) is for @xmath20 .
now we shall profit from the fact that with ( 9 ) we have a second representation of @xmath30 .
insert the expression from ( 9 ) for @xmath30 both on the left and on the right side in ( 10 ) : @xmath33 \nonumber \end{aligned}\ ] ] @xmath20 on the left cancels with @xmath12 and one integral term on the right according to ( 6 ) and we are left with @xmath34 this is an implicit equation for the unknown green s function @xmath23 . but
as @xmath23 by definition ( 7 ) does not depend on any source term , ( 11 ) must be valid for arbitrary @xmath10 .
the special choice @xmath35 gives @xmath36 which after carrying out the @xmath37-integrals and then renaming @xmath38 to @xmath37 becomes @xmath39 ( 11 ) is the general implicit integral equation for the inhomogeneous system s green s function @xmath23 and called _ dyson _ equation .
we could have derived ( 11 ) by doing all calculations from ( 8) on with only a point source @xmath40 instead of @xmath41 .
putting in the solutions ( 3 ) and ( 9 ) for this case would just have looked a bit awkward . ( 11 ) is not an explicit solution for @xmath23 , but still an implicit equation , and even though it is often written as @xmath42 one must not forget that there is a convolution - like integration over the inner @xmath3-index behind the sequence of factors @xmath43 .
the integration can either be transformed into a discrete finite sum , and thus ( 11 ) into a linear system of equations solvable by a matrix inversion ( see section ii and @xcite ) , or the convolution can be replaced by a multiplication going to frequency space by a fourier transformation ( see section iii and @xcite ) .
\(10 ) and the preparation for ( 11 ) would have looked much more elegant leaving out @xmath12 and the @xmath20-terms right from the start .
indeed , one could argue that in ( 3 ) one is only interested in the part different from the trivial homogenous solution , if no further boundary conditions have to be accounted for . with a similar argument ,
the @xmath20-part could have been dropped in ( 9 ) .
the identities @xmath44 @xmath45 @xmath46 @xmath47 would obviously also have given ( 11 ) .
leaving out @xmath12 in ( 6 ) would be precarious ( see below ) .
however , dropping it when constructing ( 11 ) along our second line of thought causes no problem . the term with @xmath10 will ensure that @xmath48 can not simply be zero .
there is some subtlety about the @xmath12 contribution in ( 6 ) .
( 6 * ) has the trivial solution @xmath49 and not necessarily another one .
surely , @xmath50 satisfies ( 5 ) , but it is not what we want .
@xmath12 is the background excitation replacing boundary conditions in this kind of problem , which shall become clear when discussing electrodynamics ( see following article ) .
@xmath12 acts as a source term which - instead of being put in @xmath10 - can be more appropriately and conveniently set as a fixed part of the @xmath6 we are looking for . inserting ( 6 ) into itself ever and ever again , @xmath20 is developed into a series in powers of @xmath51 : @xmath52 just for shorthand notation we dropped the arguments and integration variables and in the third line also the integral signs .
the resulting series is recognized from perturbation theory summing interactions to zeroth , first , second , etc . order .
if there were no contribution @xmath12 in @xmath20 , only the contribution of @xmath51 to power infinity would exist with nothing to multiply to at the end .
in other words we would have no basis on which to develop @xmath20 .
the green s functions formalism as we use it aims at solving ( 6 ) , ( 10 ) or ( 11 ) in a closed form , _ not _ cutting the series in ( 12 ) at some finite order .
the development ( 12 ) is only shown here precisely to demonstrate that interactions are included to all orders as well as to explain the importance of the homogeneous background contribution .
the dyson equation ( 11 ) can be expanded in an analogous manner to the lippmann - schwinger equation , by the way prooving the equivalence to its complementary form @xmath53 : @xmath54 ( integration over inner arguments is understood in all contributions to the sums . ) either formally or as a matrix calculation in finite discrete @xmath3-space , ( 11 ) is often solved as @xmath55 developing the @xmath56-factor in ( 14 ) into a geometric series just results in the infinite sums written in ( 13 ) .
putting together ( 12 ) and ( 13 ) we can get an alternative representation to ( 6 ) for the solution of the inhomogeneous differential equation : @xmath57 therefore in the case that should it be easier to get the green s function @xmath23 than to solve the implicit equation ( 6 ) for @xmath20 , we see that @xmath23 can be useful also for treating the equation without an external source @xmath10 .
nevertheless , ( 15 ) again illustrates that @xmath58 plays the role of the source . using green s functions
it has been shown how differential equations can be treated that differ from easily solvable ones by an additional potential term or an arbitrary source term .
the formalism as presented here is for open - boundary in contrast to boundary - value problems @xcite . from constructing the solutions of the differential equations
, we also obtained the constituting relation for the inhomogeneous system s green s function , which characterizes the response to a point source and includes all - order interactions .
examples of applications will be given in the two following sections .
the fundamental homogeneous - medium green s tensor of electrodynamics is deduced from the field of a dipole . based upon that a numerical procedure is presented to solve the wave - equation for the near - field in a scattering setup for arbitrary material distributions .
the full inhomogeneous system s green s function is not explicitly needed to get the fields , although it can be obtained by a very similar calculation and in optics can be interpreted as a density of states . the typical problem in nano - optics
@xcite is the situation that some tiny structures are illuminated by an extended source , a plane wave for example , and then one is interested in the field distribution that arises from multiple scattering @xcite , especially to identify places where the field intensity gets considerably enhanced @xcite .
a theory can be based on green s functions , however , their implication differs slightly from the standardly taught cases of fixed boundary field values @xcite or located sources .
modern optical scanning microscopes make it possible to probe and map directly even different quantities of the near - field @xcite , such as the electric and magnetic field intensities @xcite or the density of states @xcite .
applications of tayloring nano - structures with respect to optical properties include resonant particles @xcite and cavities @xcite , squeezed fields @xcite , wave guides and their adressing @xcite as well as transmission apertures @xcite and lithography masks @xcite .
we here present the green s functions formalism that forms the bases of a finite - element quite effective numerical algorithm used in current research @xcite .
this treatise is also given as an application example of the general methods to solve differential equations with certain perturbations presented in the preceeding paper .
the discussion can be reduced to monochromatic light , that is a single frequency @xmath59 and thus time dependence @xmath60 for the fields . with non - magnetic materials
the problem is to find the solution @xmath61 of the wave equation @xmath62 ( 1 ) is a differential equation of the type ( 5 ) from section i. there is no source term on the right hand side of ( 1 ) .
if the source were , for example , a dipole located at some point , a source term with its oscillation strength and direction would have to be put on the right as @xmath63 .
however , we shall see that a plane - wave source can be and is better included in ( 1 ) as it is . for simplicity
we shall assume that the background medium , in which objects with different permittivities @xmath64 are located ( fig.1 ) , is vacuum with permittivity @xmath65 .
for another embedding medium , its dielectric constant @xmath66 would take the role of @xmath65 . in ( 1 ) @xmath67 of the material distribution designates the dimensionless relative permittivity with respect to vacuum or the background medium .
to separate ( 1 ) into a part representing a homogeneous differential equation with known solution and an inhomogeneity write its as @xmath68 the correspondances to the quantities of the general formalism given in section i are @xmath69 and @xmath12 will become @xmath70 .
we can write down the solution following section i after having prepared the background green s function in the next section .
in any case we need the green s function @xmath71 of the homogeneous problem satisfying @xmath72 with @xmath73 .
@xmath4 will be a @xmath74 tensor or matrix here .
more commonly @xcite the small letter @xmath4 is used for the scalar function @xmath75 with @xmath73 .
the green s function @xmath4 from ( 2 ) is then named @xmath76 with index @xmath77 for _ homogeneous_. there are several ways to obtain @xmath4 .
one is based on the knowledge that if we have a scalar function @xmath78 solving @xmath79 , then @xmath80 and @xmath81 with a constant but arbitrary pivot vector @xmath82 will both solve the vectorial equation @xmath83 .
the tensor @xmath4 looked for in ( 2 ) can be constructed out of @xmath84 , @xmath85 , @xmath86 , @xmath87 together with @xmath88 and @xmath89 .
we shall not enter into the details of this mathematically slightly precarious approach @xcite .
a second recipe just mentioned here for completeness is given by the following statement @xcite : if @xmath90 satisfies ( 3 ) , then @xmath91 is the tensor defined by ( 2 ) . of course , because of being in homogeneous space @xmath92 and @xmath4 effectively are functions of @xmath93 alone .
@xmath94 is the unit matrix in 3 by 3 cartesian coordinate space and @xmath95 means building a matrix out of derivatives @xmath96 .
we shall deduce @xmath4 from a physical reasoning . from standard electrodynamics
@xcite one has the electric field of an oscillating dipole @xmath97 it is important to take the exact formula here including retardation in contrast to common near- or far - field approximations .
( 5 ) gives the space part , the time dependence is just @xmath60 everywhere . to get the green s tensor
, we have to evaluate from ( 5 ) what field components in x- , y- and z - direction a dipole @xmath98 at @xmath99 oriented along x would produce at @xmath100 , what components a dipole oriented along y would produce and what components a dipole along z would produce and assemble all these in a matrix .
a point dipole is the elementary excitation corresponding to the @xmath7 on the right side of ( 2 ) .
the physical meaning of @xmath4 is to tell us what field any such dipole would have .
that is shown formally in the first matrix in ( 5a ) . decomposing any @xmath98 into its cartesian components , ( 5 )
can be rewritten as @xmath101 @xmath102 @xmath103 from which we easily see that the matrix to be multiplied with @xmath98 to produce @xmath104 is @xmath105 @xmath106 without arrow means the absolut value @xmath107 and @xmath108 , @xmath109 , @xmath110 stand for @xmath111 , @xmath112 and @xmath113 , respectively .
@xmath114 is the matrix written with @xmath108 , @xmath109 and @xmath110 from ( 5a ) .
the terms from ( 5 ) vectorially oriented along @xmath98 cause the diagonal matrix contribution , those stemming from terms with @xmath115 the full matrix in ( 5a ) .
compared to the above expression @xmath116 from ( 4 ) has a minus sign and misses a factor @xmath117 . as will be discussed later ,
the source to put into equation ( 1 ) corresponding to an oscillating dipole is not the dipole moment @xmath98 itself , but @xmath118 times @xmath98 .
and because @xmath119 we have @xmath120 the formula ( 6 ) fails for @xmath121 or @xmath122 .
@xmath116 including the case @xmath122 can be represented using the principal volume method @xcite . in practice , working with finite elements , the value to put for @xmath4 of its two arguments the same place can be derived from the polarization of a dielectric body .
the discussion of @xmath123 is postponed to the next section .
the starting point to find a solution for the electric field with the objects present is eq .
( 6 ) of section i , which rewritten in the variables of our problem here reads @xmath124 where @xmath125 is a solution of @xmath126 or already assumed to be the space part of a linearly polarized plane wave @xmath127 with a fixed amplitude vector @xmath128 .
the time dependence @xmath60 can be omitted in @xmath125 as well as in @xmath61 . from ( 7 )
a numerical procedure can be deduced if the objects with @xmath129 only occupy fractions of space rather small on the scale of the wavelength , not at all principally necessarily much smaller than @xmath130 , though .
then we divide them up into finite elements ( fig.1 ) of volume @xmath131 , which we enumerate @xmath132 and associate permittivities @xmath133 or perturbances @xmath134 and local fields @xmath135 uniform over @xmath131 .
the linear sizes of the object cells should not exceed about @xmath136 .
they need not at all be placed on a regular grid . and
the only reason for demanding small enough objects is not to get too many elements @xmath137 .
the integrand in ( 7 ) only exists at places where @xmath15 does not vanish , and to evaluate the field @xmath61 also at such places @xmath138 , no values outside the objects appear in the equation . changing to finite elements
we thus get a linear system of equations for the fields in the object cells @xmath139 which can be solved by a matrix inversion .
to evaluate the resulting field at any other place , that is outside the objects , the @xmath135 just have to be inserted into the finite - element version of ( 7 ) : @xmath140 of course , we already needed @xmath141 with @xmath142 to set up the system ( 8) .
let us suppose that there is only a single cell with an @xmath64 differing from the background @xmath65 .
this is placed into a homogeneous field @xmath70 .
if the cell has the shape of a sphere , the local field throughout its inside is aligned in the direction of @xmath70 and its value is @xmath143 @xcite .
no retardation effects have to be considered here as the size of the cell can in principle be made arbitrarily small . in (
8) only keeping the term of the sum with @xmath144 , setting @xmath145 and @xmath146 gives @xmath147 from which follows that @xmath148 the factor 1/3 is also valid for cubic elementary cells , however , other shapes require different depolarization factors @xcite .
the background and the resulting field at each point are already 3-vectors . nevertheless , in order to solve ( 8) , imagine the @xmath149s for the object cells assembled into long or
double vectors of @xmath137 times 3 components @xmath150 with further the big @xmath151 matrix @xmath152 consisting of 3x3 blocks @xmath153 ( 8) then reads @xmath154 @xmath152 can be inverted using the procedure described in the appendix , however , as @xmath152 is a full matrix and the complete invers @xmath155 is needed , that is of no advantage and a standard inversion algorithm will do as well
. for the problems of a few small scatterers here by introducing finite elements the implicit integral equation ( 7 ) for the electric field has been turned into a linear system of equations that is easily solvable . even shying this effort , the coarsest ,
so - called born approximation consists in replacing @xmath149 in the integral in ( 7 ) or in the sums in ( 8) or ( 9 ) by @xmath70 . keeping the summation over finite elements to estimate the integral
we directly get @xmath156 no distinction between places @xmath100 in or outside the object cells is necessary in ( 11 ) .
the born approximation only takes into account first - order scattering off every object and thus can only be good for weak scatterers with distances between them rather large on the scale of the wavelength . producing a clearly different field pattern from the exact solution including all scattering orders , fig.2 demonstrates that the born approximation is likely to be insufficient to model near - field optics setups . in a background of @xmath157 illuminated by a p - polarized plane wave of unit amplitude with @xmath158 from the front and under an angle of @xmath159 from beneath the plane defined by the cubes
setup ( a ) and electric field intensity @xmath160 calculated via the born approximation ( b ) and the exact green s functions method ( c ) .
besides subtle differences in the near - field pattern remark that in ( b ) the grey - scale is from 0.978 to 1.022 whereas in ( c ) it is from 0.995 to 1.005 .
the intensity map is taken at a height 25 nm above the cubes and given for an area of @xmath161 .
, width=207 ] in a background of @xmath157 illuminated by a p - polarized plane wave of unit amplitude with @xmath158 from the front and under an angle of @xmath159 from beneath the plane defined by the cubes .
setup ( a ) and electric field intensity @xmath160 calculated via the born approximation ( b ) and the exact green s functions method ( c ) .
besides subtle differences in the near - field pattern remark that in ( b ) the grey - scale is from 0.978 to 1.022 whereas in ( c ) it is from 0.995 to 1.005 .
the intensity map is taken at a height 25 nm above the cubes and given for an area of @xmath161 .
, title="fig:",width=234 ] in a background of @xmath157 illuminated by a p - polarized plane wave of unit amplitude with @xmath158 from the front and under an angle of @xmath159 from beneath the plane defined by the cubes .
setup ( a ) and electric field intensity @xmath160 calculated via the born approximation ( b ) and the exact green s functions method ( c ) .
besides subtle differences in the near - field pattern remark that in ( b ) the grey - scale is from 0.978 to 1.022 whereas in ( c ) it is from 0.995 to 1.005 .
the intensity map is taken at a height 25 nm above the cubes and given for an area of @xmath161 . , title="fig:",width=234 ] adding an arbitrary source term to our original wave equation ( 1 ) changes it into @xmath162 if now we know a tensor function @xmath163 satisfying @xmath164 then obviously @xmath165 would give a special solution of ( 12 ) .
any solution of ( 1 ) could be added . as generally deduced in section i the implicit relation to get @xmath23 from
is the dyson equation @xmath166
@xmath163 for both arguments @xmath100 and @xmath99 covering all space is too much information to display at once and usually much more than what one is interested in .
the imaginary part of @xmath167 is proportional to the density of states @xmath168 @xcite .
the deduction of this statement found in quantum mechanics book @xcite , however , rather argues with a system of energy eigenstates and the variation of the green s function as well as the density of states with energy .
no real @xmath100-space is explicitly mentioned .
our interest lies in the spatial dependence of the density of states at fixed light frequency @xmath59 . even if described in terms of fields , concepts like reactance and work known from electrical circuits
may be applied @xcite .
the time average of the work done by the fields is given by @xmath169 with no other imposed fields , charges or currents than an oscillating point dipole , the latter will present the only external current @xmath170 , which will thus be located as @xmath171 . if the dipole moment oscillates as @xmath172 , the corresponding current is @xmath173 .
deducing the wave equation for time harmonic fields ( in vacuum for simplicity here ) from maxwells equations with current term @xmath174 leads to @xmath175 from which we see that the source term for the dipole has to be set as @xmath176 .
the integral ( 16 ) reduces to the value of @xmath177 at @xmath100 .
the electric field we get from ( 14 ) : @xmath178 inserting @xmath179 and @xmath149 into ( 16 ) the time factors cancel as expected for a time average and but for a factor @xmath180 we get @xmath181 choosing unit vectors along the coordinate axis for the probe dipole @xmath182 , ( 20 ) will filter out the trace elements of the matrix @xmath167 .
we associate @xmath183 , @xmath184 , @xmath185 and a total @xmath186 . a motivation for taking the negative imaginary part of @xmath167 as a measure for the presence of modes
can also be obtained by comparison to the energy resonance of a forced oscillator @xcite . for optimal excitation from the energy point of view - in contrast to amplitude resonance - the force has to be @xmath187 ahead of the elongation or in phase with the velocity of the oscillator .
@xmath167 describes the field caused by backaction of the system at the place of the probe dipole moment ( taken as reference phase zero ) , and therefore @xmath188 is the part that can in a resonant manner further enhance the dipole oscillation .
( in reality radiation out of the system will provide strong damping . )
we now intend to evaluate a map of @xmath167 on , for example , a horizontal plane .
the plane may lie above or below object cells or even cut some .
like the objects , the , of course , finite area of interest on the plane is divided into cells .
just depending on the desired resolution of the map the unit cell length of this mesh may well differ from the cell size chosen to discretize the objects ( fig.1 ) .
the list of object cell midpoints @xmath189 from the last section , which shall be called region @xmath190 , is extended by all cell midpoints from the map in the plane , which shall be called region @xmath191 and is now understood to to be included in counting @xmath192 from 1 to a new @xmath137 .
analogously to ( 8) the integral in ( 15 ) is replaced by a sum : @xmath193 it does not matter that the map @xmath191 has a different mesh from @xmath131 as for the objects @xmath190 , as @xmath194 for @xmath195 in @xmath191 , anyway .
( spatial overlap of cells from @xmath190 and @xmath191 and even coincidence of midpoints is no problem ; a place can be counted with @xmath196 in @xmath190 and without in @xmath191 . )
although only @xmath197 with @xmath198 in region @xmath191 is wanted as a final result , ( 21 ) has to be set up as an equation for a matrix of all @xmath23 with each of its arguments any cell in @xmath190 or @xmath191 , schematically sketched as @xmath199 . to solve ( 21 ) for @xmath23
we have to invert the same kind of matrix as in ( 8) , the only difference being that @xmath200 now also runs over the plane cells in addition to the object cells .
@xmath201 @xmath23 and @xmath4 are themselves matrices on contrast to vectors @xmath202 and @xmath203 .
@xmath23 consists of 3x3-blocks @xmath204 , @xmath4 is made of blocks @xmath205 and the 3x3-block at position @xmath206 in @xmath152 given by @xmath207 .
the @xmath4 in ( 21a ) and @xmath152 like the big @xmath23-matrix have the structure @xmath199 .
one could invert @xmath152 as given , for example by the procedure from appendix a. the matrix @xmath208 there is initialized with @xmath51 .
its @xmath209 and @xmath210 quadrants are zero and will stay zero throughout the procedure , @xmath211 .
this is no contradiction , as it is not @xmath152 that is singular .
quadrant @xmath212 will be needed for multiplication with @xmath4 in ( 21a ) .
however there is an even more efficient algorithm to get @xmath23 that already includes the multiplication by @xmath4 .
it directly calculates @xmath213 , which is the compact way to write ( 21a ) as the solution of ( 15 ) , also denoted @xmath42 for short .
the technical details can be found in appendix b. trace components of @xmath167 meaning densities of states for the three polarization directions ( fig.3 ) above an optical coral @xcite in analogy to a quantum coral @xcite have been measured @xcite in a so called forbidden - light near - field optical microscope @xcite .
the sample consists of a stadium arrangement of gold particles on a glass surface .
the forbidden - light setup prevents detecting light emitted from the fiber tip that has not passed through surface modes that make up the density of states for this system .
like for antinodal and nodal points in a resonator , more energy can go into the system when the excitation is placed at a point of high density of states than when coupling is bad where the density is low .
in section 4 we saw that it is convenient to start from a solution for the field in the form ( 7 ) if the excitation comes , for example , from a background field belonging to a plane wave . though the matrix to invert bore a certain similarity to the evaluation of the green s tensor in section 5 , with ( 8) and ( 9 ) we directly calculated the field .
in contrast , more adapted to localized sources , there is ( 14 ) as a solution of ( 12 ) .
if there is no additional background field to cause any excitation , no solution of the equation ( 1 ) with zero right side is to be added as further contribution and ( 14 ) is the field distribution to be observed .
( 14 ) has to be rewritten in terms of finite elements in order to be used in a numerical calculation . in the same way as the objects the source @xmath214 has to be devided into discrete cells or elementary dipoles . to distinguish their @xmath215 locations from those of the objects
we shall enumerate them as @xmath216 , @xmath217 .
the place @xmath100 to evaluate the field @xmath149 may be anywhere outside or inside the objects as well as beside or even at a source location . for the following development the dyson equation for the green s tensor is needed in a discretized form for both its variants @xmath42 and @xmath53 .
@xmath218 having in mind a region where and a resolution with which @xmath61 is to be evaluated like the discretized plane @xmath191 from the last section , it would be possible to supply @xmath23 for all needed combinations of arguments @xmath219 and calculate @xmath61 as the single sum from ( 22a ) . to weave in the influence of the objects
, @xmath23 would have to be set up as a big matrix like in the last section over all combinations of three regions @xmath190 , @xmath220 and @xmath191 here , the objects , the source and the map .
having calculated @xmath23 in the @xmath190-@xmath191-scheme from the last section , one could evaluate @xmath61 with @xmath221 as written in ( 22b ) . however , the most efficient way is given in ( 22c ) .
@xmath222 is merely needed for @xmath198 and @xmath195 from the set of object cells , keeping a matrix to be inverted as small as possible , namely of @xmath223-type .
choosing @xmath198 and @xmath224 in ( 21 ) in the object set @xmath190 , instead of in @xmath190 or @xmath191 as the equation was originally set up for , we see that ( 21 ) presents a closed system of equations for all such @xmath225 .
then for ( 22c ) more summations over products with @xmath4-functions , which are analytically known for any pair of arguments , can be considered less demanding in computing time than the inversion of large matrices . in the transformation from ( 22c ) to ( 22d ) after swapping index names @xmath192 and @xmath200 in the last sum , @xmath53 and ( 14 ) have been exploited . using ( 22d ) for @xmath61
renders an implicit equation for the field in the form ( 10 ) or ( 10 * ) from section i. @xmath61 is the equivalent of @xmath30 and as stated earlier , we assume that physically there is no background field @xmath12 that could initiate an additional field distribution @xmath20 .
although not very convenient , a plane wave as exciting field could be understood as stemming from a sufficiently long and dense array of huygens elementary dipole sources @xmath214 reasonably far away from the objects .
the other way round , for a single dipole source or a number of dipole sources distributed in space the field they would produce at any location @xmath100 in homogeneous space is the superposition of their individual fields , and putting @xmath226 the ansatz ( 7 ) can be used also for this case . like the field
@xmath149 anywhere was obtained as a straight - forward summation once having its values at the places of the object cells , finally an alternative way to the procedure from the last section to get the green s tensor @xmath23 shall be given , also requiring only the inversion of a matrix with size the number of object cells .
series expansion is used to rewrite the solution of the dyson equation : @xmath227 g = g+gv(\sum_{n=0}^{\infty } ( gv)^n)g \nonumber \\ & = & g+gv({\bf 1}-gv)^{-1}g \\ & = & g+gvg+gv(\sum_{n=0}^{\infty } ( gv)^n)gvg = g+gvg+gv\ ; ( { \bf 1}-gv)^{-1}g\ ; vg \nonumber \\ & = & g+gvg+gvgvg\end{aligned}\ ] ] designating regions the spatial arguments belong to on ( 23a ) we get for @xmath228 : @xmath229 as @xmath15 does not vanish only in region @xmath190 , the first index of @xmath230 obviously must be @xmath190 .
@xmath230 being @xmath231 , for power zero the second region index automatically is the same as the first and all other powers ending with @xmath15 imply second index @xmath190 .
it is sufficient to set up the matrix @xmath232 as an @xmath223-block and invert that .
should one prefer to evaluate a complete @xmath23 , which differs from the above inverted matrix by a factor @xmath4 , line ( 23b ) like ( 22c ) shows that it is in principle only necessary to get @xmath23 from some self - consistent implicit equation in the object region @xmath190 .
@xmath233 can be constructed applying the procedure described in appendix b to a matrix set up as @xmath223-block only .
the inversion has to be completed in this case , though .
going through the diagonal elements , all lines and columns have to be updated in each step , including the ones above and to the left of as well as the ones the respective diagonal element is in . writing ( 23b ) @xmath234 as summation over discrete elements ready for use in a calculation then reads : @xmath235 of course , summations run over all object cells here .
no numerical advantage can be drawn out of @xmath236 in ( 23b ) . with the same effort of making @xmath233
it can be used to evaluate maps of @xmath167 as well as plots of @xmath163 with @xmath99 fixed or even some function of @xmath100 .
we have presented a method to solve the problem of scattering of electromagnetic waves off an arbitrary distribution of dielectric objects , that is the exact evaluation of the field , especially in the near zone where higher - order multiple reflections can become important .
besides the field distribution we have obtained the green s tensor characterizing the system independently from the form of the excitation .
it represents the response function and also the density of states for supported electric fields . for a methodical introduction
we have restricted our considerations to dielectric materials and the electric field . without magnetic susceptibilities
the magnetic field distribution can be calculated once having the electric field inside the objets by a formula like ( 9 ) with the magnetic background field and replacing @xmath4 by a tensor including the conversion from the electric to the magnetic field by taking the rotation @xcite .
it is further possible to treat non - uniform magnetic permeabilities and even mixed systems with dielectric and magnetic objects @xcite .
the electric green s tensor presented above is then paired by a magnetic counterpart and genuine mixed response functions also exist . whereas the calculation of the field distributions even for mixed systems is quite straight forward , the construction of the green s tensor is more involved .
it lives of the idea of handling one kind of objects first and then considering this setup as the background to include the other kind .
there is no approximation or ranking in importance in this procedure .
the discussion here has only considered finite objects in a homogeneous background as well as cartesian coordinates where vector and tensor components have been written out .
cylindrical and spherical coordinates are also commonly used @xcite and the green s functins formalism has been developped for layered media @xcite . besides wave - guide applications the use for modelling typical near - field optics experiments , where the microstructures to investigate are prepared on a substrate surface , lies in putting the influence of this surface into a background green s tensor @xcite which is then implied the way we used @xmath4 here .
details of applications of the green s functions technique in electrodynamics to more complicated situations as well as beautiful results of corresponding experiments can be found in the given references .
this text focussed on calculation techniques and further intended to give an overview of slightly different formal ways to calculate green s tensors and fields of which either may be optimal for a specific problem .
suppose a complex quadratical matrix to invert is already given in the form @xmath237 or if it is not , we rewrite it like that .
there is no restriction on the values of the numbers @xmath238 . to get the inverted matrix proceed as follows : of matrix @xmath239 one by one take the diagonal elements @xmath240 and to all elements @xmath241 add @xmath242 . after having worked through the matrix for one such @xmath240 , the changed matrix values have to be taken to do so for the next , also already changed , diagonal element .
obviously @xmath137 such steps are required for an @xmath243-matrix .
this will yield @xmath244 , such that in the end 1 has to be added to all diagonal elements in order to obtain @xmath245 .
for clearness we write out the first two transformation steps of the matrix : @xmath246 the inverted matrix @xmath245 can be represented as a geometric series : @xmath247 truncating and using the sum from the right side is only possible if the series converges whereas the closed form on the left is valid in any case . in contrast to the infinite sum on the right side of ( 25 ) , our inversion procedure consists in a finite number of steps of adding contributions to the matrix elements .
nevertheless , ( 25 ) tells us that the invers @xmath245 is the sum of all powers of @xmath239 and thus each element in row @xmath248 and column @xmath249 must be the sum of all possible products @xmath250 with any number @xmath200 of inner indices including none .
( diagonal elements get an extra + 1 . )
there are @xmath137 different indices @xmath251 and they may repeat , of course . considering that @xmath252 can also be written as @xmath253 we see that the first step in ( 24 ) adds to each matrix element @xmath241 the sum of all products @xmath254 . in these at least one pair of indices 1
is squeezed between @xmath248 and @xmath249 as in @xmath255 , the contribution @xmath241 was already there . in the second step
all products @xmath256 with @xmath257 every possible sequence of 1s and 2s will be added .
the products with only indices 1 between @xmath248 and @xmath249 were there before . in the third step
every sequence of indices 1 , 2 and 3 with at least one 3-link is added .
and so on until in the end at each matrix position between outer indices @xmath248 and @xmath249 we have created all possible sequences of an endless game of dominos @xmath238 with numbers @xmath192 and @xmath200 from 1 to @xmath137 .
this argument was to proove that the result of ( 24 ) indeed gives @xmath258 .
we calculate a finite number of @xmath252 or products @xmath242 .
the sequence on the right side of ( 25 ) need not converge and the original entries in @xmath239 need not at all be small compared to 1 in their absolute values .
there is no approximation in the sense of a perturbation theory .
if the matrix @xmath259 is degenerate , the failure of the inversion will be noticed when a value @xmath260 becomes zero at some step . not to confuse notation , remark that in ( 24 ) and in products in the text like @xmath261 letters @xmath262 meant the original matrix entries whereas in expressions @xmath260 , @xmath252 and @xmath242 we referred to the entries at the respective step of the matrix transformation . in the application from the main text
@xmath263 enumerates the object cells . at position @xmath206 in @xmath230
when expanded into a series having every possible sequence @xmath264 shows that the resulting field at any place ( inside or outside the objects ) is the interference of the background field and the fields reradiated by all the object dipoles having undergone every possible scattering path between the objects ( fig.1 ) .
a complication in the electrodynamics application at this stage is the fact that each matrix element @xmath262 actually in itself is a 3x3 matrix indicating the effect of three field components at one place onto three field components at a another place .
@xmath4 on the discrete space of the objects can be written blockwise with a scheme of quadruple indices ( 17.2,10 ) ( 0,6.5)(7.75,4 ) @xmath265 ( 8,6.5)(7.75,4 ) @xmath266 ( 0,1.75)(7.75,4 ) @xmath267 ( 8,1.75)(7.75,4 ) @xmath268 ( 16,0.5)(1,1)@xmath269 ( 0,1.5)(1,0)17.2 ( 0,6.2)(1,0)17.2 ( 7.85,0)(0,1)10 ( 16,0)(0,1)10 ( 1.45,10)(2.5,0.68 ) ( 3.9,8.5)(2.5,0.68 ) ( 6.35,7)(2.5,0.68 ) ( 9.45,5.15)(2.5,0.68 ) ( 11.9,3.65)(2.5,0.68 ) ( 14.35,2.15)(2.5,0.68 ) @xmath270 in the product @xmath271 just multiplies the respective column .
one could use the inversion procedure working off the diagonal elements marked by ovals , requiring @xmath272 steps then .
however , the process is equally applicable to 3x3 blocks as marked by the dashed rectangles , since by its dimension the whole matrix can be divided up into 3x3-blocks .
then @xmath273 means the inversion of a 3x3 matrix and @xmath274 to update the blocks means the product of three 3x3 matrices .
these operations should be programmed as elementary procedures .
the given procedure to invert a matrix can become of advantage if for sparse matrices conventional routines run into numerical difficulties because of many zero values .
besides that , it can be adapted to become quite efficient if only parts of the inverted matrix are needed or for symmetry reasons it is known that blocks or patterns of matrix elements vanish and will stay zero throughout the inversion .
although for the calculation of the green s tensor a slightly modified procedure is applied that directly optimizes the numerical solution of the dyson equation ( see appendix b ) , the matrix inversion was discussed here , because it may be used for more general purposes and in other contexts as well . in the following instructions
are given how to calculate the green s tensor @xmath275 being efficient in the way that finally only @xmath167 of equal arguments on the mesh points of map
@xmath191 need to have the correct values @xcite . for a start
nevertheless consider that @xmath225 with arguments @xmath198 and @xmath224 from the big set of all object and map cell midpoints will have to be the sum of all products @xmath276 @xmath277 , @xmath278 , @xmath279 , @xmath280 can only be object cells and any sequence of them has to be created , including the empty one with no @xmath15 giving the term @xmath205 . initiate a matrix @xmath281 - for simplicity call it @xmath23 from the beginning - with line and column arguments running over all object cells in @xmath190 and all map cells in @xmath191 with @xmath205 in each 3x3-subblock . in the @xmath210-quadrant
only diagonal blocks @xmath282 will be needed , however .
work off the diagonal subblocks through the @xmath283-quadrant . for
the @xmath284th one prepare the inverted 3x3-matrix @xmath285 and to all entries from line @xmath286 and column @xmath286 on add the product given below .
@xmath287 in the following step use the updated entries in the above recipe . in the @xmath284th step
you need not update entries in line @xmath284 or above or in column @xmath284 or to the left of it , because these will not be needed as multiplication factors @xmath288 and @xmath289 for @xmath290 any more .
the parts of quadrants @xmath209 and @xmath212 remaining after these restrictions have to be changed by ( 26 ) for every @xmath284 . merely diagonal 3x3 blocks have to be done in the @xmath210-quadrant .
the first step adds all products consisting of any non - zero number of factors @xmath291 between @xmath4s , the second step all products eventually containing @xmath291 and one up to any number of @xmath292 , and so on .
the procedure is finished after @xmath293 has run down the diagonal of the @xmath283-quadrant .
all sequences of multiple scattering from the objects are then included .
we could have filled the whole @xmath210-quadrant with initial @xmath294 3x3-blocks and updated the entire @xmath210-quadrant in each step . like in the matrix inversion procedure from appendix
a we should further have worked down the complete diagonal and done ( 26 ) for every @xmath293 from region @xmath191 as well .
there will , however , be no additions as @xmath295 for @xmath293 in @xmath191 ( even if a map cell accidently coincides with an object cell ) .
this argument also reveals why non - diagonal elements in @xmath210 do not have to be evaluated .
they can never appear as multiplication factors @xmath296 or @xmath297 .
the @xmath210 diagonal is the resulting @xmath298-map we wanted .
initiating the matrix by @xmath4 and introducing the @xmath15 on treating the respective diagonal element makes the outcome of the procedure directly @xmath299 compared to initiating the matrix with @xmath51 and as an intermediate step obtaining @xmath230 .
it is demonstrated how the transport problem for two open free - electron gas reservoirs with arbitrary coupling can be solved by finding the system s green s function . in this sense
the article is an introduction on green s functions for treating interaction .
a very detailed discussion of the current formula is given on an elementary basis . despite formal resemblances the stationary transport situation , however ,
differs in its nature from introducing coupling between energy levels in a closed system where then the interest lies in modified eigenvalues and eigenstates . by modern lithography techniques so - called point contacts @xcite can be arranged between conductors . these have sufficiently small dimensions such that electronic modes get quantized .
however , coupling across such constrictions need not be so weak as to be described by a small tunnel probability , but can be influenced by coherent interference of multiple reflections .
point contacts can be obtained by indenting stm - tips into some material @xcite , electromigration @xcite or the break - junction technique @xcite . whereas for constrictions imposed by gate electrodes to a two - dimensional electron gas in semiconductors @xcite
one observes quantized conductance values in the sense of fully transmitting or totally switched - off modes , the application in mind behind this work is the type of connection like the single - atom contact , characterized by an ensemble of channels @xcite , which can also have intermediate transmission amplitudes between zero and one @xcite .
even with some fully transmitting modes the contact bears a resistance in the order of the quantum resistances @xmath300 @xcite , such that viewing the system as a left and a right side with some interaction is an appropriate picture .
furthermore contributions from several channels just add in the current .
although a transmitting channel in a point contact is the application in mind , this article shows how to set up a general procedure to solve the problem of transport for two open reservoirs with more or less strong coupling between them , and thus how the green s functions formalism from section i is implied in an area of current research @xcite . as a preparation ,
consider systems like , for example , the bulk material on the left or the right side ( fig.1a ) without coupling .
for these we suppose that we know the hamiltonian @xmath301 and the wave function @xmath302 at any given energy @xmath303 satisfying the schrdinger equation . denoting the hamiltonians @xmath304 ( doubling the index makes sense later ) , the wave functions @xmath305 and the time derivative as @xmath306 , the schrdinger equations for both sides - for the moment just formally put into matrix form
- are @xmath307 corresponding to this differential equation we have the green s functions equation @xmath308 if we knew @xmath309 , we could immediately also give a solution to ( 1 ) with a source term added @xmath310 namely @xmath311 however , we are more interested in the solution when an interaction between left and right is present , expressed through coupling hamiltonians @xmath312 and @xmath313 , @xmath314 and we shall refer to this case by @xmath315 without upper index . introducing the coupling , of course ,
was the motivation for writing the schrdinger equation in matrix form over the site space consisting of l(left ) and r(right ) .
putting the coupling terms on the right side of the equation , they mimic a source @xmath316 and following ( 4 ) the solution can formally be written as @xmath317 obtaining the implicit lippmann - schwinger equation for the wave functions .
although for the coupled system we do not expect the eigenvectors of @xmath318 to be one with only an upper component localized on the left and one with only a lower component localized on the right , it is convenient to denote these solutions as vectors @xmath319 , keeping indices l and r. in contrast to section ii , where the lippmann - schwinger equation was solved in a discretized form to obtain field distributions , for our purposes here ( 7 ) will merely serve as a formal step in the derivation of the green s function .
what we are precisely interested in from a physical point of view is the current that will flow between the left and the right side , and not necessarily explicitly evaluating the wave functions .
the green s function @xmath23 of the coupled system is inferred from an equation analogous to ( 5 ) , namely @xmath320 @xmath23 like the hamiltonian is a full matrix in site space . with @xmath23 , a solution of @xmath321 would read @xmath322 where the index cq stands for coupling and source .
as explained in section i , we could also have set up a lippmann - schwinger equation for @xmath323 as @xmath324 now from inserting ( 10 ) into ( 11 ) and for @xmath10 choosing some @xmath325 either in the l- or in the r - component the dyson equation for @xmath23 is obtained : @xmath326 the reservoirs on the left and right being bulk metal , the differential operator of the homogeneous differential equation ( 1 ) is given by the free - partical hamiltonian @xmath327 and the corresponding wave functions are @xmath328 according to the dispersion relation @xmath329 .
looking for a green s function at given frequency @xmath59 , in ( 2 ) we can replace @xmath301 by @xmath303 and thus have to solve @xmath330 we need an expression , the derivative of which produces @xmath331 .
one easily verifies that ( 14 ) is satisfied by either @xmath332 the retarded function @xmath333 only exists for @xmath334 and the advanced function @xmath335 for @xmath336 . in ( 14 ) for simplicity we restricted ourselves to a single energy and therefore ( 15 ) still contain @xmath59 as a parameter . for complete green s functions in the time domain
these terms have to be multiplied by the density of states @xmath337 and integrated over energy .
later in transport we shall be interested in the amount of charge transferred , not resolving any more , which energy levels contributions came from . thinking physically of a small contact between two metallic leads one might argue that in the constriction transverse @xmath338-vectors are quantized and a one - dimensional continuous density of states remains .
nevertheless , for not too high voltages only a certain energy range around the fermi energy will play a role in transport and thus setting the density of states equal to its value at the fermi energy is a good approximation .
anyway , a constant density of states @xmath339 with occupied states below and empty ones above the fermi level for each side , left and right , shall just enter our model here as an assumption .
we are led to the following representations of the retarded and advanced green s functions : @xmath340 taking out the factor @xmath341 , we get the dimensionless functions @xmath342 and @xmath343 in frequency space . remark that their deduction here did not consist in calculating a fourier transformation .
the phase factor @xmath344 was already there in ( 15 ) .
the @xmath59-integrals from ( 16 ) will be implemented as a useful representation of @xmath345 still in the time domain .
the @xmath346-functions have been deliberately skipped after the @xmath347-signs .
we shall however see that @xmath333 and @xmath335 finally only appear with the correct relation between their first and second time argument .
the @xmath59-integral should indeed be understood as summing over all energies and , even if @xmath348 is a constant , on no account be interpreted as this constant times @xmath349 .
@xmath350 is the projection - like conversion of the phase unit vector of an oscillation from time @xmath351 to time @xmath352 .
left(l ) and right(r ) are distinguished as the origins of wave functions constituting our basis of states , but the junction is considered point - like , that is @xmath353 for both left and right and thus @xmath354 drops out in an alike combination of @xmath355 even from l and r. the two - fold time argument suggests to take ( 16 ) formally even as double fourier transform ( however with different signs in the exponentials with @xmath352 and @xmath351 ) @xmath356 with @xmath357 , however , because different energies stay independent .
( more or less guessing the green s functions @xmath4 was easy in our normal - conducting simplest model here .
generally , it has to be found as the solution to an equation like ( 14 ) with the differential operator from the uncoupled system s schrdinger equation and an elementary perturbation @xmath7 . in the superconducting state , for example , the two - particle interaction in the hamiltonian forces the green s function to include andreev reflection @xcite , and the result is such that @xmath358 does not just vanish in the gap of the quasi - particle density of states . )
before we can solve ( 12 ) for the coupled system s green s function @xmath23 , we have to specify the coupling parts of the hamiltonian @xmath312 and @xmath313 . like with the density of states
the simplest model will assume that the coupling is energy - independent and described by a constant ( real ) interaction energy @xmath359 .
the reservoirs are unaltered by transport between them . due to the applied voltage ( fig.1b ) incoming charge carriers ( from the left on the right )
are led away and outgoing ones ( on the left to the right ) get replaced . in the contact
we do not allow relaxation or other energy - changing processes .
we choose the repective fermi levels as zeros of energy on either side ( fig.1c ) .
an electron going from right to left has to strip off its phase @xmath360 and aquire @xmath361 to fit in on the left .
an analogous argument holds for transitions from left to right , and therefore the coupling terms are @xmath362 the phase factor gives a time - dependence to @xmath312 and @xmath313 , but @xmath363 is indeed independent of energy .
for @xmath23 we make an ansatz like ( 17 ) as a two - fold fourier representation : @xmath364 from ( 12 ) , which is valid for either advanced or retarded functions , as an example , we pick the upper left component of the 2x2 matrix in lr - space and insert ( 17 ) , ( 18 ) and ( 19 ) : ( multiple integral signs are skipped from now on . )
@xmath365 the integral over @xmath366 produces @xmath367 and in @xmath368 there is @xmath369 anyway , such that the last term of ( 20 ) becomes @xmath370 with @xmath371 . strictly speaking ,
if ( 20 ) were for the retarded function , in the last term there would be @xmath372 from @xmath373 and @xmath374 from @xmath375 , and if it were for the advanced function , @xmath376 and @xmath377 , such that the integral over @xmath366 only exists between @xmath352 and @xmath351 instead of having minus and plus infinity as limits .
however , with @xmath378 and @xmath379 running over any value , one can argue that even with a finite @xmath366-integral the only remaining contribution stems from @xmath380 .
a discussion of time ordering will again appear in section 5 .
@xmath381 is the density of states per frequency interval , dividing by @xmath382 makes it the number of states per energy interval .
@xmath359 is an energy .
@xmath383 can be understood as a dimensionless transmission amplitude .
now we set up the convention that all frequency arguments of @xmath4 and @xmath23 are written with respect to the left zero level and for the case that they correspond to the right , that is an r - index , it is understood that @xmath384 is added .
( 20 ) has to hold for any @xmath352 and @xmath351 and from comparing fourier coefficients we get @xmath385 we could have inserted @xmath386 and again replaced @xmath387 and so on . instead of an implicit equation for @xmath23
this would have led to an infinite series ( see section i ) : @xmath388 \(21 ) is written for whole matrices in lr - space , calculations like ( 20 ) can be done analogously for @xmath389 , @xmath390 and @xmath391 .
( 21 ) can be read as equation in the time domain .
then @xmath392 is the matrix consisting of @xmath312 and @xmath313 and each multiplication of two following @xmath4s with @xmath392 in between means an integration over time .
however , ( 21 ) is as well valid as relation in frequency space . in this case
@xmath392 is the 2x2 matrix with just @xmath383 as off - diagonal elements .
like we have seen through evaluating the @xmath366-integral in ( 20 ) , in ( 21 ) each connection @xmath393 from @xmath392 passes the frequency argument from the @xmath4 in front to the @xmath4 behind . and as @xmath4 can only have two identical frequency arguments , no @xmath59 different from the first can ever appear , such that @xmath394 and @xmath23 also effectively is a function of only one frequency argument : @xmath395 . with green s functions of a single frequency argument ( 20a ) and
its analogues for the other three components in lr - space become a simple algebraic equation : @xmath396 ( the fact that even the green s function of the coupled system turns out to be a function of a single frequency argument is a special feature of our simple model for the normal conducting case . in the extension of this model to superconducting reservoirs and transmission processes including andreev reflection , @xmath23 becomes a function of two frequency arguments , the second , however , restricted to values differing from the first by an integer multiple of @xmath384 , such that effectively there is one continuous and one discrete frequency parameter @xcite . ) here , with @xmath397 , ( 22 ) is easily solved and @xmath398 comes out independent of frequency , too : @xmath399^{-1 } \begin{pmatrix } \mp i & 0 \cr 0 & \mp i
\cr\end{pmatrix } \nonumber \\ & = & \frac{1}{1+t^2}\begin{pmatrix } \mp i & -t \cr -t & \mp i \cr\end{pmatrix}\end{aligned}\ ] ] the same result could have been obtained from ( 21 ) by writing out a few more of the matrix multiplications and using the formula for the geometric series in each element .
we shall need two further types of green s functions .
@xmath400 will be introduced in the next section and @xmath401 and @xmath402 when calculating the current . in ( 21 ) there was a sum of products of arbitrary many factors @xmath4 and @xmath392 with outer factors @xmath4 .
product , of course , except with the green s functions taken of a single frequency parameter , in the time domain or with two - fold frequency dependence still meant a convolution - type integration over inner arguments . in analogy
we define the sum of products with outer factors @xmath392 ( as integrals in the time domain or just algebraically with @xmath358 and @xmath403 ) : @xmath404 whereas all contributions to @xmath23 in ( 21 ) began and ended with staying some time in a reservoir , described by @xmath4 - at l from @xmath352 to @xmath366 for a start in the last term of ( 20 ) , for example - each term of @xmath400 in ( 24 ) begins and ends with a transition @xmath392 and further contains at least one such hopping across the junction ( which @xmath4 does not ) .
therefore we call @xmath400 the _ transfer _ green s function . the same as for @xmath23 ,
the relation between @xmath400 as a function of times and as a function of frequency is given by @xmath405 where @xmath406 if j = l and @xmath407 if j = r and the same for @xmath408 .
( taking out @xmath409 of @xmath381 in ( 16 ) was a convention .
the prefactor of the @xmath59-integral for @xmath400 follows from consistency requirements .
( 19 ) for @xmath23 was in complete analogy to ( 16 ) for @xmath4 .
@xmath400 with two time arguments , however , has a little different character from @xmath392 with just one time parameter . ) alternatively to ( 24 ) @xmath400 could be defined through its link to @xmath23 @xmath410 be careful that replacing one by the other can introduce another internal time integration as , for example , @xmath411 .
( 24 ) and ( 26 ) hold for retarded and advanced functions . from ( 24 ) it is immediately seen that @xmath400 like @xmath23 satisfies a dyson equation @xmath412 and even the complementary forms @xmath413 are analogues . in fourier space , like ( 22 ) the @xmath400-equation ( 27 ) is an algebraic equation and the solution like @xmath414 inserting @xmath4 and @xmath392 explicitly for our model we get @xmath415^{-1 } \begin{pmatrix } 0 & t\cr t & 0\cr\end{pmatrix } \nonumber\\ & = & \frac{1}{1+t^2}\begin{pmatrix } \mp it^2 & t \cr t & \mp it^2\cr\end{pmatrix}\end{aligned}\ ] ] especially @xmath416 whereas @xmath383 is the single hopping amplitude , @xmath417 is a renormalized transfer amplitude .
one may wonder why a model for transport could not have been set up adding amplitudes for transfer processes of all orders , as the interaction ( 18 ) seems to be introduced the way it is just in order to result in powers of @xmath383 .
however , @xmath397 deduced from the schrdinger equation is decisive for the signs in ( 30 ) and ( 31 ) .
one may wonder that multiple reflections are not added as @xmath418 .
processes of different order ( fig.2 ) are not independent , but interfere .
@xmath419 is the transfer amplitude per single electron supplied on the left by the voltage source .
but for every electron that goes over to the right in an @xmath284th order process ( @xmath284 odd ) with weight @xmath420 there is one that has hopped once more to the right and back ( @xmath421 order process ) and thus is not to be newly supplied , but to be again sent through the junction .
the amplitude @xmath383 is renormalized by @xmath422 as denominator .
however , such interpretations of quantum mechanical amplitudes are precarious , and the full conversion of @xmath383 to a transmission probability will be established later .
it is quite instructive to solve ( 27 ) in a slightly different way than done in ( 30 ) .
firstly , for the four components in lr - space we have @xmath423 inserting these into each other , for example , an equation for @xmath424 alone is obtained : @xmath425 this implicit equation is the basis for calulating the transfer green s function in more complicated cases than discussed here @xcite , like for example the superconducting junction . in our model , inserting @xmath426 and @xmath427 into ( 32 ) immediately also leads to @xmath428 .
more easily than the dyson equation for the ordinary green s function @xmath23 , the one for the transfer green s function @xmath400 is illustrated as is done for the lr - component in fig.3 .
( normally indices are read from right to left such that @xmath424 is considered a transition from right to left , but it does not really matter whether they are interpreted the other way round as in fig.3 .
the actual sequence of what is earlier or later in time will be discussed when calculating the current in the next section . )
fig.3 demonstrates the implicitness of the dyson equation : any transition from left to right is either a single transfer or an electron hopping to the right and back followed by any process beginning on the left and ending on the right , no matter what happens in between .
this last part by definition is the sane as the other side of the equation , namely @xmath424 . from the heisenberg picture of quantum mechanics we know that the time derivative of a not explicitly time - dependent operator @xmath190 is given by the commutator with the hamiltonian @xcite : @xmath429\ ] ]
the operator of interest here is the projector on either side of the junction @xmath430 as explained earlier , with the junction coupling left and right together , the solution @xmath431 is not limited to one side , however , the projectors take out the respective part : @xmath432 @xmath433 and @xmath434 are proportional to the amount of charge on the left and on the right side .
their time derivatives have equal absolute values , but opposite sign and represent the current .
@xmath435 ( @xmath436 and @xmath437 are used for bra- and ket - states . here
@xmath438 means the expectation value , of course . )
@xmath439 is the charge of an electron and the sign of @xmath440 can be defined arbitrarily .
we choose to do the calculation with @xmath433 and evaluate the commutator with the hamiltonian : @xmath441=\left [ \begin{pmatrix } h_{ll } & h_{lr } \cr h_{rl } & h_{rr } \cr\end{pmatrix } , \begin{pmatrix } 1 & 0 \cr 0 & 0 \cr\end{pmatrix } \right ] = \underbrace { \begin{pmatrix } 0 & -h_{lr } \cr h_{rl } & 0 \cr\end{pmatrix}}_{\sigma_c}\ ] ] putting together ( 33 ) and ( 35 ) we obviously need the expectation value of the operator @xmath442 $ ] .
@xmath442 $ ] shall be called @xmath443 .
even if ( 33 ) stems from the heisenberg picture , it is written in such a way , that the right hand side is to be evaluated in the schrdinger system with time dependent states , and we shall here change to the interaction picture @xcite for the calculation . with @xmath444 standing for complex conjugation as well as transposition from column to line vector , the value of @xmath445 in state @xmath446 at time @xmath352
is given by @xmath447_{\mathcal{h } } \ ; \widehat{t } \ ; \exp ( \frac{1}{\hbar}\int_{\tau_0}^{\tau } ( -i ) \ ; \sigma_{\mathcal{h}}(\tau ' ) \ ; d\tau ' ) \vert \psi^0(\tau_0 ) > \nonumber \\\end{aligned}\ ] ] @xmath448 and @xmath449 denote the time occurence of @xmath352 from ( 18 ) which has to be considered as still belonging to the schrdinger picture in our case here .
@xmath450 means time ordering @xcite and @xmath451 anti - time ordering .
@xmath302 in the uncoupled system , of course , also stands for a two - vector with left and right component . replacing @xmath452 by @xmath453 in ( 37 )
we took out both the coupling as well as the time dependence from the states . accordingly the @xmath454 in the integrals , in contrast to the schrdinger - picture @xmath392 used in the preceding sections , in a heisenberg way have to include the time dependence of the uncoupled states .
@xmath455 could already be written as @xmath456 or @xmath457 for short , where bra , ket and @xmath359 still mean the respective matrices . however
, this decomposition might rather be confusing and will not be used , anyway .
the translation to the interaction picture is the following : @xmath458 or @xmath459 .
the heisenberg picture always refers the operator back to the undeveloped state at @xmath460 . on the minus and the ket @xmath461 on the plus branch
this view picks out the single transition at time @xmath352 and visualizes the calculation of @xmath462 .
each point on the contour represents an lr- or rl - transition described by @xmath392 , each line segment the development of the wave - function phase which is given @xmath4.,width=302 ] analogously @xmath443 will have to be extended by the time - dependence of the uncoupled states .
the meaning of the time - integral over @xmath454 as an exponential is best explained by writing explicitly : @xmath463 where all arguments @xmath464 have to lie between @xmath460 and @xmath352 and all products from none to arbitrarily many factors @xmath454 have to be added . with time
ordering , furthermore @xmath465 is imposed . without @xmath450 the scheme for the exponential series @xmath466 in ( 39 ) would produce @xmath467 .
but then for every fixed set @xmath468 there are @xmath469 permutations for the time values to occur such that time ordering and restricting the arguments to mutually exclusive intervals as done in ( 39 ) cancels the factorial denominators . writing out the anti - time ordered part in the same way as shown for the time - ordered part , the expression from ( 37 ) becomes @xmath470 where @xmath471 , @xmath472 and integration over all time arguments except @xmath352 and @xmath460 is understood .
like in ( 39 ) we mean the sum of all products with arbitrary many inner development factors like @xmath302 of @xmath473 and @xmath302 of @xmath474 .
spaces and different font types are just used in ( 37a ) to recognize sequential @xmath454-parts like ( 38 ) .
_ inner _ development factors like @xmath302 of @xmath473 and @xmath474 as well as @xmath475 and the inner @xmath453 have to sum over space or the basis of states and therefore should be thought of as matrices as in ( 38 ) .
we have not yet decided which components _ outer _ bra and ket with @xmath476 will project out .
all @xmath477 are the unit matrix and drop out ( @xmath302 being given by a non - normalizable @xmath439-function does not pose a problem here ) .
then @xmath478 with @xmath479 is recognized as @xmath480 and @xmath481 with @xmath482 as @xmath483 .
these @xmath484 are 2x2 diagonal matrices . with ( 21 ) in ( 37a ) the whole sequence from @xmath485 to @xmath486 can be replaced by @xmath487 and the long part from @xmath488 up to @xmath489 is just @xmath490 .
the complete operator between the outermost @xmath491 and @xmath492 in ( 37a ) therefore becomes @xmath493 \sigma_c(\tau ) \nonumber \\ \left [ { \bf 1}\ ; \delta(\tau_n-\tau)+g^r(\tau,\tau_n)\sigma(\tau_n)\right ] \vert\psi^0(\tau_n)><\psi^0(\tau_0)\vert\end{aligned}\ ] ] the @xmath94-contributions stem from the cases where there are no factors with @xmath494 or @xmath495 . we regard the coupled system as having developed out of the uncoupled system , but we are looking for a stationary state . to achieve this ,
the coupling has to have been turned on infinitely long ago , thus we let @xmath496 . the sequence of time arguments
is usually represented on the so - called keldysh contor ( fig.4 ) .
we still have to take the expectation value of ( 37 ) or ( 37a ) .
this means summing over the basis of states for the outer @xmath476 .
the uncoupled basis consists of @xmath497 and @xmath498 for a state on the left and a state on the right for each frequency @xmath59 . we shall discuss the occupation or emptiness of states shortly after having worked off some further more formal points . summing over @xmath499 and @xmath500 for
the outer @xmath476 will return the trace of the matrix @xmath239 given in ( 37b ) .
@xmath501 only being interested in the trace as a result , in a matrix multiplication the order of factor matrices can be changed cyclically . as in ( 37b )
every bra and every ket as well as each operator part between the two @xmath502 is a matrix we can rotate factors to obtain @xmath503 \quad \vert\psi^0(\tau_n)>\ ; \cdot\nonumber \\ <
\psi^0(\tau_0)\vert \psi^0(\tau_0)>\frac{i}{\hbar}<\psi^0(\tau_m)\vert \ ; \left [ { \bf 1}\ ; \delta(\tau_m-\tau)+\sigma(\tau_m)g^a(\tau_m,\tau)\right]\end{aligned}\ ] ] @xmath504 again drops out .
@xmath505 is of the structure @xmath4 , however , with no restriction as to which argument @xmath506 or @xmath507 is earlier or later in time .
we define this new type of green s function as @xmath508 ( an eventually ill - defined single point @xmath509 is irrelevant for later integrations . ) as the cases @xmath510 and @xmath511 are mutually exclusive the function can also be given by @xmath512 ( conclusions on @xmath513 from that are risky , to my opinion , though ) .
we have thus deduced the current formula @xmath514 \ ; \cdot\nonumber \\ & & \cdot \ ; g^{+-}(\tau_n,\tau_m ) \ ; \left [ { \bf 1}\delta(\tau_m-\tau)+\sigma(\tau_m ) g^a(\tau_m,\tau ) \right ] \ ; \}\end{aligned}\ ] ] ( like discussed in ( 20 ) it is rather irrelevant whether the upper integration limits are set as @xmath352 or @xmath515 ) or @xmath516 if @xmath402 is defined as @xmath517 \ ; g^{+-}(\tau_1,\tau_2 ) \cdot \nonumber \\ & & [ { \bf 1}\delta(\tau_2-\tau)+\sigma(\tau_2)g^a(\tau_2,\tau')]\end{aligned}\ ] ] or @xmath518 in short notation .
the current only needs @xmath402 of two identical time arguments .
we shall need the fourier representation of @xmath513 , but none such for @xmath402 .
although in our simple model for the normal - conducting contact the current will come out the same for any @xmath352 , @xmath519 generally does depend on time ( our expectation value is an ensemble , not a time average ) .
for the superconducting case the ac parts @xcite like the josephson current are included in the expression ( 41 ) , although one is mostly interested in the contribution in @xmath440 that is independent of @xmath352 and gives the dc part .
using @xmath520 without a third argument @xmath59 in ( 41 ) we understood integration over frequency like in ( 16 ) . but
this point requires more care as it has not been taken into account so far whether states are occupied or empty at @xmath521 . writing out the trace from ( 41a ) in lr - components reveals four contributions
@xmath522 it is clear that in contrast to the other appearing @xmath392 , in @xmath443 from ( 36 ) there is an additional relativ minus sign between the lr- and the rl - component .
obviously the net current is the difference between the current from left to right and the current from right to left . in the first - as well as the second - term in ( 43 )
a transition from r to l at @xmath352 , the time argument of @xmath443 , is picked out to be counted for the current ( see also fig.4 ) . with @xmath401 at r
, the charge carrier is supplied from a state originally located on the right .
it is an electron if the energy lies below the right - side fermi level . until @xmath352 the state has evolved to again be on the right .
the originally left behind empty state at the same energy below the fermi level must have evolved to be on the left at @xmath352 such that by the r@xmath523l transfer the electron can go into it ( fig.5a ) . or
you might say that the plus and the minus branch of the keldysh contor represent two possible parts in the evolution of an original wave function @xmath524 , between which there is a non - vanishing matrix element of the operator @xmath525 $ ] .
however , still regarding the first line of ( 43 ) , there is the further possibility that an empty state from above the fermi level on the right evolves to be at r at @xmath352 again , but its left - behind complement ( a negative charge ) has evolved to be at l. @xmath526 in this case means the transition of an empty state or positively charged particle from right to left ( fig.5b ) .
although one does not usually introduce the concept of holes with transport in normal conducting metals , it makes sense here to call unoccupied states simply
holes . in this way
the model already includes the dual nature of charge carriers needed for the superconducting case . in the normal conducting case states
do not change in nature ( electron or hole ) or energy during their evolution ( little zigzags are drawn in fig.5 only to make the multiple hoppings visible ) . at the superconducting junction
, andreev reflection can be interpreted as changing an electron into a hole or vice versa and mirroring its energy at the fermi level @xcite ( see fig.9 in section 6 ) . to have a charge carrier at a certain energy level at time @xmath352 to make a certain transition ,
it is important that there was one at the corresponding energy in the original uncoupled system .
the coupling may have changed the distribution compared to the occupation in uncoupled bulk reservoirs .
and the applied voltage imposes a non - equilibrium situation , anyway . for the evolution of a state from the right
as shown in fig.5 it does not matter whether the state at the respective energy on the left is occupied or empty .
if , for example the regarded energy lies below the fermi levels both left and right , there will be two states evolving as an electron on the plus and a hole on the minus branch of the keldysh contor , one having originated at r and the other at l. these original , uncoupled and independent states are our basis , especially for calculating an expectation value as trace .
they do not interfere .
terms with @xmath527 and @xmath528 are simply added in ( 43 ) .
schemes analogous to fig.5 could be drawn for the terms from the last three lines of ( 43 ) as well .
the conclusion of the whole argumentation of how to let the original fermi occupation function for the reservoirs left and right enter the current calculation is that @xmath528 has to change sign at the left fermi energy and @xmath527 at the right fermi energy .
let us note @xmath401 like @xmath333 and @xmath335 for any bulk reservoir with fermi level at @xmath529 .
we shall keep @xmath401 corresponding to @xmath333 and @xmath335 as in ( 40 ) for occupied electron states below the fermi level and change the sign for empty states above it . @xmath530
although it might be practical to use ( 19 ) , ( 23 ) , ( 44 ) , ( 18 ) and ( 36 ) in ( 41 ) to quite directly produce an expression that calculates the current finally as an integral over frequency and in our simple model can even be analytically evaluated , in parallel to @xcite we shall use the transfer green s functions here .
the current is best translated into an expression of the transfer functions from the form already resolved into lr - components ( 43 ) .
furthermore eliminate @xmath443 through @xmath531 and @xmath532 .
l- and r - indices follow logically from ( 26 ) , for example @xmath533 . @xmath534 as quantities here are no longer matrices , but simply functions of time or frequency , the leading @xmath535 and @xmath536 in the second and third line of ( 43a ) can be moved to the ends of the products . from ( 41 ) we remember that their argument is @xmath352 , the same as the second argument of the last factor to the right . then using relations from ( 27a ) and complementary forms the current formula simplifies to @xmath537 the terms with the factors 1 from the brackets have elegantly been made to vanish .
now we use the fourier representations for all functions in ( 43b ) .
as all terms follow the same scheme , the second is treated in an exemplary way here ( all integrals run from @xmath538 to @xmath539 ) : @xmath540 doing the @xmath541,@xmath542- and @xmath543-integrals gives @xmath544 .
then even the exponentials with @xmath352 cancel and the term simplifies to @xmath545 in our case ( 31 ) tells us that @xmath546 are all identical and real , @xmath547 is the complex conjugate of @xmath548 and @xmath549 is the complex conjugate of @xmath550 of the same @xmath59 as @xmath513 is purely imaginary , too .
thus the last two terms in ( 43b ) are the complex conjugates of the first two and thus twice the real part of these first two can be taken for @xmath551 . in the superconducting version of the model , where @xmath400 and @xmath4 actually are @xmath59-dependent , complex conjugate relations @xcite also exist between @xmath400s as well as @xmath4s , and the current formula can be reduced in the same way . just to note the quite general formula @xcite in short form : @xmath552 for the integrand from the second term in our model we get @xmath553 where the signs refer to @xmath59 greater or less than zero and come out reversed for the first term with @xmath527 , because there is @xmath335 instead of @xmath333 .
care has to be taken with the reference point for @xmath59 in both terms .
this may easily be overlooked in the normal conducting case here in contrast to the superconducting case where @xmath400 indeed is @xmath59-dependent and like @xmath23 as a function of only one argument always referred to the same fermi level ( the left , for example ) .
if we call the argument of @xmath528 from the second term @xmath59 , the one for @xmath527 in the first term is @xmath554 . on an @xmath555-axis , the second term changes sign at zero , however , the first jumps at @xmath556 ( fig.6 ) .
a shift of the integration parameter can not be made independently for both terms .
thus , for the normal conducting model here @xmath557 in principle the convention is needed , that the @xmath400-argument always refers to the left side , but for @xmath4 general formula like ( 44 ) for a single bulk with fermi level at zero frequency are applied . a more involved situation where choosing integration intervals consistently for all contributing current terms is crucial
can be found in @xcite .
the charge is transferred to the left .
( b ) evolution of an unoccupied state above the fermi level from r such that at @xmath352 it gets filled by a charge from the left or the hole state shifts to the left .
the lower pictures sketch the effect of the original uncoupled state for the current . sending electrons from the right to the left
will be compensated by electrons sent from left to right . but with levels as in fig.1c for holes sent from right to left there will be a range between the two fermi levels where there are no counterbalancing holes going the other way .
( for the left , electrons going to the right are not outweighed in this range.),width=453 ] .,width=302 ] from fig.6 it is easily seen that that the integral is twice the constant @xmath558 integrated over an interval of length @xmath384 , and outside that interval contributions cancel . with the other factor 2 from twice the real part and the prefactor @xmath559 the result for the current finally is @xmath560 the factor @xmath561 is the conductance in units of the quantum conductance or its inverse the resistance in units of the quantum resistance @xcite .
( the conductance of a channel doubles if two spin states are allowed
. then @xmath562 should be taken as resistance unit . )
@xmath346 is the transmission probability of the conductance channel through the junction we regarded .
the result that in the normal - conducting case the current is proportional to the voltage is not at all surprising , of course .
the non - trivial result is the conversion of the quantum mechanical transmission amplitude @xmath383 to the measurable transmission probability @xmath346 . @xmath563
if @xmath564 . and a totally open channel with @xmath565 has transmission probability @xmath566 .
it may seem a contradiction on the one hand calculating the current from a changing amount of charge on one side and on the other hand saying that missing charges are replaced and superfluous ones led away by the voltage source . a slightly different viewpoint may help to get convinced that the calculated quantity is indeed the current in the stationary , but non - equilibrium system .
the crucial point was putting on to evaluate @xmath567 in state @xmath568 . without further ado
we could not tell whether this state of the coupled system was occupied or not .
the schrdinger equation ( 5 ) set up the left and right material properties as well as the coupling across the junction , however , did not take any account of the effect of the voltage source . without need to specify real locations for the division ,
just principally view our structure as consisting of a junction region and leads .
the @xmath568 defined through ( 5 ) describes states in the junction region . but think of them as offered by the system and following their time development whether occupied or not .
the leads are always occupied exactly up to their fermi levels .
electrons freshly supplied by the voltage source need not be in phase with present ones . a random phase is most easily modelled by assuming the left lead wave function at any time @xmath569 without a phase in contrast to the assumed @xmath570 for the left side of the junction region . which states to which extent actually get occupied in the left side of the junction region , that is @xmath571 ,
is determined by the overlap of @xmath568 with the phaseless left lead wave function @xmath499 , which we recognize as @xmath572 .
the current then is the change in time of @xmath433 due to charge flow through the junction , that is processes inside the junction region only , corresponding to what our hamiltonian was set up for .
a similar line of thought with state overlaps can be applied to the passing of charges out of the junction region into the ( right ) lead ; it may be helpful to view this as putting empty states or holes into the junction region , though .
of course , only the l - part of @xmath568 can overlap with @xmath499 .
@xmath574 can thus be replaced by @xmath568 in our new expression for @xmath433 .
recalling the operator definition ( 34 ) the time derivative of @xmath433 @xmath575 is identical to the ansatz made by putting together ( 33 ) and ( 34 ) in ( 37 ) . at first glance the problem posed by the schrdinger equations without and with coupling , ( 1 ) and ( 5 ) ,
especially if we regard a single energy level on each side with corresponding @xmath576 and @xmath379 as in fig.1c , looks like the two - level system known from quantum mechanics textbooks @xcite . by the coupling
the two energy levels are shifted .
the new eigenstates lie further apart from one another . if the system is initiated in one of the uncoupled states , it will oscillate harmonically between the two levels , and both both the period of the oscillation as well as the maximum transition probability to the other state depend on the energy difference of the original states .
even if the time dependence of the interaction ( 18 ) can be got rid of by changing to the interaction picture , it makes no sense to numbly calculate eigenvalues and eigenvectors of a hamiltonian @xmath577 .
it is not clear from which reference levels such eigenenergies @xmath578 should be counted as zero levels are different for @xmath576 and @xmath379 .
if one tries to refer the left and right energies already of the uncoupled system to the same reference level , @xmath579 for example ( fig.8 ) , @xmath580 ( and @xmath524 ) can not be taken as eigenstates any more , because from @xmath581 with @xmath582 it does _ not _ follow that @xmath583 equals @xmath584 .
the standard treatment of the two - level system is not applicable to the non - equilibrium situation . despite having been the simplest example to introduce the green s functions scheme , applying the formalism to the normal conducting junction to get out that the current is proportional to the applied voltage was breaking a butterfly on the wheel .
although repeatedly mentioned , fully developing the extension to the superconducting case is beyond the scope of this presentation . but the resulting current - voltage characteristics shall be shown as a plea for the usefulness of the method . a different approach based on matching wave functions @xcite leads to identical results , though .
some formula shall be listed , because they are not necessarily written out in complete analogy to the presentation here in @xcite and other literature .
working in the quasiparticle picture , in the superconducting case each entry in lr - space of a green s or transfer function expands into another 2x2 matrix in nambu space over electrons and holes .
there is @xcite : @xmath585 @xmath4 is the same for ll and rr .
different signs refer to the retarded and advanced function .
@xmath586 is half the gap of the superconductor .
the small imaginary part @xmath587 is added to get the correct root besides slightly smoothening singularities .
@xmath588 still holds .
@xmath589 here signs refer to the lr- and rl - direction , respectively .
remark the reversed signs for holes in the exponential compared to electrons .
there is no electron - hole conversion during a single hopping .
the @xmath590- and @xmath591-elements of @xmath4 introduce so - called andreev reflection .
( cooper - pair tunneling @xcite is not included . )
@xmath400 is a full matrix in @xmath590-space and has to be set up as @xmath592 we skipped the distinction of reference @xmath593 as in ( 25 ) , because in practice only one type , e.g. @xmath424 , has to be calculated , as then @xmath419 follows from complex conjugation .
@xmath284 runs over all integers .
a dyson equation like ( 32 ) leads to a recursion which connects @xmath594 to @xmath595 and @xmath596 .
it can be solved by reasonably truncating the @xmath284-range .
then , just also summing over @xmath284 , the established current formula ( 45 ) can be used .
expressions like ( 32 ) have to sum over all possible combinations of @xmath439- and @xmath77-indices .
for example : @xmath597 ) [ reproduced after @xcite ] .
the normal conducting @xmath598 is added for comparison ( keep @xmath586 just for units).,width=264 ] to point out the andreev reflection in the model , we look at a term @xmath599 with alternating @xmath439- and @xmath77-indices from the development ( 21 ) . such a sequence will also be contained as parts in higher - order terms from ( 21 ) or ( 24 ) . contracting inner time arguments results in @xmath600 because the @xmath541 and @xmath542-integration give @xmath601 . in fig.9 for holes the energy axis has to be reversed as compared to electrons .
therefore the initial @xmath59 is negative . for the argument here , it does not matter if the beginning factor in ( 51 ) is @xmath602 or @xmath603 .
if it were @xmath77 , then the andreev reflection marked by the dashed arrow would follow .
the last index of the last @xmath604 could also be @xmath77 instead of @xmath439 .
whereas in the normal conducting case in multiple reflections the charge carrying particle always came back to the same energy with regard to the fermi level when coming back to the same side of the junction ( fig.1 ) , in ( 51 ) the frequency argument of the first @xmath604 is shifted by twice the applied voltage equivalent with respect to the second @xmath604 ( fig.9 ) .
such shifts enable transport even when the gaps of the left and right side are still overlapping , that is for voltages below @xmath605 ( fig.10 ) . in the @xmath598-curves steps towards lower voltages at fractions of @xmath606
are associated to ever higher - order andreev reflections .
of course , these current contributions become the more prominent the greater the channel transmission . besides the step heights and positions the curvature on each step is an important signature .
the model agrees excellently with experimental results @xcite .
the interference of transport processes of different order is correctly taken into account .
simpler models @xcite that add up tranfer probabilities proportional to @xmath607 can not reproduce these characteristics and are only valid for low transmissions @xmath608 .
it has been explained on a quite basic level how quantum transport between two reservoirs in a stationary non - equilibrium state can be modelled .
the purpose was to present a green s functions technique for handling coupling in the context of a field of current research interest , namely transport through point contacts . although besides the general formalism ( also see section i ) requiring the development of quite some more subject specific mathematical framework such as the transfer green s functions , the transport through a contact with arbitrary transmission is a suitable example to illustrate the inclusion of interaction up to all orders in the implicit dyson equations .
calculations have been carried through in every detail for the normal - conducting single junction .
basic formula and results have been given for the superconducting junction .
the decomposition into transport channels ( eigenmodes ) of a point contact can be inferred from its superconducting transport characteristics which can be taken like a pin - code @xcite .
the presented green s functions formalism has a great potential for extension .
systems of two more or less coherently linked junctions can be modelled @xcite .
models for transport through molecules @xcite or atomic chains @xcite so far mainly rely on ab initio calculations of the density of states .
time dependent density functional theory for non - equilibrium situations is also developed @xcite . the author thanks alain dereux for learning about green s functions in electrodynamics , juan - carlos cuevas for through his former work introducing her to green s functions in quantum transport , her supervisor elke scheer for the freedom to work on this subject as well as klaus - ulrich neumann for the suggestion to write a paper of this kind starting from some lecture notes .
van wees b j , van houten h , beenakker c w j , williamson j g , kouwenhoven l p , van der marel d and foxon c t 1988 _ phys.rev.lett . _
* 60 * 848 ; beenakker c w j and van houten h 1991 _ solid state physics _ * 44 * 1 scheer e , joyez p , esteve d , urbina c and devoret m h 1997 _ phys.rev.lett . _ * 78 * 3535 ; scheer e , agrait n , cuevas j c , levy - yeyati a , ludoph b , martin - rodero a , rubio - bollinger g , vanruitenbeek j and urbina c 1998 _ nature _ * 394 * 154 | this introduction to green s functions is based on their role as kernels of differential equations . the procedures to construct solutions to a differential equation with an external source or with an inhomogeneity term
are put together to derive the dyson equation for the green s function of the inhomogeneous system .
very different areas of physics such as , for example , electrodynamics and quantum transport , can profit from this green s function formalism .
the fundamental homogeneous - medium green s tensor of electrodynamics is deduced from the field of a dipole . based upon
that a numerical procedure is presented to solve the wave - equation for the near - field in a scattering setup for arbitrary material distributions .
the full inhomogeneous system s green s function is not explicitly needed to get the fields , although it can be obtained by a very similar calculation and in optics can be interpreted as a density of states .
it is demonstrated how the transport problem for two open free - electron gas reservoirs with arbitrary coupling can be solved by finding the system s green s function . in this sense
the article is an introduction on green s functions for treating interaction .
a very detailed discussion of the current formula is given on an elementary basis . |
nearly all superconductors discovered recently and some well - studied compounds ( e.g. , nb@xmath0se and v@xmath1si ) are multiband materials with complex fermi surfaces and anisotropic order parameters .
measuring temperature dependences of the london penetration depth @xmath7 , often converted to the superfluid density @xmath8 , and of the electronic specific heat @xmath2 , are among primary tests directly linked to pairing mechanisms of new superconductors . still , the methods employed to interpret the data are often empiric with simplicity as a main justification .
the most popular among practitioners
@xmath4-model " takes a shortcut by assigning the bcs temperature dependence to the two gaps @xmath9 with which to fit data on the specific heat @xcite and the superfluid density @xmath10.@xcite here , @xmath11 are evaluated with @xmath12 and @xmath13 takes account the relative band contributions . although the @xmath4-model had played an important and timely role in providing convincing evidence for the two - gap superconductivity in mgb@xmath0 , @xcite it is intrinsically inconsistent in the most important task of this procedure , namely , in describing properly the temperature dependencies of @xmath14 and @xmath2 .
in fact , one can not _ a priory _ assume temperature dependencies for the gaps in the presence of however weak interband coupling imposing the same @xmath5 for two bands .
in unlikely situation of zero interband coupling , two gaps would have single - gap bcs - like @xmath15 dependencies , but will have two different transition temperatures , see fig.[fig1 ] . the full - blown microscopic approach based on the eliashberg theory ,
on the other hand , is quite involved and difficult for the data analysis .
@xcite hence , the need for a relatively simple but justifiable , self - consistent and effective scheme experimentalists could employ .
the weak - coupling model is such a scheme . over the years
, the weak - coupling theory had proven to describe well multitude of superconducting phenomena .
similar to the weak coupling is the
renormalized bcs " model of ref .
that incorporates the eliashberg corrections in the effective coupling constants in a manner described below .
we will call our approach a weak - coupling two - band scheme " and clarify in the text below that applicability of the model for the analysis of the superfluid density and specific heat data is broader than the traditional weak coupling .
the s - wave weak - coupling multigap model has been proposed at the dawn of superconductivity theory by moskalenko @xcite and suhl , matthias , and walker@xcite when numerical tools were still in infancy . in this work
we basically follow these seminal publications to develop a self - consistent procedure for the penetration depth data analysis .
our scheme allows one to connect between two independent data sets , the superfluid density and the specific heat , thus providing a reliability check upon values of the coupling constants extracted from fitting the data .
we also discuss the suppression of the critical temperature @xmath5 by non - magnetic impurities and suggest that the data on this suppression for mgb@xmath0 are consistent with a weak repulsive interband interaction that corresponds to opposite signs of the order parameter on two bands , i.e. , to @xmath16s structure of the order parameter . to test our formal scheme , the data for two known s - wave two - gap superconductors , mgb@xmath0 and v@xmath1si , were used .
the specific heat data were taken from the ref .. the magnetic penetration depth , @xmath7 , was measured by using a self - oscillating tunnel - diode resonator ( tdr ) with resonant frequency @xmath17 mhz .
the measured quantity is the shift of this frequency , @xmath18 , where @xmath19 is the total magnetic susceptibility in the meissner state and @xmath20 is a geometric factor defined by volumes of the coil , @xmath21 , and of the sample , @xmath22 , and by the demagnetization factor @xmath23 .
@xmath24 is measured directly by pulling the sample out of the coil at the lowest temperature .
@xcite for the susceptibility we use @xmath25 - 1 $ ] .
perhaps , the simplest formally weak - coupling approach is based on the eilenberger quasiclassical formulation of the superconductivity valid for general anisotropic order parameters and fermi surfaces.@xcite eilenberger functions @xmath26 for clean materials in equilibrium obey the system : @xmath27 here , @xmath28 is the fermi momentum ; @xmath29 is the gap function which may depend on the position @xmath30 at the fermi surface in cases other than the isotropic s - wave .
further , @xmath31 is the total density of states at the fermi level per one spin ; the matsubara frequencies are defined by @xmath32 with an integer @xmath33 , and @xmath34 is the debye frequency ; @xmath35 stands for averages over the fermi surface . as a _ weak coupling _
theory the eilenberger scheme deals with the _ effective electron - electron coupling _ @xmath36 responsible for superconductivity ; properties of intermediate bosons ( phonons or other possible mediators ) enter via properly renormalized @xmath36 .
consider a model material with the gap given by @xmath37 where @xmath38 are two sheets of the fermi surface .
the gaps are constant at each band . denoting the densities of states on the two parts as @xmath39 ,
we have for a quantity @xmath40 constant at each fermi sheet : @xmath41 where @xmath42 ; clearly , @xmath43 . equations ( [ eil1 ] ) and ( [ eil3 ] ) are easily solved .
within the two - band model we have : @xmath44 where the band index @xmath45 .
the self - consistency equation ( [ self - cons ] ) takes the form : @xmath46 where @xmath47 are dimensionless effective interaction constants .
a remark is here in order on applicability of eq.([self - cons1 ] ) central for our approach .
starting with the general eliashberg formalism , nicol and carbotte derived a
renormalized bcs " self - consistency equation:@xcite @xmath48 where @xmath49 is the coupling due to electron - phonon interaction , @xmath50 describes the coulomb interaction , and @xmath51 is the strong coupling renormalization . replacing @xmath52 we obtain our eq.([self - cons1 ] ) .
one should have this in mind while interpreting the constants @xmath53 which can be obtained from fitting the data to our renormalized weak - coupling " model .
note that the notation commonly used in literature for @xmath54 differs from ours : @xmath55 .
we find our notation convenient since , being proportional to the coupling potential , our coupling matrix is symmetric : @xmath56 .
it is worth stressing that for a given coupling matrix @xmath57 , relative densities of states @xmath58 , and the energy scale @xmath59 , eq.([self - cons1 ] ) determines both @xmath5 and @xmath60 . as @xmath62 , @xmath63 , and @xmath64 .
the sum over @xmath65 in eq.([self - cons1 ] ) is readily evaluated : @xmath66 @xmath67 is the euler constant .
this relation can also be written as @xmath68 the system ( [ self - cons1 ] ) is linear and homogeneous : @xmath69 it has nontrivial solutions @xmath9 if its determinant is zero : @xmath70 the roots of this equation are : @xmath71 since @xmath72 , eq.([s - tc ] ) shows that only positive @xmath73 are admissible .
the second form ( [ s1 ] ) shows that both roots are real ; their product is @xmath74 . choosing a proper root
, one should consider various possibilities .
if @xmath75 ( that implies both @xmath76 and @xmath77 are of the same sign ) , eq.([s1 ] ) shows that both roots are positive , and one should choose the smallest ( to have maximum @xmath5 ) .
if @xmath78 ( @xmath79 that can happen ( i ) for a sufficiently strong interband coupling for both @xmath80 and @xmath81 positive or ( ii ) if one of @xmath80,@xmath81 is repulsive ) , one should take the square root with _
minus_. it is of interest to note that even for @xmath82 , the interband coupling of either sign may lead to superconductivity .
in fact , @xmath83 for a dominant interband interaction @xmath84 , eq.([det=0 ] ) yields @xmath85 .
finally , if the interband coupling is exactly zero , a quite unlikely situation , the second form of @xmath73 in eq.([s1 ] ) gives two roots @xmath86 and @xmath87 .
the smallest one gives @xmath5 , whereas the other corresponds to the temperature at which the small gap turns zero .
this situation is depicted in fig.[fig1 ] .
we conclude this incomplete list of possibilities by noting that within this model , interband coupling enters @xmath73 only as @xmath88 , i.e. , interband attraction in clean materials affects @xmath5 exactly as does the repulsion .
this is no longer true in the presence of interband scattering , the question discussed below .
denoting the properly chosen root as @xmath89 we have : @xmath90 one easily checks that for all @xmath91 s equal this yields the standard bcs result . among various possibilities we mention here the case @xmath92 for which @xmath93
this case corresponds to a popular model with factorizable coupling potential @xmath94.@xcite this potential is amenable for analytic work , but it reduces severely the richness of the two - band scheme .
since the determinant of the system ( [ systemtc ] ) is zero , the two equations are equivalent and give near @xmath5 : @xmath95 when the right - hand side is negative , @xmath29 s are of opposite signs . within the one - band bcs ,
the sign of @xmath96 is a matter of convenience ; in fact for one band the self - consistency equation determines only @xmath97 . for two bands ,
@xmath98 and @xmath99 may have opposite signs .
if the @xmath29 values are @xmath100 and @xmath101 , eq.([self - cons1 ] ) shows that @xmath102 and @xmath103 is a solution too .
one should be aware of this multiplicity of solutions when solving the system ( [ self - cons1 ] ) numerically .
the problem is even worse because @xmath104 is always a solution . .
in this calculation @xmath105 , @xmath106 , @xmath107 , and @xmath108 , see text below.,width=321 ] , @xmath105 , @xmath106 , @xmath107 , and @xmath108 , see text below.,width=321 ] turning to evaluation of @xmath109 , we note that the sum in eq.([self - cons1 ] ) is logarithmically divergent . to deal with this difficulty
, we employ eilenberger s idea of replacing @xmath59 with the measurable @xmath5 .
these are related by eq.([tc ] ) which can be written as @xmath110 now add and subtract the last sum from one in eq.([self - cons1 ] ) : @xmath111\nonumber\\ & & = \sum _ { \mu } n_\mu \lambda_{\nu\mu } \delta_\mu \left[\sum_{\omega } ^{\infty}\left(\frac{2\pi t}{\beta_\mu}- \frac{2\pi t}{\hbar\omega}\right ) + \frac{1}{{\tilde\lambda}}- \ln\frac{t}{t_c}\right ] \nonumber\\ \label{self - cons2}\end{aligned}\ ] ] the last sum over @xmath65 is fast - converging and one can replace @xmath34 with @xmath112 .
numerically , the upper limit of summation over @xmath33 can be set as a few hundreds that suffices even for low temperatures .
introducing dimensionless quantities @xmath113 with @xmath114 , we rewrite eq.([self - cons2 ] ) : @xmath115 for given coupling constants @xmath53 and densities of states @xmath58 , this system of two equations can be solved numerically for @xmath116 and therefore provide the gaps @xmath117 .
two examples of these solutions with the sets of parameters differing only in @xmath118 are given in insets to figs.[fig1 ] and [ fig2 ] .
we observe that even a small interband coupling changes drastically the behavior of the small gap .
having formulated the way to evaluate @xmath119 , we turn to the london penetration depth given for general anisotropies of the fermi surface and of @xmath29 by ( see e.g. , ref . ) : @xmath120 where @xmath121 is the fermi velocity .
we consider here only the case of currents in the @xmath122 plain of uniaxial or cubic materials with two separate fermi surface sheets , for which a simple algebra gives for the superfluid density @xmath123 : @xmath124^{-3/2}\ , , \nonumber\\ \gamma & = & \frac{{{n_1}v_{1}^2}}{{{n_1}v_{1}^2 + { n_2}v_{2}^2}}\ , .
\label{rhogamma}\end{aligned}\ ] ] where @xmath125 are averages over corresponding band of the in - plane fermi velocities .
the formal similarity of the first line here to the widely used @xmath4-model ( with an @xmath13 in place of out @xmath3 ) prompts to name our scheme the @xmath3-model " ( we do not renormalize the gaps , so no analog of @xmath4 exists in our scheme .
we note , however , that our @xmath3 that determines partial contributions from each band is not just a partial density of states @xmath126 of the @xmath4-model , instead it involves the band s fermi velocities .
we now apply the approach developed to fit the data for the superfluid density of mgb@xmath0 crystals acquired by using the tdr technique described above .
figure [ fig3 ] shows the result of the fitting with three free parameters : @xmath76 , @xmath77 , and @xmath118 . the partial density of states and the parameter @xmath3 were taken from the the literature : the two - band mapping of the four - band mgb@xmath0 gives @xmath127 and the fermi velocities @xmath128 and @xmath129@xmath130/s@xmath131.@xcite this fit requires solving two coupled nonlinear equations , eqs .
( [ .eps ] ) .
we used matlab with the optimization toolbox and utilized a nonlinear solver using direct nelder - mead simplex search method .
@xcite the result is shown in fig.[fig3 ] with the best fit parameters listed in the caption .
we show below that the same set of parameters used to calculate the free energy and the specific heat reproduces the data on @xmath2 remarkably well .
single crystal and corresponding temperature - dependent gaps ( inset ) .
the fitting parameters are : @xmath132 , @xmath133 , @xmath134 ; the partial density of states @xmath135 and @xmath136 were fixed.,width=321 ] si single crystal and corresponding temperature - dependent gaps ( inset ) .
the fitting parameters are : @xmath137 , @xmath138 , @xmath139 , @xmath140 and @xmath141.,width=321 ] our numerical experimentation shows that if , in addition to coupling constants , the partial densities of states and the fermi velocities are unknown , the fitting procedure becomes unstable and good fits can be found for various combinations of the coupling constants . in the case of mgb@xmath142 these quanities are known and our fitting is quite certain .
for v@xmath1si we do not have detailed information regarding the band structure , partial densities of states , and fermi velocities on separate sheets of the fermi surface of this material .
hence , we took all those as free parameters in the fitting procedure .
the conclusions thus are less reliable for this material than for mgb@xmath0 : being mapped onto a two - band model , v@xmath1si comes out to have two nearly decoupled bands with an extremely weak interband coupling ( still sufficient to give a single @xmath5 ) .
the results and the best - fit parameters are given in the caption to fig.[fig4 ] .
note that the long linear tail in @xmath143 as @xmath144 approaches @xmath5 is a direct manifestation of a very small gap , in this case @xmath98 , in this temperature domain .
by fitting the data for @xmath143 , we can extract the coupling constants @xmath53 along with @xmath109 .
this allows one to determine all thermodynamic properties of the material in question , of which we consider here the specific heat @xmath2and its the jump at @xmath5 . to this end , one starts with the eilenberger expression for the energy difference : @xmath145 near @xmath5 , one obtains : @xmath146 following ref . , one can look for solutions @xmath147 of eq.([.eps ] ) near @xmath5 as an expansion : @xmath148 we substitute this in eq.([.eps ] ) and compare terms of different powers of @xmath149 ; the quantity @xmath150 . in the lowest order
we obtain the system of linear homogeneous equations for @xmath151 that coincides with the system ( [ systemtc ] ) .
the same arguments that led to eq.([sign ] ) provide @xmath152 the conditions for existence of non - trivial solutions for @xmath153 in the next oder provide the second relation for @xmath154 .
we omit a cumbersome algebra and give the result : @xmath155 and @xmath77 .
@xmath156 at @xmath5 normalized on the bcs value of 1.43 is calculated using eq.([jump1 ] ) with @xmath157 , @xmath134 and two values of @xmath126 and plotted _ vs _ @xmath81 .
, width=321 ] we now obtain the energy near @xmath5 , @xmath158 and the specific heat : @xmath159 the relative jump at @xmath5 is : @xmath160 ^ 2 .
\label{jump1}\end{aligned}\ ] ] if all coupling constants are the same , @xmath161 , @xmath162 , and @xmath163 , as is should be .
we note that the sign of the interband coupling @xmath118 has no effect on the jump @xmath164 since in eq.([jump1 ] ) @xmath165 is insensitive to this sign .
it is easy to study numerically the jump dependence on the the three coupling parameters . as an example we show in fig.[fig5 ]
the jump dependence on the mismatch between @xmath76 and @xmath77 for a fixed @xmath118 for two values of @xmath126 .
one can see that for equal relative densities of states , the jump peaks at @xmath166 at the one - band bcs value of 1.43 ; with the fixed @xmath80 and changing @xmath81 the jump drops with the mismatch @xmath167 .
the peak position and the drop speed vary with varying bands contributions , so that the value of the jump _ per se _ can not be interpreted as evidence for a particular order parameter . .
symbols are the data from ref . .
the solid line shows result of the @xmath3-model , it is calculated using the energy ( [ energy ] ) with parameters determined from the fit to the data on superfluid density.,width=321 ] now we can test our theory by employing the parameters obtained from the fit to the data on superfluid density as shown in fig .
[ fig3 ] and calculate the free energy and the specific heat .
the result is shown in fig.[fig6 ] .
the dashed line shows a single - gap weak - coupling bcs results and the solid line is the result of our calculations .
note that this is not a fit , but a calculation with parameters determined in independent measurement .
the data shown by circles are taken from ref .. since this is _ not _ a fit , the agreement with the general behavior of @xmath2 and , in particular , with value of the jump at @xmath5 is remarkable .
the intraband scattering does not affect @xmath5 , so that we focus on the effect of interband scattering with an average scattering time @xmath149 .
since @xmath168 at @xmath5 , the eilenberger equations for @xmath169 in two bands read:@xcite @xmath170 ( @xmath171 ) .
this system yields : @xmath172 where @xmath173 .
the self - consistency equation , @xmath174 again reduces to a system of linear and homogeneous equations for @xmath9 , the determinant of which must be zero . omitting the algebra ,
we give the result : @xmath175 @xmath176 is defined in eq.([det=0 ] ) .
the quantities @xmath177 are given by : @xmath178 where @xmath179 with @xmath180 being the critical temperature of the clean material given in eq.([tc ] ) .
the scattering parameter @xmath181 one can easily rearrange @xmath182 to the form : @xmath183 next , one solves the quadratic eq.([eqp ] ) for @xmath182 and chooses the smaller of two roots ( with the minus sign in front of the square root ) .
denoting this root as @xmath184 where @xmath185 stands for the set of all coupling constants and of partial densities of states , we obtain an implicit equation for @xmath186 that can be solved numerically : @xmath187 for the case @xmath161 , the only root of eq.([eqp ] ) is @xmath188 since in this particular case @xmath189 , we obtain : @xmath190 one can verify that this coincides with the suppression formula obtained within the model with factorizable coupling potential , see ref . or a more general work by openov .
@xcite with the parentheses on the right - hand side equal to 1 , this is just the abrikosov - gorkov result for the @xmath5 suppression by a pair - braker with the scattering parameter @xmath191 .
thus , only if @xmath192 , or @xmath193 @xmath5 drops to zero at a finite @xmath149 .
otherwise @xmath194 is a decreasing function at all @xmath191 .
one can show numerically that these features of the @xmath5 suppression are qualitatively the same for a general two - band case : unlike formulas of preceding sections describing clean materials , the sign of the interband coupling @xmath118 does affect the @xmath5 suppression .
one can verify that the interband scattering causes faster decrease of @xmath5 if the interband coupling is repulsive , @xmath195 .
, @xmath196 are the same for all curves .
the suppression is much stronger when the sign of @xmath197 changes to negative ( repulsive ) . for a strong repulsive @xmath197 ,
@xmath5 turns zero at a finite scattering parameter @xmath191 . ]
to illustrate this point we calculate suppression of @xmath5 with the coupling parameters @xmath198 , @xmath196 , and with three different values of @xmath199 shown in figure [ fig7 ] .
whereas with positive @xmath118 the suppression is weak and similar to the case of materials having one anisotropic gap , the suppression for @xmath195 is much stronger .
this finding can be checked experimentally and , in fact , the published data on unusually fast suppression of @xmath5 with carbon , aluminum , or lithium doping@xcite imply that mgb@xmath0 might have repulsive interband coupling . otherwise , it is hard to reconcile the @xmath5 suppression by a factor of 4 by 15% of c substitution . if one interprets this effect as caused by impurities scattering , eq.([t(rho ) ] ) with @xmath200 provides @xmath201 needed for such a suppression .
this value of the scattering parameter corresponds to unrealistically short men - free path @xmath202or less . changing the sign of @xmath203 to negative ( i.e. taking the interband coupling as repulsive ) , results in @xmath204 and a reasonable estimate of @xmath205 .
the negative interband coupling would imply opposite signs of the order parameter on the two effective bands of mgb@xmath0 , i.e. @xmath206s order parameter , a proposition calling for more studies .
we have presented a two - band weak - coupling @xmath3-model that takes into account self - consistently all relevant coupling constants to evaluate temperature dependencies of the two gaps , of the superfluid density , and of the specific heat in clean s - wave materials .
the interband coupling is shown to have a strong effect an these dependencies irrespective of the sign of this coupling .
in particular , if the interband coupling is negative ( repulsive ) it may cause the two order parameters to have opposite signs , i.e. the order parameter may have the @xmath207s structure . in this case , the @xmath5 suppression by interband scattering should be very strong , the feature that can be utilized as a signature of the @xmath207s order parameter .
we speculate that a strong @xmath5 suppression by various dopands in mgb@xmath0 may signal such a possibility .
all these features make the model advantageous to the empiric and _ not self - consistent _ @xmath4-model commonly employed to interpret the data on penetration depth and specific heat of two - gap materials .
we thank r. t. gordon and h. kim for help with the experiments , j. karpinski for mgb@xmath0 and d. k. christen for v@xmath1si single crystals , p. c. canfield , a. carrington , a. v. chubukov , s. l. budko , a. j. legett , i. i. mazin , j. schmalian , m. a. tanatar , z. tesanovic and f. catus for interest and discussions .
work at the ames laboratory is supported by the department of energy - basic energy sciences under contract no .
de - ac02 - 07ch11358 .
r.p . acknowledges support of alfred p. sloan foundation . | the two gaps in a two - band clean s - wave superconductor are evaluated self - consistently within the quasiclassical eilenberger weak - coupling formalism with two in - band and one inter - band pairing potentials . superfluid density , free energy and specific heat are given in the form amenable for fitting the experimental data .
well - known two - band mgb@xmath0 and v@xmath1si superconductors are used to test the developed approach .
the pairing potentials obtained from the fit of the superfluid density data in mgb@xmath0 crystal were used to calculate temperature - dependent specific heat , @xmath2 .
the calculated @xmath2 compares very well with the experimental data . advantages and validity of this ,
what we call the @xmath3-model " , are discussed and compared with the commonly used empirical ( and _ not self - consistent _ ) @xmath4-model " .
correlation between the sign of the inter - band coupling and the signs of the two order parameters is discussed .
suppression of the critical temperature by the inter - band scattering is evaluated and shown to be severe for the inter - band repulsion as compared to the attraction .
the data on a strong @xmath5 suppression in mgb@xmath0 crystals by impurities suggest that the order parameters on two effective bands of this material may have opposite signs , i.e. , may have the @xmath6 structure similar to the current proposals in iron - based pnictide superconductors . |
we apply the linear response theory @xcite to the electric susceptibility of eit . for the perturbation hamiltonian @xmath132 ,
the linear response of any operator @xmath133 under the ground state @xmath49 of the unperturbed hamiltonian @xmath134 is ( @xmath135 ) @xmath136 \right\vert \psi_{g}\right\rangle_{h},\label{linear}\end{aligned}\ ] ] where @xmath137 is the operator defined in the heisenberg picture ( with subscript @xmath138 ) , and the same as for @xmath139 . to calculate the electric susceptibility of eit
, we consider the variation of the polarization operator , @xmath140 , \end{aligned}\ ] ] and put the above into eq .
( [ linear ] ) , we have @xmath141\right\vert \psi_{g}\right\rangle_h , \nonumber\\ & = -i\int_{-\infty}^{\infty}dt^{\prime}\theta(t - t^{\prime})\frac{d_{0}}{v}\sum_{{\mathbf{k}},{\mathbf{k}}',{\mathbf{q}}^{\prime}}\big\{\bar{\omega}_{1,{\mathbf{k}}}(t^{\prime}){\big.}_h\left\langle \psi_{g}\left\vert\left[e^{i\hat{h}t^{\prime}}\hat{e}_{{\mathbf{k}}+{\mathbf{k}}'}^{\dag}(t^{\prime})\hat{g}_{{\mathbf{k}}'}(t^{\prime})e^{-i\hat{h}t^{\prime}},e^{i\hat{h}t}\hat{g}_{{\mathbf{q}}^{\prime}}^{\dag}(t)\hat{e}_{q+{\mathbf{q}}^{\prime}}(t)e^{-i\hat{h}t}\right ] \right\vert \psi_{g}\right\rangle_h \nonumber\\ & + \bar{\omega}_{1,{\mathbf{k}}}^{\ast}(t^{\prime}){\big.}_h\left\langle \psi_{g}\left\vert\left [ e^{i\hat{h}t^{\prime}}\hat{g}_{{\mathbf{k}}'}^{\dag}(t^{\prime})\hat{e } _ { { \mathbf{k}}+{\mathbf{k}}'}(t^{\prime})e^{-i\hat{h}t^{\prime}},e^{i\hat{h}t}\hat{e}_{-{\mathbf{q}}+{\mathbf{k}}^{\prime}}^{\dag}(t)\hat{g}_{{\mathbf{q}}^{\prime}}(t)e^{-i\hat{h}t}\right ] \right\vert \psi_{g}\right\rangle_h\big\},\end{aligned}\ ] ] where in the last line we have neglected the terms of @xmath142 and @xmath143 , which are vanishing in the expectation value of ground state wavefunction , @xmath49 . for the unperturbed hamiltonian ( or probe - free hamiltonian , @xmath144 ) ,
all atoms are in the atomic ground state @xmath2 , and states @xmath4 and @xmath3 are empty .
therefore , further expansion of the above equation gives @xmath145 .\end{aligned}\ ] ] since the excited state @xmath4 and the second ground state @xmath3 are coupled by the control field , we then replace @xmath146 with the new eigen - bases @xmath147 and @xmath148 with corresponding eigenenergy @xmath149 respectively .
here @xmath150 @xmath17 @xmath151 @xmath118 @xmath152 and @xmath153 @xmath17 @xmath154 @xmath155 @xmath156 , where @xmath157@xmath17@xmath158/[\epsilon_+({\mathbf{k}})-\epsilon_-({\mathbf{k}})]}$ ] .
the associated eigenvalues are : @xmath149@xmath17@xmath159/2 $ ] , where @xmath160 , @xmath161 , and @xmath162 as the recoil momentum
. a phenomenonological spontaneous decay rate ( @xmath7 ) of the excited state can be added by replacing @xmath163 with @xmath164 . after shifting the momentum , @xmath165 and @xmath166 , we have @xmath167 \nonumber\\ & \left[\cos\phi_{{\mathbf{k}}+{\mathbf{k}}'}\hat{a}_{{\mathbf{k}}+{\mathbf{k}}'}^{\dag}(t^{\prime})-\sin\phi_{{\mathbf{k}}+{\mathbf{k}}'}\hat{b}_{{\mathbf{k}}+{\mathbf{k}}'}^{\dag}(t^{\prime})\right ] \hat{g}_{{\mathbf{k}}'}(t^{\prime})\left\vert\psi_{g}\right\rangle _ { h } -h.c.({\mathbf{q}}\rightarrow-{\mathbf{q}})\big\},\nonumber\\ & = \frac{id_{0}}{v}\int_{-\infty}^{\infty}dt^{\prime}\theta(t - t^{\prime})\sum_{{\mathbf{k}}}\big\{\bar{\omega}_{1,{\mathbf{q}}}(t^{\prime})f_{{\mathbf{k}}}(t - t^{\prime } ) \left[\cos^{2}\phi_{{\mathbf{k}}+{\mathbf{q}}}e^{-i\epsilon_-({\mathbf{k}}+{\mathbf{q}})(t - t^{\prime})}+\sin^{2}\phi_{{\mathbf{k}}+{\mathbf{q}}}e^{-i\epsilon_+({\mathbf{k}}+{\mathbf{q}})(t - t^{\prime})}\right ] \nonumber\\ & -h.c.({\mathbf{q}}\rightarrow-{\mathbf{q}})\big\},\end{aligned}\ ] ] where we have defined the ground state correlation function as @xmath168@xmath17@xmath169 .
use the fourier transform @xmath170@xmath17@xmath171 , and define the electric susceptibility as @xmath172 where @xmath173 is the fourier transform of @xmath174 , we have ( after the change of variables , @xmath175 , @xmath176 ) @xmath177.\end{aligned}\ ] ] finally , since the correlation function @xmath178 is related to the time - ordered retarded green s function via @xmath179 , where @xmath180 , we can rewrite above results to be @xmath181,\label{chi1}\end{aligned}\ ] ] where @xmath51 is the fourier transforms of the green s function @xmath182 .
the eigenvalues @xmath149 and @xmath183 have been defined before . as an example we consider a weakly interacting condensate . by separating the fluctuation from the condensate : @xmath184 @xmath17 @xmath185 @xmath118 @xmath186 with @xmath187 being the condensate particle number , we apply bogoliubov transformation to calculate the single particle green s function .
the final eit spectrum can be separated into the condensate ( c ) and the non - condensate ( nc ) parts : @xmath188 , where @xmath189 $ ] , and @xmath190 $ ] . here
@xmath191 is condensate density , @xmath192 is the phonon excitation energy , and @xmath193 , where @xmath194 , @xmath195 , and @xmath196 .
note that the non - condensate part is contributed by all momentum channels due to the quantum depletion , consistent with results from the dark state approach @xcite .
it is instructive to simplify eq .
( [ chi1 ] ) further in the strong control field limit , i.e. @xmath10 is much larger than atomic kinetic and interaction energies .
it is straightforward to have following leading order results : @xmath197 where @xmath198 , and @xmath199 .
note that @xmath200 obtained in this limit has no momentum ( @xmath201 ) dependence .
here we demonstrate how to derive the single particle green s function @xcite for a luttinger liquid ( ll ) .
first , we use the density - phase representation of low - energy bosonic field operators , @xmath202 , for the luttinger liquid @xcite , where @xmath203 and @xmath204 are density and phase fluctuation operators .
the single particle green s function becomes @xmath205 where we have used the fact that the density fluctuation is suppressed and negligible for a repulsively interacting 1d gas @xcite , and therefore only phase fluctuation is kept for the low energy effective theory .
@xmath206 is a heaviside step function .
the above equation can be simplified by using @xmath207 /2}$ ] if @xmath208 and @xmath209 are linear combination of bosonic operators and @xmath210 $ ] is a complex number .
we can also apply @xmath211 for the expectation on a bilinear hamiltonian of @xmath208 . in the low energy limit
, it has been shown that the effective hamiltonian of a 1d bosonic gas can be described by a ll model , where the phase operator @xmath204 can be calculated within the periodic boundary condition @xcite , @xmath212 , \label{phi_b}\end{aligned}\ ] ] where @xmath213 is the bosonic eigenstate field operator of the effective ll model .
a positive length scale , @xmath74 , is introduced as a cutoff length scale for the convergence of integrals , and @xmath214 is the system size . here
the ll parameter is denoted as @xmath54 , and the dispersion is @xmath215 with @xmath55 being the phonon velocity .
note that the density - phase representation in a ll is valid in the long wavelength limit @xmath216 .
the exact values of @xmath54 and @xmath55 should be determined by a more microscopic calculation or from the experimental measurement . as a result
, the single particle green s function of the ll model can be calculated to be @xmath217 /2}n\exp\left\ { -\frac{1}{2}\left\langle \hat{t}\left [ \hat{\phi}(x , t)-\hat{\phi}(0)\right ] ^{2}\theta(t)\right\rangle \right\ } , \nonumber\\ & \simeq e^{\left [ \hat{\phi}(x , t),\hat{\phi}(0)\right ] /2}n\exp\left\ { -\frac{1}{4\kappa}\ln\frac{x^2+(a+ivt)^2}{a^2}\right\}\theta(t ) , \label{g_ll1}\end{aligned}\ ] ] where @xmath218 is the time - ordered operator . this correlation function from ll model describes the behavior of quasi - long range order that has an exponent proportional to the interaction strength .
the logarithmic function inside the exponential function is derived from ( use @xmath219 and eq .
( [ phi_b ] ) ) @xmath220 ^{2}\right\rangle \nonumber\\&=\frac{\theta(t)}{4}\sum_{q\neq0}\left\vert \frac{2\pi}{ql\kappa}\right\vert e^{-a\left\vert q\right\vert}\left[2 - 2e^{i(qx-|q|vt)}\right],\nonumber\\ & = \frac{\theta(t)}{2\kappa}\int_{0}^{\infty}dq\frac{e^{-aq}}{q}\left [ 2-e^{iq(x - vt)}-e^{-iq(x+vt)}\right],\nonumber\\ & = \frac{\theta(t)}{2\kappa}\ln\frac{x^2+(a+ivt)^2}{a^2},\end{aligned}\ ] ] which can be also found in ref .
the commutation relation in the prefactor of the green s function in eq .
( [ g_ll1 ] ) can be also calculated to be @xmath221 & \nonumber\\ & = \frac{1}{4k}\int_{-\infty}^{\infty}\frac{e^{-a\left\vert q\right\vert } } { \left\vert q\right\vert } \left [ e^{i(qx-\left\vert q\right\vert vt)}-e^{-i(qx-\left\vert q\right\vert vt)}\right ] , \nonumber\\ & = \frac{1}{4k}\log\frac{\left [ a+i\left ( x - vt\right ) \right ] \left[a - i\left ( x+vt\right ) \right ] } { \left [ a - i\left ( x - vt\right ) \right ] \left [ a+i\left ( x+vt\right)\right]}=0,\end{aligned}\ ] ] where in the last line we have taken the limit @xmath222 . finally , the full dynamical correlation function of a ll model ( which is in general valid only in the low - energy and long wavelength limit of a 1d bose gas ) is then derived as @xmath223^{1/(4\kappa)}}. \label{green}\end{aligned}\ ] ]
here we proceed to calculate the electric susceptibility ( i.e. eit spectrum ) of a ll .
it is more instructive to consider the case of large control field limit ( i.e. eq . ( [ chi2 ] ) ) so that by inserting the result of eq .
( [ green ] ) , we have @xmath224^{1/(4\kappa)}}\nonumber\\ & \times\left [ \cos^{2}\phi e^{-i(\omega+\epsilon_-)t}+\sin^{2}\phi e^{-i(\omega+\epsilon_+)t}\right],\nonumber\\ & = id_{0}n\int_{0}^{\infty}dt\frac{a^{1/(2\kappa)}}{(a+ivt)^{1/(2\kappa)}}\left [ \cos^{2}\phi e^{-i(\omega+\epsilon_- ) t}\right .
\nonumber\\&\left.+\sin^{2}\phi e^{-i(\omega+\epsilon_+ ) t}\right],\nonumber\\ & = d_0n\big(\frac{a}{v}\big)^{1/(2\kappa)}\gamma(1-\frac{1}{2\kappa})\times\nonumber\\ & \bigg[\frac{\cos^2\phi } { ( \omega+\epsilon_{-})^{1 - 1/(2\kappa)}}+\frac{\sin^2\phi } { ( \omega+\epsilon_{+})^{1 - 1/(2\kappa)}}\bigg],\label{ll_analytic}\end{aligned}\ ] ] where in the last line we have used the integral property @xcite , @xmath225 under the conditions that re@xmath226>0 $ ] , im@xmath227<0 $ ] , and re@xmath228\geq 0 $ ] .
@xmath229 is the incomplete gamma function which becomes the gamma function @xmath230 when @xmath231 .
( [ ll_analytic ] ) indicates a nontrivial power law dependence in the eit spectrum , as shown in fig . 1 from strong to weak interacting regimes ( @xmath232)@xcite .
the standard ( noninteracting ) eit spectrum is similar to the weak interacting one ( fig .
1(d ) ) where the zero of the dispersion relation coincides with the transparency point ( @xmath233 ) , and symmetric ( anti - symmetric ) absorption ( dispersion ) profile is retrieved around @xmath233 . when the interaction becomes stronger ( smaller @xmath54 ) , the zero of dispersion relation moves away from the transparency point , and
the eit is destroyed for even stronger interaction .
interestingly , the eit profile for @xmath69 case ( hard core boson limit ) shows an inversion symmetry between absorption and dispersion profiles around @xmath233 , very different from the standard eit spectrum in the noninteracting limit ( @xmath234 ) .
rb atoms in the large control field limit from eq .
( [ ll_analytic ] ) .
we take a static probe field ( @xmath63 ) and a resonant control field ( @xmath64 ) with rabi frequency @xmath65 .
the excited state is chosen as low - lying rydberg transition of @xmath66 with @xmath67 .
the absorption ( re[@xmath68 , solid - blue ) and dispersion ( im[@xmath68 , dash - red ) profiles for ( a ) @xmath235 , ( b ) @xmath117 , ( c ) @xmath236 , and ( d ) @xmath70 . the horizontal line guides the eye to the zero.,width=321,height=170 ]
here we show how to derive the single particle green s function for a mott state . at zero temperature , and in the deep mott - insulating limit of average @xmath81 particle per site
, we may assume only small number fluctuations about @xmath81 so that the hilbert space can be truncated to three particle numbers per site only : @xmath237 , @xmath81 , and @xmath238 .
within such three - state model @xcite , the original bosonic field operator at site @xmath79 can be re - expressed to be : @xmath239 , where @xmath240 and @xmath241 are the raising and lowering operators .
the conservation of total number of particles provides an additional constraint on the hilbert space : @xmath242 .
since we are interested in the deep mott insulator regime with little number fluctuation , we can apply the conservation of particle number shown above to have the following approximation @xcite : @xmath243 . as a result , the single particle correlation function can then be easily expressed to the quadratic order of small fluctuations to be @xmath244 , \nonumber\\ \simeq & ( n_{0}+1)\langle \hat{t}_{1,\mathbf{r}}^{\dag}(t)\hat{t}_{1,\mathbf{r}^{\prime}}(0)\rangle + n_{0}\langle \hat{t}_{-1,\mathbf{r}}(t)\hat{t}_{-1,\mathbf{r}^{\prime}}^{\dag}(0)\rangle + \sqrt{n_{0}(n_{0}+1 ) } \nonumber\\ & \times\left[\langle \hat{t}_{1,\mathbf{r}}^{\dag}(t)\hat{t}_{-1,\mathbf{r}^{\prime}}^{\dag}(0)\rangle + \langle \hat{t}_{-1,\mathbf{r}}(t)\hat{t}_{1,\mathbf{r}^{\prime}}(0)\rangle \right ] .
\label{g_mott}\end{aligned}\ ] ] within the mi state , it has been shown that we can diagonalize the effective hamiltonian @xcite in the three state model , and change the lowering and raising operators by the quasi - hole and quasi - particle excitations : @xmath245 and @xmath246 , i.e. @xmath247 , and @xmath248 , where @xmath249 the corresponding eigenenergies of the particle and hole excitations are @xmath250+\tilde{\omega}({\mathbf{k}})$ ] , where @xmath251 @xmath17 @xmath252 , and @xmath253 for a 3d cubic lattice @xcite . as a result
, we can easily calculate a correlation function as following : @xmath254 similarly the single particle green s function in eq .
( [ g_mott ] ) becomes @xmath255^{2 } \nonumber\\&\times e^{-i{\mathbf{k}}\cdot(\mathbf{r}-\mathbf{r}^{\prime})}e^{-i\epsilon_{h}({\mathbf{k}})t}.\end{aligned}\ ] ] in the deep mott regime , we have @xmath256 and hence @xmath257 , and @xmath258 .
the single particle green s function for the mott state could be further simplified to be ( use @xmath259 as the fourier transform of @xmath260 ) @xmath261 note that only hole excitation energy appears because the green s function we need for eit spectrum is time - ordered , i.e. an atom is excited from the ground state ( @xmath2 ) to the excited state ( @xmath4 ) by the probe field , leaving a hole excitation inside the strongly interacting system .
here we show how to derive the single particle green s function for a bcs superfluid state of two - component fermi gases .
the single particle green s function in this context is defined as @xmath262 where we define the original ground state ( in the eit @xmath9 scheme ) to be spin up .
the other state of spin down is assumed not directly involved in the eit experiment . to evaluate the above expectation
, we can express the ground state in terms of bogoliubov quasi - particles ( denoted by @xmath263 and @xmath264 ) of cooper pairs in the superconducting state , @xmath265 where @xmath266 .
here @xmath267 is the excitation energy of the quasi - particles with @xmath104 .
@xmath24 is the chemical potential and is determined by the particle density .
the green s function can then be obtained to be @xmath268 99 a. l. fetter and j. d. walecka , _ quantum theory of many - particle systems _
( dover publications , 2003 ) .
h. h. jen , bo xiong , ite a. yu , and daw - wei wang , arxiv:1208.0118 ( 2012 ) .
m. a. cazalilla , j. phys .
b : at . mol .
37 * , s1 ( 2004 ) .
thierry giamarchi , _ quantum physics in one dimension _ ( oxford university press , 2003 )
. i. s. gradshteyn , and i. m. ryzhik , _ table of integrals , series , and products , seventh edition _
( elsevier inc . , 2007 ) .
note that @xmath54 can be smaller than one when considering a long - ranged repulsive interaction .
wang , phys .
a * 80 * , 063620 ( 2009 ) .
e. altman and a. auerbach , phys .
* 89 * , 25040 ( 2002 ) . | we develop a general theory to study the electromagnetically induced transparency ( eit ) in ultracold quantum gases , applicable for both bose and fermi gases with arbitrary inter - particle interaction strength .
we show that , in the weak probe field limit , the eit spectrum is solely determined by the single particle green s function of the ground state atoms , and reflects interesting quantum many - body effects when atoms are virtually coupled to the low - lying rydberg states . as an example
, we apply our theory to 1d luttinger liquid , bose - mott insulator state , and the superfluid state of two - component fermi gases , and show how the many - body features can be observed non - destructively in the unconventional eit spectrum . in the last decade ,
systems of ultracold atoms open up great opportunities to study many - body physics , which may be even not existing in traditional condensed matter systems @xcite .
huge advances in experimental techniques of tuning interaction strengths and generating optical lattice enable studying strongly correlated atomic gases with high controllability .
most of the experimental measurements rely on the interference pattern of matter wave in different circumstances , for example the time - of - flight experiment , the noise correlations measurement @xcite , and the bragg scattering spectroscopy @xcite , etc .
in - situ imaging is also shown important in characterizing the equation of the state and the critical properties near the phase transition point @xcite .
besides of the many - body problem , the quantum control and manipulation of light - atom interaction is also an extensively studied field in cold atoms . one of the most important examples is the electromagnetically induced transparency ( eit ) , which has led to several interesting subjects such as the dark - state polariton , slow light , and induced photon - photon interaction in a cold atomic gas @xcite .
eit experiments in quantum degenerate gases @xcite or in rydberg states @xcite are also recently explored by several groups .
however , so far the eit theory is still mainly based on the single - particle picture ( or within a mean field approximation ) , and therefore can not clarify the non - trivial coupling between light - atom interaction and the many - body physics .
for example , how is the eit spectrum modified when the ground state is a strongly correlated state without single particle excitations ? in what conditions can a non - trivial many - body effect be measured experimentally ?
resolving these interdisciplinary problems is not only an interesting theoretical subject itself , but may be also applied to experimentally detect some many - body properties via eit spectroscopy .
-type scheme : the control ( @xmath0 ) and the probe ( @xmath1 ) fields couple two hyperfine ground states @xmath2 and @xmath3 with the excited state @xmath4 ( detunings are @xmath5 and @xmath6 respectively ) .
@xmath7 is the spontaneous decay rate of @xmath4 .
filled circles denote atoms with mutual interaction @xmath8 in the ground state .
( b ) experimental setup for eit on a 2d array of 1d luttinger liquids in the counter - propagating scheme.,width=283,height=151 ] in this paper , we develop a general eit theory in a strongly interacting quantum gas , taking into account the full kinetics and inter - atom interaction , within the weak probe field limit .
we explicitly show how the experimental eit spectrum is directly related to the dynamical green s function , and the quantum many - body phenomena is manifested most when atoms are coupled to a low - lying rydberg state due to the sufficient recoil energy compared to the decay rate of the excited state @xcite . as an example
, we apply our theory to three important strongly correlated systems : luttinger liquid of 1d bose gases , bose - mott insulator in an optical lattice , and the superfluid state of two - component fermi gases . for luttinger liquid , we demonstrate a power law dependence of the eit spectrum near the resonance , while the significant frequency shift and the asymmetric absorption spectrum can be identified for the other cases . our results for the first time
demonstrate the strongly correlated effects in the eit experiment , suggesting a non - destructive measurement on the quantum gases .
we consider the conventional eit setup ( @xmath9 type scheme ) as shown in fig .
[ eitfig](a ) : atoms are initially prepared in the ground state @xmath2 . the probe ( @xmath1 ) and control ( @xmath10 ) fields couple the ground state to another low energy state @xmath3 and an excited state @xmath4 with detunings @xmath11 and @xmath12 respectively where we set @xmath13 . here
@xmath14 are central frequencies of probe and control fields , while @xmath15 denote the energies of atoms in different internal states . after using the dipole approximation and rotating wave approximation ,
the total hamiltonian ( @xmath16@xmath17@xmath18 ) in the rotating frame becomes : @xmath19-\sum_{\mathbf{k}}(\omega_2\hat{e}_{{\mathbf{k}}+{\mathbf{k}}_1}^\dagger\hat{s}_{{\mathbf{k}}+{\mathbf{k}}_1-{\mathbf{k}}_2}+h.c.),\nonumber\\ & \hat{h}_u=\frac{1}{2v}\sum_{c , d=\{g , s , e\}}\sum_{{\mathbf{k}},{\mathbf{k}}',{\mathbf{q}}}u_{cd}\,\hat{c}_{{\mathbf{k}}+{\mathbf{q}}}^\dagger\hat{d}_{{\mathbf{k}}'-{\mathbf{q}}}^\dagger\hat{d}_{{\mathbf{k}}'}\hat{c}_{\mathbf{k}},\nonumber\\ & \hat{h}_1(t)=-\frac{1}{v}\sum_{{\mathbf{k}},{\mathbf{q}}}\bar{\omega}_{1,{\mathbf{k}}}(t)\hat{e}_{{\mathbf{k}}+{\mathbf{q}}}^\dagger\hat{g}_{\mathbf{q}}+h.c.\end{aligned}\ ] ] here @xmath20 and/or @xmath21 denote the atomic field operators in the momentum space .
@xmath22 includes all the single particle part without the probe field for atomic mass @xmath23 , chemical potential @xmath24 , and the rabi frequency of plane wave control field @xmath10 .
all the inter - atom interactions are included in @xmath25 with the short - range interaction @xmath26 ( extension to a certain long - ranged interaction is straightforward ) .
finally , @xmath27 denotes the effect of the probe field with @xmath28 being the slow - varying rabi frequency of wavevector @xmath29 .
@xmath30 is the quantization volume and @xmath31 are momenta of the probe and control fields respectively . to derive the electric susceptibility in the linear response of probe field @xcite
, we treat @xmath32 as a time - dependent perturbation and keep the full interaction effects .
therefore , @xmath33 can be separated into two decoupled parts : @xmath34 , where @xmath35 includes all the ground state kinetic energy and interaction energy terms , while @xmath36 includes all the single particle terms of states @xmath3 and @xmath4 ( see supplemental material ( sm ) @xcite ) . in the leading order limit of a weak probe ,
all atoms are in the ground state ( @xmath2 ) and none in states @xmath4 and @xmath3 . as a result ,
the interaction between @xmath2 and the other two states are nothing but a background energy shift in the single particle energy , i.e. @xmath37 , where @xmath38 is the total particle density .
the mutual interaction between @xmath3 and @xmath4 are the second order effect and hence negligible .
similar treatment can be also easily applied to long - ranged interaction if rydberg states of high principal quantum number are involved @xcite
. a phenomenonological spontaneous decay rate ( @xmath7 ) of the excited state can be added , and we also assume negligible dephasing rate between the two hyperfine ground states . note that when considering standard d2 transition , @xmath39 mhz , is much larger than the average atomic kinetic energy , and therefore the quantum many - body effect may not be easily observed . defining the polarization operator , @xmath40 @xmath17 @xmath41 $ ] , where @xmath42 is the field operator , and @xmath43 is the dipole moment , we calculate its variation with respect to the probe field ( @xmath32 ) via the linear response theory @xcite : @xmath44 @xmath17 @xmath45 @xmath46|\psi_g\rangle_h$ ] , where @xmath47 and @xmath48 are corresponding operators in heisenberg picture , and @xmath49 is the ground state of the unperturbed hamiltonian @xmath34 . in the momentum - frequency space , the electric susceptibility can be obtained to be ( see sm @xcite ) : @xmath50.\label{chi}\end{aligned}\ ] ] here @xmath51 is the fourier transform of the green s function at zero temperature @xcite : @xmath52 , where the upper ( lower ) sign refers to bosons ( fermions ) .
we note that the above result is very general in the limit of weak probe field , and all the many - body effects are hidden in @xmath51 .
however , as pointed out in ref . @xcite , such many - body effect can be identified more easily in the counter - propagating ( rather than the co - propagating ) setup of the eit experiment , since the recoil momentum of the former is manifested more significantly in the quantum statistics . in the rest of this paper , therefore , we just consider the counter - propagating scheme in order to measure the many - body effect from the eit spectrum .
now we study eit in a luttinger liquid ( ll ) as shown in fig .
[ eitfig](b ) , which is a very general 1d effective model @xcite and has no condensate even at zero temperature . since all the elementary excitations of a ll are collective , it is therefore interesting to investigate how the probe field propagates inside such strongly correlated system .
the single particle green s function can be exactly calculated by bosonization method @xcite ( see sm @xcite ) : @xmath53^{1/4\kappa } } , \label{g_ll}\end{aligned}\ ] ] where @xmath54 is the luttinger parameter , @xmath55 is the phonon velocity , and @xmath56 is the system - dependent momentum cut - off .
@xmath57 for a short - ranged repulsive interaction , while @xmath54 can be smaller than one if the interaction is long - ranged . in figs .
[ ll](a ) and ( b ) , we show the full numerically integrated eit absorption and dispersion profiles ( from eq .
( [ chi ] ) and ( [ g_ll ] ) ) of a 1d @xmath58rb atoms ( see the caption for parameters ) in the counter - propagating scheme .
it is easy to see that the power law dependence of @xmath59 makes the dispersion highly asymmetric about the resonance points and the absorption depth ( @xmath60 ) becomes larger at the transparency point ( @xmath61 ) in the strong interacting regime ( smaller @xmath54 ) , in contrast to the standard eit profiles in the weak interacting limit ( @xmath62 ) .
this is because when a ground state atom is excited by the probe field , the transition matrix elements are composite of all the collective excitations of different energies , weighted by the density of states in a power - law distribution . in fig .
[ ll](c ) , we show the group velocity of light ( proportional to the slope of dispersion at the transparent point , where the absorption is the smallest ) as a function of @xmath54 .
one can see that for a stronger interaction ( smaller @xmath54 ) , the group velocity becomes smaller due to steeper dispersion relations , but the absorption is also larger , making the slow light propagate less effectively in a ll .
on the other hand , such unconventional eit spectrum also implies a sensitive measurement about the many - body properties .
rb atoms in the counter - propagating excitation scheme .
we take a static probe field ( @xmath63 ) and a resonant control field ( @xmath64 ) with rabi frequency @xmath65 .
the excited state is chosen as low - lying rydberg transition of @xmath66 with the spontaneous decay rate , @xmath67 .
the effective 1d interaction strength can be tuned by the confinement resonance of 1d tubes @xcite .
( a ) absorption ( re[@xmath68 ) and ( b ) dispersion ( im[@xmath68 ) profiles for @xmath69 ( solid ) and @xmath70 ( dash ) .
the horizontal line guides the eye to the zero in ( b ) .
( c ) shows the group velocity @xmath71 ( @xmath72 ) , and the absorption depth @xmath60 ( @xmath73 ) at the transparency point ( @xmath61 ) as a function of @xmath54 .
for convenience , we take the phonon velocity @xmath55=4.3 mm / s and the cut - off @xmath74=0.12 @xmath24 m within a typical experimental regime @xcite . @xmath75 is the speed of light.,width=321,height=170 ] now we consider the mott - insulator ( mi ) of strongly interacting bosons in a 3d optical lattice , which can be well described by a single band hubbard model ( hm ) @xcite .
the ground state field operator is expressed as @xmath76 where @xmath77 is the wannier function , and @xmath78 is the field operator at site @xmath79 . when deep inside the mi state , we can use the three - state model @xcite to control the small number fluctuation above the fock state , and obtain the green s function at zero temperature ( see sm @xcite ) : @xmath80 , where @xmath81 is the integer filling fraction , @xmath82 is the fourier transform of @xmath77 , and @xmath83 @xmath17 @xmath84 is the hole excitation with @xmath85 when deep inside the mott state .
here @xmath86 , and @xmath87 for a 3d square lattice of lattice constant @xmath88 . @xmath89 and @xmath8 are the tunnelling amplitude and on - site interaction . as a result , the susceptibility from eq .
( [ chi ] ) becomes @xmath90 .
\label{motteq}\end{aligned}\ ] ] rb atoms loaded in a 3d optical lattice ( lattice constant @xmath91 nm ) .
the corresponding parameters ( @xmath8 , @xmath89 ) in unit of the recoil energy @xmath92 are ( 0.5 , 0.007 ) and ( 1.05 , @xmath93 ) for @xmath94 ( solid - blue and dash - red ) and @xmath95 ( @xmath73 and @xmath72 ) respectively .
absorption ( re[@xmath68,solid - blue , @xmath73 ) and dispersion ( im[@xmath68,dash - red , @xmath72 ) profiles are plotted for various control field strengths ( a ) @xmath96 , and ( b ) @xmath97 .
all other parameters are the same as fig .
2.,width=340,height=151 ] compared to the standard eit spectrum , we find that @xmath98 is contributed by eit in different momentum states , weighting by the wannier function of the optical lattice .
when the momentum distribution is large in a strong lattice strength , the contribution from different momenta makes the eit spectrum much broadened , in contrast to the superfluid state which is similar to the standard eit spectrum in a bose - einstein condensate ( bec ) @xcite . in fig .
[ mottfig ] , we demonstrate the eit spectrum inside the mi state of @xmath58rb with unit filling for different control fields and lattice strengths . as discussed above , the momentum distribution of wannier function lead to an asymmetric and inhomogeneous broadening of the absorption profiles near the resonance .
it makes the eit window less transparent when the control field is weak ( fig .
[ mottfig](a ) ) , while it becomes less effective in the large field limit . in other words , by measuring the absorption depth one may also determine the momentum distribution of the underlying many - body system .
finally we investigate the eit in a bcs superfluid of two - component fermi gases , using similar setup @xcite by coupling one of the atomic ground state to a low - lying rydberg state @xcite . without losing generality , here
we just presume the existence of a superconducting gap , @xmath99 , not specifying the pairing mechanism .
one of the two components ( pseudo - spin up ) is our ground state ( @xmath2 ) in the eit experiment , and the other is not involved in the eit process .
the green s function can be calculated by transforming toward the bogoliubov quasi - particles ( @xmath60 and @xmath100 ) : @xmath101 , where @xmath102 . here
@xmath103 is the excitation energy with @xmath104 and @xmath24 being the chemical potential .
we find that @xmath105 ( see sm @xcite ) , and then @xmath106 note that in the limit of zero pairing gap , @xmath107 , @xmath108 becomes a step function at fermi energy , and hence @xmath109 becomes the result of a noninteracting fermi gas @xcite .
k ) .
the excited state is chosen as low - lying rydberg transition of @xmath110 with @xmath111 .
we fix the chemical potential @xmath112 @xmath113 with atomic density @xmath114 @xmath115 and resonant control field ( @xmath64 ) .
absorption ( re[@xmath68 , solid - blue , @xmath73 ) and dispersion ( im[@xmath68 , dash - red , @xmath72 ) profiles are plotted for pairing gap @xmath116 ( solid - blue and dash - red ) and @xmath117 ( @xmath73 and @xmath72 ) , while the control field strengths are ( a ) @xmath96 and ( b ) @xmath70 @xmath7 .
( c ) shows the group velocity @xmath71 ( @xmath118 ) and the absorption depth @xmath60 ( @xmath119 ) at the transparency point as a function of @xmath120 .
( d ) shows the transparency positions @xmath61 ( @xmath121).,width=321,height=177 ] in figs .
[ scfig](a ) and ( b ) , we plot the eit profiles of the bcs superfluid phase for @xmath122k in the counter - propagating excitation schemes .
different control field strengths and different gap amplitudes are shown together for comparison .
one can see that when the control field is turned smaller , the recoil energy becomes more important so that the presence of fermi sea makes the eit spectrum disappeared , similar to the noninteracting fermions @xcite .
when the control field becomes stronger , the quantum degeneracy becomes less important so that the standard eit features appear , while an asymmetric broadening appears when the gap is opened .
interestingly , we find the group velocity is larger for a larger @xmath99 ( see fig .
4(c ) ) , different from results of a ll . in ( d )
we show that the transparency points @xmath61 scales as the gap energy , and therefore can be used to measure the pairing gap .
the difference between eit approach and standard bragg spectroscopy ( bs ) has two folds .
firstly eit is nondestructive while bs operates with interference of atoms .
secondly bs measures the density - density correlation function , in contrast to the first order correlation in eit . for experimental realization
, it may involve the integration time and precision of the spectrum measurement .
with khz resolution required in our scheme , we need ms integration time compatible with the atomic lifetime of several ms , which can be fulfilled by minimizing the inelastic collisions of bec @xcite
. the contrast of susceptibility can be also enhanced by increasing the optical depth of atoms . including the thermal excitations is crucial for practical implementations and will be discussed in the future work . in summary
, we develop a general and analytic theory to investigate the eit spectrum of quantum degenerate ultracold atoms .
the spectrum is solely determined by the single particle green s function , and has most significant quantum many - body physics when the atoms are coupled to a low - lying rydberg state . in the regime of strong interaction , the absorption profile has a highly asymmetric inhomogeneous broadening with a frequency shift in detuning .
our results suggest a new non - destructive method to investigate the strongly correlated physics from the well - developed eit experiment .
we appreciate fruitful discussions with ite a. yu , y .- c .
chen , m. fleischhauer , and g. juzelinas .
this work is supported by nsc and ncts grants in taiwan . 99 for a review ,
see for example : i. boch , j. dalibard , and w. zwerger , rev .
mod .
phys . *
80 * , 885 ( 2008 ) .
e. altman , e. demler , and m. d. lukin , phys .
rev .
a * 70 * , 013603 ( 2004 ) .
m. kozuma , _ et al .
, _ phys .
rev .
lett . * 82 * , 871 ( 1999 ) ; j. stenger , _ et al .
, ibid .
_ * 82 * , 4569 ( 1999 ) ; s. b. papp , _ et al . ,
ibid . _ * 101 * , 135301 ( 2008 ) ; g. veeravalli , e. kuhnle , p. dyke , and c. j. vale , _ ibid .
_ * 101 * , 250403 ( 2008 ) ; j. m. pino , r. j. wild , p. makotyn , d. s. jin , and e. a. cornell , phys . rev .
a * 83 * , 033615 ( 2011 ) .
n. gemelke , x. zhang , c - l . hung , and c. chin , nature * 460 * , 995 ( 2009 ) .
t .-
l . ho and q. zhou , nature physics * 6 * , 131 ( 2010 ) .
m. d. lukin , , 457 ( 2003 ) .
m. fleischhauer , a. imamoglu , and j. p. marangos , , 633 ( 2005 ) . l. v. hau , s. e. harris , z. dutton , and c. h. behroozi , nature * 397 * , 594 ( 1999 ) .
u. schnorrberger , j. d. thompson , s. trotzky , r. pugatch , n. davidson , s. kuhr , and i. bloch , phys .
rev .
lett . * 103 * 033003 ( 2009 ) .
m. saffman and t. g. waller , rev .
mod .
phys . * 82 * , 2313 ( 2010 ) .
y. o. dudin and a. kuzmich , * science * 336 , 887 ( 2012 ) .
j. d. pritchard , d. maxwell , a. gauguet , k. j. weatherill , m. p. a. jones , c. s. adams , phys .
rev .
lett . * 105 * , 193603 ( 2010 ) .
t. peyronel , o. firstenberg , q. liang , s. hofferberth , a. v. gorshkov , t. pohl , m. d. lukin , and v. vuletic , nature * 488 * , 57 ( 2012 ) . in the eit theory developed by jiang , _ et .
al . _
@xcite for fermi pairing gas , they consider standard d2 transition where the recoil energy is smaller than the decay rate of the excited state . in order to observe quantum many - body effect in such setup
, the control field ( @xmath10 ) has to be tuned much smaller than the decay rate ( @xmath7 ) , making it more difficult to be experimentally observed .
l. jiang , h. pu , w. zhang , and h. y. ling phys .
rev .
a * 80 * , 033606 ( 2009 ) .
see supplemental material for detailed derivations of the linear susceptibility and single particle green s function . in this paper , we just consider low - lying rydberg states ( @xmath123p levels , @xmath124 ) @xcite of alkali atoms .
the van der waals coefficient @xmath125 for @xmath126p-@xmath126p rubidium asymptotes is in the range of @xmath127 a.u .
@xcite , giving us the blockade radius @xmath128 @xmath24 m @xcite for a density @xmath129 @xmath130 with the effective excitation rabi frequency @xmath131 khz .
this radius provides an upper estimation of the dipole blockade strength , which can be further reduced by decreasing @xmath1 .
note that the probe field has a fundamental limit of single photon shot noise which can be overcome by more experimental trials . as a result
, we neglect the blockade in our theory .
t. f. gallagher , ( cambridge university press , 1994 ) .
d. tong , s. m. farooqi , j. stanojevic , s. krishnan , y. p. zhang , r. ct , e. e. eyler , and p. l. gould , phys .
rev .
lett . * 93 * , 063001 ( 2004 ) .
k. singer , j. stanojevic , m. weidemller and r. ct , j. phys .
b : at . mol .
. phys . * 38 * , s295 ( 2005 ) .
r. lw , h. weimer , u. krohn , r. heidemann , v. bendkowsky , b. butscher , h. p. bchler , and t. pfau , phys .
rev . a. * 80 * , 033422 ( 2009 ) .
r. heidemann , u. raitzsch , v. bendkowsky , b. butscher , r. lw , l. santos , and t. pfau , phys .
rev .
lett . * 99 * , 163601 ( 2007 ) .
a. l. fetter and j. d. walecka , _ quantum theory of many - particle systems _
( dover publications , 2003 ) .
h. h. jen , bo xiong , ite a. yu , and daw - wei wang , arxiv:1208.0118 ( 2012 ) .
f. d. m. haldane , phys .
rev .
lett . * 47 * , 1840 ( 1981 ) ; f. d. m. haldane , j. phys .
c : solid state phys .
* 14 * , 2585 ( 1981 ) .
m. a. cazalilla , j. phys .
b : at . mol .
. phys . * 37 * , s1 ( 2004 ) .
m. olshanii , phys .
rev .
lett . * 81 * , 938 ( 1998 ) .
t. kinoshita , t. wenger , and d. s. weiss , science * 305 * 1125 ( 2004 ) .
m. p. a. fisher , p. b. weichman , g. grinstein , and d. s. fisher , phys .
rev .
b * 40 * , 546 ( 1989 ) .
d .- w .
wang , phys .
rev .
a * 80 * , 063620 ( 2009 ) .
e. altman and a. auerbach , phys .
rev .
lett . * 89 * , 250404 ( 2002 ) .
r. zhang , s.r .
garner , and l. v. hau , phys .
rev .
lett . * 103 * 233602 ( 2009 ) . |
cosmic inflation is the most widely known scenario proposed for resolving several problems associated with the standard model of cosmology @xcite .
a number of inflationary models have been proposed over several decades @xcite .
the recent observations on the cosmic microwave background ( cmb ) anisotropy data may constrain many of the inflationary models @xcite .
it is believed that inflation seeded the formation of the large scale structures in the universe .
inflation also predicts a nearly scale invariant spectrum for the scalar and tensor perturbations which occurred in the early universe .
the tensor perturbations of cosmological origin are known as primordial gravitational waves ( gws ) .
it is believed that the primordial gravitational waves have left its imprint on the cosmic microwave background .
therefore the primordial gws can be studied with the aid of cmb anisotropy and polarization .
the cmb is polarized in the early universe due to the thomson scattering . the density ( scalar )
fluctuations generate the @xmath0-mode polarization of the cmb , while the gravitational waves generate both @xmath0-mode and @xmath1-mode polarizations @xcite .
the primordial gravitational waves are a unique source of @xmath1-mode of cmb and its detection will help in understanding the inflation as well as the primordial gravitational waves itself .
the gravitational waves were generated during the inflation period due to the zero - point quantum oscillations @xcite .
an initial vacuum state ( no graviton ) can evolve into a multi - particle quantum state known as the squeezed vacuum state @xcite , which is a well - known state in the context of quantum optics @xcite . therefore the primordial gravitational waves are believed to exist in the squeezed vacuum state . @xcite .
the primordial gravitational waves are placed in the squeezed vacuum state and its effect on the @xmath2 mode angular correlation power spectrum of cmb is studied with wmap data @xcite .
recently , it is shown that the @xmath2 mode angular power spectrum gets enhanced at its lower multi poles by considering the primordial gravitational waves in thermal state @xcite .
these studies show that the primordial gravitational waves may exhibit both the squeezing as well as thermal features and hence it is worthwhile to examine their combined effects on the @xmath2 mode angular correlation power spectrum in light of the recent joint bicep2/keck array and planck data .
the aim of the present work is to study effect of primordial gravitational waves in the squeezed vacuum state on the @xmath2 mode angular correlation power spectrum of cmb with various slow roll inflationary models .
thus the obtained @xmath2 mode angular correlation power spectrum of cmb for the squeezed vacuum as well as the joint effect of squeezing and thermal cases are compared with the joint bicep2/keck array and planck data .
the perturbed metric for a flat friedmann - lematre - robertson - walker universe can be written as @xmath3,\ ] ] where @xmath4 is the flat space metric and @xmath5 is the tensor perturbation , @xmath6 , @xmath7 and @xmath8 is the conformal time . in quantum theory , the field @xmath9 can be written in the fourier mode as @xmath10 , \end{aligned}\ ] ] where @xmath11 is the normalization constant , @xmath12 is the planck length and @xmath13 is the wave vector .
the wave number is @xmath14 and is related to wavelength , @xmath15 .
the two polarization states @xmath16 , @xmath17 satisfy the conditions @xmath18 these linear polarizations are respectively known as plus @xmath19 polarization and cross @xmath20 polarization .
the creation ( @xmath21 ) and annihilation ( @xmath22 ) operators satisfy the relationships , @xmath23 $ ] = @xmath24 and @xmath25 $ ] = @xmath26=0.$ ] the evolution of these operators are governed by the heisenberg equations of motion , @xmath27 , \,\ , \frac{d}{d\tau}b_{k}(\tau ) = -i[b_k(\tau),h].\ ] ] the initial vacuum state @xmath28 is defined as @xmath29 under the bogoliubov transformation the creation and annihilation operators become @xmath30 @xmath31 where @xmath32 and @xmath33 are the initial values of the operators , @xmath34 and @xmath35 are complex functions and they satisfy the condition @xmath36 the coupling of the mode functions @xmath37 with @xmath38 gives @xmath39 where @xmath40 can have the following form . ]
@xmath41 which satisfies the equation of motion @xmath42 where prime indicates the derivative with respect to the conformal time @xmath43
. two - point correlation function of the tensor perturbation is given by : @xmath44 where the angle bracket denotes ensemble average and @xmath45 is known as the tensor power spectrum . using eqs .
( [ fourm ] ) and ( [ mf ] ) , taking the contribution from each polarization to be the same , @xmath46e^{i\textbf{k}.\textbf{x}}.\ ] ] the primordial gravitational waves are created due to the zero - point quantum oscillations in the early universe @xcite .
the initial vacuum state with no graviton evolves into multi - particle quantum state through parametric amplification .
hence the primordial gws are possible to be considered in the squeezed vacuum state @xcite .
the squeezed vacuum state is defined as @xcite @xmath47 where @xmath48 is the single mode squeezing operator and is given by @xmath49,\ ] ] where @xmath50 is a complex number , @xmath51 is the squeezing parameter and @xmath52 is the squeezing angle .
the unitary transformations of the squeezing operator @xmath53 on the annihilation and creation operators lead to : @xmath54 using eqs .
( [ fourm ] ) and ( [ qso ] ) , the two point correlation for the tensor perturbation in the squeezed vacuum is obtained as @xmath55 \delta^3 ( \textbf{k}-\textbf{k } ' ) .
\end{aligned}\ ] ] thus from eqs .
( [ tpc ] ) and ( [ tpg ] ) , we get the tensor power spectrum in the squeezed vacuum state as @xmath56.\end{aligned}\ ] ] in the case of quasi de sitter universe during inflation , @xmath57 , where the slow roll parameter @xmath58 is related to the scalar field potential @xmath59 and is given by , @xmath60 . for small @xmath58 , @xmath61 , and @xmath62 . if the slow roll parameter @xmath58 is considered as constant , then eq .
( [ eom ] ) gets modified as @xmath63\chi_k = 0.\ ] ] the general solution for eq .
( [ mudd ] ) is given by @xmath64,\ ] ] where @xmath65 and @xmath66 are respectively the hankel functions of the first and second kind , and @xmath67 and @xmath68 are the constants of integration . for within the horizon ( @xmath69 ) , the approximate solution is given by @xmath70 using the above approximate solution , one can obtain @xmath71 , \ , \ , c_2(k)=0.\end{aligned}\ ] ] therefore , for long wavelength limit ( @xmath72 ) , eq .
( [ gsol ] ) leads to @xmath73 using eq .
( [ musol ] ) in eq .
( [ sqzspc ] ) , the tensor power spectrum in the superhorizon limit ( @xmath72 ) is obtained as , @xmath74,\end{aligned}\ ] ] where @xmath75 is the normalization constant and @xmath76 is the hubble parameter at @xmath77 during the inflation , @xmath78 being the pivot frequency .
in most models of inflation @xcite , a homogeneous scalar field , called inflaton , is considered as candidate for the inflation .
the inflaton field @xmath79 is governed by the equation of motion given by @xmath80 where dot and prime indicate derivatives with respect to time ( @xmath81 ) and the field ( @xmath79 ) respectively .
the hubble parameter @xmath82 is determined by the energy density of the inflaton field,@xmath83 thus the friedmann equation can be written as @xmath84 in the slow - roll limit , the energy density of the inflaton field is dominated by its potential , @xmath85 . from eq .
( [ fdmeq ] ) , the hubble parameter and the inflaton potential are related by @xmath86 this condition is characterized by the slow - roll parameters which are defined in terms of the inflaton potential @xmath59 and its derivatives as @xcite @xmath87 and so on .
inflation lasts as long as the slow - roll conditions are satisfied , i.e. , @xmath88 and @xmath89 .
the duration of inflation is characterized by the e - fold number , @xmath90 , which can be written in terms of the inflaton potential , @xmath91 in the slow - roll approximation , the power spectrum of the scalar perturbations ( @xmath92 ) and the tensor perturbations ( @xmath45 ) generated outside the horizon are respectively given in terms of the potential @xcite @xmath93 where @xmath94 indicates that @xmath82 and hence , @xmath59 is evaluated at the time when the mode with wave number @xmath95 crosses the horizon . for all calculations ,
we take the scalar power spectrum to be @xmath96 .
the tensor - to - scalar ratio can be written in terms of the parameter @xmath58 as @xmath97 this parameter is often used to characterize the amplitude of the tensor perturbation at the cmb scale .
physically , _ r _ is a measure of the slope ( of the quantum hill ) down which the scalar field is rolling .
since inflation predicts a nearly scale invariant spectrum , the slope is small but not flat .
r _ is small and can differentiate between the many inflation models .
next , we obtain the slow roll parameter and tensor spectral index corresponding to various slow roll inflation models given below . for this purpose
, we use the e - fold number @xmath98 , for all the inflationary models under the present work @xcite .
the chaotic inflation model assumes that the scalar field rolls down its potential and rests at its vacuum state for a while , then after getting displaced due to some fluctuations , rolls back to its true vacuum state and the same mechanism repeats itself . for quadratic chaotic inflation , the scalar field has the potential given by @xcite @xmath99 where @xmath100 is the mass of the inflaton field and is taken to be @xmath101 gev .
using eqs .
( [ efold ] ) and ( [ ch2pot ] ) , the slow roll parameter and index of the tensor power spectrum for the quadratic chaotic inflationary model are obtained as : @xmath102 the corresponding tensor - to - scalar ratio is obtained as , @xmath103 . the quartic chaotic inflation model suggests the existence of the curvaton which generates the curvature fluctuations during inflation after the inflaton field has decayed , while it does not drive the inflation itself .
the model has the potential given by @xcite @xmath104 where normalization gives the self - coupling of the scalar field ( @xmath79 ) as @xmath105 gev , and hence @xmath106 the tensor - to - scalar ratio is for this model is found as @xmath107 .
the new inflation model @xcite is based on the coleman - weinberg potential@xcite , @xmath108 in this model , the parameters are taken to be @xmath109 , where @xmath110 is planck s mass , and @xmath111 gev .
@xmath112 is the vacuum expectation value of the scalar field at minimum and @xmath113 is the quadratic self coupling of the field .
therefore we get @xmath114 the tensor - to - scalar ratio in this case is obtained as @xmath115 . in this model ,
the inflation is driven by two scalar fields , in which one of the fields ( @xmath79 ) is responsible for the normal slow - roll inflation while the other field ( @xmath112 ) triggers the end of inflation .
the potential for this model is @xcite @xmath116 where the parameters are taken to be as follows : @xmath117 , @xmath118 are the self - coupling constants of the inflaton field and the trigger field respectively , and the masses of the fields are @xmath119 , @xmath120 gev respectively .
the @xmath79 field drives the inflation , thus determining the period of inflation while the @xmath112 field determines the rate of inflation .
therefore we get : @xmath121 the tensor - to - scalar ratio in this case is found as @xmath122 .
the cmb radiation can be polarized due to the gravitational waves @xcite called the @xmath1-mode @xcite of cmb .
the angular correlation power spectrum of the @xmath1-mode cmb is given by @xcite @xmath123\right|^2,\end{aligned}\ ] ] where @xmath124 is the visibility function , @xmath125 is the differential optical depth for the thomson scattering , @xmath126 and @xmath127 is the spherical bessel function .
the @xmath2 mode correlation angular spectrum of cmb is obtained for the gravitational waves in the squeezed vacuum and thermal vacuum states for the quadratic chaotic , quartic chaotic , new inflation and hybrid inflationary models .
the angular power spectra for the different inflation models are generated using the camb code with the tensor spectral index @xmath128 corresponding to each inflationary model .
for all the cases , the optical depth is taken as @xmath129 , the pivot wave number for tensor modes is taken as @xmath130 mpc@xmath131 and that for scalar mode it is @xmath132 mpc@xmath133 .
the tensor to scalar ratio used for each inflationary model is taken from the previous section where their values are computed .
the obtained @xmath2 mode angular power spectrum for various values of the squeezing parameter and temperature are compared with the limit of the bicep2/keck array and planck 353 ghz joint analysis data . the implemented limit ( bk x bk - @xmath134
bk x p)/(1 - @xmath134 ) at @xmath135 is evaluated @xcite from the auto - spectra and cross - spectra of the combined bicep2/keck 150 ghz maps and planck 353 ghz maps to clean out the dust contribution , bk x bk indicates the bicep2/keck auto - spectra at 150 ghz and bk x p indicates the cross - spectra of bicep2/keck maps at 150 ghz and planck maps at 353 ghz .
this combination is taken after the subtraction of the dust contribution which is 0.04 times as much in the bicep2 band as it is in the planck 353 ghz band ( see ref.@xcite for details ) .
-mode angular spectra for various inflation models with lensing for zero squeezing effect with bicep2/keck and planck joint analysis result [ f01 ] ] -mode angular spectra for various inflation models with unlensing for squeezing parameter @xmath136 and squeezing angle @xmath137 [ f02 ] ] -mode angular spectra for various inflation models with lensing for squeezing parameter @xmath136 and squeezing angle @xmath137 with bicep2/keck and planck joint analysis result [ f03 ] ] -mode angular spectra for various inflation models with lensing for squeezing parameter @xmath138 and squeezing angle @xmath137 with bicep2/keck and planck joint analysis result [ f04 ] ] -mode angular spectra for various inflation models with lensing for squeezing parameter @xmath139 and squeezing angle @xmath137 bicep2/keck and planck joint analysis result [ f05 ] ] -mode angular spectra for various inflation models with lensing for temperature @xmath140 mpc@xmath131 and squeezing parameter @xmath136 and squeezing angle @xmath137 with bicep2/keck and planck joint analysis result [ f06 ] ] -mode angular spectra for various inflation models with lensing for temperature @xmath140 mpc@xmath131 and squeezing parameter @xmath138 and squeezing angle @xmath137with bicep2/keck and planck joint analysis result [ f07 ] ] -mode angular spectra for various inflation models with lensing for temperature @xmath140 mpc@xmath131 and squeezing parameter @xmath139 and squeezing angle @xmath137 with bicep2/keck and planck joint analysis result [ f08 ] ] -mode angular spectrum for the hybrid inflation model with unlensing for various values of squeezing parameter and angles [ f09 ] ] -mode angular spectrum for the hybrid inflation model with lensing for various values of squeezing parameter and angles with bicep2/keck and planck joint analysis result [ f10 ] ] the @xmath2-mode correlation angular spectra for the different inflation models for zero squeezing with lensing effect are obtained and their corresponding plots are given in fig.([f01 ] ) . the @xmath2 mode angular spectrum for various inflationary models for the squeezing , thermal and their combined effects with lensed and unlensed cases
are also obtained and the results plots are in figs .
( [ f02]-[f10 ] ) .
it can be seen that for various values of the squeezing parameter and squeezing angle , the quadratic chaotic , quartic chaotic and new inflation models are out of the limit of bicep2/keck array and planck 353 ghz joint data . for the thermal effect with various values of the squeezing parameter and squeezing angle , the quadratic chaotic , quartic chaotic and new inflation models are also not found within the limit of bicep2/keck array and planck 353 ghz joint data . it can be observed that the @xmath2 mode angular correlation power spectrum corresponding to the hybrid inflation model for various values of squeezing parameter and squeezing angle is found within the limit of the bicep2/keck and planck joint analysis data at @xmath141 , even though the squeezing and thermal effects enhances the angular correlation power spectrum moderately .
thus the analysis of the present results show that the hybrid inflation model is most favorable by considering the primordial gravitational waves in presence of squeezing thermal and effects .
the primordial gravitational waves are very important in cosmology .
it is believed that the origin of the primordial gravitational waves are due to inflation mechanism . thus the primordial gravitational waves are placed in a special quantum state called the squeezed vacuum state , a well known state in quantum optics .
the gravitational waves have not been detected directly yet but its effect is expected to be observed through the @xmath1 mode of cmb .
if this is the case , then the role of squeezing effect is also expected to reflect on the @xmath2 mode angular correlation power spectrum of cmb .
the @xmath2-mode correlation angular power spectrum for various slow - roll inflationary models are studied by considering the primordial gravitational waves in the squeezed vacuum state .
the obtained angular spectrum for the squeezing , thermal as well as their combined effects are compared with the bicep2/keck array and planck 353 ghz joint analysis data .
the comparative study of the @xmath2 mode angular spectrum obtained for the various inflationary models with the bicep2/keck array and planck 353 ghz joint analysis data show that the hybrid inflation model is most favorable with the primordial gws in the squeezed vacuum state .
the @xmath2 mode angular correlation power spectrum can be obtained even with higher or lower values of the squeezing parameters and thermal parameter but the results do not alter the present conclusions . note that the @xmath2 mode angular correlation power spectrum for the vacuum case can be recovered in absence of squeezing and thermal effects .
further , the result of the joint analysis data does not rule out the existence of gravitational waves in the squeezed vacuum or thermal squeezed vacuum states which are in agreement with previous studies .
n m acknowledges the financial support of council of scientific and industrial research ( csir ) , new delhi .
p k s acknowledges financial the support of dst - serb , new delhi .
we would like to thank planck and bicep websites for the data .
after the big bang , the universe was in the form of plasma of hot and dense matter . during this stage ,
the light particles like free electrons , due to thomson scattering , acted as scattering centers for the surrounding radiation , keeping the universe at that time thermalized and opaque to radiations . among these particles and radiations which got thermalized and escaped after the recombination are the decoupled gravitons which left an imprint on the cmb anisotropy . due to these thermalized gravitons ,
the gravitational waves are believed to be amplified by stimulated emission into the existing thermal background of gravitational waves which changes the spectrum of the waves by temperature - dependent factor @xcite .
then the time and temperature - dependent annihilation and creation operators through the bogoliubov transformation become @xmath149 the hermitian conjugate of the above equations give rise to similar equations for @xmath150 and @xmath151 .
000 r. h. brandenberger , `` ipm school on cosmology 1999 '' : arxiv : hep - ph/9910410v1 ( 1999 ) a. d. linde , _ particle physics and inflationary cosmology _
( crc press , 1990 ) a. d. linde , lect .
notes phys .
* 738 * 1 ( 2008 ) a. h. guth , phys .
d * 23 * 347 ( 1981 ) a. d. linde , physics letters b * 129 * 177 ( 1983 ) a. h. guth and s.y.pi , phys .
. lett . * 49 * 1110 ( 1982 ) a. d. linde , phys .
d * 49 * 748 ( 1994 ) j. garcia - bellido and a. d. linde , phys .
d * 57 * 6075 ( 1998 ) p. a. r. ade et al .
, phys . rev .
lett . * 112 * 241101 ( 2014 ) p. a. r. ade et al .
, astron . astrophys .
* 571 * a22 ( 2014 ) j. martin , c. ringeval and v. vennin , arxiv:1303.3787 ( 2013 ) j. martin , c. ringeval , r. trotta and v. vennin , jcap * 03 * 039 ( 2014 ) j. martin , arxiv:1502.05733v1 ( 2015 ) l. p. grishchuk , sov .
jetp * 40 * 409 ( 1975 ) l. p. grishchuk , lect .
notes phys .
* 562 * 167 ( 2001 ) b. l. schumaker , phys .
rep . * 135 * 317 ( 1986 ) j. r. klauder and b. s. skagerstam , _ coherent states _
( world scientific , 1985 ) c. m. caves and d. f. walls , phys . rev .
lett * 57 * 2164 ( 1980 ) l. p. grishchuk , phys .
d * 53 * 6784 ( 1996 ) l. p.
grishchuk , arxiv:0707.3319v4 [ gr - qc ] ( 2010 ) l. p. grishchuk and yu .
v. sidorov , phys .
d * 42 * 3413 ( 1990 ) b. s. sathyaprakash , b. f. schutz , living rev .
relativity * 12 * 2 ( 2009 ) s. dodelson , _ modern cosmology _
( academic press , 2003 ) s. dodelson , w. h. kinney and e. w. kolb , phys.rev .
d56 3207 ( 1997 ) j. abadie et al . ,
arxiv:1203.2674v2 [ gr - qc ] ( 2012 ) b. p. abbott et al .
d * 80 * 042003 ( 2009 ) h. grote , class.quant.grav . *
27 * 084003 ( 2010 ) m. hewitson et al .
, class.quant.grav .
* 20 * s581 ( 2003 ) s. j. waldman , arxiv:1103.2728v1 [ gr - qc ] ( 2011 ) s. kuroyanagi and t. takahashi , jcap * 1110 * 006 ( 2011 ) k. bhattacharya , s. mohanty and a. nautiyal , phys .
lett . * 97 * 251301 ( 2006 ) s. kuroyanagi , t. chiba and n. sugiyama , phys .
d * 79 * 103501 ( 2009 ) y. zhang , y. yuan , w. zhao and y. t. chen , class.quant.grav .
* 22 * 1383 ( 2005 ) l. p.
grishchuk , arxiv : gr - qc/9302036v1 ( 1993 ) s. nakamura , n. yoshino and s. kobayashi , prog .
* 88 * 1107 ( 1992 ) d. f. walls , nature * 300 * 141 ( 1983 ) m. giovannini , pmc phys . a * 4 * 1 ( 2010 ) k. nakayama , s. saito , y. suwa and j. yokoyama , jcap * 0806 * 020 ( 2008 ) s. kuroyanagi , c. gordon , j. silk and n. sugiyama , phys .
d * 81 * 083524 ( 2010 ) w. zhao , d. baskaran and p. coles , phys .
b * 680 * 411 ( 2009 ) s. r. coleman and e. j. weinberg , phys .
d * 7 * 1888 ( 1973 ) m. s. turner , m. j. white and j. e. lidsey , phys .
d * 48 * 4613 ( 1993 ) w. hu , lecture notes on cmb theory : from nucleosynthesis to recombination : arxiv:0802.3688(2008 ) a. kosowsky , new astron .
* 43 * 157 ( 1999 ) y. t. lin and b. d. wandelt , astroparticle physics * 25 * 151 ( 2006 ) u. seljak and m. zaldarriaga , phys .
lett . * 78 * 2054 ( 1997 ) d. baskaran , l. p. grishchuk and a. g. polnarev , phys . rev .
d * 74 * 083008 ( 2006 ) p. a. r. ade.et al . ,
rev . lett . * 114 * 101301 ( 2015 ) s. koh , s. p. kim and d. j. song , jhep * 2004 * 060 ( 2004 ) h. umezawa , h. matsumoto and m. tachiki , _ thermo field dynamics and condensed states _
( north - holland publishing company , 1982 ) b. ghayour and p. k. suresh , class .
* 29 * 175009 ( 2012 ) | the bb - mode angular correlation power spectrum of cmb is obtained by considering the primordial gravitational waves in the squeezed vacuum state for various inflationary models and results are compared with the joint analysis of the bicep2/keck array and planck 353 ghz data .
the present results may constrain several models of inflation . |
it is in principle possible to calculate the next order correction to eq.([mod_lan ] ) by retaining terms to second order in the taylor series expansion of @xmath38 in eq.([sum_phi ] ) and eq .
( [ sum_phigrad ] ) .
furthermore since the exponent is slowly varying , the the sums over @xmath99 can be approximated by integrals , leading to : @xmath100,\ ] ] where @xmath101.\ ] ] the integrals in eq.([second_correction ] ) can not be computed exactly and approximations are also hard to come by , since we do not _ a priori _ know the magnitude of the walker s velocity @xmath102 and acceleration @xmath67 , which appear in the argument of the exponent .
the main problem here is that one has an implicit equation ( given by eq.([e1 ] ) together with eq.([second_correction ] ) ) in the walker s velocity . the only fact which can be safely deduced
is that for @xmath42 , the two integrals are approximately equal ; this implies that the second term in eq.([second_correction ] ) vanishes in the limit @xmath103 and that in high dimensions there are no further corrections to the modified langevin equation eq.([mod_lan ] ) . clearly in the formalism of the langevin equation it is not easily possible to systematically calculate corrections to the modified langevin equation eq.([mod_lan ] ) .
however it is to be emphasized that the derivation of the latter equation ( and the subsequent prediction of the phase transition ) , elude a treatment based on the equation of motion of the walker s probability density function . | we study an active random walker model in which a particle s motion is determined by a self - generated field .
the field encodes information about the particle s path history .
this leads to either self - attractive or self - repelling behavior . for self - repelling behavior
, we find a phase transition in the dynamics : when the coupling between the field and the walker exceeds a critical value , the particle s behavior changes from renormalized diffusion to one characterized by a diverging diffusion coefficient .
the dynamical behavior for all cases is surprisingly independent of dimension and of the noise amplitude .
active walkers are random walkers whose motion is determined by a potential surface which is amenable to change by the walkers themselves @xcite .
they are one of several different classes of random walks in which walkers interact with their path history .
the memory of these walkers leads to interesting properties , uncommon to pure random walks .
other common types of such self - interacting walks are the reinforced random walk @xcite and the self - avoiding random walk @xcite
. models based on active walkers have been used to study a number of phenomena , including ant foraging patterns @xcite , traffic @xcite , the formation of human and animal trail systems @xcite , animal mobility @xcite , chemotactic aggregation @xcite and the self - assembly of networks @xcite .
the advantage of formulating problems in such a formalism is that it enables analytic study using methods from statistical physics and the general theory of stochastic processes . in this article
, we analyze in detail a fairly generic type of active walker model , in which a particle interacts with its path history by means of a self - generated field .
this model is related to the self - driven many particle system first introduced by schweitzer and schimansky - geier @xcite , which studied it by means of a mean - field approximation applied on the stochastic model .
newman and grima @xcite analyzed the effect of fluctuations on the system s dynamics by means of a many - body theory approach .
the spatio - temporal correlations make the problem difficult to understand , usually confining analysis to the case of weak coupling between particles and the generated field @xcite or to the case of small noise @xcite . indeed
even the case of a single self - driven particle is non - trivial ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* see for example ) .
we study a variant of the single self - driven particle model and develop a general theory to elucidate the particle s rich and complex dynamical behavior .
this approach differs from the previous ones , in that it is valid for all coupling strengths and for both weak and strong noise .
the model is potentially applicable to understanding a chemotactic biological system under certain conditions , a topic briefly discussed at the end of this article .
consider a random walker whose motion is described by the following coupled equations : @xmath0 eq.([e1 ] ) is a langevin equation describing the motion of a walker with mass @xmath1 .
the stochastic variable @xmath2 is white noise defined through the statistical averages : @xmath3 and @xmath4 , where @xmath5 and @xmath6 refer to the spatial components of the noise vectors .
the mass is assumed to satisfy the condition @xmath7 , implying that in the absence of self - interaction the walker s dynamics are of the over - damped type .
the walker s diffusion coefficient in this case is @xmath8 .
the self - generated field is denoted by @xmath9 ; its temporal dynamics are determined by the reaction - diffusion equation eq.([e2 ] ) .
we shall refer to the field as a chemical field since this is most consistent with a physical interpretation of eq.([e2 ] ) .
then the latter equation describes the continuous local release of chemical by the walker at a rate @xmath10 , the diffusion of the chemical with diffusion coefficient @xmath11 and its decay at a constant rate @xmath12 .
the self - interaction comes from the second term on the r.h.s . of eq.([e1 ] ) .
this term implies that the walker s motion is partly determined by the local gradient of the field .
note that the field encodes information about the walker s path history , meaning that the walker s motion at any given time is a complicated function of its previous whereabouts .
the strength of the coupling between the field and the walker s motion is determined by @xmath13 .
the constant @xmath14 can take values @xmath15 or @xmath16 : for @xmath17 , the walker tends to explore regions already visited ( a self - attracting walker ) whereas for @xmath18 the opposite is true ( a self - repelling walker ) .
we assume that the initial chemical concentration is described by some function @xmath19 which is greater than zero at all points in space .
the langevin formulation is not usually considered the most convenient representation for the purposes of analytic calculations and so it is customary to derive a differential equation for the single - particle probability distribution . for the problem at hand
, this approach does not permit much analytic progress in understanding the walker s behavior . applying a mean field approximation on the differential equation for the single - particle probability distribution one obtains a fokker - planck type equation , which then permits a perturbative analysis in a coupling parameter .
this approach ignores the important spatio - temporal correlations inherent in the problem and enables one to understand the walker s behavior only when the coupling strength @xmath13 is very small .
the non - markovian nature of the problem makes its solution a challenging task .
we here present a simple method to extract the asymptotic behavior of the active walker model .
the results can also be reproduced by the method described in @xcite .
however the method to be presented here is more transparent and gives a physically tractable picture of the complex underlying dynamics .
its main advantages are that spatio - temporal correlations are not ignored and that it enables an understanding of the motile behavior for all values of the coupling strength and of the other parameters .
we start by switching to a description in discrete time @xmath20 where @xmath21 .
space is continuous .
the walker s position at time @xmath22 is determined by the gradient of the logarithm of the chemical concentration it measures at time @xmath23 . since the walker secretes an amount of chemical @xmath24 at every time step , then if the walker is at position @xmath25 at time @xmath23 , the chemical field sensed by the walker at time @xmath23 is given by : @xmath26 ^ 2}{4 d_1 n \delta t } \biggr ] = \sum_{n=1}^{t/\delta t } \phi_n,\ ] ] and the gradient of the field is given by : @xmath27}{2 d_1
n \delta t } \phi_n,\ ] ] where @xmath28 is the @xmath29 component of the particle position vector @xmath30 and @xmath31 is the dimensionality of the space in which particle movement occurs .
since the chemical decays in a time of the order @xmath32 then the concentration at time @xmath23 will be approximately determined by the previous positions of the walker at times @xmath33 .
this implies that the sum in eq.([sum_phi ] ) and eq.([sum_phigrad ] ) can be truncated at @xmath34 .
now consider the term @xmath35 .
since we are interested in the walker s behavior in the asymptotic limit @xmath36 , then @xmath37 .
thus it is possible to replace the term @xmath38 in the above two equations by its taylor series expansion . keeping terms only to first order in @xmath39 we have : @xmath40 thus from eq.([e1 ] ) and the above equation
, it follows that for long times , the behavior of the walker is dominated by the effective langevin equation : @xmath41 it is not possible to systematically calculate corrections to this equation by keeping more terms in the taylor expansion of the position terms .
it can however be shown that such corrections are negligible in high dimensions , @xmath42 .
these issues are discussed more fully in the appendix .
the modified langevin equation eq.([mod_lan ] ) is clearly valid after some time @xmath43 such that @xmath44 .
we define @xmath45 and integrate eq.([mod_lan ] ) to get an equation for the time evolution of the @xmath29 component of the velocity vector : @xmath46 } + \frac{1}{m } \int_{t^*}^{t } dt ' \exp{[-\gamma(t - t ' ) ] } \xi^i(t').\ ] ] then it follows that the velocity auto - correlation function is given by : @xmath47 } \\ & + \frac{d_0}{\gamma m^2 } \biggl [ \exp{(-\gamma|s - t| ) } - \exp{[-\gamma(s+t-2t^ * ) ] } \biggr].\end{aligned}\ ] ] using a green - kubo relation @xmath48 it is possible to determine the effective ( renormalized ) particle diffusion coefficient @xmath49 : @xmath50}(1 - \exp{[-\gamma ( t - t^ * ) ] ) } \\ & + \frac{d_0}{\gamma^2 m^2 } \biggl [ 1 + \exp{[-2\gamma ( t - t^ * ) ] } - 2 \exp{[-\gamma ( t - t^ * ) ] } \biggr],\end{aligned}\ ] ] which evaluated in the limit @xmath51 leads us to the final set of results : @xmath52 we defer a discussion of the physics behind these results for later . for the moment we focus on the numerical validation of the theory . for the case of a self - attracting walker ( @xmath17 )
it is predicted that : ( i ) the asymptotic behavior is diffusion with a renormalized diffusion coefficient whose magnitude decreases with increasing values of the coupling parameter @xmath13 , ( ii ) that the behavior is independent of dimension ( superuniversality in the velocity correlations ) .
we test these predictions by numerically integrating the model equations eq.([e1 ] ) and eq.([e2 ] ) ( see fig.1 ) .
the diffusion coefficient in all simulations is calculated from the slope of plots of the variance versus time for the time range @xmath53 .
the initial chemical concentration is a gaussian centered at the origin , though any non - zero function is suitable .
[ h ] versus the non - dimensional coupling parameter @xmath54 , where @xmath55 .
the parameter @xmath14 equals one , implying that the walker has a tendency to explore previously visited spatial regions .
the other parameters are @xmath56 , @xmath57 , @xmath58 and @xmath59 with @xmath60 samples .
as predicted , the renormalized diffusion coefficient is the same in one , two and three dimensions.,title="fig:",width=480 ] as expected , we find that the asymptotic behavior is diffusion characterized by renormalized diffusion coefficients which are relatively independent of the dimension ( fig .
1 ) . however there is some discrepancy between the theoretical values of @xmath49 and the ones obtained from the numerics .
regression of the one - dimensional data in fig .
1 . shows that the numerical data is best fit by an equation of the form ( see fig .
2 ) : @xmath61 where @xmath62 ( @xmath63 realizations ) .
the difference between this value and the theoretical value of unity , stems from a combination of the approximations used in deriving the effective langevin equation eq.([mod_lan ] ) and numerical error due to a finite time step ( note that the simulations are off - lattice and thus there is no numerical error due to a finite spatial step ) .
as discussed in the appendix , it is not possible to systematically calculate corrections to the effective langevin equation .
however by repeating the simulations with a time step an order of magnitude smaller than those in fig . 1 and fig .
2 , we find that the value of @xmath64 increases to @xmath65 ( @xmath66 realizations ) , which is closer to the theoretical value . hence it is probable that the discrepancies between numerics and theory are in significant part due to numerical error rather than to the approximations implicit in deriving eq.([mod_lan ] ) .
this is plausible since the next order correction to eq.([mod_lan ] ) is proportional to the walker s acceleration @xmath67 ( see eq.([second_correction ] ) in the appendix ) which is negligible for a self - attracting walker since the dynamics are over - damped for small coupling ( @xmath7 ) and become more strongly over - damped as the coupling increases ( this is since the coefficient of the velocity term in eq.([mod_lan ] ) increases with the coupling ) .
[ h ] versus the non - dimensional coupling parameter @xmath54 , where @xmath55 , for data obtained from 1d simulations .
parameter values as in fig .
1 . this verifies the functional form predicted by theory . the solid line is the best fit through the data points .
this line has a gradient of @xmath68 and intercept of @xmath69 theory predicts a gradient of @xmath70 and an intercept of @xmath70.,title="fig:",width=480 ] for the case of a self - repelling walker ( @xmath18 ) theory predicts that : ( i ) if the coupling is less than a critical threshold , @xmath71 , then the asymptotic behavior is diffusion with a renormalized diffusion coefficient , ( ii ) if the coupling is above this threshold , @xmath72 , the particle diffusion coefficient diverges ( iii ) the behavior is independent of dimension .
we tested these predictions by simulations .
we find that near the predicted singularity , @xmath73 , the variation of @xmath49 with coupling is given by : @xmath74 where @xmath75 in one dimension , @xmath76 in two dimensions and @xmath77 in three dimensions ( fig.3 ) .
these estimates agree well with the theoretical value of @xmath78 .
the constant of proportionality in eq.([e12 ] ) is dependent on dimension , a feature not predicted by theory interestingly , as shown in fig .
3 , the data in three dimensions is the one closest to the exact theoretical result .
however these features are not completely unexpected .
this is since the coefficient of the velocity term in eq.([mod_lan ] ) decreases with increasing coupling , meaning that the acceleration of the walker becomes a determining factor as the critical coupling is approached .
thus the next order correction to eq.([mod_lan ] ) ( see eq.([second_correction ] ) in the appendix ) is probably not negligible ( unlike the case of a self - attracting walker ) . as discussed in the appendix , these corrections are small in high dimensions and thus mostly significant in low dimensions .
these theoretical arguments support the numerical data in fig .
3 .
[ h ] versus the natural log of the non - dimensional parameter @xmath79 , where @xmath55 .
note that the @xmath54 values in this plot vary between 0.9 and 0.98 , meaning that we are exploring the walker s behavior near the theoretically predicted singularity at @xmath80 .
the parameter @xmath14 equals minus one , implying that the walker has a tendency to explore previously unvisited spatial regions .
the other parameters values are exactly as in fig .
1 .
this graph confirms that @xmath81 implying a singularity at @xmath80 in all dimensions.,title="fig:",width=480 ] [ h ] for @xmath82 using the same parameters as fig .
3 .
the model simulated is that given by eq .
( [ e13 ] ) and eq .
( [ e14 ] ) .
the solid line is the best fit through the data points , indicating that @xmath83 where @xmath64 is some positive number .
this suggests that in the limit @xmath84 , @xmath85 for @xmath86.,title="fig:",width=480 ] now we numerically explore the walker s asymptotic behavior for @xmath87 . in this regime ,
the simulations break down after a few time steps . using smaller values of the numerical time step does not help much and makes the numerical analysis computationally very expensive .
this is overcome by simulating a model given by the equations : @xmath88 since we are really interested in the case @xmath89 , we obtained data for several small values of @xmath90 in the hope that we could eventually extrapolate to the desired limit .
as shown in fig.4 , we find that for @xmath91 the renormalized diffusion coefficient clearly tends to infinity as @xmath84 .
this is found to be generally true for @xmath92 , meaning that the purported transition at @xmath93 is from finite @xmath49 to @xmath94 , in agreement with theory .
note that transitions in random walks with positive long - correlations are known @xcite , though in this case the dynamical transition is from normal - diffusion to super - diffusion in a one - dimensional space .
now we discuss the theoretical results from a physical perspective . in the absence of self - interaction , the walker s motion is determined by the frictional force which is directly proportional to its velocity rather than by inertia ( over - damped dynamics ) . for the case of a self - attracting walker ( @xmath17 )
it is expected that the walker spans space more slowly than for the case of no self - interaction : this typically means sub - diffusive behavior or diffusive behavior with a renormalized diffusion coefficient smaller than @xmath8 .
we have shown that to a first approximation obtained by truncating the walker s memory to a time of the order @xmath32 , the self - interaction leads to a renormalization of the frictional force .
this implies that the dynamics are always over - damped and that the asymptotic behavior is that of renormalized diffusion , not sub - diffusion .
for the case of a self - repelling walker ( @xmath18 ) it is expected that the walker spans space faster than for the case of no self - interaction : this typically means super - diffusive behavior or diffusive behavior with a renormalized diffusion coefficient greater than @xmath8 .
we have shown that the self - interaction leads to a renormalization of the magnitude of the frictional force experienced by the walker . the frictional force decreases with increasing coupling @xmath13 between the walker s motion and the chemical field until at a particular value of the coupling , @xmath95 , the frictional force is exactly zero and the dynamics of the walker are purely determined by the inertial force .
thus the walker s behavior changes from one characterized by a low - reynolds number for weak self - interaction to behavior characterized by a high reynolds number as one approaches the critical coupling .
when the coupling exceeds the critical value , we find that the walker experiences a force which is proportional to the velocity but which has the opposite effect of damping ; the random fluctuations in the velocity due to the stochastic force are amplified rather than suppressed and thus the velocity of the walker increases uncontrollably with time , leading to an infinite velocity .
this is the underlying reason for the divergence of the particle diffusion coefficient in this parameter regime .
it is worthwhile to compare the walker s behavior with logarithmic response to the behavior with linear response .
the linear response model was studied by grima @xcite , who found that if the particle diffusion coefficient ( in the absence of self - interaction ) is small then ( i ) for a self - attracting walker , the asymptotic behavior is that of renormalized diffusion ( ii ) for a self - repelling walker , the asymptotic behavior is renormalized diffusion below a critical threshold and ballistic diffusion above this threshold .
the first difference to be emphasized between the two models with different responses is that for the logarithmic response the results are generally valid for any value of the particle diffusion coefficient and of the other parameters , whereas for the linear response model the results are restricted to small particle to chemical diffusion coefficients ( small noise analysis ) . in other words the logarithmic response gives rise to behavior which is independent of the amplitude of the noise , an unusual property for example , for intermediate to large noise , simulations indicate that the dynamics of a particle with linear response are not that of renormalized diffusion .
it is also to be noted that the value of the critical coupling in the linear model is dimensionally dependent , unlike the superuniversal behavior in the present case .
the logarithmic response @xmath96 is weaker than the linear response @xmath97 when @xmath98 and stronger otherwise . for the case of a self - attracting walker
, the walker tends to stay in spatial regions which it has previously visited , meaning that the sampled chemical concentration @xmath9 at all times is significant and not small ; thus in this case we expect that the walker with linear response to exhibit a stronger or at least equally strong perturbation of its motion compared to that with logarithmic response .
this reasoning agrees with our results .
for the case of a self - repelling walker , the walker tends to avoid spatial regions which it has previously visited , meaning that the sampled chemical concentration tends to be very small ; thus in this case we expect the walker with logarithmic response to exhibit a stronger or at least equally strong perturbation of its motion compared to that with linear response .
this qualitatively explains the onset of the diverging diffusion regime for logarithmic response compared to the onset of ballistic diffusion for linear response .
as we mentioned in the introduction , the model has potentially some applications in biology .
the self - driven many particle system is a model for chemotactic aggregation or dispersion , behavior exhibited by a number of organisms , such as the slime mould @xcite .
the particles are then motile cells which secrete chemical and which simultaneously move up ( or down ) gradients of this chemical .
the self - driven single particle can be thought of as a model for a chemotactic cell which is away from the bulk of other cells ; in that case the cell will predominantly sense its own chemical rather than that of other cells . in the linear response models studied by several authors ,
the cell s average velocity is assumed to be linearly proportional with the gradient of the local chemical field . in our model ,
the response is logarithmic which is known to be more realistic in some specific cases .
this type of response is frequently referred to as the weber - fechner law and is thought to describe the sensory adaptation of a number of chemotactic cells to chemotactic signals , over certain concentration ranges @xcite . with this proviso
, our present model may perhaps be applicable to understanding some features of sensory adaptation at the micro - organism level .
the author would like to thank alessandro flammini and alessandro vespignani for a critical review of the manuscript and timothy newman for interesting discussions .
this work was supported by a grant from the faculty research support program from the ovpr , indiana university ( bloomington campus ) .
ohphs f. schweitzer and l. schimansky - geier , physica a * 206 * , 359 ( 1994 ) b. davis , prob .theor .
relat .
fields * 84 * , 203 ( 1990 ) r. pemantle , prob .theor . relat .
fields * 92 * , 117 ( 1992 ) d.j .
amit , g. parisi and l. peliti , phys .
rev .
b * 27 * , 1635 ( 1983 ) b.m .
schulz , s. trimper and m. schulz , phys .
letts .
a * 339 * , 224 ( 2005 ) f. schweitzer , k. lao and f. family , biosystems * 41 * , 153 ( 1997 ) d. helbing , rev .
mod .
phys . * 73 * , 1067 ( 2001 ) d. helbing , f. schweitzer , j. keltsch and p. molnar , phys . rev .
e * 56 * , 2527 ( 1997 ) w. ebeling , f. schweitzer and b. tilch , biosystems * 49 * , 17 ( 1999 ) r. grima , phys .
rev .
lett . * 95 * , 128103 ( 2005 ) t. j. newman and r. grima , phys . rev .
e.*70 * , 051916 ( 2004 ) f. schweitzer and b. tilch , phys .
rev .
e * 66 * , 026113 ( 2002 ) y. tsori and p .-
g de gennes , europhys . letts . * 66 * , 599 ( 2004 ) s. hod and u. keshet , phys .
rev .
e * 70 * , 015104 ( 2004 ) j.p .
armitage and j.m .
lackie , _ biology of the chemotactic response _ ( cambridge university press , cambridge , 1990 ) d. a. brown and h. c. berg , proc .
nat .
acad .
sci .
usa * 71 * , 1388 ( 1974 ) e. f. keller and l. a. segel , j.theor.biol * 30 * , 235 ( 1971 ) h. a. levine and b. d. sleeman , siam .
j. appl . math . *
57 * , 683 ( 1997 ) |
the existence of glueball states made of gluons is one of the important predictions of qcd .
discovery and confirmation of these glueball states would be the strong support to the qcd theory .
therefore , the search for and identifying the glueball states have been a very excited and attractive research subject .
the abundance of @xmath9 mesons and the possible mixing of glueballs and ordinary mesons make the current situation with the identification of the glueball states rather complicated .
however , some progress has been made in the glueball sector . by studying
the mixing between quarkonia and glueball to study the properties of glueballs or identify the glueball states is an appealing approach @xcite .
in contrast to the vector and tensor mesons , the identification of the scalar mesons is a long standing puzzle . in particular ,
the @xmath10 , @xmath11 sector is the most complex one both experimentally and theoretically .
the quark model predicts that there are two mesons with @xmath10 in the ground @xmath12 @xmath9 nonet , but apart from the state @xmath13 with @xmath14 or ( and ) @xmath15 , four states @xmath16 , @xmath17 , @xmath1 and @xmath2 are listed by particle data group ( pdg)@xcite . there are too many controversies about these states , especially relating to the @xmath16 and @xmath17 .
the convenient but not convincing recent tendency is to put aside the @xmath16 and @xmath17 , and focus on @xmath1 , @xmath2 and @xmath3 @xcite , although the spin - parity of the @xmath13 @xmath11 or ( and ) @xmath18 is controversial . recently , several authors have discussed the quarkonia - glueball content of the @xmath1 , @xmath2 and @xmath3 by studying the mixing of the three states @xcite .
the different assumption about the mass level order of the bare states @xmath6 , @xmath5 and @xmath4 leads to the different quantitative predictions about the glueball - quarkonia content of the three states . in ref .
@xcite @xmath19 is assumed and in ref .
@xcite @xmath20 is assumed .
the assumption , @xmath21 , is consistent with the prediction given by lattice qcd @xcite that the bare glueball state has a higher mass than the bare quarkonia states .
however , without the confirmation that which state , @xmath22 or @xmath23 , is the isovector member in the ground @xmath12 nonet , the level order of the @xmath24 and @xmath25 in the scalar sector perhaps still remains open . on one hand , if the @xmath23 is assigned as the isovector member in the ground @xmath12 nonet , since the @xmath23 with mass @xmath26 mev@xcite has a higher mass than the observed isodoublet scalar states @xmath27 with mass @xmath28 mev@xcite , according to the gell - mann- okubo mass formula@xcite one would expect @xmath29 . on the other hand ,
if the @xmath22 is assigned as the isovector member , one would expect @xmath30 , which is also consistent with that the strange quark @xmath31 has a higher mass than the non - strange quark @xmath32 or @xmath33 in a constituent quark picture .
we believe that neither @xmath30 nor @xmath29 seems to be ruled out in the scalar sector in the current situation . in this letter , based on the mixing scheme of the @xmath1 , @xmath2 and @xmath3 , we shall discuss the glueball - quarkonia content of the three states taking into account the two possible assumptions @xmath7 and @xmath34 and taking @xmath35 gev , @xmath36 gev , when @xmath37 changes from 1.2 to 1.5 gev , from eqs .
( 2)@xmath38(4 ) , we find that @xmath39 requires @xmath40 gev , which leads to @xmath41 in eqs .
( 5 ) , ( 6 ) . therefore
, the possibility of @xmath42 can be ruled out . ] .
this paper is organized as follows : in sect .
ii , the two - body hadronic decays of @xmath3 , @xmath2 and @xmath1 are investigated in the quarkonia - glueball mixing framework .
the results for two cases @xmath7 and @xmath34 are presented in sect .
our conclusions are reached in sect .
in the @xmath6 , @xmath5 , @xmath4 basis , the mass matrix describing the mixing of a glueball and quarkonia can be written as follows @xcite : @xmath43 where @xmath44 represents the flavor independent mixing strength between the glueball and quarkonia states .
the vanishing off - diagonal elements indicate that there is no direct quarkonia mixing which is assumed to be a higher order effect .
@xmath45 , @xmath24 and @xmath25 represent the masses of the bare states @xmath46 , @xmath47 and @xmath48 , respectively . here
we assume that the physical states @xmath49 , @xmath50 and @xmath51 are the eigenstates of @xmath52 with the eigenvalues of @xmath53 , @xmath54 and @xmath55 , respectively ( @xmath56 , @xmath57 and @xmath37 denote the masses of @xmath3 , @xmath2 and @xmath1 , respectively ) .
if one defines a @xmath0 unitary matrix @xmath58 which transforms the states @xmath46 , @xmath47 and @xmath48 into the physical states @xmath49 , @xmath50 and @xmath51 , then @xmath59 must be the diagonal matrix with the diagonal elements @xmath53 , @xmath54 and @xmath55 , from which one can get the following equations : @xmath60 the three physical states can be read as @xmath61 where @xmath62 with @xmath63^{-\frac{1}{2}}$ ] . in the above mixing scheme , for the hadronic decays of the @xmath1 , @xmath2 and @xmath3 , neglecting the possible glueball component in the final states , we consider the following coupling modes as shown in fig .
1 : the coupling of the quarkonia components of the three states to the quarkonia components of the final state pseudoscalar mesons , and the coupling of the glueball components of the three states to the quarkonia components of the final state pseudoscalar mesons . performing an elementary @xmath64 calculation @xcite , one can get the following equations : @xmath65 ^ 2 } { [ ( \sqrt{2}\alpha^2z_2 + 2\beta^2y_2)^2 + 2r\cos\theta(\sqrt{2}\alpha^2z_2 + 2\beta^2y_2 ) x_2+r^2x^2_2)]},\\ & ~&\frac{\gamma(f_0(1500)\rightarrow\pi^0\pi^0)}{\gamma(f_0(1500)\rightarrow \eta\eta)}=\frac{p_{\pi^0\pi^0}}{p_{\eta\eta}}~ \frac{[\frac{z^2_2}{2}+\sqrt{2}r\cos\theta z_2x_2+r^2x^2_2 ] } { [ ( \sqrt{2}\alpha^2z_2 + 2\beta^2y_2)^2 + 2r\cos\theta(\sqrt{2}\alpha^2z_2 + 2\beta^2y_2)x_2+r^2x^2_2]},\\ & ~&\frac{\gamma(f_0(1500)\rightarrow k\bar{k})}{\gamma(f_0(1500)\rightarrow
\pi\pi ) } = \frac{p_{kk}}{3p_{\pi\pi}}~ \frac{[(\frac{z_2}{\sqrt{2}}+y_2)^2 + 4r\cos\theta(\frac{z_2}{\sqrt{2}}+y_2)x_2 + 4r^2x^2_2 ] } { [ \frac{z^2_2}{2}+\sqrt{2}r\cos\theta z_2x_2+r^2x^2_2]},\\ & ~&\frac{\gamma(f_0(1710)\rightarrow\pi\pi)}{\gamma(f_0(1710)\rightarrow k\bar{k})}=3\frac{p^{\prime}_{\pi\pi}}{p^{\prime}_{kk}}~ \frac{[\frac{z^2_1}{2}+\sqrt{2}r\cos\theta z_1x_1+r^2x^2_1 ] } { [ ( \frac{z_1}{\sqrt{2}}+y_1)^2 + 4r\cos\theta(\frac{z_1}{\sqrt{2}}+y_1)+4r^2x^2_1]},\\ & ~&\frac{\gamma(f_0(1710)\rightarrow \eta\eta ) } { \gamma(f_0(1710)\rightarrow k\bar{k})}= \frac{p^{\prime}_{\eta\eta}}{p^{\prime}_{kk } } \frac{[(\sqrt{2}\alpha^2z_1 + 2\beta^2y_1)^2 + 2r\cos\theta(\sqrt{2}\alpha^2z_1 + 2\beta^2y_1)x_1+r^2x_1 ] } { [ ( \frac{z_1}{\sqrt{2}}+y_1)^2 + 4r\cos\theta(\frac{z_1}{\sqrt{2}}+y_1)x_1 + 4r^2x^2_1]},\\ & ~&\frac{\gamma(f_0(1710)\rightarrow \eta\eta^{\prime } ) } { \gamma(f_0(1710)\rightarrow k\bar{k})}= \frac{p^{\prime}_{\eta\eta^\prime}}{2p^{\prime}_{kk } } \frac{[4\alpha\beta(\frac{z_1}{\sqrt{2}}-y_1)]^2 } { [ ( \frac{z_1}{\sqrt{2}}+y_1)^2 + 4r\cos\theta(\frac{z_1}{\sqrt{2}}+y_1)x_1 + 4r^2x^2_1]},\\ & ~&\frac{\gamma(f_0(1370)\rightarrow \pi\pi ) } { \gamma(f_0(1370)\rightarrow k\bar{k})}=3\frac{p^{\prime\prime}_{\pi\pi}}{p^{\prime\prime}_{kk } } \frac{[\frac{z^2_3}{2}+\sqrt{2}r\cos\theta z_3x_3+r^2x^2_3 ] } { [ ( \frac{z_3}{\sqrt{2}}+y_3)^2 + 4r\cos\theta(\frac{z_3}{\sqrt{2}}+y_3)x_3 + 4r^2x^2_3]},\\ & ~&\frac{\gamma(f_0(1370)\rightarrow \eta\eta ) } { \gamma(f_0(1370)\rightarrow k\bar{k})}= \frac{p^{\prime\prime}_{\eta\eta}}{p^{\prime\prime}_{kk } } \frac{[(\sqrt{2}\alpha^2z_3 + 2\beta^2y_3)^2 + 2r\cos\theta(\sqrt{2}\alpha^2z_3 + 2\beta^2y_3)x_3+r^2x^2_3 ] } { [ ( \frac{z_3}{\sqrt{2}}+y_3)^2 + 4r\cos\theta(\frac{z_3}{\sqrt{2}}+y_3)x_3 + 4 r^2x^2_3]},\end{aligned}\ ] ] where @xmath66 , @xmath67 , @xmath68 is the mixing angle of isoscalar octet - singlet for pseudoscalar mesons ; @xmath69 ( @xmath70 , @xmath71 ) ( @xmath72 ) is the momentum of the final state meson in the center of mass system for the @xmath73 decays of the @xmath2 ( @xmath3 , @xmath1 ) ; @xmath74 represents the ratio of the effective coupling strength of the coupling mode ( b ) to that of the coupling mode ( a ) ; @xmath75 is the relative phase between the amplitude of the coupling mode ( b ) and that of the coupling mode ( a ) .
for the two - photon decays of the three states , one can get@xcite @xmath76
the decay data relating to the @xmath2 are as follows@xcite : @xmath79 the decay datum of the @xmath3 is @xcite @xmath80 apart from @xmath35 gev and @xmath36 gev , the central values of the masses of the @xmath3 and @xmath2 , respectively@xcite , we take the central values of the decay data mentioned above and @xmath81 @xcite as input . in this way
seven parameters , @xmath45 , @xmath37 , @xmath25 , @xmath24 , @xmath82 , @xmath75 and @xmath74 are unknown .
we perform numerically to solve the unlinear equations @xmath83 for the cases @xmath19 and @xmath8 , respectively .
the two solutions are presented in table i. table i shows that in both cases , the mass of the pure glueball is @xmath84 gev , which is in agreement with the lattice qcd simulations which give @xmath85 gev @xcite and @xmath86 gev @xcite for the scalar glueball mass .
in addition , from table i the masses of the pure @xmath48 and @xmath47 are close to the mass of the pure glueball in both cases , which implies that a large mixing would exist on @xmath1 , @xmath2 and @xmath3 .
we note that the mass of @xmath1 is determined to be the value of @xmath87 gev in the case @xmath77 and @xmath88 gev in the case @xmath78 , which is also consistent with @xmath89 gev estimated by pdg@xcite . for the case @xmath7 , the numerical form of the unitary matrix @xmath58 is @xmath90 the physical states @xmath49 , @xmath50 and @xmath51 can be read as @xmath91 which indicates that in the case @xmath77 , @xmath3 ( @xmath2 , @xmath1 ) contains about @xmath92 ( @xmath93 , @xmath94 ) glueball component , @xmath95 ( @xmath96 , @xmath97 ) @xmath98 component and @xmath99 ( @xmath100 , @xmath101 ) @xmath102 component . in the case
@xmath77 our results are consistent with the results given by ref .
@xcite . based on eqs .
( 7)@xmath38(15 ) as well as eqs .
( 19)@xmath38(21 ) , the numerical results relating to the hadronic decays of the @xmath3 , @xmath2 and @xmath1 in the case @xmath7 are shown in the table ii , and the two - photon decay width ratio for the three states is given by @xmath103 for the case @xmath8 , the numerical form of the unitary matrix @xmath58 is @xmath104 the physical states @xmath49 , @xmath50 and @xmath51 can be read as @xmath105 which indicates in the case @xmath78 , @xmath3 ( @xmath2 , @xmath1 ) contains about @xmath106 ( @xmath107 , @xmath108 ) glueball component , @xmath109 ( @xmath110 , @xmath111 ) @xmath98 component and @xmath112 ( @xmath113 , @xmath100 ) @xmath102 component .
similarly , based on eqs .
( 7)@xmath38(15 ) as well as eqs .
( 24)@xmath38(26 ) , the numerical results relating to the hadronic decays of the @xmath3 , @xmath2 and @xmath1 are shown in the table iii , and the two - photon decay width ratio for the three states is given by @xmath114 from eqs . (
18 ) and ( 23 ) , in both cases , a large mixing effect on the @xmath3 , @xmath2 and @xmath1 exists , which is consistent with our above results that in the scalar sector , pure glueball and quarkonia lie in a vicinal mass region .
the largest components of @xmath1 and @xmath2 are quarkonia and the quarkonia content of the @xmath2 and @xmath1 differs due to the different mass level order of the bar states @xmath48 and @xmath47 , which is consistent with the main property of the mass matrix ( 1 ) that upon mixing the higher mass bare state becomes more massive , while the lower mass bare state becomes less massive ( i.e. , the mass splitting between the higher and lower mass bare states increases as a result of the mixing)@xcite .
the largest component of @xmath3 is glueball , which supports the argument that @xmath13 is a mixed @xmath9 glueball having a large glueball component if its spin is determined to be 0@xcite .
furthermore , in both cases , the results exhibit destructive interference between the states @xmath48 and @xmath47 for @xmath2 while constructive interference between the states @xmath48 and @xmath47 for @xmath3 and @xmath1 , which is also consistent with the conclusion given by ref.@xcite . the predictions about the decays of the three states in two cases can provide a stringent consistency check of the two assumptions , therefore measurements of the above decay channels of the three states , especially the @xmath115 decay channel of @xmath3 and the @xmath116 decay channel of @xmath1 as well as the two - photon decays of the three states , can be most relevant in clarifying which assumption is really reasonable .
in addition , it is important to investigate the nature of the @xmath23 and @xmath22 , since the isovector scalar state and the isodoublet scalar state @xmath117 can set a natural mass scale of the ground scalar meson nonet .
in the scalar glueball - quarkonia mixing framework , we study the two - hadronic decays of @xmath1 , @xmath2 and @xmath3 considering the coupling quarkonia and glueball components of the three states to the quarkonia components of the final states pseudoscalar mesons . taking into account two possible assumptions about the mass level order of the bare states
@xmath46 , @xmath47 and @xmath48 , @xmath77 and @xmath78 , we obtain the quarkonia - glueball content of the three states by solving the unlinear equations .
our conclusions are as follows : 1 ) . in the scalar sector , the pure glueball and quarkonia have comparable masses and a significant mixing of glueball with the isoscalar mesons exists . 2 ) .
the largest component of @xmath3 is glueball ( about @xmath118 ) and the largest component of @xmath2 ( @xmath1 ) is quarkonia .
which flavor @xmath102 or @xmath98 is dominant component of @xmath2 ( @xmath1 ) depends on the mass level order of the pure states @xmath48 and @xmath47 .
the interference between @xmath48 and @xmath47 is destructive for @xmath2 while constructive for @xmath1 and @xmath3 . 4 ) .
the measurements for the @xmath115 decay channel of @xmath3 and the @xmath116 decay channel of @xmath1 as well as the two - photon decays of the three states would be relevant to judge which assumption is really reasonable .
moreover , the confirmation of the nature about the @xmath22 and @xmath23 would be useful to check the consistency of two assumptions .
we wish to thank drs .
l. burakovsky and p.r .
page for their useful comments on this work .
this project is supported by the national natural science foundation of china under grant no .
19991487 , no . 19677205 and grant no .
lwtz-1298 of the chinese academy of sciences . | based on a @xmath0 mass matrix describing the mixing of the scalar states @xmath1 , @xmath2 and @xmath3 , the hadronic decays of the three states are investigated . taking into account the two possible assumptions concerning the mass level order of the bare states
@xmath4 , @xmath5 and @xmath6 in the scalar sector , @xmath7 and @xmath8 , we obtain the glueball - quarkonia content of the three states by solving the unlinear equations .
some predictions about the decays of the three states in two cases are presented , which can provide a stringent consistency check of the two assumptions . |
we use a second - order langevin integrator @xcite to study the dynamics of the bilayer system in the classical regime .
the time evolution is given by the langevin equation @xmath100 here , @xmath101 corresponds to a wiener process and @xmath102 .
this equation can be solved numerically in second order accuracy with the following updates @xmath103 here , @xmath104 and @xmath105 are two independent noise variables from a wiener process and @xmath106 .
note that in the limit of zero - friction @xmath107 , the equations of motion reduce to a energy conserving ( micro - canonical ) propagator .
for completeness we repeat here the derivation of the numerical integrator used in the semi - classical regime given in ref .
the partition function is rewritten in terms of feynman path integrals @xmath108,\ ] ] with @xmath109 here , @xmath110 and @xmath111 .
the full hamiltonian consists of system hamiltonian and interaction hamiltonian @xmath112 the system part is @xmath113 ^ 2 \right).\ ] ] while sophisticated semi - classical methods have been developed to sample this hamiltonian without dissipation , the coupling to a thermal bath is challenging as the frequencies in the path space separate from the frequencies from the thermal bath . in the framework of path
integral langevin dynamics , this is solved by splitting the dynamics into two parts : first , the full hamiltonian evolves according to the path integral formalism in absence of a thermal bath . in a second step , the thermal bath acts on the slowest and higher normal modes of the paths separately , on the level of the equations of motion .
the energy conserving liouvillian part of the dynamics ( without dissipation ) is written as a sum of system and interaction @xmath114 here , @xmath115 $ ] and @xmath116 $ ] . using totter s theorem ,
the propagator for a small @xmath99 can be written as @xmath117 the semi - classical treatment of the heat bath requires the normal mode represent of the hamiltonian which is derived from the transformation @xmath118 where the matrix elements of @xmath119 are @xmath120 in the normal mode representation the hamiltonian reads as @xmath121 dissipation is then added on the level of the equations of motion in the space of normal modes .
this leads to the following equations of motion @xmath122 while @xmath123 and @xmath124 are associated with momentum and position dynamics in the path integral , @xmath125 is defined in the space of normal mode variables .
this requires four transformation between normal mode and path integral space in each time step .
the trotter step associated with @xmath124 is @xmath126 the evolution of the path integral with @xmath123 defined in normal mode space corresponds to @xmath127 and @xmath128 in the last step , @xmath125 describes thermalization in the normal mode space and leads to the updates @xmath129 , @xmath67 deep in the crystalline region @xmath130 ( a ) and close to the superfluid region @xmath131 ( b ) ( see fig . 2 in the main text ) .
the projection of the paths onto real space of layer @xmath19 ( red ) and layer @xmath20 ( blue ) indicate the extension of the wave function.,width=321 ] fig .
[ fig : paths]a and b show typical snapshots from a path integral simulations in the large @xmath1 limit ( classical ) compared to a phase where quantum fluctuations become important .
the dynamical slowing down of the glass phase corresponds to the development of a plateau in the time dependent intermediate scatter function @xmath132 ( classical ) and @xmath133 ( path integral ) .
this time dependent order parameter allows one to measure the relaxation time of a characteristic @xmath51-vector .
this characteristic @xmath51-vector ( @xmath134 ) is associated with the first peak in the static structure factor @xmath135 and taken as an input parameter for the time dependent intermediate scatter function of the kubo transformed density @xmath136 ( or @xmath60 ) .
the static structure factor @xmath137 is defined as @xmath138 here , @xmath139 is the position of the classical particle or in the quantum case the centroid of the path integral @xmath140 , where the average runs over all imaginary time slices . finally , the time - dependent density - density correlation function is @xmath141 here , @xmath142 and the integral runs over the imaginary path integral time .
the ensemble averages @xmath48 are taken over a large number of realizations .
: @xmath143 ( red ) . in the crystal with average particle distance @xmath144 ( density @xmath145 ) ,
the pair interactions localize the particles resulting in a decrease in the wavefunction extension.,width=302 ]
the extension of the wavefunction of a free particle with mass @xmath146 at temperature @xmath147 is @xmath148 in the path integral formalism , the average extension of the particles is calculated from @xmath149 to check for consistency we compare @xmath150 with the numerical results .
[ fig : wavelength ] depicts the particle extension as a function of the interaction strength @xmath151 . in the free particle limit
@xmath152 the result approaches exactly the expected value of @xmath150 .
the extension of the wavefunction can also be calculated from the width of the wigner function with a thermal characteristic function after integrating out the momentum .
the wigner function of a thermal state is @xcite @xmath153 \times \\ \nonumber & & \exp[-\frac{1}{2 } |\lambda|^2 ] \exp[\frac{-|\lambda|^2}{e^{\hbar \omega/ k_b t}-1 } ] \\
\nonumber & = & \frac{2}{\pi } \tanh[\frac{\hbar \omega}{2 k_b t } ] \exp\{-2 |\alpha|^2 \tanh[\frac{\hbar \omega}{2 k_b t}]\}\end{aligned}\ ] ] using @xmath154 with @xmath155 we rewrite eq .
[ equ : wigner ] in terms of position and momentum @xmath156 . integrating out
the momentum leads to @xmath157 } \exp\{-\tanh[\frac{\hbar \omega}{k_b t } ] \eta x^2\}\end{aligned}\ ] ] we interpret the wigner function as a quasi - probability and compare @xmath158 to the gaussian distribution of a thermal wavefunction @xmath159 \label{equ : px}\ ] ] expanding the @xmath160 in eq .
( [ equ : wignerx ] ) and comparing the coefficients of eq .
[ equ : wignerx ] and [ equ : px ] we find @xmath161 which is consistent with eq .
( [ equ : eta ] ) . in a crystal
the potentials of neighboring sites act as a trapping potential which localizes the wavefunctions ( see fig .
[ fig : wavelength ] ) .
the system is described by the hamiltonian eq . ( 1 ) in the main text . here
, we explicitly include the time dependent parameter @xmath162 which allows one to vary the interlayer separation after a quench from @xmath163 to @xmath164 for various quench velocities .
densities and temperature a chosen as in case anti - parallel dipoles depicted in fig . 1 in the main text .
for total quench times larger than @xmath165 the temperature is basically unchanged while a fast quench with @xmath166 leads to considerable non - adiabatic dissipative work .
the distributions is sampled from @xmath167 runs.,width=302 ] @xmath168 here , the sum over @xmath19 and @xmath20 indicates interactions between particles of the same layer and the last term the interaction between particles in different layers .
the layer separation is then linearly switched from @xmath169 to @xmath170 with @xmath171 by the protocol @xmath172 the non - equilibrium work performed during the quench is the sum of reversible work and irreversible work @xmath173 or explicitly @xmath174 the irreversible work from the quench increases with the speed of the protocol and heats up the system non - adiabatically .
the protocol is a follows : first , the molecules are prepared at temperature @xmath175 as described in the main text . by a rapid quench in the interlayer separation ,
the system is driven out of equilibrium . during the quench ,
the system evolves micro - canonically and work is performed from the change in the layer separation @xmath176 .
we define an effective temperature in the micro - canconical ensemble as the average kinetic energy @xmath177 here , @xmath178 is the kinetic energy and @xmath179 the dimensionality of the system . fig . [
fig : heating ] depicts the probability density of the effective temperature in the final state . for a slow quench with switching time larger than @xmath180
the temperature increase is negligible .
the local diffusion @xmath181 ^ 2 \rangle$ ] reveals a remarkable connection between structure and dynamics in the glass phase as was recently shown in colloidal glasses @xcite . in particular , for some parameters the mobility of individual particles @xmath182
is correlated with the local structure of the neighbors .
this local order is measured in two dimensions by the averaged hexatic order parameter @xmath183 with @xmath184 here , the sum runs over all neighbors @xmath185 including all particles within a distance @xmath186 of particle @xmath86 .
the dynamical @xmath182 and the structural order parameter @xmath187 are homogeneous in a liquid or a solid but are locally heterogeneous and anti - correlated in the glass phase .
[ fig : heterogeneity](b ) and ( c ) show the mean position of the particle wavefunctions where colors represent the value of @xmath187 and @xmath188 , respectively , with the normalized diffusion @xmath189 . in the path integral picture , structure and dynamics
is also correlated with the extension of the particles in imaginary time .
[ fig : heterogeneity](a ) shows the projection of the paths from the path integral simulation .
the extension of the quantum fluctuations are calculated from the second moment of the normal modes of the path integral @xmath190 which is consistent with the analytic result of the extension of the wigner function from a thermal quantum characteristic function ( see section iii ) .
the extension of a free particle at finite temperature is given by the thermal debroglie wavelength @xmath191^{1/2}$ ] while in a crystal the extension is smaller due to the interaction of neighboring particles acting as a trapping potential which is to first order quadratic .
in the glass , the structure and therefore the effective potential at each site is heterogeneous which is also apparent in the wavefunction extension .
10 e. vanden - eijnden and c. ciccotti , _ chem .
lett . _ * 429 * , 310 ( 2006 ) .
m. ceriotti , m. parrinello , t. e. markland , and d. e. manolopoulos , _ j. chem
. phys . _ * 133 * , 124104 ( 2010 ) . c. w. gardiner and p. zoller , _ quantum noise _ , springer ( 2010 ) .
t. e. markland , j. a. morrone , b. j. berne , k. miyazaki , e. rabani , and d. r. reichman , _ nat .
phys . _ * 7 * , 134 ( 2011 )
. t. e. markland , j. a. morrone , k. miyazaki , b. j. berne , d. r. reichman , and e. rabani , _ j. chem
. phys . _ * 136 * , 074511 ( 2012 ) .
h. tanaka , t. kawasaki , h. shintani , and k. watanabe , _ nature materials _ * 9 * , 324 ( 2010 ) .
k. watanabe , t. kawasaki , and h. tanaka , _ nature materials _ * 10 * , 512 ( 2011 ) . | we study the dynamics of a bilayer system of ultracold polar molecules , which exhibits classical and quantum glassy behavior , characterized by long tails in the relaxation time and dynamical heterogeneity . in the proposed setup ,
quantum fluctuations are of the order of thermal fluctuations and the degree of frustration can be tuned by the interlayer distance .
we discuss the possible observation of a glassy anomalous diffusion and dynamical heterogeneity in experiment using internal degrees of freedom of the molecules in combination with optical detection .
the recent experimental realization of cold ensembles of polar molecules @xcite has opened a new pathway to explore the dynamics of quantum many body systems with strong , long - range and anisotropic polar interactions@xcite . in combination with low dimensional trapping geometries , this allows the realization of stable many body phases , for example in the form of 1d wires or 2d pancakes @xcite , or as stacked pancakes representing coupled multilayer systems @xcite .
most of the experimental and theoretical studies of polar molecular gases have focused so far on the quantum degenerate regime , and on equilibrium phenomena , including superfluid and crystal phases , quantum magnetism , and topological phases ( for a review see @xcite ) . instead
, our interest below will be on non - equilibrium many - body dynamics .
we will show that a bilayer setup of ultracold polar bosonic molecules can feature a _ glass phase _ , and we present methods to prepare this phase and measure the relevant order parameters with tools available with present experimental setups . the unique feature and the theoretical challenge of glass physics with cold molecular ensembles is the possibility to study the crossover from classical to quantum glasses . .
( c ) sketch of the phase diagram of a single layer of bosonic polar molecules with the ratio between kinetic energy and interaction @xmath0 and @xmath1 from ref .
@xcite .
note , that the limit @xmath2 corresponds to the classical limit as the extension of the wavefunction vanishes with @xmath3 .
the arrow indicates the parameter region of interest .
( d ) for parallel dipoles [ panel ( a ) ] , the effective volume and symmetry of the defects ( dimers , trimers ) depends on the interlayer separation .
the ground state from a classical simulated annealing calculation at @xmath4 results in patterns include triangular , cubic and asymmetric structure where particles in the crystal layer ( red ) are displaced by the molecules in the defect layer ( blue ) .
( e ) for anti - parallel dipoles [ panel ( b ) ] defect patterns can range from triangular , cubic to fivefold symmetries.,width=283 ] a glass phase is characterized by a plateau in the relaxation time - scale , known as _ aging _ , with exponentially increasing tails that prevent the system from reaching its equilibrium state @xcite . in a structural glass , global reorganization to the equilibrium is prevented from geometric frustration as a result of the dynamics @xcite . while large reorientation is slow , on a local scale , relaxation can be fast , a phenomenon known as _ dynamical heterogeneity _
@xcite . in a classical glass
this relaxation dynamics is dominated by thermal fluctuations , and a considerable understanding of the glass transition has been gained from various theoretical methods @xcite and experimental model systems @xcite .
in contrast , the question of the influence of quantum fluctuations on the glass relaxation dynamics is far less well understood .
recent theoretical studies indicate , based on analytical @xcite and numerical @xcite methods , that quantum fluctuations can enhance but also inhibit the glass transition . in this letter , we propose and analyze a protocol to prepare and measure a glass phase in a bilayer setup of cold polar molecules ( see figs .
[ fig : illustration]a and b ) in the regime of the cross over from a classical to a quantum glass .
we assume ultracold molecules prepared in their electronic and rovibrational ground states , where a static electric field @xmath5 oriented perpendicular to the trapping layers in the @xmath6-plane induces an electric dipole moment @xmath7 .
thus the molecules will interact according strong , long - range and anisotropic dipole - dipole interaction .
the stacked pancake potential of figs .
[ fig : illustration]a and b can be realized with a 1d optical lattice with layer separation @xmath8 controllable by laser parameters . for strong confinement
the motion of the molecule within each layer can be described as an effective 2d dynamics , while the tunneling between adjacent layers will be suppressed by a sufficiently high barrier .
the hamiltonian for our bilayer system thus has the form @xmath9 . here
@xmath10 ( @xmath11 ) is the hamiltonian for the intralayer dynamics , @xmath12as sum of kinetic energy and a tranverse trapping potential , and the ( purely repulsive ) dipolar interaction with in - plane distance @xmath13 .
the interlayer interactions are @xmath14with @xmath8 the layer separation and @xmath15 the distance between the molecules with @xmath16 .
note , that our model allows for different dipole moments @xmath17 and @xmath18 in layers @xmath19 and @xmath20 , respectively .
different effective dipole moments in the two layers can be achieved e.g. by a gradient in the electric field @xmath21 or by engineering of the interactions using internal rotational degrees of freedom @xcite .
as we will see below , this feature is essential to observe clear signatures of a glass phase .
the basic steps of our protocol to study glassy dynamics are as follows .
( i ) we start with two uncoupled layers ( @xmath8 large ) , where layer @xmath19 is in a high density 2d crystal phase and layer @xmath20 in a low density gas phase . to prepare the initial state for the dynamics
, the layer separation is quenched to a small distance @xmath8 , which leads to the formation of defects with various symmetries and patterns depending on the value of @xmath8 ( see fig . [
fig : illustration ] ) .
( ii ) we evolve the system to find glassy dynamics , identified by dynamical heterogeneity , the deviation from the linear diffusion law and a plateau in the relaxation ( see fig . [
fig : resultspanel ] ) .
( iii ) in order to measure these features in a possible experiment with cold molecules we introduce _ marker molecules _ , i.e. molecules prepared in a different internal state , which allow tracking the time evolution of the particles ( see fig .
[ fig : marker ] ) . _ preparation of the initial state _ - we first prepare molecules in two uncoupled layers : a dipolar crystal of polar molecules in layer @xmath19 and a low density phase of defects in layer @xmath20 . for uncoupled layers ,
the phase of layer @xmath19 is described by parameters @xmath22 and @xmath23 , where @xmath24 . the first parameter is the ratio of the dipolar interaction @xmath25 for a given mean intralayer distance between the particles @xmath26 , and the kinetic energy @xmath27 , which in the dipolar crystal phase is @xmath28 .
the second parameter measures the temperature in units of the interaction . for bosons the relevant phase diagram is sketched in fig .
[ fig : illustration](c ) @xcite , which at low temperatures shows a low density ( 2d ) superfluid phase , a high - density crystal phase and a high temperature fluid phase .
the theoretically predicted conditions for the formation of a self - assembled crystal are , in the case of bosons , @xmath29 and @xmath30 @xcite .
we note that these conditions for @xmath31 and @xmath0 are a requirement for temperature , density and dipole moment . while in present experiments with krb the crystalline phase has not been realized , the ongoing effort in the laboratory to prepare cold ensembles of lics molecules @xcite with its large electric dipole moment @xmath32 and mass @xmath33 provides a promising candidate to obtain dipolar crystals . in this case for a given temperature @xmath34 the requirement on @xmath1 and @xmath0 corresponds to an effective ( induced ) dipole moment of @xmath35 and a lattice spacing of the self - assembled triangular crystal @xmath36 .
we assume that the density in layer @xmath20 is lower than that in layer @xmath19 while all other parameters are the same . in the following
we find it convenient to define length , energy and time in the reduced units @xmath26 , @xmath37 , @xmath38 , respectively . to prepare the initial state for the glassy dynamics , the two layers are brought together by a quench in the layer separation .
the speed of the quench is chosen such , that non - adiabatic effects are negligible @xcite .
we consider two cases : the dipoles can be aligned parallel [ see fig .
[ fig : illustration](a ) ] or anti - parallel [ see fig . [
fig : illustration](b ) ] . fig .
[ fig : illustration](d ) depicts the resulting defect patterns for parallel dipoles @xmath39 where molecules belonging to different layers attract each other .
this attraction leads to stable dimers or trimers @xcite if the two layers are separated by @xmath40 .
these composite defects are reminiscent of interstitials in a crystal . here
, the effective volume of the interstitials can be tuned by the ratio of the dipole moments @xmath41 and the interlayer separation @xmath8 .
for various layer separations ranging from @xmath42 to @xmath43 the resulting patterns include hexagonal , cubic and asymmetric defects .
we note that the binding energy of similar impurities has been recently studied in a fermionic system @xcite . in the anti - parallel case ,
@xmath44 , the interlayer interaction is repulsive . this can be achieved using internal rotational degrees of freedom @xcite .
the defect patterns depicted in fig .
[ fig : illustration](e ) include cubic , five - fold symmetric and triangular symmetries reminiscent of vacancy defects .
again , the effective volume of the defects can be tuned by the layer separation . in both cases ,
parallel and anti - parallel bilayer setup , the combined system is a mixture of effective defects and molecules .
this is an experimental realization of a binary mixture of dipoles which is a well known glass - forming liquid in the classical regime @xcite .
we note , that the anti - parallel setup allows one to induce defects with five - fold symmetry which may allow one to implement a classical spin liquid model , where fivefold symmetries are studied @xcite , in the quantum regime . _ glass dynamics _ - below we study the relevant glass order parameters from the dynamics of the proposed setup after the quench .
in the classical regime we use molecular dynamics simulations . in the quantum regime
we employ the recently developed dynamical path integral methods , which has been applied to the glass transition in ref .
@xcite ( for details on the numerical algorithms see sm @xcite ) .
we note that these methods include quantum effects but ignore the exchange statistics between the particles and that we are interested mainly in the transition from the classical phase to a phase where quantum fluctuations become important .
a glass phase is identified by a dramatic increase in the relaxation time which divergences with @xmath45 $ ] when approaching the ideal glass temperature @xmath46 @xcite .
this corresponds to an extended plateau in the self - intermediate scattering function @xcite , a two - time correlation function defined as @xmath47 } \rangle .
\label{equ : fskt}\ ] ] here , angular brackets @xmath48 denote ensemble averages over many realizations of the experiment after the quench .
the sum runs over all particles in both layers @xmath49 and @xmath50 is the absolute value of the characteristic @xmath51-vector of the coupled system corresponding to the first peak in the static structure factor @xmath52 } \rangle$ ] where @xmath13 is the distance in the @xmath6-plane .
note that eq .
( [ equ : fskt ] ) bears some similarity with the fourier transform of the density - density correlation function , however here , the sum runs over individual particles at different times . in the path integral
picture , the analogous order parameter to the classical case eq .
( [ equ : fskt ] ) reads as @xcite @xmath53 here , @xmath54 and the integral runs over the imaginary path integral time .
the semi - classical approximation allows one to follow the real time propagation of the path integral @xcite neglecting exchange ( for details see sm @xcite ) .
a crucial dynamical characteristic of the glass phase is its deviation from the linear diffusion laws .
in particular , the local diffusion of individual particles is spatially heterogeneous , a characteristic of the glass phase known as _ dynamical heterogeneity_. the local diffusion of particle @xmath55 is @xmath56 and the mean squared displacements @xmath57 . in the presence of quantum fluctuations the analogous measure in the path integral picture
is the root mean squared displacement of the projections of the path integral with @xmath58 and @xmath59 . of the combined system .
( b ) the relaxation of the time dependent structure factor @xmath60 for @xmath61 with @xmath62 ( black ) and @xmath63 ( red ) diverges when approaching the glass phase [ symbols as in ( a ) ] .
( c ) relaxation dynamics with initial temperature @xmath64 in the anti - parallel setup for various choices of effective @xmath1 of the combined system .
parameters are chosen such that the classical limit @xmath65 enters a glass phase and for smaller @xmath1 the system melts due to large quantum fluctuations ( blue ) .
this allows one to study the cross - over from a classical glass to a quantum phase .
( d ) relaxation dynamics as a function of the layer separation with effective @xmath66 and @xmath67 in the anti - parallel setup .
while for distance @xmath68 the system is a liquid due to large quantum fluctuations ( blue ) one reaches the glass phase when approaching @xmath69 ( black ) .
the number of particles used was @xmath70 and @xmath71 in the classical simulations and @xmath72 and @xmath73 in the path integral simulations using @xmath74 time - slices per particle and periodic boundary conditions .
averages are taken from @xmath75 independent runs . for lics
a time of @xmath76 in reduced units corresponds to @xmath77 milliseconds.,width=321 ] the unique feature of the bilayer system of ultracold molecules is the possibility to study glassy dynamics in both the classical and quantum regime , which are characterized by the dominance of thermal vs. quantum fluctuations , respectively .
this transition can be controlled by three tunable parameters : @xmath8 representing the layer separation , the dimensionless temperature @xmath0 , and the dipolar crystal parameter @xmath1 .
the case of large @xmath78 corresponds to the classical limit [ compare fig . [
fig : illustration](c ) ] .
the associated classical glass transition can be studied by varying @xmath0 with fixed @xmath1 , while varying the initial defects with @xmath8 [ see fig . [
fig : illustration](d ) and ( e ) ] .
below we will discuss this limit in fig .
[ fig : resultspanel](a , b ) .
the transition from the classical to the quantum regime is achieved by decreasing @xmath31 .
this increasing role of quantum fluctuations on the glassy dynamics is summarized in fig .
[ fig : resultspanel](c ) for various @xmath31 , while fig .
[ fig : resultspanel](d ) discusses the effect to defect patterns , as controlled by @xmath8 [ compare fig .
[ fig : illustration](d ) and ( e ) ] . the averaged root mean square displacement as a function of time is shown in fig .
[ fig : resultspanel](a ) .
the dynamics does not follow the einstein diffusion law @xmath79 , where @xmath80 is the dimension of the system , but develops a plateau larger than the sampling time , which corresponds to a divergence in the thermodynamic limit , for an effective @xmath0 of the combined system smaller than @xmath81 .
fig .
[ fig : resultspanel](b ) shows the according relaxation dynamics @xmath82 for a final layer separation @xmath61 in the classical regime .
both setups with parallel and anti - parallel dipoles show a growing plateau and enter a glass phase . including quantum fluctuations the relaxation dynamics of the system changes considerably .
fig .
[ fig : resultspanel](c ) depicts the relaxation dynamics @xmath83 calculated from a path integral simulation as a function of @xmath1 at a fixed @xmath0 . with increasing @xmath1 ,
the dynamics approaches that of the classical molecular dynamics simulation corresponding to @xmath84 where the system is glassy . by lowering @xmath1
, the system melts due to large quantum fluctuations . in the bilayer system
, one can also tune the defect size with the layer separation ( see fig .
[ fig : illustration ] ) .
fig .
[ fig : resultspanel](d ) shows a set of parameters , where one can reach the glass phase by variation of the layer separation .
in contrast to the classical model system of colloids @xcite , the actual dynamics after the initial thermalization corresponds to that of an isolated system ( micro - canonical ) . however , this does not influence the long tail part in the relaxation which is independent of the microscopic dynamics @xcite .
( red ) and layer @xmath20 ( blue ) depicted as the projection of the paths onto real space .
( b ) the parameter @xmath85 measures the relative order in the vicinity of particle @xmath86 ( for definition and details see sm @xcite ) .
here , regions of large order ( @xmath87 ) are regions of small classical diffusion and a small extension of the particle wavefunction .
regions of low order ( @xmath88 ) feature large diffusion and larger quantum fluctuations .
note , that dynamical heterogeneity is a more general feature of glasses , which is also present in systems without apparent partial crystallinity ( see sm @xcite ) .
( c ) snapshots of two possible initial choices of marker molecule positions ( green and orange ) at @xmath89 with @xmath90 .
the particle positions indicated by spheres represent the maximum of the path probability in layer @xmath19 ( gray ) and layer @xmath20 ( lightblue ) .
( d ) after evolving the system in time , the radius of gyration of marker molecules after time @xmath91 ( @xmath92 ) differs depending on the initial position . in the glass phase , due to dynamical heterogeneity , the growth of @xmath93 strongly depends on the choice of the initial position .
the cloud of maker molecules placed in an inactive region ( green ) of the glass follows the dynamics of an amorphous crystal @xmath94 while molecules in a mobile region ( orange ) show a fast increase in size ( inset).,width=302 ] _ measurement of dynamical heterogenity _ - one way to measure the anomalous diffusion as a glass feature is to consider a subensemble of polar molecules in a small spatial region ( _ marker molecules _ ) , which are addressed with a laser and transferred to another internal ( e.g. hyperfine ) state @xcite .
the position of this molecular cloud at a later time can be measured with optical techniques .
the ultimate tool for such position measurements will be provided by the quantum gas microscope for molecules , which - following the atomic case - will provide single site / single molecule detection with a resolution of the single site of an optical lattice @xcite .
dynamical heterogeneity can then be measured as follows : we mark molecules in the crystal layer a in a small region with laser waist - size @xmath95 .
after the system evolves in time , the positions of the _ marker molecules _ is measured , and the extension of the cloud of marker molecules is given by the radius of gyration @xmath96 . in a liquid a cloud of tagged particles spreads out linearly with time while in an amorphous solid it is constant @xmath97 .
most important , in a liquid and crystal , the dynamics is independent of the initial position of the makers molecules as the dynamics is spatially homogeneous .
however , in the glass phase due to dynamical heterogeneity the scaling differs significantly depending on the initial position ( see fig .
[ fig : marker ] ) .
as the initial size is @xmath98 from the laser waist , a series of single ( destructive ) measurements after time @xmath99 allows one to distinguish a glass from liquid or amorphous solid ( see fig . [
fig : marker ] ) . note that the marker molecules are quantum mechanically distinguishable which changes the dynamics in the deep quantum regime when exchange statistics is included , which is not relevant in the present case . in conclusion , we have shown that a bilayer system of polar molecules can , with properly chosen parameters , exhibit a glass phase in a regime where quantum fluctuations are of the order of thermal fluctuations .
thus ultracold ensembles of polar molecules could provide a widely tunable paradigmic model system for classical and quantum glass physics , where theory and experiment can meet in a yet unexplored parameter regime of glass physics , providing in particular a stimulus for the theoretical developments .
we thank k. binder and m. baranov for fruitful discussions . work supported by the austrian science fund ( fwf ) : p 25454-n27 , sfb foqus , marie curie initial training network coherence , and erc synergy grant uquam .
99 d. s. jin and j. ye , chem . rev . * 112 * , 4801 ( 2012 ) .
l. d. carr , d. demille , r. v. krems , and j. ye , new .
j. phys .
* 11 * 055049 ( 2009 ) .
k .-
k .
ni , s. ospelkaus , m. h. g. de miranda , a. peer , b. neyenhuis , j. j. zirbel , s. kotochigova , p. s. julienne , d. s. jin , and j. ye , science * 322 * , 231 ( 2008 ) .
s. ospelkaus , k .- k .
ni , g. quemener , b. neyenhuis , d. wang , m. h. g. de miranda , j. l. bohn , j. ye , and d. s. jin phys .
rev .
lett . * 104 * , 030402 ( 2010 ) .
b. yan , s. a. moses , b. gadway , j. p. covey , k. r. a. hazzard , a. m. rey , d. s. jin , j. ye , arxiv:1305.5598 ( 2013 ) .
j. deiglmayr , a. grochola , m. repp , k. mrtlbauer , c. glck , j. lange , o. dulieu , r. wester , and m. weidemller , phys .
rev .
lett .
* 101 * , 133004 ( 2008 ) .
j. g. danzl , e. haller , m. gustavsson , m. j. mark , r. hart , n. bouloufa , o. dulieu , h. ritsch , and h .- c .
ngerl , science * 321 * , 1062 ( 2008 ) .
m. a. baranov , m. dalmonte , g. pupillo and p. zoller , chem .
rev . * 112 * , 5012 ( 2012 ) .
t. lahaye , c. menotti , l. santos , m. lewenstein , and t. pfau , rep .
. phys . * 72 * , 126401 ( 2009 ) .
h. p. bchler , e. demler , m. lukin , a. micheli , n. prokofev , g. pupillo , and p. zoller , phys .
rev .
lett . * 98 * , 060404 ( 2007 ) .
g. e. astrakharchik , j. boronat , i. l. kurbakov , and y. e. lozovik , phys .
rev .
lett .
* 98 * , 060405 ( 2007 ) .
n. r. cooper and g. v. shlyapnikov , phys .
rev .
lett .
103 , 155302 ( 2009 ) .
m. lewenstein , a. sanpera , v. ahufinger , `` ultracold atoms in optical lattices : simulating quantum many - body systems '' , oxford university press , oxford ( 2012 ) .
g. pupillo , a. griessner , a. micheli , m. ortner , d .- w .
wang , and p. zoller , phys .
rev .
lett . * 100 * , 050402 ( 2008 ) .
d .- w .
wang , m. d. lukin , and e. demler , phys .
rev .
lett . * 97 * , 180413 ( 2006 ) .
a. pikovski , m. klawunn , g. v. shlyapnikov , and l. santos , phys .
rev .
lett .
105 , 215302 ( 2010 ) .
a. c. potter , e. berg , d .- w .
wang , b. i. halperin and e. demler , phys .
rev .
lett . * 105 * , 220406 ( 2010 ) .
m. knap , e. berg , m. ganahl , and e. demler , phys .
rev .
b * 86 * , 064501 ( 2012 )
. k. binder and w. kob , `` glassy materials and disordered solids '' , world scientific publishing , london ( 2011 ) .
p. g. debenedetti and f. h. stillinger , nature * 410 * , 259 ( 2001 ) .
h. shintani and h. tanaka , nature phys .
* 2 * , 200 ( 2006 ) . for glasses in amo setups as a result of disorder see : t. roscilde and j. i. cirac , phys .
rev .
lett . * 98 * , 190402 ( 2007 ) , l. sanchez - palencia and m. lewenstein , nature phys . * 6 * , 87 ( 2010 ) , s. gopalakrishnan , b. l. lev , and p. m. goldbart , phys .
rev .
lett . * 107 * , 277201 ( 2011 ) .
d. chandler and j. garrahan , annu .
rev .
phys . chem . *
61 * , 191 ( 2010 ) .
h. tanaka , t. kawasaki , h. shintani , and k. watanabe , nature mat .
* 9 * , 324 ( 2010 ) .
g. h. fredrickson and h. c. andersen , phys .
rev .
lett . * 53 * , 1244 ( 1984 ) .
s. whitelam , l. berthier , and j. p. garrahan , phys .
rev .
e * 71 * , 026128 ( 2005 ) .
b. coluzzi , g. parisi and p. verrocchio , phys .
rev .
lett . * 84 * , 306 ( 2000 ) .
v. lubchenko , and p. g. wolynes , adv .
chem .
phys .
* 136 * , 95 ( 2007 ) .
g. l. hunter and e. r. weeks , rep .
. phys . * 75 * 066501 ( 2012 ) .
e. r. weeks , j. c. crocker , a. c. levitt , a. schofield , d. a. weitz , science * 287 * , 627 ( 2000 )
. f. ebert , p. keim , and g. maret , eur .
phys .
j. e * 26 * , 161 ( 2008 ) . v. n. novikov and a. p. sokolov , phys .
rev .
lett . * 110 * , 065701 ( 2013 ) .
e. rabani and d. r. reichman , annu .
rev .
phys . chem . *
56 * , 157 ( 2005 ) .
t. e. markland , j. a. morrone , b. j. berne , k. miyazaki , e. rabani , and d. r. reichman , nature phys .
* 7 * , 134 ( 2011 ) .
t. e. markland , j. a. morrone , k. miyazaki , b. j. berne , d. r. reichman , and e. rabani , j. chem . phys . * 136 * , 074511 ( 2012 ) .
m. boninsegni , l. pollet , n. prokofev , and b. svistunov , phys .
rev .
lett . * 109 * , 025302 ( 2012 ) .
m. repp , r. pires , j. ulmanis , r. heck , e. d. kuhnle , and m. weidemller , and e. tiemann , phys .
rev .
a * 87 * , 010701 ( 2013);s .- k .
tung , c. parker , j. johansen , c. chin , y. wang , and p. s. julienne , phys .
rev .
a * 87 * , 010702 ( 2013 ) .
see supplementary material .
n. matveeva and s. giorgini , _ arxiv_:1306.5588 ( 2013 ) .
m. ceriotti , m. parrinello , t. e. markland , and d. e. manolopoulos , j. chem . phys . * 133 * , 124104 ( 2010 ) .
t. gleim , w. kob , and k. binder , phys .
rev .
lett . * 81 * , 4404 ( 1998 ) . j. ye , private communication .
for the atomic quantum gas microscope see : w. s. bakr , j. i. gillen , a. peng , s. flling and m. greiner , nature * 462 * , 74 ( 2009 ) ; c. weitenberg , m. endres , j. f. sherson , m. cheneau , p. schau , t. fukuhara , i. bloch , s. kuhr , nature * 471 * , 319 ( 2011 ) * supplemental material to + _ from classical to quantum glasses with ultracold polar molecules _ * |
1 truecm * fig . 1 :* diquark breaking component for nucleon - nucleus scattering and two inelastic collisions . * fig .
2 :* dp component for nucleon - nucleus scattering for two inelastic collisions . *
3 :* baryon yields for @xmath0 and @xmath1 collisions at 160 gev per nucleon calculated from eqs .
( [ 1e]-[5e ] ) using the values in table 1 before ( dotted line ) and after ( full line ) final state interaction .
the dashed lines are the results with final state interaction using the monte - carlo results for the conventional dp and string breaking components .
the data are from the wa 97 collaboration .
5 truemm * fig .
4 :* same as in fig . 3 for the ratios @xmath74 . 2 truecm
1 truecm 1 truecm values of the rapidity densities at @xmath27 in eqs .
( [ 1e],[3e]-[5e ] ) .
the values in the first lower rows are for central @xmath1 collisions ( @xmath75 , @xmath76 ) and those in the last two rows for @xmath28 .
three numbers in the first row and one in the second one have been adjusted to the data ( see main text ) . | we introduce a new formulation of the diquark breaking mechanism which describes with no free parameters the huge nuclear stopping observed in central nucleus - nucleus collisions . supplemented , in the dual parton model , with strings originating from diquark
antidiquark pairs from the nucleon sea , it gives a substantial increase of hyperon and antihyperon yields from @xmath0 to central @xmath1 collisions . compared to data , this increase is slightly underestimated for @xmath2 s and @xmath3 s but is five times too small for @xmath4 s . introducing final state interactions ,
a reasonable description of all baryon and antibaryon yields is achieved .
= 22 truecm = 15 truecm = -0.5 truecm = -2 truecm to 1 truecm * baryon and antibaryon production in hadronic and nuclear interactions * + 8 truemm * a. capella , e. g. ferreiro and c. a. salgado * + laboratoire de physique thorique et hautes energies + universit de paris xi , btiment 210 , f-91405 orsay cedex , france + 1 truecm 2 truecm lpthe orsay 99/06 january 1999 = 24 pt enhancement of strange baryon and anti - baryon production in central nucleus - nucleus collisions has been observed experimentally @xcite @xcite . in particular , a spectacular enhancement of @xmath3 s and @xmath4 s from @xmath0 to central @xmath1 collisions has been measured @xcite .
this phenomenon has been proposed as a signal of quark gluon plasma formation @xcite .
on the other hand explanations based on string models have also been proposed @xcite . in ref .
@xcite enhancement of @xmath2 s and @xmath5 s was described in the dual parton model ( dpm ) @xcite with final state interaction .
an important drawback of @xcite is the lack of nucleon stopping which would increase the net baryon yield at mid rapidities . as a consequence a very strong effect due to final state interaction was needed , as well as a rather large value of the fraction of strange quarks in the nucleon sea .
a mechanism of nuclear stopping based on diquark breaking ( db ) has been known for a long time @xcite-@xcite .
recently , it has been introduced in nucleus - nucleus interactions @xcite-@xcite .
the purpose of this paper is threefold .
first , we propose a new formulation of the db stopping mechanism of @xcite-@xcite , which allows to describe net baryon stopping without any dynamical parameter governing the db probability .
second , we show that the dpm with db and strings with diquarks antidiquark pairs ( @xmath6-@xmath7 ) from the sea at their ends , allows to describe most of the observed yields of @xmath2 , @xmath5 , @xmath3 and @xmath8 but significantly underestimates the production of @xmath4 s and @xmath9 s .
third , using the corresponding baryon densities as initial conditions for final state interaction we achieve a reasonable description of all baryon and antibaryon production .
5 truemm we describe baryon stopping via the mechanism shown in fig . 1 the same as fig . 5 of @xcite .
however , contrary to @xcite we assume here that there is no dynamical suppression associated to db i.e . the weight of the db diagram of fig .
1 is the same as the conventional dp one of fig . 2 .
more precisely , in the case of a single inelastic collision per nucleon ( two string configuration ) , two valence quarks of the nucleon are sitting together at a string end .
we assume that they form a sort of compact object which hadronizes in the conventional way ( dp mechanism ) .
however , in the presence of several inelastic collisions we assume that the net baryon number ( string junction ) has equal probability of finding itself in any of them . if it follows the valence diquark ( fig .
2 ) we have the conventional dp situation . in the other cases it joins with sea quarks producing a much slower baryon ( db mechanism of fig . 1 ) .
the rapidity distribution of the net baryon production @xmath10 in @xmath11 collisions is then given by [ 1e ] dn^aab dy ( y ) = |n_a |n . here
@xmath12 is the average number of participants in each nucleus and @xmath13 the average number of collisions .
@xmath14 is given by the conventional dp hadronization mechanism and @xmath15 is the rapidity density given by the db component which behaves @xcite as @xmath16 $ ] . at high energies has been proposed in @xcite , @xcite .
this has very important consequences for the net baryon number at @xmath17 at lhc @xcite . a new component consisting on a diquark which contains a quark from the nucleon sea also has been introduced in @xcite .
however , in this case the diquark hadronizes as a valence diquark producing baryons mainly in the fragmentation regions . ]
both distributions depend , due to energy - conservation , on the average number of collisions per nucleon @xmath18 .
note that for @xmath19 , only the dp component is present and we recover the usual scaling in the number of participants .
baryon number conservation implies that both dp and db rapidity distributions are normalized in such a way that their integral over @xmath20 is equal to two .
following @xcite we have [ 2e ] dn_db^b dy ( y ) = c_|n/|n_a ( z_+^1/2 ( 1 - z_+)^_|n/|n_a + z_-^1/2 ( 1 - z_-)^_|n/|n_a ) where @xmath21 . the same shape will be used for all baryon species .
two different ansatzs for the power @xmath22 were proposed in @xcite which give for central @xmath1 collisions @xmath23 a value of @xmath24 at @xmath17 in the range 0.5 to 0.60 . in order to reproduce the data of the na49 collaboration @xcite for @xmath25-@xmath26 at @xmath27
, we need a value 0.48 .
our analysis will concentrate on mid - rapidities and this value will be used throughout this paper .
eq .
( [ 1e ] ) also applies to @xmath28 collisions putting @xmath29 .
however , this applies only at low energies where the two string configuration dominates . at high energies
the second term in ( [ 1e ] ) is present and increases the net baryon stopping .
unfortunately , the only available data at isr do not allow to either confirm or rule out the presence of the db component .
however , the effect of this component is quite important in @xmath30 collisions where the @xmath0 data of @xcite confirm its presence ( see below ) . in dpm , the extension of eq .
( [ 1e ] ) to @xmath30 collisions is straightforward .
a good approximation to the exact formula , valid at @xmath31 , is [ 3e ] dn^pa b dy(0 ) = |n 2 dn^ppb dy(0 ) + 1 2 ( dn^b dy ( 0 ) ) _ |n . here
@xmath32 and @xmath33 is given by eq .
( [ 1e ] ) with @xmath34 .
eq .
( [ 3e ] ) is self - evident . on the nucleus side
we have @xmath13 diquarks which hadronizes as in @xmath28 collisions ( the average number of collisions per nucleon at present energies is close to one in both cases ) . on the proton side
, we have one diquark which hadronizes as in @xmath11 collisions .
( the average number of collisions of this nucleon in @xmath0 and in central @xmath1 collisions is practically the same ) .
eq .
( [ 3e ] ) is very useful since it gives the rapidity density of the net baryon in @xmath0 , in terms of the same quantities which appear in @xmath1 ( eq . ( [ 1e ] ) ) plus the corresponding one for @xmath28 .
5 truemm following @xcite the rapidity distributions of antibaryons in @xmath11 collisions is given by [ 4e ] dn^aa |b dy(y ) = |n_a ( dn^|b_string dy(y ) ) _
|n/|n_a + ( |n - |n_a ) ( dn^|b_sea dy(y ) ) _ |n/|n_a .
here @xmath35 denotes the conventional pair production resulting from @xmath6-@xmath7 production in the string breaking process . at present energies the @xmath36-@xmath37 strings have too little energy for baryon pair production ; practically all their production occurs in @xmath38-@xmath36 strings , hence the scaling in the number of participants .
this gives a much too small antibaryon yield .
for this reason we have introduced @xcite the second term in ( [ 4e ] ) which corresponds to pair production in strings with a @xmath6 or @xmath7 from the nucleon sea at one of their ends .
the average number of @xmath6-@xmath7 pairs scales as @xmath13-@xmath12 .
note that for @xmath39 , the scaling in the number of participants is converted into a scaling in the number of collisions . for @xmath30 collisions
the equivalent of expression ( [ 3e ] ) , valid at @xmath27 , is [ 5e ] dn^pa|b dy(0 ) = |n 2 dn^pp |b dy(0 ) + 1 2 ( dn^|b dy ( 0))_|n here @xmath32 and @xmath40 is given by eq .
( [ 4e ] ) with @xmath34 .
5 truemm let us concentrate first on the non - conventional contributions ( second terms in eqs .
( [ 1e]-[3e ] ) and ( [ 4e]-[5e ] ) ) which are numerically very important since they contain the factor @xmath13-@xmath12 .
actually , @xmath41 has been computed in @xcite .
however , as stated there , there is a huge uncertainty on its absolute normalization . in view of that
, we are going to determine it from @xmath30 data .
since data @xcite for multi - strange hyperons are only available in the rapidity range @xmath42 we shall restrict our analysis to this region .
turning to the relative yields of the different baryon species , we note that in @xmath41 the baryons and antibaryons are produced from three sea quarks .
therefore the factors controlling the relative yields are : @xmath43 , @xmath43 , @xmath44 , @xmath45 , @xmath45 and @xmath46 for @xmath47 , @xmath48 , @xmath49 , @xmath50 , @xmath51 and @xmath4 , respectively .
moreover , we use for both baryons and antibaryons the conventional relation @xmath52 , which is obtained after resonance decay .
we take @xmath53 and @xmath54 . in this way
the strangeness suppression factor has the conventional value @xmath55 and the sum of all baryon yields add up to @xmath56 .
however , this fixes only the relative normalization of the various ( anti ) baryon species and there is one free parameter which provides the absolute normalization which is fixed by the data .
the obtained values are given in table i. the situation is quite different for the new component @xmath57 . here
net baryon conservation fixes the absolute normalization and therefore there is in principle no free parameter since the strangeness suppression factors are the same as above .
however , it turns out that proceeding in this way the strange baryon yields in @xmath30 are overestimated especially @xmath3 s and @xmath4 s . a possible explanation of this fact is that , at present energies , it may happen that the baryon is not formed out of three sea quarks as in fig . 1 , but , due to lack of phase space , the valence quark at the string end
is picked up instead in agreement with tevatron @xcite and sps collider @xcite data . ] .
in this case , the relative yields of @xmath2 s of @xmath3 s will be reduced and that of @xmath4 s will be zero . in any case , we take here a phenomenological attitude . we determine the strange net baryon yields at @xmath27 from the data .
we have thus three free parameters @xmath2-@xmath5 , @xmath51-@xmath58 and @xmath4-@xmath9 plus a forth parameter for the absolute normalization of antibaryons as discussed above .
we arrive in this way to the values listed in table i. finally , the conventional components @xmath14 and @xmath59 have been obtained in two different ways : using the formalism of ref . @xcite and using the qgsm @xcite monte carlo without string fusion collisions @xcite , have been tuned up , and its results in @xmath28 collisions are now in agreement with the calculations of @xcite and with the experimental data ( see table 1 in ) . ] . in the former case ,
the obtained values are given in table i. it should be noted that the contribution of these conventional components to the final results is significantly smaller than the one of the new components , since the latter are proportional to @xmath13-@xmath12 . with the values given in tables i
we obtain the results shown in figs . 3 and 4 .
the @xmath30 data are well reproduced for all baryon species . in @xmath1 ,
the @xmath60 and most of the @xmath61 yield are reproduced .
however , the @xmath62 yield is somewhat underestimated and the @xmath63 one is five times smaller than the measured one . in an attempt to solve this problem , we use our expressions ( [ 1e ] ) and ( [ 4e ] ) as initial baryon densities for final state interaction .
we consider the following channels [ 6e ] n k , n k , k , k , k , plus the corresponding ones for antiparticles - channel .
they have not been included .
all reactions corresponding to quark diagrams with light quark annihilation and @xmath64-@xmath65 production have been included with identical cross - section .
we have neglected all strangeness exchange reactions @xmath66 , etc .
although , at threshold , the cross - sections are larger than those of the non - strangeness exchange reactions ( ) , this is no longer the case for the averaged cross - sections @xmath67 ( see ) . channels ( ) are thus dominant due to the relations @xmath68 between baryon densities . ] .
the final state interaction is treated in the same way as in ref .
@xcite .
the products of densities in ( [ 6e ] ) are taken at fixed values of @xmath69 and @xmath70 .
for the baryon densities , the dependence on these variables is contained in the geometrical factors @xmath12 and @xmath13 in eqs .
( [ 1e ] ) and ( [ 4e ] ) .
they are computed in the glauber model with saxon - woods profiles .
the pion densities and the initial and final times are the same as in @xcite .
the only difference resides in the fact that , in the case of @xmath71 suppression , there is a single channel and the @xmath71 appears both in the initial and final state . in this case , the differential equations for the gain and loss of particles can be solved analytically and a simple exponential form is obtained . in the present case , the gain and loss differential equations for the solution of the coupled channels ( [ 6e ] ) has to be obtained numerically .
we have only one free parameter : the cross - section @xmath67 averaged over the momentum distribution of the incoming particles which is taken to be the same for all reactions ( the relative velocity between the interacting particles is included in @xmath67 ) .
we take @xmath72 fm .
( a comparable value has been obtained in @xcite ) .
it is clear that the relative enhancement of the various baryon and antibaryon yields resulting from the interactions ( [ 6e ] ) will be the largest for @xmath4 s and the smallest for @xmath2 s .
this is due to the fact that @xmath2 s are produced in some reactions and destroyed in others . on the contrary @xmath4 s
are produced and never destroyed .
the results are shown in figs . 3 and 4 .
a reasonable description of all baryon and antibaryon is obtained .
however , the ratio @xmath73 in @xmath1 collisions is slightly underestimated .
5 truemm in conclusion , we have introduced a new formulation of the diquark breaking mechanism which describes baryon stopping with a diquark breaking probability equal to one . incorporated in dpm , together with strings originating from diquark - antidiquark pairs from the nucleon sea and final state interactions
we have obtained a reasonable description of all baryon and antibaryon yields at mid - rapidities in @xmath28 , @xmath0 and @xmath1 collisions .
3 truemm it is a pleasure to thank a. kaidalov for many useful discussions and suggestions .
we also thank c. pajares for comments on the idea of this work .
one of the authors ( a.c . )
thanks j. a. casado for a correspondence on subjects related to the present work .
he also acknowledges partial support from a nato grant outr.lg 971390 .
e.g.f .
thanks fundacin ramn areces from spain and c.a.s fundacin caixa galicia from spain for financial support . |
several recent results show a relationship between properties of a polynomial or of a formal power series in say @xmath1 non - commuting indeterminates and the properties of the matrix - valued functions on @xmath1-tuples of matrices of all sizes obtained from this polynomial or this formal power series by matrix substitution ; e.g. , see @xcite .
the objective of this paper is to establish a result of this type in the framework of positive kernels @xcite .
we recall that if @xmath2 is any set and @xmath3 is a hilbert space , an operator valued function @xmath4 denotes the @xmath5-algebra of bounded linear operators on @xmath3 . ]
is called a positive @xmath6-valued kernel on @xmath2 if for any integer @xmath7 , any points @xmath8 and any vectors @xmath9 , one has @xmath10 alternatively , there exists an auxilliary hilbert space @xmath11 and an operator valued function @xmath12 denotes the banach space of bounded linear operators from @xmath11 to @xmath3 . ]
so that the factorization @xmath13 holds for all @xmath14 .
let @xmath15 denote the free semigroup with @xmath1 generators @xmath16 and unit element @xmath17 . for @xmath18 and @xmath19 from @xmath15 ( words in the alphabet
@xmath20 the semigroup multiplication is the concatenation : @xmath21 and @xmath22 plays a role of the empty word : for any @xmath23 , @xmath24 for a word @xmath25 define its transpose @xmath26 and its length @xmath27 , and define @xmath28 .
let @xmath29 be non - commuting indeterminates , and @xmath30 .
set @xmath31 for any vector space @xmath32 we denote by @xmath33 the space of all formal power series in @xmath29 with coefficients in @xmath32 . a formal power series in @xmath0 non - commuting indeterminates @xmath34 and @xmath35 , @xmath36 is called a _ positive non - commutative kernel _
if @xmath37 is a positive @xmath6-valued kernel on @xmath38 , i.e. , for any integer @xmath7 , any words @xmath39 and any vectors @xmath9 , one has @xmath40 alternatively , there is an auxiliary hilbert space @xmath11 and a non - commutative formal power series @xmath41 such that @xmath42 where we use the convention @xmath43 this notion of positivity was studied extensively in @xcite .
commutes with @xmath44 . for our purposes
it is more convenient to deal with @xmath0 non - commuting indeterminates ; notice that it does not really matter since in all our formulae every @xmath44 appears to the right of every @xmath45 , in every monomial . ]
let @xmath46 , i.e. , @xmath47 is an @xmath1-tuple of @xmath48 matrices with complex entries , and @xmath25 .
set @xmath49 let @xmath50 .
define the set @xmath51 of @xmath1-tuples @xmath47 of @xmath48 matrices with complex entries , which are _ jointly nilpotent of rank at most _
@xmath52 , i.e. , @xmath53 for all @xmath23 such that @xmath54 .
define also the set @xmath55 of _ jointly nilpotent _
@xmath1-tuples of @xmath48 matrices .
[ main ] a formal power series @xmath56 as in is a positive non - commutative kernel if and only if @xmath57 where @xmath58 , @xmath59 ( notice that the sum is therefore finite ! ) , is a positive @xmath60-valued kernel on @xmath61 for every @xmath62 , i.e. , for any integer @xmath7 , any jointly nilpotent @xmath1-tuples of matrices @xmath63 and any vectors @xmath64 , one has @xmath65
one direction of the theorem is obvious .
if @xmath66 is a positive non - commutative kernel then we have a factorization @xmath67 for some formal power series @xmath68 as in and some hilbert space @xmath11 .
therefore for any @xmath69 we have @xmath70 where @xmath71 ( notice that the sum is finite ) .
therefore for each @xmath62 we have that @xmath72 is a positive @xmath60-valued kernel on @xmath73 .
section [ s : proof1 ] contains the proof of the other direction of the theorem .
notice that by using nilpotent matrix substitutions one avoids any convergence assumptions on the formal power series .
there is also a `` convergent '' version of theorem [ main ] .
a formal power series @xmath74 as in is called _ convergent _ if for every @xmath62 there exists a connected open neighbourhood @xmath75 of @xmath76 in @xmath77 such that the series @xmath72 as in converges uniformly on compact subsets of @xmath78 in the norm of @xmath79 ( we assume that @xmath80 and @xmath81 are ordered , say , lexicographically ) .
[ conv ] a convergent formal power series @xmath56 ( with respect to a set of neighbourhoods @xmath75 of @xmath76 in @xmath82 ) is a positive non - commutative kernel if and only if @xmath72 is a positive @xmath60-valued kernel on @xmath75 for every @xmath62 .
we prove this theorem in section [ s : proof2 ] . in the `` nilpotent '' case
there is a considerable strengthening of theorem [ main ] .
the authors thank the anonymous referee for suggesting this result and its proof .
[ impr ] a formal power series @xmath56 as in is a positive non - commutative kernel if and only if for every @xmath62 and @xmath83 , @xmath84 is a positive semidefinite operator , i.e. , for any @xmath85 one has @xmath86 one direction of the theorem is obvious due to the remark following the formulation of theorem [ main ] and the fact that the operator @xmath87 is positive semidefinite .
section [ s : proofs3,4 ] contains the proof of the other direction of theorem [ impr ] .
while theorem [ impr ] implies theorem [ main ] , section [ s : proof1 ] contains an independent proof of theorem [ main ] since the argument of this proof is also used in the proof of theorem [ conv ] .
let us remark also that the statement analogous to the one of theorem [ impr ] does not hold in the `` convergent '' case .
indeed , set @xmath88 ( here @xmath89 ) .
clearly , @xmath90 for a matrix @xmath47 ( of any size ) close to the origin , however @xmath66 is not a positive kernel since its matrix of coefficients @xmath91 is indefinite .
theorem [ impr ] implies a factorization result for a class of non - commutative polynomials . for a vector space @xmath92
we denote by @xmath93 the space of all polynomials in non - commuting indeterminates @xmath29 with coefficients in @xmath92 , i.e. , a subspace in @xmath94 consisting of formal power series with a finitely supported set of coefficients .
an operator - valued polynomial in @xmath0 non - commuting indeterminates @xmath95 and @xmath96 is called called _ hereditary _ if it is of the form @xmath97 where @xmath98 is a hilbert space , i.e. , every @xmath44 appears to the right of every @xmath45 , in every monomial .
[ fact ] a hereditary polynomial @xmath99 as in satisfies @xmath100 for all @xmath101 ( @xmath102 ) if and only if there exist an auxiliary hilbert space @xmath103 ( of dimension at most @xmath104 ) and a polynomial @xmath105 such that @xmath106 furthermore , for this factorization to hold it is enough to assume that @xmath100 holds for all @xmath107 .
we deduce this theorem in section [ s : proofs3,4 ] .
theorem [ fact ] is an analogue of the main result of @xcite ( for related results and generalizations see @xcite and @xcite ) , for hereditary non - commutative polynomials and jointly nilpotent test matrices . in view of the results of this paper
it is of interest to investigate the relationship between the non - commutative formal reproducing kernel hilbert space associated to a positive non - commutative kernel ( such spaces are studied in @xcite ) and the reproducing kernel hilbert spaces associated to the kernels obtained from the non - commutative kernel by matrix substitutions .
for the proof of theorem [ main ] we need the following lemma .
[ lem ] let @xmath108 be a hilbert space and let an @xmath109-valued polynomial function @xmath110 be a positive kernel on @xmath111 . here for @xmath112 and @xmath113 we set @xmath114 and @xmath115 .
then the matrix @xmath116 ( for a certain order on @xmath117 ) defines a positive semidefinite operator on @xmath118 .
this holds also for sesquianalytic rather than polynomial kernels ( of course in this case one has to consider the matrices @xmath119 for any integer @xmath120 ) and the converse is true as well ; notice that these facts are a `` commutative convergent '' analogue of theorem [ main ] .
they are well known for @xmath89 ( see , e.g. , ( * ? ? ?
* lemma 1.1.5 ) ) and the proofs in the general case are analogous .
we provide a proof of the lemma for the sake of completeness . for an arbitrary @xmath121
we have @xmath122 where @xmath123 is the normalized lebesgue measure on @xmath124 . for an arbitrary @xmath62 ,
let @xmath125 be a partition of @xmath124 such that @xmath126 and let @xmath127 , @xmath128 , be arbitrary points .
then @xmath129 as desired . as we already said in section [ s : intro ] , one direction of this theorem is obvious , and thus only the other one is left to prove .
assume that a formal power series @xmath130 as in is such that the function @xmath72 as in is a positive @xmath60-valued kernel on @xmath61 for every @xmath62 .
let @xmath131 , @xmath132 .
set @xmath133 .
define a linear operator @xmath134 on the space @xmath135 by its action on the basis vectors @xmath136 where @xmath137 , @xmath138 are the standard basis vectors in @xmath139 : @xmath140 the operator @xmath134 is represented in the basis @xmath141 , by a @xmath48 permutation matrix .
let @xmath142 , @xmath143 , be matrices of the form @xmath144 define @xmath145 for @xmath112 , set @xmath146 for @xmath147 , one has @xmath148 ( in the case @xmath149 , the last term in the tensor product above disappears , and @xmath150 ) , or , on the basis vectors , @xmath151 let us remark that since @xmath152 , @xmath153 , are scalars , and thus commute , we may rewrite @xmath154 as @xmath155 , where @xmath156 is the _ abelianization map _ : @xmath157 with the non - negative integer number @xmath158 equal to the numbers of times that the letter @xmath159 appears in the word @xmath160 .
thus , we have @xmath161 for @xmath162 , one has @xmath163 since @xmath164 for any @xmath165 .
thus , @xmath166 , and by the assumption , @xmath167 is a polynomial @xmath168-valued function which is a positive kernel on @xmath111 . by lemma [ lem ] ,
the coefficients of this kernel serve as the operator blocks @xmath169 of the matrix @xmath116 ( for some order on @xmath117 ) which represents a positive semidefinite operator on the hilbert space @xmath170 . therefore , for arbitrary vectors @xmath171 where @xmath172 we can write down : @xmath173 let us calculate each of the sums @xmath174 , @xmath175 and @xmath176 separately : @xmath177
@xmath178 @xmath179 now , for @xmath180 , and @xmath181 , set @xmath182 and @xmath183 then the inequality @xmath184 turns into @xmath185 since the integer @xmath132 and the vectors @xmath186 are chosen arbitrarily , this means that @xmath66 is a positive non - commutative kernel ( for the cases @xmath187 and @xmath188 the inequality above follows now automatically ) .
for the proof of theorem [ conv ] we need the following lemma .
( for a result of a similar nature in a commutative setting , see ( * ? ? ?
* proposition 2.4 ) . )
[ l : conv ] an arbitrary convergent positive non - commutative kernel @xmath56 admits a factorization @xmath189 with a convergent formal power series @xmath190 , for an auxiliary hilbert space @xmath103 .
( a formal power series @xmath190 is called convergent if for every @xmath62 there exists a connected open neighbourhood @xmath191 of @xmath76 in @xmath77 such that the series @xmath192 as in converges uniformly on compact subsets of @xmath191 in the norm of the banach space @xmath193 . )
let @xmath194 , where @xmath75 is a neighbourhood of @xmath76 in @xmath77 associated with @xmath66 as in the definition of a convergent non - commutative formal power series which was given in section [ s : intro ] .
then the series @xmath195 converges in the operator norm , and hence there exists a constant @xmath196 such that @xmath197 for all @xmath198 , where @xmath199 , and for a @xmath200 the @xmath1-tuple of integers @xmath201 is defined in section [ s : proof1 ] .
set @xmath202 , and @xmath203 .
since @xmath66 is a non - commutative positive kernel , it admits a factorization @xmath189 with a formal power series @xmath190 , for an auxiliary hilbert space @xmath103 .
this means that @xmath204 for every @xmath198 .
therefore we have @xmath205 we will use now an argument similar to the one used in the proof of abel s lemma in the setting of several complex variables ( see , e.g. , @xcite ) .
denote by @xmath191 the open neighbourhood of @xmath76 in @xmath77 defined by the inequality @xmath206 .
then for @xmath207 we have : @xmath208 therefore for every @xmath62 the power series @xmath209 converges absolutely and uniformly on compact subsets of @xmath191 in the norm of @xmath210 , i.e. , @xmath68 is a convergent formal power series .
[ r : nbhd ] it is clear from the proof of lemma [ l : conv ] that the assumption of convergence of the series @xmath211 at only one point @xmath212 is already sufficient for the conclusion of this lemma . moreover , this guarantees the existence of neighbourhoods @xmath213 of the same size " , i.e. , with the same bounds on the norms of @xmath214 , such that the series @xmath192 converges uniformly and absolutely on compact subsets of @xmath191 .
assume that the convergent formal power series @xmath56 ( with respect to a set of neighbourhoods @xmath75 of @xmath76 in @xmath82 ) is a positive non - commutative kernel .
then by lemma [ l : conv ] @xmath66 admits a factorization @xmath67 for some convergent formal power series @xmath68 as in , i.e. , for every @xmath62 there exists a neighbourhood @xmath191 of @xmath76 in @xmath77 , such that the series @xmath215 converges uniformly on compact subsets of @xmath191 ( in the norm of the banach space @xmath216 ) . since we can replace @xmath191 by @xmath217 for all @xmath62 , without loss of generality we may assume that @xmath218 . in this case
@xmath72 is a positive @xmath79-valued kernel on @xmath75 .
( we first see that @xmath219 is a positive kernel on @xmath191 , then observe that due to uniform convergence of the series for @xmath72 on compact subsets of @xmath75 the function @xmath72 is sesquianalytic , and then use the fact that for sesquianalytic kernels the positivity property propagates to an arbitrarily large region on which the kernel is sesquianalytic ; see ( * ? ? ?
* theorem 1.1.4 ) in the one dimensional case
the general case is completely analogous . )
conversely , if @xmath74 is a convergent formal power series , and if @xmath72 is a positive @xmath60-valued kernel on @xmath75 for every @xmath62 , then @xmath66 is a positive non - commutative kernel .
the proof is the same as in section [ s : proof1 ] , except that the polynomial @xmath220 is guaranteed to be a positive kernel only on a neighbourhood of @xmath76 in @xmath111 ; it is clear from the proof of lemma [ lem ] that this still implies positive semidefiniteness of the block operator matrix @xmath221 ( alternatively , one can use the propagation property of positivity for sesquianalytic kernels , mentioned above , to show first that @xmath220 is a positive kernel on all of @xmath111 ) .
as we already said in section [ s : intro ] , one direction of this theorem is obvious , and thus only the other direction is left to prove .
assume that a formal power series @xmath222 as in is such that for every @xmath62 and @xmath83 the operator @xmath223 as in is positive semidefinite .
fix @xmath131 , and for arbitrary @xmath224 denote by @xmath225 the hilbert space with orthogonal basis equal to the set @xmath226 of words @xmath200 of length at most @xmath120 , normalized by @xmath227 .
let @xmath228 denote the backward shift @xmath1-tuple on @xmath225 defined as @xmath229 if @xmath230 , and @xmath231 otherwise .
clearly , @xmath232 . given vectors @xmath233 , compute : @xmath234 by hypothesis
, @xmath235 is a positive semidefinite operator on @xmath225 for each @xmath224 .
letting @xmath236 tend to infinity gives @xmath237 which , in virtue of arbitrariness of @xmath131 and of vectors @xmath238 , means that @xmath66 is a positive non - commutative kernel .
the statement follows from theorem [ impr ] , since in this case @xmath239 is a positive non - commutative kernel .
the estimate for @xmath240 is easily obtained from the fact that the matrix @xmath241 with blocks from @xmath242 is of size @xmath243 and positive semidefinite .
the last statement of the theorem ( about the size and the joint rank of nilpotency of the test matrices ) follows from the proof of theorem [ impr ] .
d. alpay , a. dijksma , j. rovnyak , and h. de snoo , _ schur functions , operator colligations , and reproducing kernel pontryagin spaces _ , operator theory : adv
96 , birkh " auser verlag , basel , 1997 .
mr 2000a:47024 d. alpay and d.s .
kalyuzhny - verbovetzki , _ on the intersection of null spaces for matrix substitutions in a non - commutative rational formal power series _ , c. r. math .
paris * 339 * ( 2004 ) , no . 8 , 533538 .
j.a . ball and v.
vinnikov , _ formal reproducing kernel hilbert spaces : the commutative and noncommutative settings _ , in _ reproducing kernel spaces and applications _
, 77134 , operator theory adv .
143 , birkh " auser verlag , basel , 2003 .
mr 2019348 b.v .
shabat , _ introduction to complex analysis .
part ii _ , translations of mathematical monographs 110 , functions of several variables , translated from the third ( 1985 ) russian edition by j. s. joel , american mathematical society , providence , ri , 1992 . mr 93g:32001 | we show that a formal power series in @xmath0 non - commuting indeterminates is a positive non - commutative kernel if and only if the kernel on @xmath1-tuples of matrices of any size obtained from this series by matrix substitution is positive .
we present two versions of this result related to different classes of matrix substitutions . in the general case
we consider substitutions of jointly nilpotent @xmath1-tuples of matrices , and thus the question of convergence does not arise . in the `` convergent '' case
we consider substitutions of @xmath1-tuples of matrices from a neighborhood of zero where the series converges .
moreover , in the first case the result can be improved : the positivity of a non - commutative kernel is guaranteed by the positivity of its values on the diagonal , i.e. , on pairs of coinciding jointly nilpotent @xmath1-tuples of matrices .
in particular this yields an analogue of a recent result of helton on non - commutative sums - of - squares representations for the class of hereditary non - commutative polynomials .
we show by an example that the improved formulation does not apply in the `` convergent '' case . |
the hecke algebras @xmath8 of the symmetric groups are classical objects of study and the most important open problem in their representation theory is to determine the structure of the specht modules @xmath9 where @xmath4 is a partition of @xmath10 . in this area
there are many obvious questions that remain unanswered .
for example , we rarely know the composition factors of @xmath9 or their multiplicities . however , information on the specht modules may be obtained by computing @xmath11 for @xmath5 and @xmath4 partitions of @xmath10 .
an approach to this problem using the kernel intersection theorem was suggested by james , who gave an easy classification of @xmath12 ( * ? ? ?
* theorem 24.4 ) . this approach has subsequently been developed .
in particular , results of fayers and martin @xcite have given us techniques to compute @xmath13 for more general @xmath4 ( which they used in the same paper to give an elementary proof of the carter - payne theorem ) .
in this paper , we extend the most useful of their results to the hecke algebra @xmath0 .
this enables us to give an algorithm , easily implemented on a computer , which will compute certain homomorphism spaces . using this method we completely classify the homomorphism space @xmath11 where @xmath4 has at most two parts , @xmath7 and @xmath0 is defined over a field of characteristic zero .
the paper is organised as follows .
we begin with the definition of the hecke algebras and some background discussion .
we then state our main results and give some examples and applications .
the proofs of the main results , theorem [ hdtthm ] , theorem [ lemma7 ] and propositions [ first ] to [ last ] are deferred to the next section ; in fact , we give only an indication of the proof of the last propositions , for reasons we discuss in section [ spaces ] .
we end with a brief discussion about homomorphisms between the specht modules of dipper and james .
the definitions in this section are standard and may all be found the the book of mathas @xcite . for each integer @xmath14 , let @xmath15 be the symmetric group on @xmath10 letters .
if @xmath16 is a ring and @xmath2 an invertible element of @xmath16 then the hecke algebra @xmath17 is defined to be the unital associative @xmath16-algebra with generators @xmath18 subject to the relations @xmath19 so that if @xmath20 then @xmath21 . if @xmath22 can be written as @xmath23 where @xmath24 is minimal , we define @xmath25 . then @xmath0 is a free @xmath16-module with a basis @xmath26 .
an expression @xmath23 with @xmath24 minimal is known as a reduced expression for @xmath27 and we define the length of @xmath27 by @xmath28 .
recall that a composition of @xmath14 is a sequence @xmath29 of non - negative integers that sum to @xmath10 and a partition is a composition with the additional property that @xmath30 .
if @xmath5 is a partition of @xmath10 , write @xmath31 .
we define a partial order @xmath32 on the set of compositions of @xmath10 by saying that @xmath33 if @xmath34 for all @xmath35 .
if @xmath33 and @xmath36 , write @xmath37 .
let @xmath5 be a composition of @xmath10 .
define the corresponding young diagram @xmath38 $ ] by @xmath39 = \{(r , c ) \mid 1 \leq c \leq \mu_r \}.\ ] ] a @xmath5-tableau @xmath40 is a map @xmath41 \rightarrow \{1,2,\ldots\}$ ] ; we think of this as a way of replacing the nodes @xmath42 of @xmath38 $ ] with the integers @xmath43 and so may talk about the rows and columns of @xmath40 .
( note that we use the english convention for writing our diagrams . )
let @xmath44 be the set of @xmath5-tableaux such that each integer @xmath45 appears exactly once and let @xmath46 be the tableau with @xmath47 entered in order along the rows of @xmath38 $ ] from left to right and top to bottom . if @xmath48 and @xmath49 , say that @xmath50 is in row - order in @xmath51 if for all @xmath52 with @xmath53 either @xmath35 lies in a lower row than @xmath54 ( that is , a row of higher index ) ; or @xmath54 and @xmath35 lie in the same row , with @xmath35 to the right of @xmath54 .
then @xmath55 is the unique tableau in @xmath44 with @xmath47 in row - order .
the symmetric group @xmath15 acts on the right on the elements of @xmath44 by permuting the entries in each tableau .
if @xmath48 , let @xmath56 be the permutation such that @xmath57 .
let @xmath58 denote the row - stabilizer of @xmath55 , that is , the set of all permutations @xmath27 such that each @xmath59 lies in the same row of @xmath55 as @xmath60 .
we say that @xmath48 is row - standard if the entries increase along the rows and standard if @xmath5 is a partition and the entries increase both along the rows and down the columns .
let @xmath61 denote the set of row - standard @xmath5-tableaux and , if @xmath5 is a partition , let @xmath62 denote the set of standard @xmath5-tableaux .
define @xmath63 and set @xmath64 to be the right @xmath0-module @xmath65 define @xmath66 to be the anti - isomorphism determined by @xmath67 and if @xmath68 define @xmath69 then @xmath70 is a cellular basis of @xmath0 with respect to the partial order @xmath32 and the anti - isomorphism @xmath71 . in accordance with the theory of cellular algebras ,
if @xmath72 we define @xmath73 to be the free @xmath16-module with basis @xmath74 then @xmath73 is a two - sided ideal of @xmath0 . following graham and lehrer @xcite
, we define the cell module @xmath9 , also known as a specht module , to be the right @xmath0-module @xmath75 and define @xmath76 to be the natural projection determined by @xmath77 .
these specht modules are the main objects of interest in the study of the representation theory of the hecke algebras @xmath0 and the symmetric groups @xmath15 .
one of the most important open problems in representation theory is to determine the decomposition matrices for the hecke algebra @xmath0 , that is , compute the composition factors of the specht modules . in this paper
, we study a closely related problem .
we consider homomorphisms between specht modules @xmath78 and @xmath9 , for @xmath5 and @xmath4 partitions of @xmath10 .
suppose @xmath4 is a partition of @xmath10 and let @xmath40 be a @xmath4-tableaux .
say that @xmath40 is of type @xmath5 if @xmath5 is the composition such that each integer @xmath79 appears @xmath80 times in @xmath40 .
let @xmath81 denote the set of @xmath4-tableaux of type @xmath5 .
we say that @xmath82 is row - standard if the entries are non - decreasing along the rows and is semistandard if it is row - standard and the entries are strictly increasing down the columns .
let @xmath83 denote the set of row - standard @xmath4-tableaux of type @xmath5 and @xmath84 denote the set of semistandard @xmath4-tableaux of type @xmath5 .
if @xmath85 , define @xmath86 to be the tableau obtained by replacing each integer @xmath87 with its row index in @xmath55 .
suppose that @xmath4 is a partition of @xmath10 and that @xmath5 is a composition of @xmath10 . if @xmath88 define @xmath89 to be the homomorphism determined by @xmath90 let @xmath91 be the subspace of @xmath92 consisting of homomorphisms @xmath93 such that @xmath94 for some @xmath95 . by construction , if @xmath88 then @xmath96 . if @xmath97 , say that a partition @xmath4 is @xmath98-restricted if @xmath99 for all @xmath54 .
[ ssbasis ] the maps @xmath100 are a basis of @xmath101 . furthermore , unless @xmath102 and @xmath103 is not 2-restricted @xmath104 note that this implies that if @xmath105 then @xmath33 .
+ now suppose @xmath5 is a partition and let @xmath106 be the set of maps @xmath107 with the property that @xmath108 .
again , @xmath109 unless @xmath102 and @xmath4 is not 2-restricted .
apart from the fact that our techniques are well - adapted to determining @xmath106 , we have another reason to want to study this space . [ weyl ]
let @xmath110 denote the @xmath2-schur algebra and let @xmath111 and @xmath112 be the weyl modules corresponding to the partitions @xmath5 and @xmath4 . then @xmath113 fix a pair of partitions @xmath4 and @xmath5 of @xmath10 .
we want to compute @xmath106 . if @xmath114 then @xmath115 can be pulled back to give a homomorphism @xmath116 .
conversely , @xmath117 factors through @xmath78 if and only if @xmath118 for all @xmath119 such that @xmath120 .
m^&^ _ & s^ + & _ & _ | + & & s^ we would therefore like to make it easier to check this condition . if @xmath121 is any composition and @xmath122 , let @xmath123 and @xmath124 . for @xmath125
, let @xmath126 denote the symmetric group on the letters @xmath127 and let @xmath128 be the set of minimal length right coset representatives of @xmath129 in @xmath126 .
( hence if @xmath51 is the @xmath130-tableau with the numbers @xmath131 entered in order along its rows then @xmath132 if and only if @xmath133 is row - standard . ) set @xmath134 if @xmath135 then for @xmath136 and @xmath137 , define @xmath138 let @xmath139 . then @xmath140 [ hdtthm ] let @xmath141 be the right ideal generated by @xmath142 . then @xmath143 we prove theorem [ hdtthm ] in section [ hdtproof ] .
[ simple ] suppose that @xmath144 .
then @xmath118 for all @xmath119 such that @xmath120 if and only if @xmath145 for all @xmath146 and @xmath137 .
we have chosen to work with the specht modules which arise as the cell modules for the murphy basis , rather than the specht modules of dipper and james .
this is consistent , for example , with the work of corlett on homomorphisms between specht modules for the ariki - koike algebras @xcite . as such ,
the kernel intersection theorem ( * ? ? ?
* theorem 3.6 ) does not apply , and so theorem [ hdtthm ] has been created to take its place . in section [ djspecht ] , we show that working in either world gives the same results . we have shown that determining @xmath106 is equivalent to finding @xmath147 bearing in mind that @xmath91 has a basis indexed by semistandard @xmath4-tableaux of type @xmath5 .
if @xmath148 is a tableau and @xmath149 and @xmath150 are sets of positive integers we define @xmath151 be the number of entries in row @xmath152 of @xmath148 , for some @xmath153 , which are equal to some @xmath154 .
we further abbreviate this notation by setting @xmath155}$ ] , @xmath156 and so on . now if @xmath157 define @xmath158 = 1+q+\ldots + q^{m-1 } \in r.\ ] ] let @xmath159!=1 $ ] and for @xmath160 , set @xmath161 ! = [ m][m-1]!$ ] .
if @xmath162 , set @xmath163!}{[k]![m - k]!}.\ ] ] if @xmath164 and any of the conditions @xmath162 fail , set @xmath165 .
we record some results which we will need later .
the first is well - known .
[ gausssum ] suppose @xmath166 .
then @xmath167 [ gausslemma ] suppose @xmath168 .
then , @xmath169 the lemma is true for @xmath170 so suppose @xmath171 and that the lemma holds for @xmath172 . then using lemma [ gausssum ] and the inductive hypothesis , @xmath173 the following result may be seen by applying ( * ? ? ?
* equation 4.6 ) and the anti - isomorphism @xmath71 to ( * ? ? ?
* proposition 2.14 ) .
[ combtheorem1b ] suppose that @xmath174 .
choose @xmath175 with @xmath176 and @xmath177 with @xmath137 .
let @xmath178 be the set of row - standard tableaux obtained by replacing @xmath177 of the entries in @xmath40 which are equal to @xmath179 with @xmath175 .
each tableaux @xmath180 will be of type @xmath181 where @xmath182 recall that @xmath183 and @xmath184
. then @xmath185 so if @xmath116 then we may write @xmath186 where @xmath187 is a linear combination of homomorphisms indexed by @xmath4-tableaux of type @xmath181 , with known coefficients . however , since these tableaux may not be semistandard , the corresponding homomorphisms may not be linearly independent and so we can not say immediately whether @xmath145 .
we would therefore like a method of writing a map @xmath188 as a linear combination of homomorphisms indexed by semistandard tableaux .
unfortunately , we do not have an algorithm for this process .
however , we do have a way of rewriting homomorphisms .
the following result is due to fayers and martin , and holds when @xmath189 .
it was probably the strongest combinatorial result they used to give their elementary proof of the carter - payne theorem @xcite .
recall that if @xmath121 is any sequence of integers then @xmath190 .
[ fm : lemma ] suppose @xmath189 and that @xmath191 is a partition of @xmath10 and @xmath192 a composition of @xmath10 .
suppose @xmath193 .
choose @xmath194 with @xmath195 and @xmath196 and @xmath175 with @xmath197 .
let @xmath198 for @xmath199 , let @xmath200 be the row - standard tableau formed by moving all entries equal to @xmath175 from row @xmath201 to row @xmath202 and for @xmath203 moving @xmath204 entries equal to @xmath54 from row @xmath202 to row @xmath201 ( where we assume we may reorder the rows if necessary ) .
then @xmath205 since fayers and martin work in the setting by james @xcite , it is not immediate that their result carries over to our cellular algebra setting .
see , however , section [ djspecht ] . unfortunately , the obvious @xmath2-analogue of proposition [ fm : lemma ] , that is , in the notation above , that the @xmath0-homomorphism @xmath206 can be writen as a linear combination of maps @xmath207 where @xmath199 , is false .
the following identity can be checked by hand .
we identify a tableau @xmath208 of type @xmath209 with the image @xmath210 .
let @xmath211 and @xmath212 .
then @xmath213 now @xmath214 so that if @xmath215 , we can not write @xmath216 as a linear combination of @xmath217 and @xmath218 . we do however have the following weaker analogue of proposition [ fm : lemma ] . [ lemma7 ]
suppose @xmath191 is a partition of @xmath10 and @xmath192 is a composition of @xmath10 .
let @xmath193 . 1 .
suppose @xmath219 and that @xmath197 .
let @xmath220 for @xmath199 , let @xmath200 be the row - standard tableau formed by moving all entries equal to @xmath175 from row @xmath221 to row @xmath152 and for @xmath203 moving @xmath204 entries equal to @xmath54 from row @xmath152 to row @xmath221
. then @xmath222 2 .
suppose @xmath219 and @xmath223 and that @xmath197 .
let @xmath224 for @xmath199 , let @xmath200 be the row - standard tableau formed by moving all entries equal to @xmath175 from row @xmath152 to row @xmath221 of @xmath148 and for @xmath203 moving @xmath204 entries equal to @xmath54 from row @xmath221 to row @xmath152 .
then @xmath225 the proof of theorem [ lemma7 ] is both technical and long , so we postpone it until the next section and give some examples . as above , we identify a tableau @xmath40 of type @xmath226 with @xmath227 .
set @xmath228 with @xmath229 if @xmath230 for all @xmath231 , and recall that if @xmath16 is a field then @xmath0 is ( split ) semisimple if and only if @xmath232 .
suppose @xmath16 is a field .
let @xmath233 and @xmath234 . if @xmath235 then @xmath93 is determined by @xmath236 for some @xmath237 . then applying proposition [ combtheorem1b ] and theorem [ lemma7 ] , @xmath238\beta \tab(11123,22)+ q[2]\beta \tab(11122,23 ) + q^2 \gamma \tab(11123,22 ) , \\
\theta(m_\mu h_{2,2 } ) & = \alpha \tab(11122,22)+ q[2]^2 \beta \tab(11122,22 ) + q^4 \gamma \tab(11122,22 ) , \\ \theta(m_\mu h_{1,1 } ) & = [ 4 ] \alpha \tab(11112,33 ) + [ 4 ] \beta \tab(11113,23 ) + \beta \tab(11123,13 ) + \gamma
\tab(11133,12 ) \\ & = [ 4 ] \alpha \tab(11112,33 ) + [ 4 ] \beta \tab(11113,23 ) - \beta \tab(11113,23 ) - [ 2 ] \beta \tab(11112,33 ) - q \gamma \tab(11113,23 ) , \\
\theta(m_\mu h_{1,2 } ) & = { { { 5 } \brack { 2 } } } \alpha \tab(11111,33 ) + [ 4 ] \beta \tab(11113,13 ) + \gamma \tab(11133,11 ) \\ & = { { { 5 } \brack { 2 } } } \alpha \tab(11111,33 ) -[2][4]\beta \tab(11111,33 ) + q \gamma \tab(11111,33 ) .
\end{aligned}\ ] ] since @xmath239 if and only if @xmath145 for all @xmath240 above , we see that @xmath241 and hence @xmath242 is 1-dimensional if @xmath243 and zero otherwise .
moreover if @xmath243 , the space @xmath241 is spanned by the map @xmath93 determined by @xmath244 \tab(11122,33 ) -q^2 \tab(11123,23)+ [ 2 ] \tab(11133,22).\ ] ] suppose that @xmath20 and that @xmath16 is a field of characteristic 2 .
let @xmath245 and @xmath246 . then @xmath247 .
if @xmath248 then the space @xmath241 is spanned by the maps @xmath249 and @xmath250 .
it has recently been shown that for fixed parameters @xmath98 and @xmath251 there exist homomorphism spaces of arbitrarily high dimension @xcite .
take @xmath252 and @xmath253 . then @xmath254 and @xmath255 but we now have no obvious way of using theorem [ lemma7 ] to write @xmath256 in terms of homomorphisms indexed by semistandard tableaux .
we are therefore most interested in pairs of partitions @xmath4 and @xmath5 where proposition [ combtheorem1b ] and theorem [ lemma7 ] can give an algorithm for computing @xmath106 .
suppose @xmath257 . if @xmath176 and @xmath137 then theorem [ ssbasis ] and proposition [ combtheorem1b ] show that there exist unique @xmath258 such that @xmath259 where @xmath181 is defined in proposition [ combtheorem1b ] .
now suppose @xmath16 is a field .
if @xmath260 is the matrix whose columns are indexed by tableaux @xmath261 and rows by tableaux @xmath262 for some @xmath176 and @xmath137 , with entries @xmath263 as in equation [ mut ] then , by corollary [ simple ] , @xmath264 so the outstanding problem is to determine an explicit formula for @xmath263 . for the remainder of section [ homsection ] , suppose that @xmath16 is a field and that @xmath191 and @xmath265 satisfy @xmath266 since @xmath267 if @xmath268 , we may also assume that @xmath269 .
let @xmath270 and @xmath137 .
if @xmath257 then say that @xmath271 if * @xmath272 ; * @xmath273 for all @xmath274 and all @xmath35 ; * @xmath275 for all @xmath35 .
[ abig ] let @xmath257 .
suppose that @xmath276 and @xmath137 .
then @xmath277 and if @xmath278 then @xmath208 is semistandard . since @xmath271 precisely when @xmath208 is a row - standard tableau formed from @xmath40 by changing @xmath177 entries equal to @xmath179 into @xmath175 ,
the first part of the lemma is a restatement of proposition [ combtheorem1b ] .
so suppose @xmath271 . if @xmath279 , then @xmath280 for @xmath281 , since @xmath40 is semistandard , and so @xmath282 where the last inequality comes from our assumption on @xmath4 and @xmath5 .
each row @xmath283 therefore contains at least @xmath284 entries equal to @xmath152 and so each entry equal to @xmath179 in @xmath40 is either in the top row or lies in row @xmath221 for some @xmath285 and so is directly below a node of residue @xmath286 . hence a row - standard tableau obtained by changing entries equal to @xmath179 into @xmath175 is semistandard .
let @xmath287 and @xmath137 .
if @xmath257 then say that @xmath271 if @xmath272 and @xmath288 of course , some of the conditions in the table above are redundant since they are implied by the others .
let @xmath257 .
suppose that @xmath289 and @xmath137 . then @xmath290 and if @xmath278 then @xmath208 is semistandard . by proposition [ combtheorem1b ] , @xmath291 where the sum is over row - standard tableaux formed by changing @xmath177 entries equal to @xmath179 in @xmath40 into @xmath175 .
suppose @xmath148 is such a tableau . by theorem [ lemma7 ] ,
if @xmath292 then @xmath293 ; otherwise @xmath188 is a linear combination of maps @xmath294 indexed by row - standard tableaux @xmath208 where @xmath208 is formed from @xmath148 by moving @xmath295 entries equal to @xmath175 from row @xmath179 to row @xmath175 and replacing them with entries not equal to @xmath175 from row @xmath175 .
if @xmath208 has this form , then clearly @xmath278 so that @xmath296 for some @xmath297 .
so suppose @xmath278 .
each of the intermediate tableaux @xmath148 were formed from @xmath40 by changing entries of @xmath40 from @xmath179 to @xmath175 . then @xmath298 entries were changed in row @xmath35 for @xmath299 , and for some @xmath300 , @xmath301 entries were changed in row @xmath175 and @xmath302 entries in row @xmath179 . therefore ,
summing over all @xmath301 with @xmath303 , we have @xmath304 now we change the limits on the sum and apply lemma [ gausslemma ] .
@xmath305 hence @xmath306 it remains to show that if @xmath278 then @xmath208 is semistandard . from the proof of lemma [ abig ] , we have that @xmath307 for @xmath283 , so the only way that @xmath208 can fail to be semistandard is if there is an entry in row @xmath179 which is as big or bigger than the entry directly below it .
however , @xmath308 so this is not possible .
let us summarize the results above .
[ algorithm ] suppose @xmath191 and @xmath265 are partitions of @xmath10 with the property that @xmath309 for @xmath310 .
let @xmath260 be the matrix whose columns are indexed by tableaux @xmath261 and rows by tableaux @xmath262 for some @xmath176 and @xmath137 , where @xmath311 then @xmath312 .
we have taken a hard problem in representation theory and reduced it to a combination of combinatorics and linear algebra .
however it should however be noted that in doing so we have lost some algebraic information .
for example , @xmath313 unless @xmath78 and @xmath9 lie in the same block .
proposition [ algorithm ] does not seem to make use of this fact .
we note that proposition [ combtheorem1b ] and theorem [ lemma7 ] can be used to compute homomorphism spaces other than those we have considered above . for example , the proof of the one - node carter - payne theorem in @xcite relied on proposition [ combtheorem1b ] and some special cases of theorem [ lemma7 ] , but a one - node carter - payne pair @xmath5 and @xmath4 do not necessarily satisfy @xmath314 for @xmath315 .
while a computer can use proposition [ algorithm ] to solve individual problems , it is more satisfying to have explicit results .
this is the purpose of the next section . in this section
we show that a lower bound on @xmath316 can be obtained by looking at the algebra @xmath317 and we then give the dimension of @xmath318 where @xmath6 and @xmath7 .
we would like to thank meinolf geck and lacrimioara iancu for pointing out the proof of proposition [ biggerdim ] .
fix a field @xmath24 of characteristic @xmath319 and let @xmath320 .
let @xmath321 where we assume that @xmath322 .
let @xmath323 be a primitive @xmath324 root of unity in @xmath325 .
let @xmath326 $ ] denote the ring of laurent polynomials in the indeterminate @xmath327 .
if @xmath328 is a field and @xmath2 an invertible element of @xmath328 , define @xmath329 to be the ring homomorphism which sends @xmath327 to @xmath2 .
if @xmath330 is a ring and @xmath331 , then @xmath332 is the greatest order of any non - zero minor of @xmath333 .
if @xmath334 is a ring homomorphism , define @xmath335 to be the matrix with @xmath336-entry @xmath337 . for @xmath338
set @xmath339 .
let @xmath340 $ ] denote the @xmath324 cyclotomic polynomial .
observe that @xmath341 and @xmath342 so that the maps @xmath343 and @xmath344 both factor through @xmath345 / ( \phi_e({\hat{q}}))$ ] , that is , there exist ring homomorphisms @xmath346 and @xmath347 such that the following diagram commutes .
& & k + & ^_k , & _ _ k , + = [ , ^-1 ] & & [ , ^-1]/(_e())=r + & _ _ , & _
_ , + & & [ lemf ] suppose @xmath348
. then @xmath349 .
this follows since if @xmath350 is any @xmath351 submatrix of @xmath333 then @xmath352 .
. then @xmath353 .
this follows from the proof of lemma [ lemf ] and the fact that @xmath354 is injective .
[ rank ] suppose @xmath355 .
then @xmath356 now let @xmath357 .
if @xmath328 is a field and @xmath2 an invertible element of @xmath328 then @xmath358 , where @xmath327 acts on @xmath328 as multiplication by @xmath2 .
if @xmath1 is a @xmath359-module then we define the @xmath360-module @xmath361 .
[ biggerdim ] suppose @xmath1 and @xmath362 are @xmath363-modules which free as @xmath364-modules of finite rank .
then @xmath365 choose bases @xmath366 of @xmath1 and @xmath367 of @xmath362 and let @xmath368 be the corresponding basis of @xmath369 .
then @xmath370 lies in @xmath371 if and only if the coefficients @xmath372 satisfy a system of equations of the form @xmath373 for @xmath374 , some @xmath375 .
if we let @xmath333 be the matrix whose columns are indexed by @xmath376 and rows by @xmath374 and which has entries @xmath377 then @xmath378 . furthermore , @xmath379 by corollary [ rank ] .
in particular , we may take @xmath1 and @xmath362 to be specht modules .
[ dimcorollary ] suppose that @xmath4 and @xmath5 are partitions of @xmath10
. then @xmath380 the following result is not implied by corollary [ dimcorollary ] if @xmath381 .
it could also be proved by giving an analogue of proposition [ biggerdim ] for the @xmath2-schur algebra and using theorem [ weyl ] .
[ charp ] suppose that @xmath4 and @xmath5 are partitions of @xmath10 .
then @xmath382 let @xmath260 be the matrix with entries in @xmath364 given by equation [ mut ] .
then @xmath383 now fix @xmath384 and let @xmath385 where @xmath323 is a primitive @xmath324 root of unity in @xmath325 .
if @xmath209 is a partition of @xmath10 , let @xmath386 be the number of non - zero parts of @xmath209 .
we describe @xmath387 where @xmath388 and @xmath7 .
where @xmath389 , please note that the homomorphism space @xmath390 has been computed for arbitary hecke algebras of type @xmath1 , even for partitions where @xmath391 @xcite .
[ first ] suppose that @xmath392 and that @xmath265 is a partition of @xmath10 .
then @xmath393 for the remainder of section [ spaces ] , suppose @xmath4 and @xmath5 are such that @xmath394 , @xmath395 and @xmath7 . since @xmath396 if @xmath397 we also assume @xmath398 and @xmath399 . for @xmath400 ,
let @xmath401 . if @xmath125 , let @xmath402 be the integer such that @xmath403 and @xmath404 .
[ onedim]we have that @xmath405 suppose that @xmath406 .
let @xmath407 and @xmath408 .
then @xmath409 in fact , this is obvious in our setup since the matrices @xmath260 obtained in both cases are the same .
suppose that @xmath410 and @xmath411 .
then @xmath412 if and only if * @xmath413 and @xmath414 .
[ @xcite ] suppose that @xmath415 and @xmath411
. then @xmath412 if and only if * @xmath416 and @xmath414 and @xmath417 ; or * @xmath418 and @xmath414 and @xmath419 and @xmath420 and @xmath421 ; or * @xmath422 and @xmath423 and @xmath420 ; or * @xmath422 and @xmath423 and @xmath424 and @xmath425 ; or * @xmath422 and @xmath426 and @xmath427 and @xmath428 and @xmath420 and @xmath413 .
suppose that @xmath429 , that @xmath411 and that @xmath430 . then @xmath431 if and only if @xmath432 and * @xmath416 and @xmath414 ; or * @xmath418 and @xmath414 and @xmath433 and @xmath434 and @xmath435 ; or * @xmath422 and @xmath428 and @xmath433 and @xmath434 and @xmath435 ; or * @xmath422 and @xmath436 and @xmath433 and @xmath437 .
now say that the partition @xmath5 has a good shape if it satisfies the following properties .
* @xmath429 ; and * @xmath438 ; and * @xmath424 ; and * @xmath439 ; and * if @xmath440 then @xmath441 and @xmath442 .
if @xmath5 has a good shape define @xmath443 to be the partition given by @xmath444 let @xmath445 with @xmath446 if no such integer exists and let @xmath447 with @xmath448 if no such integer exists .
( we assume @xmath449 . ) if @xmath226 and @xmath450 are two compositions , define the composition @xmath451 by @xmath452 for all @xmath54 .
suppose that @xmath411 , that @xmath5 has a good shape and that @xmath453 .
set @xmath454 then we have the following homomorphisms .
* suppose that @xmath455 or @xmath456 .
then @xmath457 if and only if @xmath458 .
* suppose that @xmath459 .
then @xmath457 if and only @xmath458 or @xmath460 .
* suppose that @xmath461 and @xmath462 .
then @xmath457 if and only @xmath460 .
suppose that @xmath411 , that @xmath5 has a good shape and that @xmath463 .
suppose @xmath464 are such that @xmath465 set @xmath466 then we have the following homomorphisms .
* if @xmath467 then @xmath457 if and only @xmath468 or @xmath469 for some @xmath470 as above .
* if @xmath471 then @xmath457 if and only @xmath460 or @xmath472 for some @xmath24 as above .
[ last ] suppose that @xmath429 , that @xmath473 , that @xmath474 and that @xmath5 does not have a good shape . then @xmath475 . combining propositions [ first ] to [ last ] above completely classifies the homomorphism space @xmath106 where @xmath476 , @xmath477 and @xmath7 .
our proof of these results is obtained via case - by - case analysis ; we are doing nothing more than solving systems of homogeneous linear equations . unfortunately , there are many cases to check and the resulting computations are repetitive and formulaic .
we do not , therefore , propose to prove all the propositions in this paper . in section [ spaceproof ]
, we highlight the methods used and illustrate them with some examples .
the computations that helped lead us to these results were carried out using gap @xcite .
in this section , we give the proofs of theorem [ hdtthm ] and theorem [ lemma7 ] and indicate the proof of propositions [ first ] to [ last ] . for obvious reasons ,
this section is more technical than those preceeding it .
let @xmath17 and fix a partition @xmath478 .
for @xmath479 and @xmath137 , recall that @xmath480 and that @xmath141 is the right ideal of @xmath0 generated by @xmath481 we prove theorem [ hdtthm ] , that is , that @xmath482 let @xmath483 be a set of minimal length right coset representatives for @xmath484 in @xmath15 and recall ( * ? ? ?
* propn .
3.3 ) that @xmath485 .
[ ( * ? ? ?
3.4 ) ] the module @xmath64 is a free @xmath16-module with basis @xmath486 with the action of @xmath0 determined by @xmath487 where @xmath488 . if @xmath489 and @xmath490 then @xmath491 and @xmath492 . therefore if @xmath22 then @xmath493 for some @xmath494 and @xmath490 so that @xmath495 .
equally , if @xmath496 , let @xmath497 be the row - standard tableau obtained by rearranging the entries in each row of @xmath498 . then @xmath499 and @xmath500 .
if @xmath209 is a composition of @xmath10 , define an equivalence relation @xmath501 on @xmath502 by saying that @xmath503 if @xmath504 for all @xmath505 . recall that
if @xmath506 , then @xmath507 is the tableau obtained by replacing each integer @xmath87 with its row index in @xmath55 . now
if @xmath508 and @xmath509 , define @xmath510 [ ( * ? ?
* thm . 4.9 ) ] [ mstbasis ] the right ideal @xmath511 has a basis @xmath512 [ ( * ? ? ?
* eqn . 4.6 ) ] [ leftright ] suppose @xmath209 is a composition of @xmath10 and @xmath508 .
then @xmath513 let @xmath514 and @xmath515 .
let @xmath516 .
then @xmath517 [ minihigh ] suppose @xmath209 is any composition of @xmath10 .
let @xmath4 be the partition obtained by rearranging the parts of @xmath209 .
if @xmath518 then @xmath519 is an @xmath16-linear combination of elements of the form @xmath520 where @xmath521 and @xmath522 are row - standard @xmath4-tableaux . [ higher ] suppose @xmath209 is any composition of @xmath10 such that @xmath523 and @xmath524 . then @xmath525 .
this follows from lemma [ minihigh ] , noting that @xmath526 .
now suppose @xmath136 and @xmath137 and let @xmath527 be the composition given by @xmath528 let @xmath529 be the the row - standard @xmath209-tableau such that @xmath530 , @xmath531 and @xmath532 for all @xmath203 .
by lemma [ leftright ] , @xmath533 .
the next result then follows by lemma [ higher ] , noting that @xmath511 is a right ideal of @xmath0 .
[ halfsub ] we have @xmath534 now we introduce some new notation which will help us describe the elements @xmath535 . if @xmath536 , let @xmath537 be the row - standard @xmath5-tableau in which @xmath54 is in row @xmath152 if the place occupied by @xmath54 in @xmath538 is occupied by @xmath152 in @xmath148 . if @xmath539 then define @xmath540 . [
describes ] suppose @xmath541 and @xmath542 .
then @xmath543 recall that the length @xmath544 of a permutation @xmath22 may be determined by @xmath545 and if @xmath546 is such that @xmath547 then @xmath548 . if @xmath549 and @xmath550 define @xmath551 note that the last identity holds since @xmath552 .
observe that if @xmath553 then @xmath554 [ getorder ] suppose @xmath541 and @xmath542 .
write @xmath555 .
let @xmath556 and , for @xmath557 , let @xmath558 where @xmath559 occupies the same position in @xmath560 that @xmath54 occupies in @xmath498 .
then @xmath561 is the row - standard @xmath5-tableau with the entries @xmath562 occupying the same positions that they occupy in @xmath498 and with all other entries in row order , and furthermore @xmath563 in particular , @xmath564 the description of @xmath561 is easily seen by induction on @xmath54 .
equation [ eqni ] follows , using induction or otherwise , by observing that @xmath565 for @xmath566 define @xmath567 let @xmath157 and let @xmath568 be a composition .
then @xmath569 we may assume @xmath570 .
let @xmath571 . if @xmath572 then by equation [ length ] , @xmath573 so that @xmath574
. now recall that @xmath575 if and only if @xmath576 and observe that @xmath577 is precisely the row - standard @xmath568-tableau with the entries @xmath578 in the second row . [ rearrangecor ] suppose that @xmath566 and @xmath579 .
then @xmath580 @xmath581 [ pullsthrough ] suppose @xmath566 and @xmath582 .
then @xmath583 if @xmath584 , define @xmath585 by setting @xmath586 for all @xmath24 .
we claim that @xmath587 for all @xmath588 .
let @xmath589 and let @xmath590 be the tableau containing @xmath591 in the first row and @xmath592 in the third row .
let @xmath27 be the permutation such that @xmath593 .
it is sufficient to check that @xmath594 and that @xmath595 which is a routine exercise .
[ pullsthrough2 ] suppose @xmath596 .
then @xmath597 the proof is similar to the proof of lemma [ pullsthrough ] ; we consider tableaux of shape @xmath598 .
[ intothree ] let @xmath125 and @xmath121 be a composition .
suppose @xmath599 .
then @xmath600 again , we may assume that @xmath601 .
let @xmath590 and let @xmath27 be the permutation such that @xmath602 .
suppose that the entries in rows @xmath603 of @xmath51 are @xmath604 and that the entries in rows @xmath605 are @xmath606 .
let @xmath607 be the permutation of @xmath608 which sends @xmath609 to @xmath610 , where @xmath611 occupies the same position in @xmath51 that @xmath612 occupies in @xmath613 .
similarly , let @xmath614 be the permutation of @xmath615 which sends @xmath616 to @xmath610 , where @xmath617 occupies the same position in @xmath51 that @xmath612 occupies in @xmath613 .
finally let @xmath618 .
it is clear that @xmath619 and @xmath620 .
we leave it as an exercise to check that @xmath621 and that @xmath622 .
the proof of lemma [ intothree ] follows by counting the number of terms on both sides of the equation .
[ termstofront ] suppose @xmath209 is a partition of @xmath10 with @xmath523 and @xmath623
. choose @xmath24 minimal such that @xmath624 and @xmath625 minimal such that @xmath626 .
then @xmath627 for some @xmath119 .
using lemma [ describes ] and lemma [ intothree ] , @xmath628 as in lemma [ getorder ] , and keeping the notation of that lemma , we may write @xmath629 where @xmath630 commutes with @xmath631 . the result follows .
[ killsall ] suppose that @xmath632 and @xmath633 .
then @xmath634 it is straightforward to see that the proof for arbitrary @xmath24 is identical to the proof for @xmath635 , so we assume that @xmath635 .
we now prove that the lemma holds for all @xmath636 .
if @xmath637 and @xmath638 then @xmath639 so now suppose that @xmath640 and that lemma [ killsall ] holds for all @xmath641 . choose @xmath642 with @xmath643 . recall that if @xmath644 then @xmath645 we have @xmath646 where @xmath647 so that it is sufficient to show that @xmath648 now for @xmath649 we may write @xmath650 so that , substituting into equation [ toshow ] and commuting terms to the left where possible , we must show that @xmath651 consider @xmath652 . by the inductive hypothesis ,
@xmath653 so that it is sufficient to prove that for @xmath654 we have @xmath655 now , using corollary [ rearrangecor ] , lemma [ pullsthrough ] and lemma [ intothree ] , @xmath656 and by the inductive hypothesis again , @xmath657 [ allequal ] we have @xmath658 by corollary [ halfsub ] , @xmath659
. now by lemma [ mstbasis ] , @xmath660 has a basis @xmath661 if @xmath523 and @xmath662 then , since @xmath141 is a right ideal , it follows from lemmas [ termstofront ] and [ killsall ] that @xmath663 and so @xmath664 .
hence @xmath665 . in this section ,
we give the proof of theorem [ lemma7 ] .
let @xmath327 be an indeterminate over @xmath666 and let @xmath667 $ ] .
let @xmath668 .
we prove theorem [ lemma7 ] for @xmath669 ; the general result follows by specialization .
let @xmath191 be a partition of @xmath10 , @xmath192 a composition of @xmath10 and @xmath193 where we assume that @xmath670 .
our aim is to write @xmath671 as a linear combination of homomorphisms indexed by tableaux @xmath672 . as in the previous examples , we identify @xmath672 with @xmath673 .
[ niceex ] let @xmath674 and @xmath675 .
recall that @xmath676 then @xmath677 if @xmath672 , define @xmath678 to be the tableau with @xmath679 and @xmath680 in row order , for all @xmath681 .
the following result is given by the definition of the map @xmath294 .
( observe also the ` reverse ' statement in corollary [ describes ] . ) [ reverse ] let @xmath672 .
then @xmath682 we begin by considering the case where @xmath6 .
if @xmath193 is such that @xmath683 and @xmath684 for @xmath685 then we will represent @xmath148 by @xmath686 if a number @xmath54 does not appear in the top ( resp .
bottom ) row it should be understood that @xmath687 ( resp .
@xmath688 ) . any such representation containing an index @xmath689
should be taken to be zero , so for example in corollary [ basic5 ] below , if @xmath690 then the first term on the right - hand side of the equation should be ignored .
we continue to identify @xmath148 with @xmath691 .
[ type1n ] let @xmath692 and @xmath693 and let @xmath694 be the tableau with @xmath695 in the top row ( and @xmath696 in the second row ) .
for @xmath697 let @xmath698 be the row - standard tableau obtained from @xmath148 by swapping @xmath699 with @xmath54 and rearranging the rows if necessary .
then @xmath700 we use lemma [ pullsthrough2 ] and note that @xmath701 . if @xmath702 then @xmath703 else if @xmath704 then @xmath705 using the definition of the maps @xmath294 , we have the following corollary .
[ basic5 ] we have that @xmath706 [ cosetattack ] suppose @xmath197 is such that @xmath707 and @xmath708
. then @xmath709 { \begin{array}{l}{1^{\alpha_1}2^{\alpha_2 } \ldots
i^{\alpha_i-1 } \ldots d^{1 } \ldots b^{\alpha_b}}\\{1^{\beta_1}2^{\beta_2}\ldots i^{\beta_i+1 } \ldots d^0 \ldots b^{\beta_b}}\end{array}}\\ - \sum_{i = d+1}^{b } { \hat{q}}^{-\bar\beta_{d}+\bar\beta_{i-1}}[\beta_i+1 ] { \begin{array}{l}{1^{\alpha_1}2^{\alpha_2}\ldots d^{1 } \ldots i^{\alpha_i-1 } \ldots b^{\alpha_b}}\\{1^{\beta_1}2^{\beta_2}\ldots d^0 \ldots i^{\beta_i+1 } \ldots b^{\beta_b}}\end{array}}.\end{gathered}\ ] ] note that @xmath710 where @xmath711 let @xmath712 and let @xmath713 be the unique tableau such that @xmath714 .
choose @xmath35 in the top row of @xmath51 and let @xmath715 be the tableau obtained by swapping @xmath35 and @xmath716 and rearranging the rows .
by lemma [ type1n ] , @xmath717 that is @xmath718 now choose @xmath54 with @xmath719 .
we claim that @xmath720 { \begin{array}{l}{1^{\alpha_1}2^{\alpha_2 } \ldots i^{\alpha_i-1 } \ldots d^{1 } \ldots b^{\alpha_b}}\\{1^{\beta_1}2^{\beta_2}\ldots i^{\beta_i+1 } \ldots d^0 \ldots b^{\beta_b}}\end{array}}.\ ] ] to prove the claim , choose @xmath35 with @xmath721 .
let @xmath498 be the row - standard @xmath4-tableau containing entries from the set @xmath722 .
then @xmath723 where @xmath724 acts on @xmath498 as follows .
suppose that @xmath642 occupies the same position in @xmath498 that @xmath35 occupies in @xmath51 .
then @xmath724 moves @xmath725 into the first row , such that the row is in increasing order , moves @xmath642 into the far left of the second row and pushes the entries in the second row , up to the entry immediately to the left of @xmath725 in @xmath498 , one box to the right . if @xmath497 is the row - standard tableau obtained by rearranging the rows of @xmath726 , then @xmath727 where @xmath728 has length @xmath729 .
furthermore , using equation [ length ] , we may check that @xmath730 .
it therefore follows that @xmath731 where @xmath732 is the tableau obtained from @xmath498 by swapping @xmath642 and @xmath725 and rearranging the rows . if we let @xmath733 then using lemma [ intothree ] , @xmath734 m_{\lambda}t_{d(\dot{\mathfrak{s}}(x_0 ) ) } c(\bar\nu_{i-1};\alpha_{i}-1,\beta_i+1 ) \\\end{aligned}\ ] ] the claim then follows since @xmath735 a similar proof shows that if @xmath736 then @xmath737 { \begin{array}{l}{1^{\alpha_1}2^{\alpha_2 } \ldots d^{1 } \ldots i^{\alpha_i-1 } \ldots b^{\alpha_b}}\\{1^{\beta_1}2^{\beta_2}\ldots d^0 \ldots i^{\beta_i+1 } \ldots b^{\beta_b}}\end{array}},\ ] ] completing the proof of lemma [ cosetattack ] .
[ stage3 ] suppose @xmath738 where @xmath739 .
let @xmath740 for @xmath199 , let @xmath200 be the row - standard tableau obtained from @xmath148 by moving all entries equal to @xmath175 from row 2 to row 1 , and for @xmath741 moving down @xmath204 entries equal to @xmath54 from row 1 to row 2 .
then @xmath742 the case that @xmath708 is lemma [ cosetattack ] .
so assume @xmath743 and that the lemma holds when @xmath744 .
we first consider @xmath745 consider the map @xmath746 . if @xmath747 is the tableau @xmath748 of type @xmath749 then @xmath750 .
let @xmath751 for @xmath199 , let @xmath752 be the tableau obtained from @xmath753 by moving all entries equal to 3 or 4 from the second row to the first and moving @xmath754 entries equal to 1 and @xmath755 entries equal to 5 from the first to the second . for @xmath756 , let @xmath757 be the tableau obtained from @xmath753 by moving all entries equal to 4 from the second row to the first and moving @xmath758 entries equal to 1 and @xmath759 entries equal to 5 from the first to the second .
note that @xmath760 , where @xmath200 is as defined in the statement of the lemma . applying the inductive hypothesis repeatedly to @xmath747
, we have @xmath761{\hat{q}}^{g_4(\beta_2+\beta_3 ) } \theta_{\dot\u_g } \\ & \quad + \sum_{g \in \mathcal{g ' } } ( -1)^{\beta_3 } { \hat{q}}^{-\beta_2}{\hat{q}}^{-\binom{\beta_3}{2}+g'_1}{\hat{q}}^{-(\beta_3 - 1)(\beta_2 + 1)}[g_1']{\hat{q}}^{\beta_2 + 1}{\hat{q}}^{g'_4(\beta_2+\beta_3 ) } \theta_{\dot\u_{g ' } } \\ & \quad + \sum_{g \in \mathcal{g } } ( -1)^{\beta_3 } { \hat{q}}^{\beta_3 - 1}{\hat{q}}^{-\binom{\beta_3}{2}+g_1 } { \hat{q}}^{-(\beta_3 - 1)\beta_2 } { \hat{q}}^{(g_4 - 1)(\beta_2+\beta_3 - 1)}[g_4]\theta_{\dot\u_g } \\ & \quad + \sum_{g ' \in \mathcal{g ' } } ( -1)^{\beta_3 } { \hat{q}}^{\beta_3 - 1}{\hat{q}}^{-\binom{\beta_3}{2}+g_1'+1}{\hat{q}}^{-(\beta_3 - 1)\beta_2}{\hat{q}}^{\beta_2}{\hat{q}}^{(g_4'-1)(\beta_2+\beta_3 - 1)}[g_4']\theta_{\dot\u_{g ' } } \\ & = \sum_{g \in \mathcal{g } } ( -1)^{\beta_3 } { \hat{q}}^{-\binom{\beta_3 + 1}{2}+g_1-\beta_2\beta_3 } { \hat{q}}^{g_4(\beta_2+\beta_3 ) } [ \beta_3 ] \theta_{\dot\u_g } \\ & \quad + \sum_{g ' \in \mathcal{g ' } } ( -1)^{\beta_3 } { \hat{q}}^{-\binom{\beta_3}{2}+g'_1}{\hat{q}}^{g'_4(\beta_2+\beta_3 ) } { \hat{q}}^{-(\beta_2 + 1)(\beta_3 - 1)}{\hat{q}}[\beta_3 - 1 ] \theta_{\dot\u_{g'}}. \end{aligned}\ ] ] applying the inductive hypothesis again , we also have @xmath762 so , substituting the two values of @xmath763 into the equation below , we get that @xmath764\theta_\s(m_\nu ) & = \theta_{\dot\s}(m_{\dot\nu } ) + { \hat{q}}[\beta_3 - 1]\theta_{\dot\s}(m_{\dot\nu } ) \\ & = [ \beta_3 ] \sum_{g \in \mathcal{g } } ( -1)^{\beta_3}{\hat{q}}^{-\binom{\beta_3 + 1}{2}+g_1-\beta_2\beta_3}{\hat{q}}^{g_4(\beta_2+\beta_3 ) } \theta_{\dot\u_g}(m_{\dot\nu } ) \\ & = [ \beta_3 ]
\sum_{g \in \mathcal{g } } ( -1)^{\beta_3}{\hat{q}}^{-\binom{\beta_3 + 1}{2}+g_1-\beta_2\beta_3}{\hat{q}}^{g_4(\beta_2+\beta_3 ) } \theta_{\u_g}(m_{\nu } ) . \end{aligned}\ ] ] since we are working in @xmath765 , we may cancel the terms @xmath766 $ ] on both sides of the equation .
this completes the proof of lemma [ stage3 ] when @xmath148 has the form @xmath767 .
the proof of lemma [ stage3 ] for general @xmath148 follows in the same way as the end of the proof of lemma [ cosetattack ] .
[ main2part ] suppose @xmath768 where @xmath739 .
let @xmath740 for @xmath199 , let @xmath200 be the tableau obtained from @xmath148 by moving all entries equal to @xmath175 from row 2 to row 1 , and for @xmath741 moving down @xmath204 entries equal to @xmath54 from row 1 to row 2 .
then @xmath769 the case that @xmath707 is precisely lemma [ stage3 ] .
so suppose @xmath770 and that the lemma holds when @xmath771 .
let @xmath772 and suppose they are of type @xmath749 so that @xmath773 let @xmath774 for @xmath199 , let @xmath752 be the tableau obtained from @xmath775 by moving all entries equal to @xmath175 from row 2 to row 1 , and for @xmath776 moving down @xmath204 entries equal to @xmath54 from row 1 to row 2 and for @xmath777 moving @xmath778 entries equal to @xmath54 from row 1 to row 2 . for @xmath756 , let @xmath757 be the tableau obtained from @xmath779 by moving all entries equal to @xmath175 from row 2 to row 1 , and for @xmath776 moving down @xmath780 entries equal to @xmath54 from row 1 to row 2 and for @xmath777 moving @xmath778 entries equal to @xmath54 from row 1 to row 2 . then , using the inductive hypothesis , @xmath781 we now move on to the more general case where @xmath4 may have more than 2 parts .
[ somemoreparts ] suppose @xmath193 where @xmath191 and @xmath670 .
choose @xmath152 with @xmath285 and suppose that @xmath148 satisfies the following conditions : there exists @xmath24 with @xmath782 such that * all entries of @xmath148 in rows @xmath783 are equal to @xmath35 .
* all entries of @xmath148 in rows @xmath152 and equal to one of @xmath784 and all entries in row @xmath221 are equal to one of @xmath785 , * all entries of @xmath148 in rows @xmath786 are equal to @xmath787 .
choose @xmath175 with @xmath788 .
let @xmath789 for @xmath199 , let @xmath200 be the row - standard tableau obtained from @xmath148 by moving all entries equal to @xmath175 from row @xmath221 to row @xmath152 , and for @xmath741 moving down @xmath204 entries equal to @xmath54 from row @xmath152 to row @xmath221 .
then @xmath790 the proof of lemma [ somemoreparts ] is identical to the proof of lemma [ main2part ] , except for the change in notation .
we chose to give the proof of lemma [ main2part ] rather than proving lemma [ somemoreparts ] itself because the notation was easier to control .
now suppose that @xmath193 .
choose @xmath152 with @xmath279 and define @xmath753 to be the @xmath4-tableau such that * each row @xmath783 contains @xmath791 entries equal to @xmath35 . *
each row @xmath792 contains @xmath791 entries equal to @xmath793 . *
each row @xmath794 contains @xmath795 entries equal to @xmath796 , for @xmath685 .
note that @xmath753 satisfies the conditions of lemma [ somemoreparts ] .
suppose that @xmath797 and let @xmath637 .
then @xmath798 [ littleperms ] suppose @xmath799 and @xmath800 is a permutation of @xmath801 .
suppose @xmath27 is a permutation such that @xmath802 for all @xmath803
. then @xmath804 .
the proof follows by applying equation [ length ] .
[ newperms ] let @xmath193 .
choose @xmath152 with @xmath283 and define @xmath753 and @xmath749 as above . if @xmath27 is the permutation such that @xmath805 then @xmath806 . furthermore if @xmath807 is such that @xmath148 and @xmath40 are identical on all rows except possibly rows @xmath152 and @xmath221 then @xmath808 ( so that @xmath809 ) .
note that @xmath810 and @xmath27 satisfy the conditions of lemma [ littleperms ] so that @xmath811 .
it is straightforward to see that the permutation @xmath27 works for both @xmath148 and @xmath40 .
[ morethanthree ] let @xmath157 and @xmath812 be a composition such that @xmath670 .
choose @xmath152 with @xmath283 .
then @xmath813 as usual , we may assume @xmath570 .
applying lemma [ intothree ] repeatedly , we get @xmath814 [ commutes ] let @xmath815 .
suppose @xmath27 is a permutation such that for all @xmath816 we have @xmath817 for some @xmath818 .
then @xmath819 if @xmath800 is any permutation of @xmath820 , let @xmath821 be the permutation of @xmath822 which sends @xmath823 to @xmath824 .
then clearly @xmath825 and @xmath826 by lemma [ littleperms ] .
the result follows since @xmath827 is a sum of basis elements indexed by permutations of @xmath820 .
[ badperm ] let @xmath193 .
choose @xmath152 with @xmath283 and define @xmath828 and @xmath27 as in lemma [ newperms ]
. then @xmath829 applying lemmas [ reverse ] , [ newperms ] , [ morethanthree ] and [ commutes ] , we have that @xmath830 we may now combine the previous results .
let @xmath831 so that @xmath832
. then @xmath833 \tab(1111,2334,224,5)-[2]\tab(1111,2233,244,5 ) \big ) t_{(2,6,3,7,4,10,11,12,5 ) } \operatorname{c}(0;1,3)\operatorname{c}(0;4,1 ) c(5;2,2 ) \operatorname{c}(9;1,2 ) \\ & = -q^{-1}[2 ] \tab(1223,1223,113,1)- [ 2 ] \tab(1223,1122,133,1)\end{aligned}\ ] ] [ part1 ] suppose @xmath191 is a partition of @xmath10 and @xmath192 is a composition of @xmath10 .
let @xmath193 .
suppose @xmath219 and that @xmath197 .
let @xmath834 for @xmath199 , let @xmath200 be the row - standard tableau formed by moving all entries equal to @xmath175 from row @xmath221 to row @xmath152 and for @xmath203 , moving @xmath204 entries equal to @xmath54 from row @xmath152 to row @xmath221
. then @xmath835 using lemmas [ somemoreparts ] , [ newperms ] and [ badperm ] , and keeping the notation of lemma [ badperm ] , @xmath836 the proof of proposition [ part1 ] gives the first half of the proof of theorem [ lemma7 ] .
since the proof of the second half follows along identical lines , we omit most of it and give only the proof of the analogue of lemma [ type1n ] , where the difference is non - trivial .
let @xmath674 and @xmath837 .
observe that @xmath838 by lemma [ minihigh ] . then @xmath839 [ type1nb ] let @xmath840 .
let @xmath841 and @xmath842 and let @xmath694 be the tableau with @xmath843 in the top row ( and @xmath844 in the second row ) . for @xmath845 let @xmath698 be the row - standard tableau obtained by swapping @xmath699 with @xmath54 and rearranging the rows if necessary .
then @xmath846 we use lemma [ pullsthrough ] and note that @xmath847 by lemma [ minihigh ] . if @xmath848 then @xmath849 else if @xmath850 then @xmath851 let @xmath476 where @xmath2 is a primitive @xmath852 root of unity in @xmath325 for some @xmath384 . in propositions [ first ] to [ last ]
we described the homomorphism space @xmath106 where @xmath4 has at most two parts and @xmath7 .
as previously mentioned , the proof of these results is given by case - by - case analysis , where the calculations consist of solving systems of homogeneous linear equations .
these very many calculations do not belong in a research paper ; they are both trivial and lengthy .
we therefore begin this section by giving some identities which enabled us to solve the equations , followed by the proof of the results in some specific cases .
the reader who wishes to use one or more of our propositions and who prefers not to rely on results which are not explicitely proved would be best advised to construct their own proofs ; we would say this is more tedious than difficult .
we note that our proof of proposition [ onedim ] , which states that the homomorphism space is at most 1-dimensional , relies on looking at cases individually ; we do not know of a direct proof .
[ divzero ] suppose that @xmath853 .
then @xmath854}{[be ] } = \frac{a}{b}.\ ] ] we have that @xmath855}{[be ] } & = \frac{1+q+\ldots + q^{(a-1)e-1}}{1+q+\ldots+q^{(b-1)e-1 } } \\ & = \frac{(1+q^e+\ldots+q^{(a-1)e})[e]}{(1+q^e+\ldots+q^{(b-1)e})[e ] } \\ & = \frac{a}{b}. \end{aligned}\ ] ] [ whenzero ] suppose that @xmath162 .
write @xmath856 and @xmath857 where @xmath858
. then @xmath859 if and only if @xmath860 .
we have that @xmath861[m-1]\ldots[m - k+1]}{[k][k-1]\ldots[1]}\ ] ] which , using lemma [ divzero ] is zero if and only if there are more terms divisible by @xmath98 in the numerator than the denominator , that is if and only if @xmath860 .
[ e ] let @xmath862 .
then @xmath863 let @xmath864 .
then @xmath865 for all @xmath866 if and only if @xmath867 and @xmath868 .
we have @xmath869 if @xmath98 is even then @xmath870 [ repeat ] suppose @xmath168 .
then @xmath169 [ notrepeat ] suppose @xmath871 and @xmath872 .
then @xmath873 we use induction on @xmath10 .
the lemma is true for @xmath170 so suppose that @xmath171 and that the lemma holds for @xmath172
. then using lemma [ gausssum ] , @xmath874 by the inductive hypothesis .
armed with these results , we are ready to start solving some equations .
let us begin with the simplest non - trivial case which is when @xmath5 has three parts .
take @xmath6 and @xmath875 to be partitions of @xmath10 where @xmath876 .
our aim is to determine @xmath3 by finding a basis for @xmath241 . for @xmath877 let @xmath878 be given by @xmath879 where we use the notation of section [ lemma7proof ] .
then @xmath241 is the vector space of homomorphisms @xmath880 for @xmath881 , which satisfy @xmath145 for @xmath882 and @xmath137 .
note that for such @xmath240 , we have @xmath883 where the sums are over all @xmath35 such that @xmath884 and all @xmath24 such that @xmath885 .
we must therefore solve the systems of equations @xmath886 for @xmath887 within the limits described above .
let us redefine @xmath888 to be the solution space of the equations @xmath889 and @xmath890 and @xmath891 to be the solution space of the equations @xmath890 . finally , for any @xmath157
, we define @xmath892 and @xmath402 by writing @xmath893 where @xmath894 .
the equivalence relation @xmath895 is assumed to be equivalence modulo @xmath98 .
[ goodlemma ] suppose @xmath896 and that @xmath897 is a non - zero solution to the system of equations @xmath898a_{j+1 } & = 0 , & & m \leq j \leq n-1,\end{aligned}\ ] ] where @xmath899 .
if there do not exist either @xmath900 with @xmath901 and @xmath902 or @xmath175 with @xmath903 and @xmath904 then @xmath905 for all @xmath906 .
otherwise @xmath905 only if @xmath907 and furthermore if @xmath908 are such that @xmath909 and @xmath910 then @xmath905 if and only if @xmath911 .
if neither @xmath900 nor @xmath175 as above exist then all terms @xmath912,\ldots,[n],[k - m],\ldots,[k - n+1]$ ] are non - zero and the result is clear .
so suppose at least one of @xmath913 exists .
let @xmath914 and suppose @xmath915 is such that @xmath916 . if @xmath917 then the equations @xmath898a_{j+1 } & = 0 , & & i^\ast e-1 \leq j \leq i-1,\end{aligned}\ ] ] ensure that @xmath918 .
similarly if @xmath919 then the equations @xmath898a_{j+1 } & = 0 , & & i \leq j \leq i^\ast e+b-1,\end{aligned}\ ] ] ensure that @xmath918 .
but if @xmath920 and @xmath921 then we can not find @xmath900 or @xmath175 as above .
hence if @xmath916 then @xmath918 . now suppose @xmath908 are such that @xmath922 and @xmath910 .
assume @xmath923 .
the equations @xmath898a_{j+1 } & = 0 , & & i \leq j \leq k-1,\end{aligned}\ ] ] ensure that @xmath918 if and only if @xmath924 . [ lemmalow ]
suppose @xmath925 .
then @xmath926 suppose @xmath927 .
setting @xmath928 , we have the equations @xmath929 a_0 & = 0 , & & \\ [ j+1]a_j + q^{j+1}[\mu_2-j]a_{j+1 } & = 0 , & & 0 \leq j \leq { \lambda}_2 - 1.\end{aligned}\ ] ] let @xmath930 . if @xmath931 then there can be no non - zero solution to these equations .
otherwise , applying lemma [ goodlemma ] , @xmath905 only if @xmath932 and in this case @xmath905 if and only if @xmath933
so that if @xmath934 then @xmath935 . suppose @xmath936 and consider the equation @xmath937 where @xmath938 . from this , and using lemma [ whenzero ] , @xmath939 so that , applying lemma [ e ] , @xmath940 if and only if @xmath941 . hence if @xmath942 then @xmath935 and if @xmath943 then @xmath944 .
but in this case we obtain a contradiction by considering equation @xmath945 and using lemma [ e ] : @xmath946 since @xmath947 for @xmath948 and the second gaussian polynomial is zero for @xmath949 .
so @xmath950 if @xmath942 .
suppose that @xmath925 .
then @xmath951 let @xmath930 .
by lemma [ lemmalow ] , we have @xmath952 and if @xmath953 then @xmath954 ; furthermore if @xmath955 then @xmath905 if and only if @xmath956
. then if @xmath957 then it also satisfies @xmath958a_k & = [ { \lambda}_2-k]a_{k+1 } , & & 0 \leq k < { \lambda}_2 , \\ [ \mu_1 + 1-{\lambda}_2]a_{{\lambda}_2 } & = 0 . & & \end{aligned}\ ] ] therefore if @xmath953 then @xmath414 , and so @xmath959 so that by equation @xmath960 we have @xmath418 and @xmath961 .
so if @xmath962 , then @xmath963 is spanned by @xmath964 where @xmath965 for all @xmath54 .
suppose @xmath966 and that @xmath414 .
set @xmath967 for all @xmath54 .
now if @xmath968 then by using lemma [ e ] and considering equation @xmath969 and lemma [ repeat ] we obtain the contradition @xmath970 so we must have if @xmath953 then @xmath971 . we have shown that if @xmath972 then @xmath973 , @xmath416 and @xmath414 and the space @xmath963 is spanned by the solution @xmath964 where @xmath965 for @xmath974 .
we now show these conditions are sufficient .
first we must show that the equations @xmath890 where @xmath975 and @xmath976 are satisfied by the solution @xmath965 for all @xmath54 . using lemma [ notrepeat ] we
have @xmath977 by lemma [ whenzero ] , since @xmath416 , so these equations do hold .
now consider the equations @xmath889 where @xmath978 and @xmath979 . by lemma [ repeat ]
we have @xmath980 again by lemma [ whenzero ] , since @xmath981 , so these equations are also satisfied .
we believe this proof should convince the reader that solving all the equations required for propositions [ first ] to [ last ] is a lengthy business ; but we hope that we have also convinced them that it is not particularly difficult .
a strategy that works is as follows : write down a system of equations @xmath982 which need to be solved , then use the equations @xmath983 and @xmath984 to come up with necessary conditions for the space @xmath241 to be non - zero .
these conditions turn out to be sufficient .
we conclude by making a connection with the specht modules of dipper and james . recall that for each partition @xmath4 of @xmath10 , dipper and james ( * ? ? ?
* section 4 ) defined a @xmath0-module which they called a specht module and which we shall denote @xmath985 .
the connection with the modules @xmath9 is that @xmath986 where @xmath987 denotes the dual of a @xmath0-module @xmath333 ( * ? ? ?
* theorem 5.3 ) .
the first and last equations follow from equation @xmath989 above . now ,
following ( * ? ? ?
* lemma 2.3 ) , we define a @xmath0-automorphism @xmath990 . for a @xmath0-module @xmath333 , define @xmath991 to be the right @xmath0-module with action defined by @xmath992 for all @xmath993 and @xmath994 .
then @xmath995 for any partition @xmath209 of @xmath10 ( * ? ? ?
* theorem 5.2 ) , and the middle equation follows .
if @xmath102 then we might also wish to replace @xmath996 with @xmath997 above .
we now use theorem [ weyl ] .
the middle equation follows since @xmath998 see for example ( * ? ? ?
* lemma 3.4 ) . recall that for each partition @xmath4 of @xmath10 , dipper and james @xcite defined a @xmath178-module which they called a weyl module and which we shall denote @xmath999 .
the relationship with the weyl modules of theorem [ weyl ] is that @xmath1000 .
futhermore @xmath1001 and so @xmath1002 as required .
since many authors prefer to work in the dipper - james world , it is worth remarking , once and for all , that all the relevant combinatorics can be translated backwards and forwards without change .
in particular , there exists an analogue of theorem [ lemma7 ] .
suppose @xmath4 is a partition of @xmath10 and @xmath209 a composition of @xmath10 .
recall that for each @xmath193 dipper and james @xcite defined a homomorphism from @xmath9 to @xmath1003 which we shall now denote @xmath1004 . 1 .
suppose @xmath219 and that @xmath197 .
let @xmath220 for @xmath199 , let @xmath200 be the row - standard tableau formed by moving all entries equal to @xmath175 from row @xmath221 to row @xmath152 and for @xmath203 moving @xmath204 entries equal to @xmath54 from row @xmath152 to row @xmath221 .
then @xmath1005 2 .
suppose @xmath219 and @xmath223 and that @xmath197 .
let @xmath224 for @xmath199 , let @xmath200 be the row - standard tableau formed by moving all entries equal to @xmath175 from row @xmath152 to row @xmath221 of @xmath148 and for @xmath203 moving @xmath204 entries equal to @xmath54 from row @xmath221 to row @xmath152 .
then @xmath1006 the proof of theorem [ dj : lemma7 ] involves working through the same steps as the proof of theorem [ lemma7 ] . rather than working modulo @xmath73
, we now use the fact that @xmath985 is in the kernel of any homomorphism @xmath1007 where @xmath1008 .
we leave the details to the reader .
let @xmath674 and @xmath675 ( as in example [ niceex ] ) . recall that @xmath985 is generated by @xmath1009 for a certain @xmath1010 ( see @xcite for details ) and
that , by the kernel intersection theorem , @xmath1011 let @xmath1012 .
then @xmath1013 where @xmath1014 | we study the homomorphism spaces between specht modules for the hecke algebras @xmath0 of type @xmath1 .
we prove a cellular analogue of the kernel intersection theorem and a @xmath2-analogue of a theorem of fayers and martin and apply these results to give an algorithm which computes the homomorphism spaces @xmath3 for certain pairs of partitions @xmath4 and @xmath5 .
we give an explicit description of the homomorphism spaces @xmath3 where @xmath0 is an algebra over the complex numbers , @xmath6 and @xmath5 is an arbitrary partition with @xmath7 . |
a language is _ definite _ if it can be decided whether a word @xmath7 belongs to it simply by examining the suffix of @xmath7 of some fixed length .
the class of definite languages was the very first subclass of regular languages to be considered : it was introduced in 1954 in the classic paper by kleene @xcite .
it was then studied in 1963 by perles , rabin , and shamir @xcite , and brzozowski @xcite , in 1966 by ginzburg @xcite , and later by several others .
definite languages were revisited in 2009 by bordihn , holzer and kutrib @xcite in connection with state complexity . _ reverse definite _ languages were first studied by brzozowski @xcite . here
membership of @xmath7 can be determined by its prefix of some fixed length .
the class of _ finite and cofinite _ languages is the intersection of the definite and reverse definite classes . here testing for membership
can be done by checking all words shorter than some fixed length .
these three classes appear at the bottom of the dot - depth hierarchy @xcite of star - free languages , below generalized definite languages and locally testable languages .
all three classes are boolean algebras .
the semigroup @xmath8 of a finite / cofinite language is _ nilpotent _ : it has a single idempotent @xmath9 which is a zero , and is characterized by the equations @xmath10 . for definite ( reverse definite ) languages
every idempotent @xmath9 is a right zero , that is , @xmath11 ( respectively , a left zero , that is , @xmath12 ) .
we study the sizes of syntactic semigroups of finite / cofinite , definite , and reverse definite languages .
if @xmath13 is a regular language over alphabet @xmath14 , its syntactic semigroup is defined by the myhill congruence @xcite @xmath15 : for @xmath16 , @xmath17 the set @xmath18 of equivalence classes of the relation @xmath19 is the _
syntactic semigroup _ of @xmath20 .
it is well - known that this semigroup is isomorphic to the semigroup @xmath21 of transformations performed by non - empty words in the minimal deterministic finite automaton ( dfa ) recognizing @xmath20 @xcite , and it is usually convenient to deal with the latter semigroup .
it is obvious that the transformation semigroup of the minimal dfa of @xmath20 is identical to that of the minimal dfa of @xmath22 , the complement of @xmath20 .
the _ syntactic complexity _
@xmath23 of a language @xmath20 is the size of its syntactic semigroup , and @xmath24 , where @xmath25 denotes the cardinality of a set @xmath8 .
syntactic complexity can vary significantly among languages with the same state complexity @xcite , where the _ state complexity _ of a language is the number of states in its minimal dfa .
the observation that @xmath26 is a tight upper bound on the size of the transformation semigroup of a dfa with @xmath0 states was first made by maslov @xcite in 1970 , although this follows immediately from a 1935 result of piccard @xcite , who showed that three generators suffice to produce all transformations of a set of @xmath0 elements .
the interest in syntactic complexity of _ subclasses _ of regular languages is new . in 20032004 holzer and knig @xcite , and krawetz ,
lawrence and shallit @xcite studied _ unary _ and _ binary _ languages . in 2011
brzozowski and ye @xcite showed the following bounds : _
right ideals_tight upper bound @xmath27 ; _ left ideals_lower bound @xmath28 ; _ two - sided ideals_lower bound @xmath29 . in 2012 brzozowski , li and ye @xcite found the following bounds : _ prefix - free _ languages tight upper bound @xmath30 ; _ suffix - free _ languages lower bound @xmath31 ; _ bifix - free _ languages lower bound @xmath32 ; _ factor - free _ languages lower bound @xmath33 .
also in 2012 tight upper bounds were found for three subclasses of star - free languages by brzozowski and li @xcite : _ monotonic _ languages@xmath34 ; _ partially monotonic _
languages@xmath35 ; _ nearly monotonic _
languages@xmath36 , where @xmath37 is the binomial coefficient _
@xmath38 choose @xmath39_. it was conjectured in @xcite that the bound for nearly monotonic languages is also a tight upper bound for star - free languages . that bound is asymptotically @xmath40 .
we prove that @xmath1 is a tight upper bound for finite / cofinite languages , and that a growing alphabet of size at least @xmath2 is required to reach the bound . for reverse definite languages
the bound is also @xmath1 , but the alphabet size is now @xmath3 . we show that @xmath4 is a lower bound for definite languages , and that it can be reached with an alphabet of size @xmath41 .
we conjecture that this is also an upper bound , and prove the conjecture for @xmath6 .
there is a lack of left - right symmetry in several results for syntactic complexity in spite of the fact that the syntactic congruence is symmetric .
thus , in the case of ideals @xcite , it was easy to find a tight upper bound for right ideals , but no tight upper bound is known for left ideals .
it was easy to find a tight upper bound for prefix - free languages , but no tight upper bound is known for suffix - free languages @xcite .
this happens again here .
we have a tight upper bound for reverse definite languages , but no tight upper bound for definite languages .
section [ sec : prelims ] contains some preliminary material .
sections [ sec : fc][sec : def ] discuss the syntactic complexity of finite / cofinite , reverse definite , and definite languages , respectively , and section [ sec : conc ] concludes the paper .
a _ transformation _ of a set @xmath42 is a mapping of @xmath42 into itself .
we consider only transformations of finite sets , and assume without loss of generality that @xmath43 .
if @xmath44 is a transformation of @xmath42 , and @xmath45 , then @xmath46 is the image of @xmath38 under @xmath44 .
an arbitrary transformation can be written in the form @xmath47 where @xmath48 , @xmath49 , and @xmath50 .
we also use the notation @xmath51 $ ] for the transformation @xmath44 above .
if @xmath52 is a subset of @xmath42 , then @xmath53 , and the _ restriction _ of @xmath44 to @xmath52 , denoted by @xmath54 , is a mapping from @xmath52 to @xmath55 such that @xmath56 for all @xmath57 .
a _ permutation _ of @xmath42 is a mapping of @xmath42 _ onto _ itself .
a transformation @xmath44 is _ permutational _ if there exists some @xmath58 with @xmath59 such that @xmath54 is a permutation of @xmath52 .
otherwise , @xmath44 is _ non - permutational_. a _ constant _ transformation , denoted by @xmath60 , has @xmath61 for all @xmath38 .
the _ composition _ of two transformations @xmath62 and @xmath63 of @xmath42 is a transformation @xmath64 such that @xmath65 for all @xmath45 .
we usually omit the composition operator . a _ deterministic finite automaton _ ( dfa ) is a quintuple @xmath66 , where @xmath42 is a finite , non - empty set of _ states _ , @xmath14 is a finite non - empty _ alphabet _ , @xmath67 is the _ transition function _ , @xmath68 is the _ initial state _ , and @xmath69 is the set of _ final states_. we extend @xmath70 to @xmath71 in the usual way .
the dfa @xmath72 _ accepts _ a word @xmath73 if @xmath74 .
the set of all words accepted by @xmath72 is the language @xmath75 of @xmath72 .
two states of a dfa are _ distinguishable _ if there exists a word @xmath7 which is accepted from one of the states and rejected from the other .
otherwise , the two states are _
a dfa is _ minimal _ if all of its states are reachable from the initial state and no two states are equivalent .
all the minimal dfa s of a given language @xmath20 are isomorphic .
the notion of a dfa @xmath72 connects transformations to regular languages .
given a regular language @xmath20 , its minimal dfa @xmath76 , and a word @xmath77 , the transition function @xmath78 is a _ transformation _ of @xmath42 , the transformation caused by @xmath7 .
when convenient , we identify a word with its corresponding transformation .
the ( _ left _ ) _ quotient _ of a language @xmath79 by a word @xmath80 is the language @xmath81 .
note that @xmath82 , where @xmath83 is the empty word .
the _ quotient dfa _ of a regular language @xmath20 is @xmath66 , where @xmath84 , @xmath85 , @xmath86 , and @xmath87 .
the quotient dfa is isomorphic to the minimal dfa accepting @xmath20 .
one of the simplest classes of regular languages is the class of finite and cofinite languages , where a language is _ cofinite _ if its complement is finite .
since the syntactic complexity bounds for finite and cofinite languages are identical , we restrict our analysis here to finite languages .
let @xmath20 be a regular language and @xmath88 be its minimal dfa .
it is well - known that @xmath20 is finite / cofinite if and only if there exists a numbering @xmath89 on @xmath42 so that for all @xmath73 , @xmath90 implies that @xmath91 or @xmath92 .
we define the set @xmath93 of transformations on @xmath94 with these properties : @xmath95 it is clear that @xmath93 is a semigroup under composition of size @xmath1 .
[ thm : fc ] let @xmath20 be a finite or cofinite language with state complexity @xmath0
. then the syntactic complexity of @xmath20 satisfies @xmath96 and this bound is tight .
let @xmath88 be the minimal dfa of @xmath20 .
the above discussion implies that we may label the states @xmath42 so that @xmath21 is a subsemigroup of @xmath93 .
therefore the bound holds .
let @xmath97 and @xmath98 .
let @xmath72 be a @xmath99 with states numbered @xmath100 , initial state @xmath101 , sink state @xmath0 , and a final state @xmath102 .
for each transformation @xmath103 , assign a letter in @xmath104 whose input transformation on @xmath72 is exactly @xmath44 . to show that @xmath72 is minimal , note that state @xmath105 is reached from the initial state by the transformation @xmath106 $ ] .
also , if @xmath38 and @xmath39 are two states and @xmath107 , then the transformation @xmath103 that has @xmath108 , and @xmath109 for all other @xmath110 , distinguishes the two states .
hence @xmath72 is minimal and accepts a finite language .
therefore the bound is tight .
a natural question is the minimal size of the alphabet required to achieve the upper bound .
let @xmath72 be the minimal dfa of a finite language @xmath20 with @xmath111 .
for any state @xmath45 and @xmath112 , it is clear that @xmath113 or @xmath114 .
it follows that if an input transformation @xmath103 satisfies @xmath115 for _ some _ @xmath116 , then any word @xmath7 corresponding to @xmath44 must have length 1 , that is , @xmath7 must be in @xmath104 .
[ thm : finalpha ] let @xmath117 be a finite or cofinite language with state complexity @xmath118 , and suppose that @xmath119
. then @xmath120 and this bound is tight . by theorem [ thm : fc ] , we may assume that @xmath121 the preceding discussion implies that @xmath122 is at least the number of transformations which satisfy @xmath123 for some @xmath124 .
let @xmath125 be the set of these transformations .
if we place the restriction @xmath126 for all @xmath116 then there are @xmath127 choices for these @xmath46 , and hence a total of @xmath128 such transformations .
therefore @xmath129 now let @xmath130 \in a_n$ ] be arbitrary .
let @xmath131 and @xmath132.$ ] then @xmath133 and @xmath134^k $ ] .
thus @xmath135 generates @xmath93 , and the bound is tight .
[ ex : fc ] for @xmath136 , the largest semigroup is @xmath137},{\bf [ 2,4,4,4]},{\bf [ 3,3,4,4]},[3,4,4,4],{\bf [ 4,3,4,4]},[4,4,4,4]\},\ ] ] and its minimal generating set is shown in boldface .
a _ reverse definite language _ is a language @xmath117 of the form @xmath138 , where @xmath139 and @xmath140 are finite languages . because reverse definite languages are characterized by prefixes of a fixed length , their minimal dfas ( and hence syntactic complexity bounds )
are very similar to those of finite / cofinite languages . if @xmath20 has state complexity 1 , then either @xmath141 or @xmath142
. since both these languages are in the finite / cofinite class , the bound @xmath1 of theorem [ thm : fc ] applies . for state complexities
@xmath143 , we note first that if @xmath144 is not a quotient of @xmath20 , then @xmath20 is cofinite .
otherwise , @xmath144 and @xmath145 are both quotients of @xmath20 .
let @xmath88 be the minimal dfa of @xmath20 , and label the states corresponding to @xmath144 and @xmath145 with @xmath102 and @xmath0 , respectively .
one can number the other states in @xmath42 so that for all words @xmath73 , if @xmath146 then @xmath147 with equality if and only if @xmath148 . the syntactic complexity results for reverse definite languages now follow directly from the finite / cofinite results .
let @xmath149 be a reverse definite language with state complexity @xmath97 .
then @xmath150 and this bound is tight .
moreover , if this language @xmath20 achieves this upper bound and @xmath151 , then @xmath152 , and this bound is tight .
first , if @xmath144 is not a quotient of @xmath20 , then @xmath20 is cofinite and hence has the same bounds as in the previous section . to find a cofinite witness @xmath20 meeting the bound @xmath1 ,
first find a finite witness @xmath22 as in the proofs of theorems [ thm : fc ] and [ thm : finalpha ] , and then interchange its final and non - final states .
otherwise , let @xmath72 be the minimal dfa recognizing @xmath20 , and let the states be totally ordered as in the preceding discussion .
define the set of transformations analogous to the finite case : @xmath153 then @xmath154 , which a straightfoward calculation shows to be a semigroup .
clearly , @xmath155 , thus proving the bound . to find a witness for this case ,
start with the finite witness as in the proofs of theorems [ thm : fc ] and [ thm : finalpha ] , make all transitions from state @xmath102 to go to itself , and make state @xmath0 the only final state .
for the minimal size of the alphabet , we define @xmath156 to be the set of transformations @xmath44 in @xmath157 satisfying @xmath123 for some @xmath158 . as in section
[ sec : fc ] , these transformations must correspond to individual letters in @xmath104 , hence proving the bound .
the same indirect counting argument shows that for @xmath159 , @xmath160 a similar argument also shows that @xmath161 generates @xmath157 ( using the transformations @xmath162 $ ] and @xmath163 $ ] in place of @xmath164 $ ] ) .
therefore the alphabet size bound is tight .
[ ex : rd ] for @xmath136 , the finite witness meeting the bound @xmath1 has the transformation set given in example [ ex : fc ] .
we modify this set by making @xmath102 the sink state , thus obtaining @xmath165},{\bf [ 2,4,3,4]},[3,3,3,4],[3,4,3,4],[4,3,3,4],[4,4,3,4]\},\ ] ] where the generators are in boldface , and state 4 is final .
a _ definite language _ is a language @xmath117 of the form @xmath166 , where @xmath139 and @xmath140 are finite languages . like finite / cofinite and reverse definite languages ,
definite languages are characterized by their transformation semigroups . in this case
, every transformation of the minimal dfa of a regular language must be non - permutational .
conversely , if the transformation semigroup of a minimal dfa contains only non - permutational transformations , then it accepts a definite language .
our goal for this section is to find the maximal size of a non - permutational transformation semigroup , that is , one which contains only non - permutational transformations .
there is a straightforward bijection between such transformations on @xmath167 and simple labeled forests on @xmath102 nodes .
this can be seen by constructing the graph on @xmath0 nodes with edges
@xmath168 representing @xmath169 , and then removing the unique node for which @xmath170 . then cayley s theorem @xcite shows that there are @xmath27 non - permutational transformations of @xmath167 .
identifying non - permutational transformations is not sufficient to find a syntactic complexity bound , as the set of such transformations does not form a semigroup for @xmath171 .
for example , the composition of @xmath172 $ ] and @xmath173 $ ] is @xmath174 $ ] , which is permutational .
two transformations _ conflict _ if there exists a permutational transformation in the semigroup that they generate .
we exhibit the following sets of non - permutational transformations which do not conflict ; they are similar to the semigroup @xmath93 from section [ sec : fc ] .
[ thm : def ] let @xmath143 , and define the following sets of transformations : @xmath175 then the set of transformations @xmath176 is a maximal non - permutational semigroup of size @xmath177 .
one can check that each @xmath178 is a semigroup .
let @xmath179 and @xmath180 , with @xmath91 .
a direct computation shows that @xmath181 , and @xmath182 ; hence @xmath183 is a semigroup .
moreover , for all @xmath184 , @xmath185 , and so all of the transformations are non - permutational .
a simple counting argument shows that @xmath186 since the @xmath178 are disjoint , @xmath187 for the maximality of @xmath183 , we show that adding any other non - permutational transformation creates a conflict .
let @xmath188 be non - permutational , with @xmath170 .
first suppose that there exists a @xmath189 with @xmath190 .
since @xmath44 is non - permutational , we may assume @xmath191 .
then there exists a @xmath192 with @xmath193 ; then @xmath194 and @xmath195 , and so @xmath44 and @xmath196 conflict .
if no such @xmath39 exists , then there must exist a @xmath197 with @xmath198 .
consider the sequence defined by @xmath199 , @xmath200 .
if there exists an @xmath201 such that @xmath202 , let @xmath201 be the minimal one .
let @xmath203 with @xmath204 and @xmath205 .
then @xmath206 , @xmath207 , and so @xmath208 is permutational .
now suppose all @xmath209 .
since @xmath44 is non - permutational , @xmath38 must appear in the sequence ; moreover , since @xmath210 , we can pick @xmath211 so that @xmath212 . since @xmath213 , we may find a transformation @xmath214 with @xmath215 and @xmath216 .
then @xmath217 , @xmath218 , and @xmath219 is permutational . to compute the generators of @xmath183
, we require the following definition .
let @xmath220 be the set of all transformations @xmath221 \in b_n$ ] with all @xmath222 .
define the function @xmath223 by @xmath224 $ ] , and also @xmath225 clearly , @xmath226 is a bijection between @xmath220 and @xmath227 .
. then 1 .
@xmath229 is the minimum set of generators for @xmath183 .
2 . @xmath230 for ( 1 ) , note that @xmath231 \in h_n$ ] . for any @xmath232
, we can write @xmath233^k$ ] with @xmath234 and @xmath235 , as in the proof of theorem [ thm : finalpha ] . therefore @xmath229 generates @xmath183 .
now let @xmath179 and @xmath180 , with @xmath236 .
we consider @xmath237 , and use the fact that each transformation @xmath184 satisfies @xmath238 .
there are two cases : 1 . if @xmath239 , then @xmath240 , hence @xmath241 .
2 . if @xmath242 , then @xmath243 , hence @xmath244 .
it follows that @xmath245 ; a similar argument shows that @xmath246 .
consequently , no transformation in @xmath229 is a composition of two others in @xmath183 , and so @xmath229 is the minimum generating set of @xmath183 . for ( 2 ) ,
we calculate @xmath247 , or equivalently @xmath248 because @xmath226 is a bijection .
a counting argument shows that @xmath249 .
therefore @xmath250 the following corollary establishes a direct connection with definite languages . for all @xmath143 ,
there exists a definite language @xmath20 with state complexity @xmath0 , syntactic complexity @xmath251 and alphabet size @xmath252 let @xmath253 be a dfa with @xmath254 , @xmath255 , @xmath256 , and @xmath257 with each letter representing a different transformation in @xmath229 , so that the transformation semigroup of @xmath72 is @xmath183 .
we claim that this is a minimal dfa of a definite language .
first , all the states are reachable by the constant transformations @xmath258 . also , any two states @xmath259 with @xmath260 are distinguishable by the transformation @xmath232 which acts as @xmath261 for @xmath262 , and @xmath109 for @xmath263 .
state @xmath0 is distinguishable from every other state because it is the only final state .
hence @xmath72 is minimal .
then by theorem [ thm : defequiv ] , @xmath72 accepts a definite language .
let @xmath20 be a definite language with state complexity @xmath143 .
then @xmath264 and if equality holds then @xmath265 [ ex : d ] for @xmath136 we have the following transformations in @xmath183 : @xmath266}\ } , \nonumber\\ b_{4,2 } & = \ { [ 2,2,2,2 ] , \bf{[3,2,2,2 ] } , { \bf [ 4,2,2,2 ] } \ } , \nonumber\\ b_{4,3 } & = \{{\bf[2,3,3,3]},{\bf [ 2,4,3,3]},[3,3,3,3],{\bf [ 3,4,3,3 ] } , [ 4,3,3,3],{\bf [ 4,4,3,3]}\ } , \nonumber\\ b_{4,4 } & = \{{\bf [ 2,3,4,4]},{\bf [ 2,4,4,4 ] } , { \bf [ 3,3,4,4 ] } , [ 3,4,4,4],{\bf [ 4,3,4,4]},[4,4,4,4]\}. \nonumber\end{aligned}\ ] ] the generators are shown in boldface .
though we have found tight upper bounds on the syntactic complexity of finite / cofinite and reverse definite languages , we have only conjectured the bounds on the syntactic complexity and the corresponding alphabet size for definite languages .
the conjecture has been verified through computational enumeration for @xmath267 , but remains unproven for @xmath268 . also , syntactic complexity bounds have yet to be found for the related higher classes in the dot - depth hierarchy of star - free languages , namely the generalized definite and locally testable languages .
it is possible that the technique used in this paper characterize allowable transformations in the syntactic semigroup and apply combinatorial arguments to count them can be used to find bounds for these languages as well .
brzozowski , j. : canonical regular expressions and minimal state graphs for definite events . in : proceedings of the symposium on mathematical theory of automata .
volume 12 of mri symposia series , polytechnic press , polytechnic institute of brooklyn , n.y .
( 1963 ) 529561 brzozowski , j. , li , b. : syntactic complexities of some classes of star - free languages . in kutrib , m. , moreira , n. , reis , r. ,
proceedings of the 14th international workshop on descriptional complexity of formal systems @xmath269dcfs@xmath270 .
volume 7386 of lncs , springer ( 2012 ) 117129 brzozowski , j. , ye , y. : syntactic complexity of ideal and closed languages . in mauri ,
g. , leporati , a. , eds . : 15th international conference on developments in language theory , dlt 2011 .
volume 6795 of lncs , springer ( 2011 ) 117128 | we study the syntactic complexity of finite / cofinite , definite and reverse definite languages .
the syntactic complexity of a class of languages is defined as the maximal size of syntactic semigroups of languages from the class , taken as a function of the state complexity @xmath0 of the languages .
we prove that @xmath1 is a tight upper bound for finite / cofinite languages and that it can be reached only if the alphabet size is greater than or equal to @xmath2 .
we prove that the bound is also @xmath1 for reverse definite languages , but the alphabet size is @xmath3 .
we show that @xmath4 is a lower bound on the syntactic complexity of definite languages , and conjecture that this is also an upper bound , and that the alphabet size required to meet this bound is @xmath5 .
we prove the conjecture for @xmath6 .
* keywords : * definite , finite automaton , finite / cofinite , regular language , reverse definite , syntactic complexity , syntactic semigroup |
nonleptonic hyperon decays were very important in early studies of parity violation in particle physics . in a hyperon decay ,
@xmath6 ( @xmath7:hyperon , @xmath8:baryon ) , the angular distribution of the baryon in the center - of - mass ( cm ) system of @xmath7 takes the form @xmath9 , where @xmath10 is the polarization vector of the hyperon , @xmath11 is the momentum unit vector of the baryon and @xmath12 is the hyperon decay parameter , which characterizes the parity violation in hyperon nonleptonic decays . for the decay @xmath13 ,
the decay parameter is @xmath14 @xcite ; for the counterpart @xmath15 , the decay parameter was only measured by the dm@xmath16 collaboration with large uncertainty : @xmath17 @xcite .
a precise measurement of the @xmath1 decay parameter allows a more sensitive search for @xmath18 symmetry violation in @xmath19 decays @xcite .
a @xmath18-odd observable , @xmath20 , may be defined as @xmath21 if @xmath18 is conserved , @xmath22 and this observable vanishes ; while a nonzero value of @xmath20 implies evidence for @xmath18 asymmetry in @xmath19 decays .
previous analyses to search for @xmath18 asymmetry in @xmath19 nonleptonic decays have been performed at @xmath23 colliders by the r608 @xcite and ps185 @xcite collaborations , and at an @xmath24 collider by the dm2 collaboration @xcite , but the precisions of the measurements are limited by low statistics .
a precise @xmath1 decay parameter is also essential in determining the @xmath25 or @xmath26 decay parameters .
nonpolarized @xmath25 or @xmath26 decays can produce polarized @xmath1 particles , so in the @xmath1 rest frame , the angular distribution of the final state antiproton , takes the form @xmath27 @xcite . to extract @xmath28 from the product @xmath29 the value of @xmath2 is required .
a similar argument holds for @xmath26 decays .
pair production of @xmath30 in @xmath31 decays allows one to study the @xmath1 decay parameter .
although the @xmath19 and @xmath1 particles are nonpolarized in @xmath32 decays , their helicities are correlated by helicity conservation .
hence , the @xmath1 decay parameter can be extracted from the helicity correlation between @xmath19 and @xmath1 .
experimentally , this decay is very clean and can be reconstructed with high efficiency by selecting events with four charged tracks .
the @xmath33 @xmath31 decays used for this analysis were taken with the besii detector at the bepc storage ring at a center - of - mass energy corresponding to _
@xmath34_. they offer an opportunity to measure a more precise @xmath1 decay parameter .
the beijing spectrometer ( bes ) detector is a conventional solenoidal magnet detector that is described in detail in ref .
@xcite ; besii is the upgraded version of the bes detector @xcite . a @xmath35-layer vertex chamber ( vtc ) surrounding the beam pipe provides trigger and track information .
a 40-layer main drift chamber ( mdc ) , located radially outside the vtc , provides trajectory and energy loss ( @xmath36 ) information for charged tracks over @xmath37% of the total solid angle .
the momentum resolution is @xmath38 ( @xmath39 in gev/@xmath40 ) , and the @xmath36 resolution for hadron tracks is @xmath41 % .
an array of @xmath42 scintillation counters surrounding the mdc measures the time - of - flight ( tof ) of charged tracks with a resolution of @xmath43 ps for hadrons .
radially outside the tof system is a @xmath35 r.l .
, lead - gas barrel shower counter ( bsc ) .
this measures the energies of electrons and photons over @xmath44% of the total solid angle with an energy resolution of @xmath45 ( @xmath46 in gev ) .
outside of the solenoidal coil , which provides a @xmath47 t magnetic field over the tracking volume , is an iron flux return that is instrumented with three double layers of counters that identify muons of momentum greater than @xmath48 gev/@xmath40 .
a geant@xmath49 based monte carlo ( mc ) program with detailed consideration of detector performance ( such as dead electronic channels ) is used to simulate the besii detector .
the consistency between data and monte carlo has been carefully checked in many high purity physics channels , and the agreement is quite reasonable @xcite .
several @xmath50 mc samples are generated and used for determining the detection efficiency , determining the normalization factor in background subtraction , and performing an input - output check .
the @xmath19 is reconstructed under the assumption of @xmath51 decay , and @xmath1 under the @xmath52 hypothesis .
a candidate track is required to have a good helix fit , the polar angle must satisfy @xmath53 , and transverse momentum @xmath54gev / c .
events are required to have four charged tracks with total charge of zero . protons and antiprotons
are identified using tof and @xmath36 information with the requirement that their confidence levels be larger than @xmath55 . a four constraint ( 4c )
kinematic fit is applied under the @xmath56 hypothesis , and @xmath57 is required .
a comparison of @xmath58 distributions of data and mc simulation is shown in fig .
[ scaterplot ] ( a ) . the scatter plot of @xmath59 versus @xmath60 is shown in fig .
[ scaterplot ] ( b ) , and @xmath61 is clearly seen .
the @xmath19 and @xmath1 are selected by requiring @xmath62 and @xmath63 .
the mass distributions of @xmath64 and @xmath65 are shown in fig .
[ mlambda ] . after applying the above selection criteria ,
@xmath66 @xmath50 events are selected .
the angular distribution of @xmath19 in the @xmath31 rest frame is found to be consistent with the dedicated analysis in ref .
@xcite within uncommon statistical errors .
backgrounds are studied with mc simulations .
the main backgrounds come from @xmath67 ( 28 events ) , @xmath68 ( 36 events ) , @xmath69 ( 37 events ) , @xmath70 ( 10 events ) , @xmath71 ( 5 events ) , @xmath72 ( 1 events ) , and @xmath73 ( 8 events ) .
the fraction of remaining background events in the data sample is about @xmath74 .
background decays to intermediate states including @xmath19 or @xmath1 are generated with the helicity amplitude method @xcite , in which hyperon decay parameters are set according to pdg values @xcite .
the effect of the background contamination on the physics results is included as one source of systematic error .
the angles used in this analysis are defined in the helicity frame , shown in fig .
[ frame ] . for @xmath75 ,
the @xmath76 axis of the @xmath31 rest frame is along the @xmath19 out - going direction , and the solid angle @xmath77 is between the @xmath78 direction and the @xmath19 out - going direction . for @xmath79 , the solid angle of the daughter particle @xmath80
is referred to the @xmath19 rest frame , and the @xmath76 axis is also the @xmath19 out - going direction .
for @xmath1 we use symbols with bars . for @xmath81 ,
the partial decay rate is @xcite @xmath82\nonumber\\ & -&(1+\alpha)(1+\cos^2\theta)(\alpha_\lambda\alpha_{\bar{\lambda}}\cos\theta_1\cos\bar{\theta}_1 - 1),\end{aligned}\ ] ] where @xmath83 , @xmath84 is the angular distribution parameter for @xmath19 , and @xmath85 is the @xmath86 decay parameter . equation .
( [ ampllb ] ) only allows one to obtain the product @xmath87 from a fit to data . to extract the value of @xmath2
, one needs to fix the @xmath19 decay parameter @xmath88 , say , at the world average value @xmath89 @xcite . to determine the @xmath18-odd observable defined in eq .
( [ cpa ] ) , one may make a replacement @xmath90 in eq .
( [ ampllb ] ) .
an unbinned maximum likelihood method is used to fit the data . as widely used in partial wave analyses@xcite ,
a normalized probability density function for @xmath91 is defined as @xmath92 where @xmath93 , and @xmath84 and @xmath2 are parameters to be determined .
@xmath94 is given by eq .
( [ ampllb ] ) .
@xmath95 is the total cross section given by @xmath96 where @xmath97 is the detection efficiency .
the total cross section @xmath95 can be determined by mc numerical integration with a phase space generator over the allowed kinematic region of @xmath98 , i.e. @xmath99 where @xmath100 is the number of selected mc events .
the maximum likelihood function is given by the joint probability density of the selected @xmath101 events @xmath102 where @xmath103 is the number of selected data events .
to determine the unknown parameters @xmath84 and @xmath2 , the function @xmath104 , which is defined as @xmath105 is minimized . to check the goodness of fit in our analysis
, we define @xmath106 over the scatter plot of @xmath107 versus @xmath108 . here
@xmath109 is the number of cells , @xmath110 and @xmath111 are the numbers of events in the @xmath112th cell of the scatter plot for data and mc simulation , respectively .
such a variable should be distributed according to the @xmath113 distribution with @xmath114 degrees of freedom , where @xmath115 is the number of parameters to be determined in our fit . in this experiment ,
data and mc events are divided into @xmath116 cells in the scatter plots .
fitting the data using eq .
( [ minimum ] ) , we obtain @xmath117 @xmath118 while fitting the data using eq .
( [ minimum ] ) with parameters @xmath84 and @xmath20 , we obtain @xmath119 where the errors are statistical only . comparisons between the data and the fit results are shown in fig .
[ asydis ] ( a ) and ( b ) .
the fit yields @xmath120 .
if we force @xmath121 , we obtain @xmath122 , where the error is statistical only .
to validate the fitting procedure , a mc sample of @xmath123 @xmath124 events is produced according to eq .
( [ ampllb ] ) . the input parameters are @xmath125 and @xmath126 .
the mc sample is required to pass the same selection criteria as used for data selection , and the same fitting procedure is applied to the selected events with @xmath127 fixed to @xmath128 .
the fit yields @xmath129 and @xmath130 , consistent with the input values to within the @xmath131 statistical errors .
contamination from background channels is studied using mc samples , including @xmath132 and @xmath133 .
background channels with @xmath134 intermediate states are generated with full helicity amplitude information , and the decay parameters for @xmath135 and their antiparticles are fixed at the world average values @xcite .
these samples are required to pass the same selection criteria as used for the the signal channel @xmath136 .
the selected background events are subtracted from the data in the fit .
the difference of the @xmath1 decay parameter @xmath2 from the fit without background events , @xmath137 , is taken as the uncertainty from background contamination . the systematic error due to the uncertainty on the @xmath19 decay parameter @xcite is determined . in fitting to data ,
the central value of @xmath2 is obtained by fixing @xmath89 ; the difference from the central value , @xmath138 , is obtained by changing @xmath127 by 1 standard deviation ( @xmath139 ) .
the consistency between data and the mc simulation of the detector response for @xmath140 events can be determined using the channel with the same final states @xmath141 after rejecting @xmath19 , @xmath1 , @xmath142 and @xmath143 intermediate states . therefore fitting to this sample with eq .
( [ ampllb ] ) should yield @xmath144 .
the difference from zero gives the systematic error due to the mc simulation of the detector response . in fitting this sample with eq .
( [ ampllb ] ) , angles for @xmath19 or @xmath1 are replaced with ones for the quasi - two body systems ( @xmath64 and @xmath65 ) . the fit yields @xmath145 ; the uncertainty from mc simulation and detector response is taken as @xmath146 , the linear sum of the deviation and its uncertainty .
the resulting distributions of @xmath107 and @xmath147 are shown in fig .
[ controlsample ] .
the systematic error associated with mc simulation of hadronic interactions of final state particles with the detector is estimated with two models , gcalor @xcite and fluka @xcite .
the central value is obtained with gcalor model , while the difference from the central value using fluka model , @xmath148 , is taken as the systematic error . the uncertainty associated with the simulation of the mdc wire resolution
is estimated with two different versions of mc simulation software , which yields @xmath149 .
table [ sys ] summarizes all systematic errors .
the total systematic error is estimated to be @xmath150 , assuming all the sources are independent and summing them in quadrature .
.systematic errors in @xmath2 . [ cols="<,^ " , ]
the bes collaboration thanks the staff of bepc and computing center for their hard efforts .
this work is supported in part by the national natural science foundation of china under contracts no .
10491300 , no . 10225524 , no .
10225525 , no . 10425523 , no .
10625524 , no . 10521003 , no .
10821063 , no . 10825524 and no . 10225522 ( tsinghua university ) , the chinese academy of sciences under contract no .
kj 95t-03 , the 100 talents program of cas under contracts no .
u-11 , no .
u-24 , no .
u-25 , and the knowledge innovation project of cas under contracts no .
u-602 , no .
u-34 ( ihep ) , and the department of energy under contract no
. de - fg02 - 04er41291 ( u. hawaii ) .
99 t. d. lee and c. n. yang , phys . rev . * 108 * , 1645 ( 1957 ) ; + t. d. lee _ et al .
_ , phys . rev . * 106 * , 1367 ( 1957 ) . c. amsler
_ et al . _
( particle data group ) , phys .
b * 667 * , 1 - 5 ( 2008 ) .
m. h. tixier _
( dm2 collaboration ) , phys .
b * 212 * , 523 ( 1988 ) .
a. w. chan _
( e756 collaboration ) , phys .
d * 58 * , 072002 ( 1998 ) .
a. pais , phys .
lett . * 3 * , 242 ( 1959 ) .
p. chauvat _
( r608 collaboration ) , phys .
b * 163 * , 273 ( 1985 ) . p. d. barnes _ et al .
c * 54 * , 1877 ( 1996 ) .
y. c. chen _ et al .
_ ( hypercp collaboration ) , phys .
d * 71 * , 051102(r ) ( 2005 ) .
j. z. bai _ et al . _
( bes collaboration ) , nucl .
instr . meth .
a * 344 * , 319 ( 1994 ) .
j. z. bai _ et al . _
( bes collaboration ) , nucl .
instr . meth .
a * 458 * , 627 ( 2001 ) .
m. ablikim _
( bes collaboration ) , nucl .
instr . meth .
a * 552 * , 344 ( 2005 ) .
m. ablikim _
( bes collaboration ) , phys .
b * 632 * , 181 ( 2006 ) .
hong chen and r. g. ping , phys .
d * 76 * , 036005 ( 2007 ) ; + b. zhong , r. g. ping and z. j. xiao , chin .
c * 32 * , 692 ( 2008 ) .
m. ablikim _
( bes collaboration ) , phys .
b * 645 * , 19 ( 2007 ) . c. zeitnitz and t. a. gabriel , nucl .
instr . meth .
a * 349 * , 106 ( 1994 ) .
k. hanssgen , h. j. mohring and j. ranft , nucl .
* 551 * , 88 ( 1984 ) ; + j. ranft and s. ritter , z. phys .
c * 20 * , 347 ( 1983 ) ; + a. fasso _
et al . _ , in proceedings of the workshop on simulating accelerator radiation enviroments(sare 1 ) , santa fe , new mexico , 1993 ( los alamos report no . la-12835-c , 1994 ) . | based on a sample of @xmath0 decays collected with the besii detector at the bepc , the @xmath1 decay parameter @xmath2 for @xmath3 is measured using about 9000 @xmath4 decays .
a fit to the joint angular distributions yields @xmath5 , where the first error is statistical , and the second systematic . |
the fibonacci sequence origin is attributed and referred to the first edition ( lost ) of `` fiber abaci '' ( 1202 ) by leonardo fibonacci [ pisano ] ( see second edition from 1228 reproduced as il liber abaci di leonardo pisano publicato secondo la lezione codice maglibeciano by baldassarre boncompagni in scritti di leonardo pisano vol . 1 , ( 1857 ) , rome ) .
this note is inspired by @xcite and by primary binomial summation formula ( [ summ1 ] ) for @xmath1 : @xmath2 it is the famous and popular now identity ( [ summ2 ] ) propagated via internet @xcite .
here it is : @xmath3 this respectable relation and the property of pascal triangle is illustrated by the figure 1 . _
( see : @xcite ) + the connection constants @xmath14 from relation ( [ contriad ] ) of the triad @xcite under consideration are interpreted @xcite as the number of @xmath15-dimensional subspaces in @xmath16-th dimensional space @xmath17 over galois field @xmath18 . restarting this in incidence algebras language @xcite
we conclude that in the lattice @xmath19 of all subspaces of @xmath17 : @xmath20 @xmath21:=the\;number\;of\;distinct\;elements\;\textrm{z}\;in\;a\;segment\ ; [ \textrm{x , y}]\ ] ] @xmath22\;is\;of\;type\;\beta\;while\ ; [ \textrm{z , y}]\;is\;of\;type\;\gamma,\ ] ] where @xmath23 and the type of @xmath24 $ ] is equal to dim[x / y ] .
( here [ y / x ] denotes the subtraction of linear subspaces @xcite ) . _
the corresponding to ( [ contriad ] ) @xmath0-gaussian - pascal triangles have analogous summation formula property as the one given by ( [ summ1 ] ) and illustrated by the figure 1 .
let it be now illustrated for the case @xmath25 by the figure 2 showing the first elements of the sequence @xmath26}\}$ ] ( @xmath27 , @xmath25 ) : @xmath28}=\sum \limits_{k\leq n } \binom{n - k}{k}_q(q^j)^{k-1},\quad n , j\geq 0,\;\;q\geq 1}\\ { f_0^{[q^j]}=0 } \end{array } } } \right.\ ] ] we shall call the sequence of sequences defined by ( [ f - g ] ) and with given `` the fibonacci @xmath0-gauss family '' of @xmath29-pascal @xmath0-gauss sequences . for @xmath4 ( [ f - g ] ) becomes ( [ summ2 ] ) and we get the pascal triangle case with the resulting fibonacci sequence via ( [ summ1 ] ) or ( [ summ2 ] ) .
using the recurrence ( [ contriad ] ) one derives - in a way one does that in @xmath4 case - the following recurrence relations for the fibonacci @xmath0-gauss family sequences : @xmath30}=f_{n+1}^{[q^{j+1}]}+f_{n}^{[q^j]},\quad n , j \geq 0,\;q\geq 1,}\\ { f_0^{[q^j]}=0}\\ { f_1^{[q^j]}=q_0^j } \end{array } } } \right.\ ] ] 10 a. k. kwaniewski : _ on duality triads _ , bull . soc .
letters d * 53 * rech .
* 42 * ( 2003 ) , 11 .
arxiv : math.gm/0402260 v1 feb .
2004 a. k. kwaniewski : _ on fibonomial and other triangles versus duality triads _ , bull .
letters d * 53 * rech .
* 42 * ( 2003 ) , 27 .
arxiv : math.gm/0402288 v1 feb .
2004 dunlap-54 ( corrected ) , vajda-54 ( corrected ) : _ summations with binomial coefficients _ , http://www.mcs.survey.ac.uk/personal/r.knott/fibonacci/fibformulae.html http://www.mcs.survey.ac.uk/personal/r.knott/fibonacci/fibmaths.html e. spiegel , gh .
j. odonnell : _ incidence algebras _ , marcel dekker , inc basel 1997 . | the summation formula within pascal triangle resulting in the fibonacci sequence is extended to the @xmath0-binomial coefficients @xmath0-gaussian triangles @xcite .
ams classification numbers : 11b39 , 11a39 . |
f. b. jones @xcite proved in 1942 , in a famous paper , the existence of additive discontinuous functions @xmath1 whose graph @xmath3 is connected , and characterized them .
these functions are extraordinary since their graphs are dense connected subsets of the plane , containing exactly one point in each vertical line @xmath4 @xcite . in his paper
the author also stated , without proof , that the graph of a discontinuous additive function must be connected or totally disconnected . for this result
he just referenced another famous paper , by hamel @xcite , but the proof is not there .
indeed , up to our knowledge , a proof of this dichotomy result has never appeared in the literature . in this note
we prove that the graph of a discontinuous monomial is either connected or totally disconnected , and we characterize the discontinuous monomial functions @xmath1 with connected graph by means of a big graph property .
we also study the connected components of the graphs of additive functions @xmath2 for @xmath5 .
these results should be a good starting point to prove that , for larger classes of functions , such as the generalized polynomials or exponential polynomials over the real line , the graphs of the elements of these sets are either connected or totally disconnected . to find examples of both situations
is an easy corollary of the structure of these functions and jones s existence result of additive discontinuous functions with connected graph .
recall that @xmath1 is an @xmath0-monomial function if it is a solution of the so called monomial functional equation @xmath6 it is known that @xmath7 satisfies @xmath8 if and only if @xmath9 for a certain multi - additive and symmetric function @xmath10 , and that @xmath7 is a polynomial function of degree at most @xmath0 ( i.e. , @xmath7 solves frchet s functional equation @xmath11 ) if and only if @xmath12 , where @xmath13 is a @xmath14-monomial function for @xmath15 .
( see , for example , @xcite , @xcite , for the proofs of these claims ) .
[ dichotomy , for monomial functions @xmath1 ] [ main1 ] let @xmath1 be a @xmath0-monomial function .
then @xmath17 is either connected or totally disconnected .
furthermore , both cases are attained by concrete examples of discontinuous @xmath0-monomials @xmath1 , for every @xmath18 .
let @xmath7 be an @xmath0-monomial .
suppose that @xmath17 is not totally disconnected .
then there exists a connected component @xmath19 containing at least two different points @xmath20 and @xmath21 . clearly , @xmath22 . hence ,
if @xmath23 denotes the horizontal projection of the plane , @xmath24 , then @xmath25 is a connected subset of @xmath26 which contains two distinct points .
hence @xmath27 is an interval with non - empty interior and @xmath28 .
set @xmath29 and @xmath30 . obviously , @xmath31 and @xmath32 .
in particular either @xmath33 or @xmath34 .
assume , with no loss of generality , that @xmath33 ( the other case can be treated with similar arguments , or reduced to this one by using that every monomial is either even or odd , since @xmath35 for all rational number @xmath36 and all @xmath37 , which implies that @xmath38 for all @xmath37 ) .
given @xmath39 we consider the maps @xmath40 then @xmath41 is continuous . hence @xmath42 is a connected subset of the plane for every @xmath43 .
given @xmath43 , we have that @xmath44 if and only if for some @xmath45 the equality @xmath46 holds .
in other words , this intersection is nonempty if and only if @xmath47 forcing equality between the first components of these vectors we get @xmath48 , which means that @xmath49 .
furthermore , under this restriction , we get the equality between the second components of the vectors for free , since @xmath50 thus , we have demonstrated that @xmath44 if and only if there exist @xmath51 such that @xmath52 . in particular , when this holds true , the property is satisfied for all @xmath43 simultaneously and @xmath53 is connected .
furthermore , @xmath54 implies that @xmath55 , so that @xmath56 for all @xmath43 .
hence @xmath57 is always a subset of @xmath17 .
assume , by the moment , that @xmath58 and @xmath59 , and take @xmath60 such that @xmath61 .
then @xmath58 and @xmath61 imply that @xmath62 .
take @xmath63 such that @xmath64 .
then @xmath65 and @xmath66 , so that @xmath67 too .
thus , in this case , @xmath57 is a connected subset of @xmath17 .
but we also have , in this case , that @xmath68 .
this implies that @xmath69 , which contradicts both @xmath70 and @xmath71 . hence either @xmath72 or @xmath73 . if @xmath74 and @xmath75 , then for any @xmath76 , @xmath77 and @xmath78 both belong to @xmath27 , so that @xmath79 is connected , which contradicts @xmath58 . if @xmath72 and @xmath80 , then for any @xmath76 , @xmath49 and @xmath77 both belong to @xmath27 , so that @xmath57 is connected and contains @xmath81 .
this forces @xmath73 again . if , in particular , @xmath82 , we have @xmath83 , and we obtain , analogously to the previous arguments , that @xmath84 . in this case
we thus have @xmath85 and @xmath86 .
finally , let us consider the case @xmath87 . then either @xmath88 or @xmath89 . in the former case @xmath90
is connected .
furthermore , if we define @xmath91 , it is clear that @xmath92 is also a connected subset of @xmath17 .
furthermore , @xmath93 , so that @xmath94 is connected . in the latter case , when @xmath89 , we have that @xmath95 is connected and @xmath96 is disconnected . hence there
exist open sets @xmath97 and @xmath98 such that @xmath99 , @xmath100 , @xmath101 , and @xmath102 .
we may assume , with no loss of generality , that @xmath103 . then @xmath104 .
since @xmath105 is an accumulation point of @xmath106 , we obtain @xmath107 as well .
this yields that @xmath95 is disconnected , which is a contradiction .
till now , we have demonstrated that , if @xmath1 is a monomial function , then @xmath17 is either connected or totally disconnected .
let us now show that both cases are attained by concrete examples . for totally disconnected graphs the example can be easily constructed .
indeed , given @xmath108 any hamel basis of @xmath26 satisfying @xmath109 and let @xmath110 be a positive integer , we consider @xmath111 , the unique @xmath112-linear map which satisfies @xmath113 and @xmath114 for every @xmath115 . obviously @xmath116 is an @xmath0-monomial and @xmath117 .
hence the graph of @xmath118 is totally disconnected .
the existence of discontinuous @xmath0-monomials with connected graph follows from the existence of discontinuous additive functions @xmath1 with connected graph @xmath17 , a fact that was demonstrated by jones by using a nontrivial set theoretical argument on ordinals ( * ? ? ?
* theorems 4 and 5 ) .
indeed , assume that @xmath1 is additive , discontinuous , and @xmath17 is connected .
then @xmath119 is a discontinuous @xmath0-monomial function with connected graph , since the function @xmath120 given by @xmath121 is continuous and transforms the graph of @xmath7 onto the graph of @xmath122 .
[ dichotomy , for additive functions @xmath1 ] [ dichotomy ] let @xmath1 be an additive function . then @xmath17 is connected or totally disconnected .
furthermore , there exists discontinuous additive functions @xmath1 with connected graph @xmath17 .
* proof . *
any additive function is a @xmath123-monomial function .
@xmath124 if @xmath17 is connected , we have two cases : either @xmath7 is continuous and @xmath125 is a one - dimensional vector space , or @xmath17 is a connected dense additive subgroup of @xmath126 .
the following theorem may be also of interest : [ @xmath127-chotomy property of additive functions ] [ d ] if @xmath2 is an additive function , then : * there exists @xmath128 such that the connected component @xmath129 of @xmath17 which contains the point @xmath130 is a dense subgroup of an @xmath131-dimensional vector subspace of @xmath132 .
furthermore , every connected component of @xmath17 results from @xmath129 by a translation . *
all cases described in @xmath133 are attained by concrete examples
. * proof .
* @xmath133 previous to introduce the main argument , it is necessary to recall two basic facts about topological groups . concretely ,
if @xmath129 is a topological group , and @xmath134 denotes the identity component of @xmath129 ( i.e. , the biggest connected subset of @xmath129 which contains the identity @xmath135 ) , then @xmath134 is a closed normal subgroup of @xmath129 . furthermore , the elements of the quotient group @xmath136 are just the connected components of @xmath129 @xcite .
we consider @xmath2 an additive function and we set @xmath137 .
obviously , the additivity of @xmath7 implies that @xmath17 is an additive subgroup of @xmath132 .
let @xmath129 be the connected component of @xmath17 which contains the zero element .
then every connected component of @xmath17 results from @xmath129 by a translation .
@xmath129 is a connected additive subgroup of @xmath132 .
hence , its topological closure @xmath138 is also a connected subgroup of @xmath132 .
it is known that the topological closure of any additive subgroup @xmath139 of @xmath132 satisfies @xmath140 for a certain vector subspace @xmath141 of @xmath132 and a discrete additive subgroup @xmath142 of @xmath132 ( see ( * ? ? ?
* theorem 3.1 ) for a proof of this fact ) .
it follows that @xmath143 for a certain vector subspace @xmath141 of @xmath132 .
hence every connected component of @xmath17 is the translation @xmath144 of a dense connected additive subgroup @xmath129 of the vector space @xmath141 for some @xmath145 .
note that , if @xmath146 then @xmath17 is totally disconnected and , if @xmath147 , then @xmath17 is a connected dense additive subgroup of @xmath132 .
all the other cases represent an intermediate situation .
for example , if @xmath7 is continuous , then @xmath125 is a @xmath148-dimensional vector subspace of @xmath132 .
@xmath149 all cases described by theorem [ d ] can be constructed easily , since all functions @xmath2 of the form @xmath150 , with @xmath151 additive for each @xmath14 , are additive , and we can use the dichotomy result for each one of these functions @xmath152 , @xmath153 .
@xmath124 while searching in the literature for a demonstration of corollary [ dichotomy ] , the first author commented this question to professor lszl szkelyhidi , who also was unable to find the proof nowhere .
then , he got a very nice independent proof of the result @xcite .
indeed , for @xmath154 we get the dichotomy result as follows ( this is szkelyhidi s idea ) : let @xmath155 denote the horizontal projection @xmath24 , and let @xmath156 be the projection of the connected component @xmath129 of @xmath17 which contains the zero element
. then @xmath157 or @xmath158 , since the only connected subgroups of the real line are @xmath159 and @xmath26 .
thus , if @xmath17 is not totally disconnected , then @xmath158 , which implies @xmath160 and hence , @xmath17 is connected .
unfortunately , this simple proof seems to be very difficult to generalize for the case of monomial functions @xmath1 , since the graph of an @xmath0-monomial function is in general not an additive subgroup of @xmath126 .
we hope this justifies to introduce the proof of theorem [ main1 ] .
recently , almira and abu - helaiel characterized the topological closures of the graphs of monomial functions as follows ( * ? ? ? * theorem 2.7 ) : * if @xmath166 and @xmath167 , then @xmath168 . * if @xmath166 and @xmath169 , then @xmath170 if @xmath171 is an even number , and @xmath172 if @xmath173 is an odd number . in particular ,
if @xmath174 , we get the half space @xmath175 for @xmath171 and the union of the first and third quadrants @xmath176 , for @xmath173 .
* if @xmath177 and @xmath167 , then @xmath178 if @xmath171 is an even number , and @xmath179 if @xmath173 is an odd number . in particular , if @xmath180 , we get the half space @xmath181 for @xmath171 and the union of the second and fourth quadrants @xmath182 , for @xmath173 .
[ main ] let @xmath1 be a discontinuous @xmath0-monomial function and let @xmath161 and @xmath185 .
then @xmath17 is connected if and only if @xmath17 intersects all continuum @xmath186 which touches two distinct vertical lines . the proof follows the very same arguments used by jones @xcite in his original proof for the case of additive functions .
the main difference is that , for additive functions , the closure of the graph of a discontinuous additive function is the all plane and , for monomials , the corresponding sets are those shown in theorem [ graph_monomials ] .
assume that @xmath17 is not connected .
then @xmath187 with @xmath188 open subsets of the real plane , @xmath189 , @xmath190 and @xmath191 .
we can assume that @xmath192 is connected , since connected components of open subsets of @xmath126 are open sets .
indeed , by making @xmath192 or @xmath141 bigger and bigger , just deleting properly some parts of the borders @xmath193 or @xmath194 , we can assume that both @xmath192 and @xmath141 are connected and share a common border @xmath195 .
furthermore , this common frontier is necessarily connected , since the connectedness of the boundary of an open domain in @xmath126 is equivalent to the connectedness of its complement ( indeed , this result holds true for domains in @xmath196 for all @xmath197 @xcite ) .
now , the density of @xmath17 in @xmath198 implies that @xmath199 . to prove this
, we first observe that @xmath200 , since @xmath201 is an open set in the relative topology of @xmath202 which has empty intersection with @xmath203 .
thus , if @xmath204 , then there exist @xmath205 and @xmath206 such that @xmath207 , which contradicts that @xmath187 , since @xmath17 has at least one point in @xmath208 .
now we can use the characterization of the sets @xmath202 given in theorem [ graph_monomials ] to claim that @xmath209 contains a continuum which intersects two distinct vertical lines , since otherwise @xmath193 should contain the intersection of a vertical line with @xmath202 , a fact which leads to a contradiction , since @xmath17 is a graph and hence intersects all vertical lines .
this proves that , if @xmath17 intersects all continuum @xmath186 which touches two distinct vertical lines , then @xmath17 is connected .
let us now assume that @xmath17 is connected and let @xmath186 be a continuum which touches two distinct vertical lines .
if @xmath210 has non - empty interior then @xmath211 , since @xmath17 is dense in @xmath202 .
if @xmath212 and @xmath213 with @xmath214 , then , @xmath215\times\mathbb{r})$ ] separates @xmath216\times\mathbb{r})\cap \gamma_f$ ] in two ( or more ) components , since @xmath210 does not intersect the frontier of @xmath202 .
now , @xmath17 contains at least a point of each one of these components , since @xmath17 is dense in @xmath202 .
it follows that @xmath217 , since @xmath17 is connected , by hypothesis .
hence , if @xmath17 is connected , then @xmath17 intersects every continuum @xmath186 which touches two distinct vertical lines .
the implication @xmath17 connected implies @xmath218 connected is trivial .
let us prove the other implication .
indeed , assume that @xmath218 is connected with @xmath219 , @xmath220 a @xmath221-monomial function .
let @xmath210 be a continuum included into @xmath222 which touches two distinct vertical lines
. then @xmath223 is a continuum , @xmath224 , and @xmath122 touches two distinct vertical lines .
hence theorem [ main ] and the connectedness of @xmath218 imply that there exists @xmath225 such that @xmath226 .
thus @xmath227 and @xmath17 contains a point of @xmath210 .
it follows , again from theorem [ main ] , that @xmath17 is connected .
let us thus assume ( with no loss of generality ) that @xmath171 is even .
thanks to theorem [ graph_monomials ] we can also assume with no loss of generality that @xmath168 or @xmath228 for a certain @xmath169 , since the other cases have analogous proofs .
if @xmath17 is not connected , there exist a continuum @xmath186 with empty interior and two points @xmath213 with @xmath214 , such that @xmath229 .
obviously , the continuum @xmath210 separates @xmath230x_0,x_1[\times \mathbb{r } ) \cap \gamma_f$ ] in several disjoint open subsets of @xmath202 ( with the relative topology ) .
hence we can assume that @xmath231x_0,x_1[\times \mathbb{r})\cap\gamma_f \setminus k = u_k\cup v_k,\ ] ] with @xmath232 disjoint open subsets of @xmath202 , @xmath233 connected , and @xmath234 for certain @xmath235x_0,x_1[$ ] and @xmath33 .
let us set @xmath236 , @xmath237 .
then @xmath238 are open connected subsets of @xmath202 , @xmath239 , @xmath240x_0,x_1[\times \mathbb{r } \subseteq u_k^*\cup v_k^*$ ] , @xmath241 is a continuum which separates @xmath242x_0,x_1[\times \mathbb{r } \cap \gamma_f$ ] in exactly two disjoint open connected subsets of @xmath202 , @xmath243 and @xmath244 . @xmath245 and @xmath246 .
furthermore , the relation @xmath247 for all @xmath37 and all @xmath248 implies that @xmath249 for all rational number @xmath250 , where @xmath251 .
let us prove that the connected component of @xmath17 which contains the point @xmath252 with @xmath253x_0,x_1[$ ] , is the set @xmath254 . to prove this , we note that the sets @xmath238 separate any of these points from the points @xmath255 of the graph satisfying @xmath256x_0,x_1[$ ] .
thus it is only necessary to consider , to prove our claim , the following two cases : hence @xmath262 is a sequence of connected subsets of the plane which separates the point @xmath252 from any other point @xmath255 with @xmath263 , @xmath264x_0,x_1[$ ] , and @xmath265 for all @xmath0 ( see the figure ) .
it follows that @xmath254 is the connected component which contains the point @xmath252 .
the proof ends now easily .
indeed , if @xmath252 is any point of @xmath17 , there exists @xmath267 such that @xmath253\lambda x_0,\lambda^nx_1[$ ] ( since @xmath112 is a dense subset of @xmath26 ) and we can use the arguments above with @xmath268 instead of @xmath269 . | we prove that the graph of a discontinuous @xmath0-monomial function @xmath1 is either connected or totally disconnected .
furthermore , the discontinuous monomial functions with connected graph are characterized as those satisfying a certain big graph property .
finally , the connectedness properties of the graphs of additive functions @xmath2 are studied . |
data storage demands have been increasing very rapidly , leading to a lot of interest in distributed storage systems @xcite .
erasure - coded distributed storage systems provide reliable data storage at a a fraction of the cost as compared to the systems based on simple data replication .
characterizing the reliability of the storage system is an important problem .
reliability has been explored for exponential repair times for serial repair @xcite and parallel repair @xcite models .
recently , there has been interest to characterize the reliability for deterministc repair times @xcite , where geometric interpretations are provided . however , there is no complete characterization of reliability for parallel and serial repair models .
the authors of @xcite considers availability of the k - out - of - n systems with general repair times while this paper considers the reliability of k - out - of - n systems . in this paper
, we give the laplace stieltjes transform of the probability of data loss for both parallel and serial repair models . using the laplace stieltjes transform formulas , mean time to data loss can be found . even though expressions are difficult to evaluate exactly , we can find the dominant terms for the case when repair rate is much higher than the failure rate and find that the dominant term for the serial repair is the same as in the case of exponential repair in @xcite .
however , we also find that the parallel repair also gives the same dominant term as the serial repair , and thus not achieving the @xmath0 factor improvement in mean time to data loss as given in @xcite for the parallel exponential repair .
the results are compared in simulations , where we see that the mean time to data loss with parallel repair is higher , but the asymptote as the repair rate increases is the same .
the rest of the paper is organized as follows . in section 2
, we give our results for the serial repair model .
section 3 presents our results for the parallel repair model .
section 4 gives the numerical results and section 5 concludes this paper .
the serial repair model can be described with the following assumptions .
1 . @xmath1 components ( disks ) are identical and independent .
the failure rate for each disk is constant , and equals @xmath2 ( failure time is exponentially distributed for each disk ) .
3 . the repair times are independent and fixed , and equals @xmath3 .
4 . the repair for the previous broken components must be re - initialized when a new component fails , and only the newest broken component is repaired . 5 . the system stops working when @xmath4 components have failed for @xmath5 , @xmath6 .
suppose that the probability of data loss in time @xmath7 is @xmath8 .
then the laplace stieltjes transform of reliability is given as @xmath9 $ ] , where @xmath10 is the random variable representing the time to data loss , and @xmath11 .
the next result gives the laplace stieltjes transform of reliability for serial repair .
the laplace stieltjes transform of reliability for serial repair is given by @xmath12 where @xmath13 let @xmath14 be the one - step transition probability from state @xmath15 to state @xmath16 in time @xmath7 , in which one - step implies that there is only one repair or failure during an interval of time @xmath7 .
then , @xmath14 is given as @xmath17 the laplace - stieltjes transform is given by @xmath13 let @xmath18 be the first recurrence time from state @xmath15 to state @xmath16 .
then , @xmath18 is given as @xmath19 the laplace - stieltjes transform is then given by @xmath20 the above can be used to write @xmath21 let @xmath22 be the probability that the system is at state @xmath15 at time @xmath7 starting from state @xmath23 at time @xmath24 , and let @xmath25 be the laplace - steiltjes transform .
then , the probability of data loss is given by @xmath26 .
thus , it is enough to find @xmath27 . from equation
, we have @xmath28 using @xmath29 , we can find the mean time to data loss as @xmath30 .
given the exact expression of @xmath29 , we can evaluate @xmath30 .
however , the expression is cumbersome and thus omitted .
we consider however the scenario where @xmath31 and give the mean time to data loss in that scenario . the mean time to data loss for serial deterministic repair for @xmath31
is approximately given as @xmath32 the proof follows by using expression for @xmath30 and ignoring all terms of lower order .
the detailed steps are omitted .
we note that for exponential serial repair model with rate of repair @xmath33 , mean time to data loss was characterized in @xcite , and is given as @xmath34 with an approximation given as @xcite @xmath35 we note that the mtdl approximation for serial repair with deterministic repair is the same as the approximation with exponential repair for @xmath36 .
the parallel repair model can be described with the following assumptions .
1 . @xmath1 components ( disks ) are identical and independent .
the failure rate for each disk is constant , and equals @xmath2 ( failure time is exponentially distributed for each disk ) .
3 . the repair times are independent and fixed , and equals @xmath3 .
4 . the repair for the previous broken components must be re - initialized when a new component fails , and all the broken components are simultaneously repaired . 5 . the system stops working when @xmath4 components have failed for @xmath5 , @xmath6 .
suppose that the probability of data loss in time @xmath7 is @xmath37 .
then the laplace stieltjes transform of reliability is given as @xmath38 $ ] , where @xmath10 is the random variable representing the time to data loss , and @xmath39 .
the next result gives the laplace stieltjes transform of reliability for parallel repair .
the laplace stieltjes transform of reliability for parallel repair is given by @xmath40 where @xmath41 the above can be simplified to @xmath42 the proof steps are very similar to that for serial repair model , where the difference is that all the failed components get repaired simultaneously thus going from state i to state 0 on repair rather than to state i-1 for the serial repair model .
using @xmath43 , we can find the mean time to data loss as @xmath44 . given the exact expression of @xmath43
, we can evaluate @xmath44 .
however , the expression is cumbersome and thus omitted .
we consider however the scenario where @xmath31 and give the mean time to data loss in that scenario . the mean time to data loss for parallel deterministic repair for @xmath31
is approximately given as @xmath45 the proof follows by using expression for @xmath44 and ignoring all terms of lower order .
the detailed steps are omitted .
we note that the approximation is the same , thus giving the same asymptotic expression as in the case of serial repsir model .
for exponential parallel repair model with rate of repair @xmath33 , mean time to data loss was characterized in @xcite , and is given as @xmath46 with an approximation given as @xcite @xmath47 thus , even though exponential parallel repair has @xmath0 times higer mtdl than serial repair , the mtdl approximations for the serial and parallel deterministic repair models is the same .
we assume @xmath48 and @xmath49 .
we assume that the failue rate is @xmath50 per year .
we plot the increase in mean time to data loss with decreasing time to repair in figure [ fig : awesome_image ] .
we note that for serial repair model , the mean time to data loss is lower for the deterministic repair model as compared to the serial exponential repair model .
we further see that parallel deterministic repair has higher mean time to data loss as compared to serial deterministic repair .
finally , we note that the approximation result approximates all these three mean time to data loss expressions as the repair time becomes smaller .
this paper gives laplace stieltjes transform of the data loss for the k - out - of - n distributed storage system with deterministic repair times .
two repair models , namely the serial and parallel repair are considered .
we show that for failure rate much lower than the repair rate , mean time of data loss for the two models is the same unlike the case for exponential repair models .
l. mearian ,
world s data will grow by 50x in next decade , idc study predicts , " computerworld , jun 2011 .
p. m. chen , e. k. lee , g. a. gibson , r. h. katz and d. a. patterson , raid : high - performance , reliable secondary storage , " journal of the acm volume 26 , issue 2 ( 1994 ) , p. 145
j. e. angus , on computing mtbf for a k - out - of - n : g repairable system , " ieee transactions on reliability volume 37 , number 3 ( 1988 ) , p. 312 - 313 .
m. s. moustafa , time dependent availability of k - out - of - n : g systems with exponential failures and general repairs , " economic quality control , vol .
59 - 66 , 2003 .
v. aggarwal , c. tian , v. a. vaishampayan , and y. r. chen , distributed data storage systems with opportunistic repair , " proc .
ieee infocom 2014 .
a. campello and v. a. vaishampayan , reliability of erasure coded systems : a geometric approach , " proc .
ieee international conference on bigdata 2013 , santa clara , ca , oct . 2013 . v. a. vaishampayan and a. campello , reliability of erasure coded storage systems : a combinatorial - geometric approach , " ieee transactions on information theory , vol .
5795 - 5809 , nov . | in this paper , we find the laplace stieltjes transform of the probability of data loss for the k - out - of - n distributed storage system with deterministic repair times .
we consider two repair models , namely the serial and parallel repair .
we show that for failure rate much lower than the repair rate , mean time of data loss for the two models is the same unlike the case for exponential repair models .
k - out - of - n storage systems , parallel repair , serial repair , reliability , mean time to data loss |
multiferroics have attracted great attention recently because of their promising new functionality and intriguing fundamental science .
a multiferroic material with a large ferroic polarization , high ordering temperature , and strong coupling between the ferroic orders is ideal for applications .
so far , those desired properties have not been realized in a single phase material .
multiferroics like bifeo@xmath2 where the magnetic and electric orders originate from different part of the structure have high ordering temperatures but weak coupling between different orders.@xcite other materials like tbmn@xmath0o@xmath3 exhibiting ferroelectricity due to the broken symmetry caused by the spiral magnetic moment have strong magneto - electric coupling.@xcite however , here the ordering temperature is very low and the electric polarization is small .
lufe@xmath0o@xmath1 contains layers of fe@xmath0o@xmath0 with a triangular lattice that are sandwiched by luo@xmath0 layers .
combined with the mixed valance of fe , the fe@xmath0o@xmath0 layers in the triangular lattice form a charge ordered state at @xmath4=320 k , followed by a ferrimagnetic order at @xmath5=240 k.@xcite significant changes in dielectric properties have been observed upon application of a small magnetic field at room temperature.@xcite the relatively high transition temperature , large polarization , high magnetic coercivity and the strong magneto - electric coupling make lufe@xmath0o@xmath1 a unique multiferroic material .
recently , the possibility of fast switching and high tunability of lufe@xmath0o@xmath1 due to the electronic origin of its charge order was demonstrated.@xcite compared to the large amount of effort to study bulk lufe@xmath0o@xmath1 , there are only a couple of reported attempts to grow lufe@xmath0o@xmath1 thin films on @xmath6-al@xmath0o@xmath2 ( 001 ) and on si substrates using pulsed laser deposition ( pld).@xcite liu et al found that the growth of lufe@xmath0o@xmath1 on @xmath6-al@xmath0o@xmath2 ( 001 ) ( with a target consisting of a sintered mixture of lu@xmath0o@xmath2 and fe@xmath0o@xmath2 ) needs substrate temperatures as high as 850 @xmath7c.@xcite in addition , a significant deviation of the lu : fe stoichiometry from 1:2 was observed , which was attributed to different ablation efficiencies of lu and fe in the target .
this problem was circumvented by enriching the fe concentration of the target material . however , as a result , fe@xmath2o@xmath1 and fe@xmath0o@xmath2 impurities were introduced as intermediate layers between the lufe@xmath0o@xmath1 film and the @xmath6-al@xmath0o@xmath2 substrate . in this paper , we present a comprehensive study on the growth of lu - fe - o compound thin films on mgo ( 111 ) substrates using pulsed laser deposition ( pld ) .
the experimentally constructed growth diagram shows that the parameter space for growing epitaxial lufe@xmath0o@xmath1 thin films turns out to be a narrow window of temperature and o@xmath0 pressure , which creates significant experimental difficulty .
based on these results we have gained fundamental understanding of the growth of lu - fe - o compound films : the growth temperature needs to be high enough to stabilize the lufe@xmath0o@xmath1 phase ; on the other hand the loss of fe at high temperature also produces phases other than lufe@xmath0o@xmath1 .
these two effects cause narrow window of the growth condition producing lufe@xmath0o@xmath1 .
typical lufe@xmath0o@xmath1 films appear to be superparamagnetic , which is consistent with the fact that the lufe@xmath0o@xmath1 in the film is epitaxially sandwiched by an impurity phase of hexagonal lufeo@xmath2 ( h - lufeo@xmath2 ) .
the current demonstration of epitaxial growth of lufe@xmath0o@xmath1 thin films opens up new possibilities for studying multiferroicity of low dimensional lufe@xmath0o@xmath1 , tuning of its properties , and eventual functionalization .
the paper is organized as the following : section ii describes the experimental conditions used in this work ; section iii presents the experimental results including the growth diagram , structural characterizations and magnetism ; explanations of the observed growth diagram and magnetism of the films are proposed in section iv .
lu - fe - o compound thin films were grown using pld with a krf ( @xmath8=248 nm ) laser .
the energy density of the laser is 2.5 j @xmath9 with a repetition rate of 1 hz . the target - substrate distance was 3.5 cm .
the thickness of the films grown in this study is approximately 100 nm .
the substrates are mgo ( 111 ) single crystals annealed in o@xmath0 for 24 hours at 1100 @xmath7c .
the target material used is polycrystalline lufe@xmath0o@xmath1 , whose properties are verified using powder x - ray diffraction ( xrd ) and a superconducting quantum interference device ( squid ) .
after growth , the sample heating is turned off so that the sample cools to below 200 @xmath7c at the same pressure as that of the growth condition within 5 minutes .
the substrates were clamped on a heater with a pt foil in between .
the sample temperature was measured by a pyrometer using emissivity of 0.3 . in principle , all the parameters described above will have to be scanned and optimized in order to realize the growth of high quality lufe@xmath0o@xmath1 thin films . in this work , we are more focused on elucidating the mechanism of the growth . therefore , fine scans of the substrate temperature and the o@xmath0 pressure were carried out to map out the growth diagram involving the growth of more than one hundred samples , while all the other parameters were kept constant .
in this work , we start from the ternary phase diagram of the bulk lu - fe - o system , a section of which is shown in fig .
[ fig_phasegrowth](a ) at 1200 @xmath7c.@xcite this system belongs to the d - type of lanthanoid - fe - o compounds for which there are four stable three - element phases : lufe@xmath0o@xmath1 ( a ) and lu@xmath0fe@xmath2o@xmath10 ( b ) , lufeo@xmath2 ( perovskite or p ) , and lu@xmath2fe@xmath3o@xmath11 ( garnet or g).@xcite in principle , one way to form a single lufe@xmath0o@xmath1 phase is to keep atomic ratio lu : fe=1:2 and vary the o@xmath0 pressure , as shown as a thick dashed line in fig . [
fig_phasegrowth](a ) .
figure [ fig_xrd](a ) presents the xrd data of films grown at 1050 @xmath7c in various o@xmath0 pressures . the lufe@xmath0o@xmath1 phase is not observed .
in addition , the lu : fe stoichiometry of the films is very different from that of the target .
the dominant phase is always lu@xmath0o@xmath2 ( l ) .
the concentration of lufeo@xmath2 rises with increasing o@xmath0 pressure . at high enough o@xmath0 pressure , h - lufeo@xmath2 compounds start to form.@xcite in order to further elucidate the mechanism of the growth of lu - fe - o compound films , we carried out fine scans of the substrate temperature and the o@xmath0 pressure to map out the growth diagram .
figure [ fig_phasegrowth](b ) is the resulting experimental growth diagram .
the important observations can be summarized as follows : 1 ) in the low temperature region the growth follows more or less the behavior predicted by the bulk phase diagram fig .
[ fig_phasegrowth](a ) : at high pressure , the existing phases are lufeo@xmath2 , lu@xmath0o@xmath2 and h - lufeo@xmath2 ; when the pressure is decreased , the fe@xmath2o@xmath1 phase starts to appear .
this is consistent with the fact that lufe@xmath0o@xmath1 and lu@xmath0fe@xmath2o@xmath10 phases are not stable at low temperature.@xcite 2 ) in the high temperature region , the growth deviates strongly from the thick dashed line in the bulk phase diagram fig .
[ fig_phasegrowth ] ( a ) in that the lu : fe stoichiometry differs dramatically from that of the polycrystalline lufe@xmath0o@xmath1 target .
the formation of lu - fe - o compounds in the films qualitatively follows the dash - dot line in fig . [ fig_phasegrowth](a ) .
3 ) only in the small range of pressure and temperature indicated by the elliptical area in fig .
[ fig_phasegrowth](b ) , growth of lufe@xmath0o@xmath1 is the most effective . in this case
, the growth follows qualitatively the dotted line in fig . [
fig_phasegrowth](a ) .
typical xrd data are diaplayed in fig .
[ fig_xrd](c ) showing both lufe@xmath0o@xmath1 and h - lufeo@xmath2 , indicating a deviation of lu : fe stoichiometry from that of the target even in this narrow window.@xcite .morphology and epitaxial relations of compounds grown on mgo ( 111 ) substrates . [ cols="^,^,^",options="header " , ] the combination of in - situ structural characterization using reflection high energy electron diffraction ( rheed ) and ex - situ characterization by xrd allows assignment of the epitaxial relation between the existing phases and the substrates .
the results are given in table i. from the rheed image , one can measure the in - plane lattice constant for the grown film .
if three dimensional ( 3d ) island growth occurs , the rheed images correspond to the diffraction pattern of the transmitted electron beam which contains more structural information .
figure [ fig_rheed ] shows the rheed images of the mgo ( 111 ) substrates and the lufe@xmath0o@xmath1 films with the electron beams directed along mgo [ 1 - 10 ] or mgo [ 11 - 2 ] . the strong lufe@xmath0o@xmath1 ( 003 ) , ( 006 ) and ( 009 ) peaks observed in fig .
[ fig_xrd](b ) indicate that the epitaxial relation is lufe@xmath0o@xmath1 [ 001]//mgo [ 111 ] , which is expected because both faces have 3-fold rotational symmetry .
the streaky rheed patterns in fig . [ fig_rheed](c ) and ( d ) suggest quasi-2d growth of lufe@xmath0o@xmath1 .
the in - plane lattice constants of the film can be calculated from the separation of the streaks calibrated by the rheed pattern of the mgo substrates .
it is consistent with the lufe@xmath0o@xmath1 lattice constant 3.44 @xmath12 within the experimental uncertainty of 2% .
hence , the in - plane epitaxial relation is lufe@xmath0o@xmath1 [ 100]//mgo [ 1 - 10 ] .
this is unexpected from the point of view of lattice matching , which predicts lufe@xmath0o@xmath1 [ 100]//mgo [ 11 - 2 ] because a @xmath13 supercell of lufe@xmath0o@xmath1 with 30 degree rotation along the [ 001 ] direction has less than 0.1% mismatch with a 1@xmath141 of mgo ( 111 ) surface.@xcite contrasting with the apparent quasi-2d growth of lufe@xmath0o@xmath1 , lu@xmath0o@xmath2 forms quasi-3d structures on the substrate.@xcite however , the rheed pattern suggests a face centered cubic structure with a lattice constant half of that of bulk lu@xmath0o@xmath2 .
the detailed structure is not clear at present . at low pressure and low temperature ,
the rheed signal is dominated by the diffraction pattern of 3d fe@xmath2o@xmath1 islands along the [ 11 - 2 ] direction , with the [ 111 ] direction perpendicular to the substrate surface .
this is consistent with the xrd data .
the lattice constants are the same as that of bulk fe@xmath2o@xmath1 within the experimental uncertainty of 2% .
high resolution transmission electron microscopy ( hrtem ) reveals the detailed structure of the lufe@xmath0o@xmath1 films . as shown in fig .
[ fig_tem](a ) , a layered structure of the film is obvious with some variation at different locations .
the fast fourier transforms ( fft ) of the hrtem image at different locations confirm the epitaxial relation observed from rheed images : the fft of the substrate ( fig .
[ fig_tem](d ) ) indicates the reciprocal lattice of mgo viewed from [ 11 - 2 ] direction .
the fft of the majority of the film ( fig .
[ fig_tem](c ) ) is consistent with the reciprocal lattice of lufe@xmath0o@xmath1 viewed from the [ 1 - 10 ] direction , while at some locations ( fig . [ fig_tem](b ) ) it suggests h - lufeo@xmath2 viewed from the [ 100 ] direction .
these two phases lufe@xmath0o@xmath1 and h - lufeo@xmath2 were further confirmed by direct observation using atomic - resolution z - contrast imaging , which is shown in fig . [ fig_tem](e ) and ( f ) .
the luo@xmath0-feo - feo - luo@xmath0 ordering in the lufe@xmath0o@xmath1 phase and the luo@xmath0-feo - luo@xmath0 ordering in the h - lufeo@xmath2 phase are clearly observed .
although the intensity of the xrd peaks originating from the h - lufeo@xmath2 phase seems comparable to that of lufe@xmath0o@xmath1 phase , the actual dominant phase is still lufe@xmath0o@xmath1 due to the lower x - ray scattering cross section of the lufe@xmath0o@xmath1 as compared with that of the h - lufeo@xmath2 phase .
this is consistent with the low population of the h - lufeo@xmath2 phase in the hrtem image .
in addition , the rheed patterns of h - lufeo@xmath2 and lufe@xmath0o@xmath1 are supposed to be different according to their structures.@xcite the fact that the observed rheed patterns do not show any indication of h - lufeo@xmath2 within the detection limit also suggests a low concentration of the h - lufeo@xmath2 phase in the films .
ferrimagnetism , large magnetization and giant coercivity are of the key properties of lufe@xmath0o@xmath1.@xcite this makes the study of the magnetic properties of lufe@xmath0o@xmath1 films critical . as shown in fig .
[ fig_mag4](a ) , little hysteresis is observed for these lufe@xmath0o@xmath1 films , in contrast to the bulk.@xcite when magnetization is plotted against magnetic field over temperature ( @xmath15 ) , the data of 12 , 50 , 100 and 150 k fall on top of each other ( fig . [ fig_mag4](b ) ) , indicating superparamagnetic behavior.@xcite assuming superparamagnetism to be present , one can calculate the magnetic moment from the slope of the low field magnetization data with @xmath16 where @xmath17 , @xmath18 , @xmath19 and @xmath20 are the moment of the superparamagnetic domains , number of the domains per unit volume , the vacuum permeability and the boltzmann constant , respectively.@xcite the magnetic moments normalized to their maximum value as a function of temperature are plotted in fig .
[ fig_mag4](c ) , which follow the temperature dependence of the bulk saturation magnetization closely , suggesting that the neel temperature of the films is not very different from the bulk value of 240 k.
the most surprising observation of the growth diagram is that the lu - fe - o compound formation at high temperature deviates strongly from the lu : fe stoichiometry of the target .
here we propose an explanation in terms of competition between nucleation and desorption of adatoms and its dependence on temperature and supersaturation .
the residence time @xmath21 of an adsorbed atom is given by : @xmath22 where @xmath23 is the vibrational frequency and @xmath24 is the desorption energy . clearly , the residence time of an adatom is shorter at high temperature due to the higher desorption rate .
the observed loss of fe atoms suggests a smaller desorption energy ( higher desorption rate ) for fe atoms . at low temperature , because @xmath25 is large for both lu and fe adatoms , the lu : fe stoichiometry can be close to that of the target . the nucleation speed of deposited adatoms is : @xmath26 where @xmath27 is the effective supersaturation ( molar bulk gibbs free energy change with surface energy consideration ) , while @xmath28 is proportional to the square of the edge energy of the nuclei per unit length.@xcite therefore , at the high temperature limit , the nucleation speed decreases with temperature and a high supersaturation favors a high nucleation speed .
consider the reaction@xcite @xmath29 which takes place under thermodynamic equilibrium during the annealing time in between the laser pulses , the supersaturation of o@xmath0 is related to the o@xmath0 pressure as : @xmath30 where @xmath31 is the avogadro constant .
( [ eq_muo2 ] ) suggests that higher o@xmath0 pressure always corresponds to larger supersaturation , resulting in faster nucleation and better lu : fe stoichiometry .
combining eqs .
( [ eq_muo2 ] ) and ( [ eq_jnuc ] ) , one has the analytical relation between the nucleation speed and the o@xmath0 pressure : @xmath32^{1/2 } \nonumber\\ & & * \exp[-\frac{\kappa}{\delta\mu^ * _ { 0}(t)kt+\frac{3}{4}n_a ln(p_{o_2})(kt)^2 } ] , \label{eq_jnuco2}\end{aligned}\ ] ] fig [ fig_xrd](b ) shows the xrd intensity ( peak area ) of the lufeo@xmath2 phase relative to lu@xmath0o@xmath2 ( @xmath33 ) as a function of the o@xmath0 pressure at 1050 @xmath7c . assuming that the nucleation speed is proportional to the xrd intensity
, one can fit experimental data with eq .
( [ eq_jnuco2 ] ) .
the result shows that @xmath34 = 269 kj mol@xmath35 , similar to the bulk value found as @xmath36 j mol@xmath35 , taking the @xmath37 = -41.8 kj mol@xmath35 and @xmath38 = -121.4 j mol@xmath35 k@xmath35 and @xmath39 = 1050 @xmath7c.@xcite in the above analysis , the assumptions we made are : 1 ) the nucleation speed is proportional to the xrd intensity ; 2 ) at high temperature the thermodynamic equilibrium gained during the annealing between the laser pulses determines the growth .
these assumptions appear to be valid because the thermo - chemical parameters extracted from the model quantitatively agree with those from the literature .
in other words , the growth of lu - fe - o at 1050 @xmath7c can be described using equilibrium thermodynamics , presumably due to the thermodynamic equilibration that occurs in between the laser pulses . here
the competition between the desorption and nucleation determines the lu : fe stoichiometry .
when the temperature is high enough , the time scales of the nucleation and desorption are comparable . in this case ,
change of nucleation speed ( due to the change of supersaturation which is a function of o@xmath0 pressure ) affects the lu : fe stoichiometry dramatically . based on this analysis , we expect the optimal growth conditions for lufe@xmath0o@xmath1 films to be a narrow temperature and pressure window considering the necessary high temperature for the stability of lufe@xmath0o@xmath1 phase that sets a lower limit , and the different desorption speed of lu and fe adatoms which sets an upper limit to the temperature . as we have shown in section iii
, this is indeed what has been observed in our experiments .
the observation of superparamagnetism in the lufe@xmath0o@xmath1 films is unusual considering the bulk magnetic properties of lufe@xmath0o@xmath1 : an easy axis along the [ 001 ] direction with anisotropy energy as large as 100 k / spin and gigantic coercivity ( 9 t at 4 k).@xcite these unique bulk properties were attributed to the significant contribution of orbital magnetic moments ( 0.8 @xmath40/f.u . )
plus the collective freezing of magnetic domains with the size of approximately 100 nm in the fe@xmath0o@xmath0 layer and 30 nm along the [ 001 ] direction .
@xcite the following scenario may explain the reduction of coercivity qualitatively : the structure of lufe@xmath0o@xmath1 and h - lufeo@xmath2 both consist of layers of triangular lattice that can be epitaxial to each other nicely .
for lufe@xmath0o@xmath1 , the stacking is fe@xmath0o@xmath0/luo@xmath0 while for h - lufeo@xmath2 , feo layers replace fe@xmath0o@xmath0 layers .
@xcite from xrd data , one can see the co - existence of both lufe@xmath0o@xmath1 and h - lufeo@xmath2 phases .
hrtem indicates that the lufe@xmath0o@xmath1 layers are divided into clusters ( much smaller than the magnetic domain size in bulk ) by the h - lufeo@xmath2 layers and defects .
according to a recent study , h - lufeo@xmath2 is weakly ferromagnetic , i.e. much less magnetic than lufe@xmath0o@xmath1 @xcite therefore , when these lufe@xmath0o@xmath1 clusters are much smaller than the dimensions of the magnetic domain in the bulk , one expects to see a reduction in coercivity . on the other hand , given the large anisotropy energy 100 k / spin , the observed hysteresis is too small even for clusters having a size as small as a few nanometers .
further study on the microscopic magnetic structure is needed to understand the difference between the bulk and films .
in conclusion , we studied the growth dynamics of lufe@xmath0o@xmath1 films on mgo ( 111 ) substrates and constructed the growth diagram . according to our understanding , application of the correct thermochemistry
is the key to preferential formation of the lufe@xmath0o@xmath1 phase : 1 ) at low temperature , lufe@xmath0o@xmath1 is not a thermodynamically stable phase ; 2 ) at high temperature , the lu : fe stoichiometry is off by so much due to the faster desorption of fe adatoms that lufe@xmath0o@xmath1 can not be formed ; 3 ) in a narrow range of substrate temperature and o@xmath0 pressure , lufe@xmath0o@xmath1 dominates the grown phases with some h - lufeo@xmath2 phase epitaxially sandwiched in between due to the loss of fe atoms .
superparamagnetism is observed in the film of lufe@xmath0o@xmath1 containing h - lufeo@xmath2 impurities .
the extracted neel transition temperature is similar to that of bulk .
this work reveals the growth mechanism of lu - fe - o compound thin films , paving the way to the growth of high quality lufe@xmath0o@xmath1 thin films and offers an approach to tuning their properties
. this will be critical for future applications using lufe@xmath0o@xmath1 , a unique multiferroic material with large polarizations , high ordering temperatures , and strong magneto - electric coupling .
research supported by the u.s .
department of energy , basic energy sciences , materials sciences and engineering division and performed in part at the center for nanophase materials sciences ( cnms ) ( z.g . , j.d.f ) and ornl s shared research equipment ( share)(m.c . ) user facility , which are sponsored at oak ridge national laboratory by the office of basic energy sciences , u.s .
department of energy . x.s .
xu acknowledges his research performed as a eugene p. wigner fellow and staff member at the ornl , managed by ut - battelle , llc , for the u.s .
doe under contract de - ac05 - 00or22725 .
we also acknowledge partial funding supports from the national basic research program of china ( 973 program ) under the grant no .
2011cb921801 ( j.s . ) and the us doe office of basic energy sciences , the us doe grant de - sc0002136 ( w.b.w ) .
j. wang , j. b. neaton , h. zheng , v. nagarajan , s. b. ogale , b. liu , d. viehland , v. vaithyanathan , d. g. schlom , u. v. waghmare , n. a. spaldin , k. m. rabe , m. wuttig , and r. ramesh , science * 299 * , 1719 ( 2003 ) .
several phases do not show up in the xrd data including feo and lu@xmath2fe@xmath3o@xmath11 .
the fact that there are also no rheed patterns related to these phases suggests that their population is minimal if not zero .
the absence of those phases and the presence of the h - lufeo@xmath2 which is unstable in bulk may be related to the epitaxial preference .
w. wu , v. kiryukhin , h .- j .
ko , j .- h . park , w. ratcliff ii , p. a. sharma , n. harrison , y. j. choi , y. horibe,1 s. lee,1 s. park,1 h. t. yi,1 c. l. zhang,1 and s .- w .
cheong , physical review letters , * 101 * , 137203 ( 2008 ) . at 12 k ,
the remanence is 7.5 emu @xmath42 and the coercive field is approximately 300 oe .
this can be attributed to the existence of a trace amount of fe@xmath2o@xmath1 islands because their temperature dependences are very similar to those of the samples with larger fe@xmath2o@xmath1 populations ( data not shown ) and much larger remanence .
this analysis is only sensible at high temperature where desorption is significant . at high temperature , there are basically only two reactions left : eq .
( [ eq_lfo113 ] ) and @xmath44 ( i.e. other lu - fe - o phases are not thermodynamically stable at high temperatures ) . unfortunately the thermo - chemical properties of h - lufeo@xmath2 are not available .
therefore , only the analysis based on eq .
( [ eq_lfo113 ] ) can give us a degree of understandingof the growth . | a growth diagram of lu - fe - o compounds on mgo ( 111 ) substrates using pulsed laser deposition is constructed based on extensive growth experiments .
the lufe@xmath0o@xmath1 phase can only be grown in a small range of temperature and o@xmath0 pressure conditions .
an understanding of the growth mechanism of lu - fe - o compound films is offered in terms of the thermochemistry at the surface .
superparamagnetism is observed in lufe@xmath0o@xmath1 film and is explained in terms of the effect of the impurity h - lufeo@xmath2 phase and structural defects . |
an important research area of real algebraic geometry is concerned with representations of polynomials positive on a basic semi - algebraic set @xmath2 where @xmath3 $ ] , @xmath4 .
an important result in this vein is schmdgen s positivstellensatz @xcite which states that if @xmath1 is compact and @xmath5 $ ] is positive on @xmath1 then @xmath6 belongs to the preordering @xmath7 generated by the @xmath8 s ; bounds on the degrees in the representation are even provided in schweighofer @xcite . under a rather weak additional assumption on the @xmath8 s , putinar s refinement @xcite states that @xmath6 even belongs to the quadratic module @xmath9 generated by the @xmath8 s .
the above mentioned representation results do not specialize when @xmath6 is convex and the @xmath8 s are concave ( so that @xmath1 is convex ) a highly important case , particularly in optimization . also , as soon as @xmath1 is not compact any more then negative results , notably by scheiderer @xcite , exclude to represent _ any _ @xmath6 positive on @xmath1 as an element of @xmath7 or @xmath9 ( except perhaps in low - dimensional cases ) . for more details ,
the interested reader is referred to the nice survey @xcite .
however , inspired and motivated by some classical results from convex optimization , we show that specialized representation results are possible when @xmath6 is convex and the @xmath8 s are concave , in which case @xmath0 is a closed ( not necessarily compact ) convex basic semi - algebraic set .
namely , a specific subset @xmath10 of the quadratic module @xmath9 is such that @xmath11 is _ dense _ ( for the @xmath12-norm of coefficients ) in the convex cone @xmath13 of convex polynomials , nonnegative on @xmath1 .
let @xmath14 $ ] be the ring of real polynomials in the variables @xmath15 , and let @xmath16 $ ] be the subset of sums of squares ( sos ) polynomials . if @xmath5 $ ] , write @xmath17 , and denote its @xmath12-norm by @xmath18 ) .
if one is interested in representation of polynomials nonnegative on @xmath1 , the first polynomial to consider is of course @xmath31 where @xmath32 .
indeed , any other positive polynomial is just @xmath33 with @xmath34 .
and so , if @xmath31 belongs to some preordering or some quadratic module , then so does @xmath6 . from proposition [ prop1 ]
it is easy to establish the following result .
[ coro1 ] let assumption [ ass1 ] hold and let @xmath5 $ ] be convex and such that @xmath24 for some @xmath25 .
if the nonnegative polynomial @xmath35 of ( [ lagrangian ] ) is sos then @xmath36 for some convex sos polynomial @xmath37 and some nonnegative scalars @xmath38 , @xmath4 .
that is , @xmath39 , with @xmath9 as in ( [ quad ] ) .
in addition , the sos weights associated with the @xmath8 s are just nonnegative constants , and @xmath40 is convex .
[ def1 ] a polynomial @xmath5 $ ] is sos - convex if its hessian @xmath41 is a sum of squares ( sos ) , that is , there is some integer @xmath42 and some matrix polynomial @xmath43^{p\times n}$ ] such that @xmath44 if @xmath6 is sos - convex and @xmath45 is sos - convex for every @xmath4 , then @xmath39 . more precisely , ( [ main ] ) holds for some convex sos polynomial @xmath37 and some nonnegative scalars @xmath38 , @xmath4 . from proposition
[ prop1 ] , let @xmath35 be as in ( [ lagrangian ] ) . as @xmath6 and @xmath45 are sos convex , write @xmath46 for some @xmath43^{p\times n}$ ] and some @xmath47^{p_j\times n}$ ] , @xmath4 .
hence , @xmath48 with @xmath49 $ ] , and so @xmath35 is sos - convex . as ( [ kkt2 ] )
holds , by lemma 3.2 in helton and nie @xcite , the polynomial @xmath35 is sos , and so , by corollary [ coro1 ] , the desired result ( [ main ] ) holds .
next , consider the subset @xmath50 defined by : @xmath51 the set @xmath10 is a specialization of @xmath9 to the convex case , in that the weights asociated with the @xmath8 s are nonnegative scalars , i.e. , sos polynomials of degree 0 , and the sos polynomial @xmath40 is convex .
[ thmain2 ] let assumption [ ass1 ] hold , and let @xmath10 be as in ( [ qc ] ) .
let @xmath52 $ ] be the convex cone of convex polynomials nonnegative on @xmath1 .
then @xmath11 is dense in @xmath13 for the @xmath12-norm @xmath53 . in particular , if @xmath54 ( so that @xmath13 is now the set of nonnegative convex polynomials ) , then @xmath55 is dense in @xmath13 .
let @xmath56 and let @xmath57 .
given @xmath58 , let @xmath59 $ ] be the polynomial @xmath60 for every @xmath61 , the polynomial @xmath62 is convex and nonnegative on @xmath1 , i.e. , @xmath63 .
in addition , @xmath64 for some @xmath65 . indeed , the level set @xmath66 is compact for every @xmath67 , and so , @xmath68 attains its minimum on @xmath1 .
obviously , we also have @xmath69 as @xmath70 .
next , let @xmath71 be as in ( [ lagrangian ] ) , i.e. , @xmath72 for some nonnegative vector @xmath73 . as @xmath74 on @xmath75 , by corollary 3.3 in lasserre and netzer @xcite , there exists @xmath76 such that for every @xmath77
, @xmath78 is sos .
that is , @xmath79 and so @xmath80 notice that by definition , @xmath37 is convex .
next , as @xmath81 , @xmath82 , and so , equivalently , @xmath83 .
one may also replace @xmath91 in ( [ theta ] ) with the new perturbation @xmath92 this perturbation also preserves convexity .
in addition , not only @xmath93 as @xmath70 , but the convergence @xmath94 is also _ uniform _ on compact sets ! 1 j.w .
helton , j. nie , _ semidefinite representation of convex sets _ , technical report , mathematics dept .
, university of california at san diego , usa , 2007 . ` arxiv:0705.4068 ` .
lasserre , t. netzer , _ sos approximations of nonnegative polynomials via simple high degree perturbations _ , math . z.
* 256 * ( 2006 ) , 99112 .
polyak , _ introduction to optimization _ , optimization software inc . , new york , 1987 .
m. putinar , _ positive polynomials on compact semi - algebraic sets _
univ . math . j.
* 42 * ( 1993 ) , pp . 969984 . c. scheiderer ,
_ positivity and sums of squares : a guide to recent results _ , in : _ emerging applications of algebraic geometry _ , m. putinar and s. sullivant ( eds . ) , i m a proceedings , institute of mathematics and its applications , minneapolis , usa . to appear .
k. schmdgen , _ the @xmath95-moment problem for compact semi - algebraic sets _ , math . ann .
* 289 * ( 1991 ) , pp .
203 - 206 .
m. schweighofer , _ on the complexity of schmdgen s positivstellensatz _ , j. of complexity * 20 * ( 2004 ) , 529543 . | we provide a specific representation of convex polynomials nonnegative on a convex ( not necessarily compact ) basic closed semi - algebraic set @xmath0 .
namely , they belong to a specific subset of the quadratic module generated by the concave polynomials that define @xmath1 . [ section ] [ thm]corollary [ thm]lemma [ thm]proposition [ thm]definition [ thm]assumption [ thm]remark |
switching theory has developed in the 50 s and the 60 s as a common effort of the mathematicians and the engineers of studying the switching circuits ( = asynchronous circuits ) from digital electrical engineering .
after 1970 we do not know to exist any mathematical published work in what we call switching theory .
the published works are written by engineers and their approach is always descriptive and unacceptable for the mathematicians .
the label of _ switching theory _ has changed to _ asynchronous systems _ ( or _ circuits _ ) _ theory_. one of the possible motivations of the situation consists in the fact that the important producers of digital equipments have stopped the dissemination of such researches .
our interest in asynchronous systems had bibliography coming from the 50 s and the 60 s , as well as engineering works giving intuition , as well as mathematical works giving analogies .
an interesting _ rendezvous _ has happened when the asynchronous systems theory has met the dynamical systems theory , resulting the so called _ regular _ autonomous systems = boolean dynamical systems ; the vector field is @xmath0 time is discrete or real and we obtain the _ unbounded delay model _ of computation of @xmath1 , suggested by the engineers .
the _ synchronous _ iterations of @xmath2 of the dynamical systems are replaced by _
_ iterations in which each coordinate @xmath3 is iterated independently on the others , in arbitrary finite time .
we denote with @xmath4 the binary boolean algebra , together with the discrete topology and with the usual algebraical laws:@xmath5{cc } & \overline{\ \ } \\ 0 & 1\\ 1 & 0 \end{array } , \;\;\;\;\;\;\begin{array } [ c]{ccc}\cdot & 0 & 1\\ 0 & 0 & 0\\ 1 & 0 & 1 \end{array } , \;\;\;\;\;\;\begin{array } [ c]{ccc}\cup & 0 & 1\\ 0 & 0 & 1\\ 1 & 1 & 1 \end{array } , \;\;\;\;\;\;\begin{array } [ c]{ccc}\oplus & 0 & 1\\ 0 & 0 & 1\\ 1 & 1 & 0 \end{array } \\
& \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\ \ \ \ \ table\;1\end{aligned}\ ] ] we use the same notations for the laws that are induced from @xmath6 on other sets , for example @xmath7@xmath8@xmath9 etc . in figure [ simetrie30 ] [ ptb ]
we have drawn at a ) the logical gate not , i.e. the circuit that computes the logical complement and at b ) a circuit that makes use of logical gates not .
the asynchronous system that models the circuit from b ) has the state portrait drawn at c ) . in the state portraits ,
the arrows show the increase of ( the discrete or continuous ) time .
the underlined coordinates are these coordinates for which @xmath10 and they are called _ excited _ , or _ enabled _ , or _ unstable_. the coordinates @xmath11 that are not underlined fulfill by definition @xmath12 and they are called _ not excited _ , or _ not enabled _ , or _ stable_.
the existence of two underlined coordinates in @xmath13 shows that @xmath14 may be computed first , @xmath15 may be computed first , or @xmath16 may be computed simultaneously , thus when the system is in @xmath17 it may run in three different directions , non - determinism . our present purpose is to define the symmetry of these systems .
we denote @xmath18 .
@xmath19 is the notation of the characteristic function of the set @xmath20 : @xmath21{c}0,if\ t\notin a\\ 1,if\ t\in a \end{array } \right .
.$ ] we denote with @xmath22 the set of the sequences @xmath23 @xmath24 @xmath25 the set of the real sequences @xmath26 that are unbounded from above is denoted with @xmath27 we use the notation @xmath28 for the set of the functions @xmath29 having the property that @xmath30 and @xmath31 exist with @xmath32@xmath33 [ def39]let @xmath34 be a function .
for @xmath35 we define the function @xmath36 by @xmath37@xmath38 [ def43]let be @xmath39 the function @xmath40 defined by @xmath41@xmath42{c}\widehat{\phi}^{\alpha}(\mu,-1)=\mu,\\ \widehat{\phi}^{\alpha}(\mu , k+1)=\phi^{\alpha^{k+1}}(\widehat{\phi}^{\alpha } ( \mu , k ) ) \end{array } \right .
\label{ite2}\ ] ] is called * discrete time * @xmath43**semi - orbit * * * of * @xmath44 we consider also the sequence @xmath31 and the function @xmath45 from ( [ sym16 ] ) , for which the function @xmath46 is defined by : @xmath47@xmath48@xmath49 @xmath50 is called * continuous time * @xmath51**semi - orbit * * * of * @xmath44 [ def62_]the * discrete time * and the * continuous time universal semi - regular autonomous asynchronous systems * are defined by@xmath52@xmath53 @xmath54 @xmath55 and @xmath1 are usually identified .
[ exa2]in figure [ simetrie23 ] we have drawn at a ) [ ptb ] the and gate that computes the logical intersection , at b ) a circuit with two gates and at c ) the state portrait of @xmath56 we conclude that@xmath57@xmath58 since the first coordinate might finally decrease its value and the second coordinate might finally increase its value , but the order and the time instant when these things happen are arbitrary .
[ def15]let be @xmath59 and @xmath45 from ( [ sym16 ] ) .
the function @xmath60 is defined by : @xmath61@xmath42{c}^{\ast}\widehat{\phi}^{\alpha}(\mu,-1)=\mu,\\ \phi^{\alpha^{k+1}}(^{\ast}\widehat{\phi}^{\alpha}(\mu , k+1))=\text { } ^{\ast } \widehat{\phi}^{\alpha}(\mu , k ) \end{array } \right . \label{ite6}\ ] ] and @xmath62 is defined by : @xmath63@xmath64@xmath65 @xmath66 is called * discrete time * @xmath67**anti - semi - orbit * * * of * @xmath68 while @xmath69 is called * continuous time * @xmath51**anti - semi - orbit * * * of * @xmath44 we compare the semi - orbits and the anti - semi - orbits now and see that they run both from the past to the future , but the relation cause - effect is different : in @xmath70 the cause is in the past and the effect is in the future , while in @xmath71 @xmath72 the cause is in the future and the effect is in the past . [ def17]the * discrete time * and the * continuous time universal anti - semi - regular autonomous asynchronous systems * are defined by@xmath73@xmath74 [ exa16]in figure [ simetrie24 ] [ ptb ] we have drawn at a ) the circuit and at b ) the state portrait of @xmath75 @xmath76 for which@xmath77@xmath78 the arrows in figures [ simetrie23 ] c ) and [ simetrie24 ] b ) are the same , but with a different sense and we note that @xmath79 @xmath80 where @xmath1 is the one from example [ exa2 ] .
[ def21]let be @xmath81 it defines the functions @xmath82@xmath83 @xmath84@xmath85{c}(0, ... ,0),if\;\rho(t)=(0, ...
,0),\\ g(\rho(t)),otherwise \end{array } \right.\ ] ] and @xmath86@xmath87 [ the29]let be the functions @xmath88 the following statements are equivalent : \a ) @xmath89 the diagram@xmath90{ccc}\mathbf{b}^{n } & \overset{\phi^{\nu}}{\rightarrow } & \mathbf{b}^{n}\\ g\downarrow\ ; & & \;\downarrow g\\ \mathbf{b}^{n } & \overset{\psi^{g^{\prime}(\nu)}}{\rightarrow } & \mathbf{b}^{n}\end{array}\ ] ] is commutative ; \b ) @xmath91@xmath92 \c ) @xmath93@xmath94 and @xmath95@xmath96 a)@xmath97b ) we fix arbitrarily @xmath98 @xmath30 and we use the induction on @xmath99 . for @xmath100 ,
b ) becomes @xmath101 , thus we suppose that it is true for @xmath102 and we prove it for @xmath103:@xmath104@xmath105 b)@xmath97c ) the first statement results from b ) if we take @xmath106 and @xmath107 in order to prove the second statement , let @xmath108 and @xmath45 be arbitrary , thus ( [ sym16 ] ) holds with @xmath109 if @xmath110 the statement to prove takes the form @xmath101 so that we can suppose now that a finite or an infinite number of @xmath111 are @xmath112 in the case @xmath113 that does not restrict the generality of the proof , we have that@xmath114 is an element of @xmath28 and@xmath115@xmath116@xmath117@xmath118@xmath119@xmath120 c)@xmath97a ) let @xmath121 be arbitrary and fixed and we consider @xmath45 given by ( [ sym16 ] ) , with @xmath31 fixed , @xmath122 and @xmath123 .
we have@xmath124@xmath125@xmath126
case i ) @xmath127 the commutativity of the diagram is equivalent with the first statement of c ) .
case ii ) @xmath128@xmath129@xmath130@xmath131 and from ( [ ite5 ] ) , for @xmath132 we obtain@xmath133 [ def64]we consider the functions @xmath134 if @xmath135 bijective exist such that one of the equivalent properties a ) , b ) , c ) from theorem [ the29 ] is satisfied , then we say that the couple @xmath136 defines an * isomorphism * from @xmath137 to @xmath138 or from @xmath55 to @xmath139 or from @xmath1 to @xmath140 we use the notation @xmath141 for the set of these couples and we also denote with @xmath142 the set of the * automorphisms * of @xmath54 @xmath143 or @xmath144 [ the28]for @xmath145 , the following statements are equivalent : \a ) @xmath89 the diagram@xmath90{ccc}\mathbf{b}^{n } & \overset{\phi^{\nu}}{\rightarrow } & \mathbf{b}^{n}\\ g\downarrow\ ; & & \;\downarrow g\\ \mathbf{b}^{n } & \overset{\psi^{g^{\prime}(\nu)}}{\longleftarrow } & \mathbf{b}^{n}\end{array}\ ] ] is commutative ; \b ) @xmath91@xmath146 \c ) @xmath93@xmath94 and @xmath95@xmath147 a)@xmath97b ) we fix arbitrarily @xmath98 @xmath30 and we use the induction on @xmath148 in the case @xmath100 the equality to be proved is satisfied@xmath149 thus we presume that the statement is true for @xmath102 and we prove it for @xmath150 we have:@xmath151@xmath152 the proof is similar with the proof of theorem [ the29 ] .
[ def65]let be the functions @xmath153 if @xmath154 bijective exist such that one of the equivalent properties a ) , b ) , c ) from theorem [ the28 ] is fulfilled , we say that the couple @xmath136 defines an * anti - isomorphism * from @xmath155 to @xmath156 or from @xmath55 to @xmath157 or from @xmath1 to @xmath140 we use the notation @xmath158 for these couples and we also denote with @xmath159 @xmath160 the set of the * anti - automorphisms * of @xmath54 @xmath55 or @xmath1 .
the fact that @xmath161 implies @xmath162 but all of @xmath163 and @xmath164 may be empty .
[ def66]let be @xmath165 @xmath166 if @xmath167 , then @xmath54 @xmath168
@xmath169 @xmath170 are called * symmetrical * , or * conjugated * ; if @xmath171 , then @xmath54 @xmath172 @xmath173 @xmath170 are called * anti - symmetrical * , or * anti - conjugated*. if @xmath174 then @xmath54 @xmath55 and @xmath1 are called * symmetrical * and if @xmath175 then @xmath54 @xmath55 and @xmath1 are called * anti - symmetrical*. [ rem33]the symmetry of @xmath170 means that @xmath176 maps the transfers @xmath177 in transfers @xmath178 the situation when @xmath1 is symmetrical and @xmath179 is similar .
anti - symmetry may be understood as mirroring : @xmath180 @xmath158 maps the transfers ( or arrows ) @xmath181 in transfers @xmath182 and similarly for @xmath180 @xmath183 [ the32]let be @xmath134
\a ) if @xmath176 , then @xmath184 \b ) if @xmath180 @xmath158 , then @xmath185 @xmath186 \a ) the hypothesis states that @xmath89 the diagram@xmath90{ccc}\mathbf{b}^{n } & \overset{\phi^{\nu}}{\rightarrow } & \mathbf{b}^{n}\\ g\downarrow\ ; & & \;\downarrow g\\ \mathbf{b}^{n } & \overset{\psi^{g^{\prime}(\nu)}}{\rightarrow } & \mathbf{b}^{n}\end{array}\ ] ] commutes , with @xmath187 bijective@xmath188 we fix arbitrarily @xmath189 we denote @xmath190 and we note that@xmath191 as @xmath192 were chosen arbitrarily and on the other hand , when @xmath193 runs in @xmath194 runs in @xmath195 and when @xmath196 runs in @xmath197 runs in @xmath195 , we infer that ( [ ite11 ] ) is equivalent with the commutativity of the diagram@xmath90{ccc}\;\;\;\;\;\mathbf{b}^{n } & \overset{\psi^{\nu^{\prime}}}{\rightarrow } & \mathbf{b}^{n}\;\;\;\\ g^{-1}\downarrow & & \downarrow g^{-1}\\ \;\;\;\;\;\mathbf{b}^{n } & \overset{\phi^{g^{\prime-1}(\nu^{\prime})}}{\rightarrow } & \mathbf{b}^{n}\;\;\ ; \end{array}\ ] ] for any @xmath198 we have proved that @xmath199 \b ) by hypothesis @xmath89 the diagram@xmath90{ccc}\mathbf{b}^{n } & \overset{\phi^{\nu}}{\rightarrow } & \mathbf{b}^{n}\\ g\downarrow\ ; & & \;\downarrow g\\ \mathbf{b}^{n } & \overset{\psi^{g^{\prime}(\nu)}}{\longleftarrow } & \mathbf{b}^{n}\end{array}\ ] ] is commutative , @xmath187 bijective and we prove that @xmath200 the diagram@xmath90{ccc}\;\;\;\;\;\mathbf{b}^{n } & \overset{\psi^{\nu^{\prime}}}{\rightarrow } & \mathbf{b}^{n}\;\;\;\\ g^{-1}\downarrow & & \downarrow g^{-1}\\ \;\;\;\;\;\mathbf{b}^{n } & \overset{\phi^{g^{\prime-1}(\nu^{\prime})}}{\longleftarrow } & \mathbf{b}^{n}\;\;\ ; \end{array}\ ] ] is commutative .
[ the36]@xmath201 is a group relative to the law : @xmath202 @xmath203@xmath204 the fact that @xmath205 is proved like this : @xmath89@xmath206@xmath207 the fact that @xmath161 was mentioned before ; and the fact that @xmath208 @xmath209 was shown at theorem [ the32 ] a ) .
any subgroup @xmath210 with @xmath211 is called a * group of symmetry * of @xmath54 of @xmath55 or of @xmath144
[ exa6_]@xmath212 are given by , see figure [ echiv1 ] [ ptb ] @xmath213@xmath214 and the bijections @xmath215 are @xmath216@xmath217@xmath218 ( in order to understand the choice of @xmath219 to be remarked in figure [ echiv1 ] the positions of the underlined coordinates for @xmath1 and @xmath220 ) .
@xmath1 and @xmath221 are conjugated .
[ exe2]the system from figure [ simetrie15 ] is symmetrical and a group of symmetry is generated by the couples @xmath222 see table 2 ; @xmath223 are transpositions that permute the isolated fixed points @xmath224 [ ptb ] @xmath5{ccccc}(\mu_{1},\mu_{2},\mu_{3 } ) & 1_{\mathbf{b}^{3 } } & g & u & v\\ ( 0,0,0 ) & ( 0,0,0 ) & ( 0,0,0 ) & ( 0,0,0 ) & ( 0,0,0)\\ ( 0,0,1 ) & ( 0,0,1 ) & ( 0,0,1 ) & ( 0,0,1 ) & ( 0,0,1)\\ ( 0,1,0 ) & ( 0,1,0 ) & ( 0,1,0 ) & ( 0,1,0 ) & ( 0,1,0)\\ ( 0,1,1 ) & ( 0,1,1 ) & ( 0,1,1 ) & ( 0,1,1 ) & ( 0,1,1)\\ ( 1,0,0 ) & ( 1,0,0 ) & ( 1,0,0 ) & ( 1,0,1 ) & ( 1,1,1)\\ ( 1,0,1 ) & ( 1,0,1 ) & ( 1,1,1 ) & ( 1,0,0 ) & ( 1,0,1)\\ ( 1,1,0 ) & ( 1,1,0 ) & ( 1,1,0 ) & ( 1,1,0 ) & ( 1,1,0)\\ ( 1,1,1 ) & ( 1,1,1 ) & ( 1,0,1 ) & ( 1,1,1 ) & ( 1,0,0 ) \end{array } \\ & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;table\;2\end{aligned}\ ] ] [ exa11]the function @xmath225 defined by @xmath226 fulfills for @xmath227@xmath228@xmath229@xmath230@xmath231@xmath232@xmath233@xmath234 thus @xmath235 @xmath164 and @xmath1 is anti - symmetrical .
the state portrait of @xmath1 was drawn in figure [ simetrie30 ] c ) .
let @xmath236 be a bijection .
we use the notation @xmath237 for the bijection given by @xmath93@xmath238 any of @xmath54 @xmath55 and @xmath34 is called * symmetrical relative to the coordinates * if the bijection @xmath239 exists , @xmath240 such that @xmath241 [ exa17]we consider the function @xmath242 defined by @xmath243 @xmath244 and the permutation @xmath245 @xmath246{ccc}1 & 2 & 3\\ \sigma(1 ) & \sigma(2 ) & \sigma(3 ) \end{array } \right ) = \left ( \begin{array } [ c]{ccc}1 & 2 & 3\\ 3 & 1 & 2 \end{array } \right ) .$ ] a group of symmetry of @xmath55 is represented by @xmath247 @xmath248 we have given in figure [ simetrie14 ] the state portrait of @xmath144 [ ptb ] for @xmath249 we denote by @xmath250 the translation of vector @xmath251 @xmath93@xmath252 if @xmath253 holds for some @xmath254 , we say that any of @xmath54 @xmath55 and @xmath1 is * symmetrical relative to translations*. [ exa13]in figure [ simetrie10 ] [ ptb ] we have the system with @xmath1 given by table 3@xmath5{cc}(\mu_{1},\mu_{2},\mu_{3 } ) & \phi\\ ( 0,0,0 ) & ( 0,0,0)\\ ( 0,0,1 ) & ( 0,0,1)\\ ( 0,1,0 ) & ( 0,1,1)\\ ( 0,1,1 ) & ( 0,1,0)\\ ( 1,0,0 ) & ( 0,1,1)\\ ( 1,0,1 ) & ( 0,1,0)\\ ( 1,1,0 ) & ( 1,0,0)\\ ( 1,1,1 ) & ( 1,0,1 ) \end{array } \\ & \;\;\;\;\;\;\;\;\;table\;3\end{aligned}\ ] ] and @xmath255 as resulting from the state portrait .
[ exa14]in table 4 we have a function @xmath256 for which four functions @xmath257 exist:@xmath5{cccccc}(\mu_{1},\mu_{2 } ) & \phi & g_{1}^{\prime } & g_{2}^{\prime } & g_{3}^{\prime } & g_{4}^{\prime}\\ ( 0,0 ) & ( 0,0 ) & ( 0,0 ) & ( 1,0 ) & ( 0,0 ) & ( 1,0)\\ ( 0,1 ) & ( 0,1 ) & ( 0,1 ) & ( 0,1 ) & ( 1,1 ) & ( 1,1)\\ ( 1,0 ) & ( 1,1 ) & ( 1,0 ) & ( 0,0 ) & ( 1,0 ) & ( 0,0)\\ ( 1,1 ) & ( 1,0 ) & ( 1,1 ) & ( 1,1 ) & ( 0,1 ) & ( 0,1 ) \end{array } \\ & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;table\;4\end{aligned}\ ] ] such that @xmath258 the state portrait of @xmath1 is drawn figure [ simetrie8 ] .
[ ptb ] [ exe1]the system from figure [ simetrie1 ] is symmetrical relative to translations , [ ptb ] since it has the group of symmetry @xmath259 @xmath1 is self - dual @xmath260 where the dual @xmath261 of @xmath1 is defined by @xmath262 [ exa15]functions @xmath225 exist , see figure [ simetrie7 ] [ ptb ] that are invariant relative to the translations with any @xmath263 thus their group of symmetry is @xmath264 the fact that @xmath265 shows that all these functions : @xmath266 @xmath267 @xmath268 @xmath269 are self - dual , @xmath270 [ exe4]the group of symmetry @xmath271 of the system from figure [ simetrie4 ] [ ptb ] has four elements given by@xmath5{ccccc}(\mu_{1},\mu_{2 } ) & 1_{\mathbf{b}^{2 } } & g & h & \theta^{(1,1)}\\ ( 0,0 ) & ( 0,0 ) & ( 0,1 ) & ( 1,0 ) & ( 1,1)\\ ( 0,1 ) & ( 0,1 ) & ( 1,1 ) & ( 0,0 ) & ( 1,0)\\ ( 1,0 ) & ( 1,0 ) & ( 0,0 ) & ( 1,1 ) & ( 0,1)\\ ( 1,1 ) & ( 1,1 ) & ( 1,0 ) & ( 0,1 ) & ( 0,0 ) \end{array } \\ & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;table\;5\end{aligned}\ ] ] and we remark that @xmath272 hold . on the other hand@xmath5{ccccc}(\nu_{1},\nu_{2 } ) & ( _ { 1_{\mathbf{b}^{2}}})^{\prime } & g^{\prime } & h^{\prime } & ( \theta^{(1,1)})^{\prime}\\ ( 0,0 ) & ( 0,0 ) & ( 0,0 ) & ( 0,0 ) & ( 0,0)\\ ( 0,1 ) & ( 0,1 ) & ( 1,0 ) & ( 1,0 ) & ( 0,1)\\ ( 1,0 ) & ( 1,0 ) & ( 0,1 ) & ( 0,1 ) & ( 1,0)\\ ( 1,1 ) & ( 1,1 ) & ( 1,1 ) & ( 1,1 ) & ( 1,1 ) \end{array } \\ & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;table\;6\end{aligned}\ ] ] @xmath271 has a proper subgroup @xmath273 showing that @xmath274 like previously .
the paper defines the universal semi - regular autonomous asynchronous systems and the universal anti - semi - regular autonomous asynchronous systems .
it also defines and characterizes the isomorphisms ( automorphisms ) and the anti - isormorphisms ( anti - automorphisms ) of these systems .
symmetry is defined as the existence of such isomorphisms ( automorphisms ) , while anti - symmetry is defined as the existence of such anti - isomorphisms ( anti - automorphisms ) .
many examples are given .
a by - pass product in this study is anti - symmetry , that is related with systems having the cause in the future and the effect in the present .
another by - pass product consists in semi - regularity , since important examples of isomorphisms ( automorphisms ) are of semi - regular systems only , they do not keep progressiveness and regularity @xcite , @xcite . | the regular autonomous asynchronous systems are the non - deterministic boolean dynamical systems and universality means the greatest in the sense of the inclusion .
the paper gives four definitions of symmetry of these systems in a slightly more general framework , called semi - regularity and also many examples .
* msc * : 94c10 * keywords * : asynchronous system , symmetry |
ever since the advent of the local metropolis algorithm ( lmc ) @xcite , monte carlo simulations of systems with many degrees of freedom have played an important role in statistical physics . near phase transitions , lmc is severely hampered by dynamical arrest phenomena such as critical slowing down for second - order transitions , nucleation and coarsening at first - order transitions , and glassy behavior in disordered systems .
a number of specialized algorithms then allow to speed up the sampling of configuration space , namely the swendsen
wang @xcite and the wolff @xcite cluster algorithms , the multicanonical method @xcite and the exchange monte carlo method @xcite based on extended ensembles . the above algorithms respect detailed balance , a sufficient condition for the convergence towards the equilibrium boltzmann distribution .
recently , algorithms breaking detailed balance but satisfying the necessary global - balance condition have been discussed @xcite . among them ,
the event - chain monte carlo ( ecmc ) algorithm @xcite has proven useful in hard - sphere @xcite and more general particle systems @xcite , allowing to equilibrate larger systems than previously possible @xcite . it has also been applied to continuous spin systems @xcite . ecmc uses a factorized metropolis filter @xcite and relies on an additional `` lifting '' variable to augment configuration space @xcite .
it is rejection - free and realizes an irreversible markov chain .
so far , however , the speedup realized by ecmc with respect to lmc has always represented a constant factor in the thermodynamic limit , although larger gains are theoretically possible @xcite . in this paper , we apply ecmc to the three - dimensional ferromagnetic heisenberg model , defined by the energy @xmath2 where @xmath3 is the unit of the energy , @xmath4 is a three - component unit vector and the sum runs over all neighboring pairs of the @xmath5 sites of a simple cubic lattice of linear extension @xmath6 . in our simulations , we consider the critical inverse temperature @xmath7 @xcite . to describe the dynamics of the system , we compute the autocorrelation functions of the energy , the system magnetization @xmath8 and the magnetic susceptibility @xmath9 the energy and the susceptibility are both invariant under global rotations of the spins @xmath10 around a common axis , whereas the magnetization follows the rotation .
we will argue that the energy and the susceptibility are slow variables , that is that their slowest time constant describes the correlation ( mixing ) time of the underlying markov chain . under this hypothesis
, we will present evidence that the ecmc for the three - dimensional heisenberg model reduces the dynamical critical exponent from the lmc value of @xmath1 to @xmath11 .
this considerable reduction of mixing times with respect to the lmc may well be optimal within the lifting approach @xcite .
the observed reduction is all the more surprising as in the closely related @xmath12 model @xcite , where the spins are two - dimensional unit vectors , the ecmc realizes speedups by two orders of magnitude with respect to lmc , but does not seem to lower the dynamical critical exponent .
applied to the heisenberg model , the ecmc augments the physical space of spin configurations by a lifting variable @xmath13 which specifies the considered infinitesimal counterclockwise rotation of spin @xmath14 about the axis @xmath15 . by virtue of the factorized metropolis filter
, this physical move can only be rejected by a single neighboring spin , @xmath16 , and the lifting variable will then be moved as @xmath17 , keeping the sense of rotation , but passing it on to the spin responsible for the rejection . in the augmented space ,
the rejections are thus supplanted by events , namely the lifting moves for arrested physical states .
the ecmc , for a given axis @xmath15 , breaks detailed balance , yet satisfies global balance , as the probability flow into each lifted configuration equals the flow out of it , to first order in the time increment @xmath18 . for the @xmath12 model of planar rotators , @xmath15 is uniquely defined as the axis perpendicular to the sense of rotation . for this reason ,
the ecmc around this axis is irreducible , and the chain length @xmath19 in this model is best taken equal to the simulation time @xcite . for the heisenberg model
, spin rotations must be about at least two axes , in order to reach the entire configuration space .
the resampling of the rotation axis is performed after the cumulative rotation angles about the previous axis reaches the chain length @xmath19 .
all configurations of the chain sample the equilibrium distribution and any uniform subset of them yield valid observable averages .
observables may be integrated during the continuous evolution or _
e.g. _ retrieved at regular intervals independent of the lifting events . for a fixed rotation axis @xmath15 , the ecmc algorithm for the heisenberg model reduces to the one of the @xmath12 model : with @xmath20 the spherical coordinates of a spin @xmath14 in a system where the @xmath21-axis is aligned with @xmath15 , the pair energy @xmath22 between spins
@xmath14 and @xmath16 is @xmath23 both @xmath24 and @xmath25 depend only on the polar angles @xmath26 and remain unchanged along the event chain . the azimuthal - angle dependence in eq .
( [ hxy_e ] ) is @xmath27 , as in the @xmath12 model .
the azimuthal angle @xmath28 increases for each ecmc chain from its initial value @xmath29 until one of its neighbors , @xmath16 , triggers a lifting @xmath30 at @xmath31 .
the latter is sampled with a single random number in the event - driven approach @xcite .
precisely , @xmath32 is given by the sampling of the positive pair energy increase : @xmath33/ \beta = \\ -j'\int^{\phi_{l , \text{event}}}_{\phi_0 } \max\left(0 , \frac{\mathrm{d}\cos(\phi_{{\boldsymbol{v } } , k } - \phi_{{\boldsymbol{v } } , l})}{\mathrm{d}\phi_{{\boldsymbol{v } } , k } } \right ) \mathrm{d}\phi_{{\boldsymbol{v } } , k } , \label{sampling_phi}\end{gathered}\ ] ] where @xmath34 is a uniform random number between @xmath35 and @xmath36 .
to solve eq .
( [ sampling_phi ] ) for @xmath32 , one first slices off any full rotations ( these @xmath37 rotations by @xmath38 yield an energy increase of @xmath39 ) , leaving a value @xmath40 , @xmath41 where @xmath42 the true lifting event corresponds to the earliest of the independent event times sampled for all the neighbors of the spin @xmath14 . in ecmc ,
monte carlo time is continuous and proportional to the total displacement of the spins .
we have checked the correctness of the ecmc , and obtained perfect agreement for the mean energy , the specific heat and the susceptibility with the heat - bath algorithm @xcite modified with the exchange monte carlo method ( or `` parallel tempering '' ) @xcite ( see fig . [
fig : static ] ) .
at the critical temperature for various system sizes . the inset shows the spin autocorrelation function of a trivial algorithm that only performs global rotations in spin space along the two axes . ] at the critical temperature @xmath43 , the correlation length @xmath44 of a model undergoing a second - order phase transition equals the system size @xmath6 and the autocorrelation time of slow variables @xmath45 diverges as @xmath46 , where @xmath21 is the dynamical critical exponent .
we measure time in sweeps : one ecmc sweep corresponds , in average , to @xmath47 lifting events and one lmc sweep to @xmath47 attempted moves .
time autocorrelation functions are defined by @xmath48 where the brackets @xmath49 indicate the thermal average and @xmath50 is set sufficiently large for equilibration .
the dynamical critical exponent of lmc for the three - dimensional heisenberg model was estimated from the autocorrelation function of the magnetization @xmath51 as @xmath52@xcite .
the over - relaxation algorithm @xcite seems to give @xmath53 @xcite which was obtained from the autocorrelation function of the magnetization , and the wolff algorithm is believed to yield a value close to zero : @xmath54 , a value obtained from the susceptibility autocorrelation function @xcite . to evaluate the correlation time and the dynamical critical exponent for the ecmc , one must pay attention to the irreversible nature of the underlying markov chain . during one event chain ,
spins all rotate in the same sense , and the system undergoes global rotations with taking into account the thermal fluctuation .
this results in fast oscillations of the magnetization @xmath51 and a quick decay of its autocorrelation function that is insensitive to the system size ( see fig . [ fig : autocorrelation ] ) , and even to the temperature .
however , this effect is also visible for a trivial algorithm , which simply performs global rotations ( see the inset of fig . [
fig : autocorrelation ] ) .
the trivial algorithm satisfies global balance , but its correlation time is infinite , as it does not relax the energy .
a similar effect appears in the ecmc for particle systems @xcite , that likewise is not characterized by the mean net displacement of particles . to characterize the speed of the ecmc
, we consider the energy density and the susceptibility that we conjecture to be slow variables at the critical temperature .
both @xmath55 and @xmath56 are insensitive to global rotations and do not oscillate . as shown in fig .
[ f : ec_aco ] , the autocorrelation functions both of the energy density and of the susceptibility are well approximated as a single exponential decay @xmath57 on essentially the same timescales .
furthermore , the finite - size behavior of the autocorrelation times indicates @xmath11 dynamical scaling .
this @xmath21 value is significantly less than for the lmc and very similar to the one obtained for over - relaxation methods , although the @xmath58 value of the cluster algorithm is not reached .
the earliest application of lifting @xcite , the motion of a particle on a one - dimensional @xmath47-site lattice with periodic boundary conditions , already featured the decrease of the dynamical scaling exponent from @xmath59 to @xmath60 ( the reduction of the mixing time from @xmath61 to @xmath62 ) . to reach such reductions
, the markov chain must be irreversible .
it was pointed out that the `` square - root '' decrease of the critical exponent was the optimal improvement @xcite .
the concepts of factorized metropolis filters and of infinitesimal moves bring irreversible lifting algorithms to general @xmath47-body systems , although only finite speed - ups were realized so far in the @xmath63 limit .
the three - dimensional heisenberg model seems to be a first such ecmc application with a lowered critical dynamical exponent .
our observation relies on the hypotheses that the energy and the susceptibility are indeed `` slow '' variables , and that the observed decay of the autocorrelation function continues for larger times .
however , in fig .
[ f : ec_aco ] , a crossover from @xmath60 back to @xmath59 as it was observed in the @xmath12-model after @xmath64 sweeps @xcite appears unlikely to arise after hundreds of sweeps .
the dynamical critical exponent @xmath0 represents a maximal improvement with respect to the @xmath65 of lmc , supposing again that the theorems of ref .
@xcite apply to infinitesimal markov chains . in summary ,
we have successfully applied ecmc to the heisenberg model in three dimensions .
ecmc shows considerable promise for spin models , and the numerical data presented in this paper allow us to formulate the exciting conjecture that the dynamical critical exponent for the heisenberg model is @xmath66 .
ecmc is also applicable to frustrated magnets and spin glasses , which involve antiferromagnetic interactions and/or quenched disorder .
our preliminary study indicates that the ecmc algorithm is also useful for a heisenberg spin glass model .
ecmc can be easily combined with other algorithms such as the exchange monte carlo method and the over - relaxation algorithm in the usual manner .
this may allow to investigate the three - dimensional heisenberg spin glass model in the low - temperature region .
large - scale simulations in this direction are currently in progress
. it would be very interesting to understand why ecmc is so much more successful in the heisenberg model than both in hard and soft disks and in the @xmath12 model .
yn and kh thank s. hoshino and m. j. miyama for useful discussions , and j. takahashi and y. sakai for carefully reading the manuscript .
this research was supported by grants - in - aid for scientific research from the jsps , japan ( nos .
25120010 and 25610102 ) , and jsps core - to - core program `` nonequilibrium dynamics of soft matter and information . ''
this work was granted access to the hpc resources of mesopsl financed by the region ile de france and the project equip@meso ( reference anr-10-eqpx-29 - 01 ) of the programme investissements davenir supervised by the agence nationale pour la recherche . | we apply the event - chain monte carlo algorithm to the three - dimensional ferromagnetic heisenberg model .
the algorithm is rejection - free and also realizes an irreversible markov chain that satisfies global balance .
the autocorrelation functions of the magnetic susceptibility and the energy indicate a dynamical critical exponent @xmath0 at the critical temperature , while that of the magnetization does not measure the performance of the algorithm .
this seems to be the first report that the event - chain monte carlo algorithm substantially reduces the dynamical critical exponent from the conventional value of @xmath1 . |
in the standard model there is a single source of cp violation , that is a cp violating phase @xmath0 , in consequence of which all the cp observables are strongly correlated .
the cp violation phase is introduced as a free parameter in the standard model . from the point of view of the fundamental theory of quarks and leptons
this situation is not satisfactory and we would like to see the theoretical origin of the quark mixing matrix phase @xmath0 which describes cp violation .
+ the purpose of the present article is to show that maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath0 are phenomenologically viable .
the jarlskog invariant @xmath2 is related to cp violation and is a function of the cp violating phase @xmath0 .
it is well known that the jarlskog invariant @xmath2 is obtained from the quark mass matrices @xmath4 . without loss of generality we can consider quark mass matrices @xmath4 to be hermitian .
the jarlskog invariant for hermitian mass matrices @xmath4 is given as @xcite , @xcite @xmath5)/2f\ ] ] where @xmath6 here @xmath7 are the quark mass ratios @xmath8 with subscripts @xmath9 referring to @xmath10 in the u - type sector and @xmath11 in the d - type sector , @xmath12 is the top quark mass and @xmath13 is the beauty quark mass . for three quark families ,
@xmath4 are @xmath14 hermitian mass matrices , @xmath15 the hermitian mass matrix @xmath4 may be written in terms of a real symmetric matrix @xmath16 and a diagonal matrix of phases @xmath17 as follows @xmath18 where @xmath19 and @xmath20 the phases in the hermitian matrix @xmath4 may be written as , @xmath21 that is , @xmath22 and @xmath23 are correlated . the real symmetric matrix @xmath16 may be brought to a diagonal form by means of an orthogonal transformation @xmath24 where @xmath25 are orthogonal matrices and @xmath26.\ ] ] + after diagonalization of the mass matrices @xmath27 , one obtains the mixing matrix @xmath28 . for hermitian mass matrices
the quark mixing matrix @xmath28 can be written as @xmath29 where @xmath30 is the diagonal matrix of relative phases , @xmath31 since in the standard model there is a single source of cp violation , only one phase difference in eq .
( [ eq : i.9 ] ) is observable . using the freedom in choosing the unobservable phases of the quark fields ,
it is always possible to take @xmath32 . without loss of generality
we can choose the diagonal matrix given in eq .
( [ eq : i.9 ] ) , with the relative phases as follows , @xmath33 hence , the quark mixing matrix @xmath28 is a function of only one free cp violating phase @xmath0 .
+ we can compute the commutator of the quark mass matrices directly from eqs .
( [ eq : i.3])-([eq:1.7 ] ) and using the phase relations given in eqs .
( [ eq:1.14 ] ) and ( [ eq:1.23 ] ) , it is easy to see that the jarlskog invariant is given as ; @xmath34 here @xmath35 and @xmath36 depend only on the mass matrix elements given in eq .
( [ eq:1.11 ] ) ; the strength of cp violation depends on @xmath0 . in most parameterizations
@xcite the jarlskog invariant @xmath2 is proportional to @xmath37 . in consequence
, there is cp violation if the phase @xmath0 is different from zero or @xmath38 .
it is clear that @xmath2 will have a maximum or a minimum value if the first derivative of @xmath2 with respect to @xmath0 is zero , this implies @xmath39 . furthermore , from eqs .
( [ eq : i.7])-([eq:1.23 ] ) and computing the first and second derivatives of @xmath40 with respect to @xmath0 one obtains that all of the squared quark mixing moduli @xmath40 have an inflection point for @xmath39 , that is , democratic cp violation is realized in nature . in this way
, the fact that there is only one cp violating phase @xmath0 implies that if @xmath2 takes a maximum or a minimum value for one particular value of the phase @xmath0 , @xmath41 , then the quark mixing matrix moduli entries should also take a special value for this particular value of @xmath42 .
the hypothesis of maximal jarlskog invariant and democratic cp violation fixes the cp violating phase @xmath0 to @xmath39 .
+ pushing forward this idea , we are naturally led to hermitian quark mass matrix textures which allow maximal cp violation and maximal jarlskog invariant @xmath2 .
we find phenomenologically viable mass matrix textures which yield a maximum for the jarlskog invariant @xmath2 when the phase @xmath0 takes the value @xmath41 .
+ flavor permutational symmetry of the standard model and an assumed symmetry breaking pattern from which the phenomenologically allowed hermitian quark mass matrix textures are derived can be the theoretical origin of the quark mixing matrix phase @xmath42 .
+ this paper is organized as follows : in section [ sec : ii ] , we find the conditions to obtain the hermitian quark mass textures which give maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath0 . in section [ sec : iii ] we show that all hermitian mass matrices can be obtained from flavor permutational symmetry breaking . in section [ sec : v ]
we study the mass patterns derived from maximal jarlskog invariant and maximal cp violation .
we determine that four of the so - called ramond - roberts - ross ( rrr ) @xcite patterns are derived from these conditions . in section [ sec : vi ] , we undertake a chi square @xmath43 fit of the quark mixing matrix to show that maximal cp violation is phenomenologically viable .
finally , in section [ h3 ] we give our conclusions .
the jarlskog invariant @xmath2 takes its maximum or minimum value when the first derivative of @xmath2 with respect to the cp violating phase @xmath0 vanishes ; @xmath44 in general , the maximum or minimum value of @xmath2 is not explicitly exhibited if we compute the first and second derivatives of @xmath2 with respect to the cp violating phase @xmath0 . in order to have maximal jarlskog invariant @xmath2 , we demand that the first derivative of the jarlskog invariant @xmath2 with respect to the cp violating phase @xmath0 vanishes for @xmath45 , that is ; @xmath46 three solutions for eq .
( [ eq:1.39 ] ) are interesting : * @xmath47 and @xmath48 ; in this case @xmath2 has an inflection point , cp is an exact symmetry and the jarlskog invariant @xmath2 is zero .
that is , the jarlskog invariant @xmath2 as a function of the cp violating phase @xmath0 does not take a maximal or minimal value for @xmath49 . *
@xmath50 and @xmath51 ; this solution corresponds to maximal jarlskog invariant @xmath2 and non - maximal cp violating phase @xmath0 . *
@xmath52 and @xmath53 ; this solution corresponds to maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath0 .
we can satisfy the solutions of eq .
( [ eq:1.39 ] ) by taking particular values for the hermitian quark mass matrix entries and look for the phenomenological implications of the exact zeros .
+ this means that the basis in which the jarlskog invariant takes its maximum or minimum value is a basis in the space of flavors in which the mass matrices exhibit exact texture zeros for both up and down type quark sectors .
we can study the implications of maximal cp violation in the quark mixing matrix . in the mass basis ,
cp violation is related to the quark mixing matrix , this implies that the quark mixing matrix is also a function of the cp violating phase @xmath0 .
we can always write the diagonal phase matrix @xmath30 given in eq .
( [ eq : i.9 ] ) as @xmath54 where @xmath55 and @xmath56 are diagonal matrices .
the quark mixing matrix can be written as @xmath57 where @xmath58 only the second term on the right hand side of eq .
( [ eq:1.43 ] ) depends on the cp violation phase @xmath0 . when the flavor mixing matrix @xmath28 is written as in eq .
( [ eq:1.43 ] ) the complex phase @xmath0 describing cp violation appears in all elements of @xmath28 .
the invariant measures of the quark mixing matrix are the moduli of its elements , i.e. the quantities @xmath59 , the jarlskog invariant @xmath2 and the inner angles of the unitarity triangles .
computing the derivative of the quark mixing matrix elements given in eq .
( [ eq:1.43 ] with respect to the cp violation phase @xmath0 , we obtain @xmath60 for the second derivative we have @xmath61 from eqs .
( [ eq:1.47 ] ) and ( [ eq:1.49 ] ) we have for i = k , j = l three non trivial cases * when the cp violating phase takes the values @xmath49 , cp is an exact symmetry , all of the elements @xmath62 of the quark mixing matrix are real and all of the moduli of the quark mixing matrix @xmath63 take a minimum or a maximum value for @xmath49 .
* when the cp violating phase takes the value @xmath50 the first and second derivatives of @xmath59 are equal , but it does not have any further implications . * when the cp violating phase takes the values @xmath52 all of the elements of the quark mixing matrix moduli @xmath63 have an inflection point . this is a very general property and implies that none of the quark mixing matrix moduli are preferred by nature , in consequence none of the quark mixing matrix moduli entries are maximal or minimal when the cp violation phase @xmath0 takes its maximal value , that is @xmath41 . in a theory with hermitian mass matrices and only one cp violating phase ,
all of the quark mixing matrix moduli elements @xmath63 have an inflection point for @xmath52 . in this picture of the quark mixing matrix the magnitude of the cp violating phase @xmath0
can be fixed demanding that the jarlskog invariant @xmath2 takes its maximum value .
this gives the solution with @xmath52 and @xmath53 and means that the basis in which the jarlskog invariant @xmath2 takes its maximum value is a basis in the space of flavors in which @xmath53 imposes exact zeros for the mass matrices , the cp violating phase @xmath0 takes its maximal value and all of the quark mixing matrix moduli elements have an inflection point .
this particular property of the quark mixing matrix @xmath28 and the jarlskog invariant @xmath2 is the main result of this paper .
in the standard model , prior to the introduction of the higgs boson and mass terms , the lagrangian is chiral and invariant with respect to any permutation of the left and right quark fields , as a consequence of this the left and right quark fields are transformed independently . that is @xcite @xmath65 where @xmath66 transforms the left fields , and @xmath67 transforms the right fields . in this expression ,
@xmath68 and @xmath69 denote the left and right quark @xmath70- and @xmath71-fields in the current or weak basis respectively , @xmath72 is a column matrix , its components @xmath73 are the quark dirac fields , @xmath74 is the flavor index .
+ in this way , the flavor permutational symmetry group of the bilinear form , @xmath75 is @xmath76 , with elements all the pairs @xmath77 with @xmath78 and @xmath79 .
+ the charged currents are transformed under the flavor group @xmath76 as : @xmath80 from this expression it is clear that the charged currents @xmath81 are invariant under the transformations of the flavor symmetry group if and only if @xmath82 and @xmath83 are the same matrix . thus , the charged current invariance condition under the family symmetry group implies that the up and down quark fields are transformed with the same group .
+ when the gauge symmetry is spontaneously broken , the quarks and leptons acquire mass and the chiral symmetry of the theory is broken
. then left and right fields are not independent .
the quark mass terms in the lagrangian gives rise to quark mass matrices @xmath84 and @xmath85 , @xmath86 we assume that under the flavor symmetry group the quark fields are transformed in the following way : @xmath87 the left and right components of the same field are transformed with the same group element .
the flavor symmetry group of the bilinear form ( [ eq:1.3 ] ) is the group @xmath88 whose elements are the pairs @xmath89 with first element @xmath78 and second element @xmath90 , and @xmath91 .
clearly , @xmath92 .
+ the mass term in the yukawa coupling ( [ eq:1.3 ] ) , transforms under the family symmetry group @xmath76 , in the following way : @xmath93 from here we obtain that under the flavor symmetry group @xmath94 , the mass matrices @xmath95 are transformed according to the rule : @xmath96 the yukawa sector is invariant under the flavor family group @xmath94 if , @xmath97 thus , if the mass matrix @xmath98 commutes with all the elements of the flavor group @xmath99 , the yukawa sector of the standard model will have this symmetry .
a number of authors @xcite-@xcite have pointed out that realistic quark mass matrices result from the flavor permutational symmetry @xmath64 and its spontaneous or explicit breaking . as an attempt to provide a realistic and general symmetry breaking pattern which clarifies the possible origin of the texture zeros in the hermitian mass matrices , we suggest an @xmath100 matrix ansatz and its spontaneous or explicit breaking . under exact @xmath64 symmetry , the mass spectrum , for either up or down quark sectors , consists of one massive particle ( top and bottom quarks ) in a singlet irreducible representation and a pair of massless particles in a doublet irreducible representation .
the group @xmath101 treats three objects symmetrically , while the hierarchical nature of the mass matrices is a consequence of the representation structure @xmath102 of @xmath101 , which treats the generations differently .
the group @xmath101 is non - abelian and has a doublet @xmath103 and @xmath104 and a singlet @xmath105 , @xmath106 hermitian mass matrix @xmath98 contains 9 real free independent parameters , namely , @xmath107 and @xmath108 .
these free parameters emerge in different stages of the flavor permutational symmetry breaking .
+ in the weak basis , the mass matrix with the exact @xmath64 symmetry reads @xmath109 where @xmath110 denotes the mass of the third family quark , @xmath111 or @xmath112 . to generate masses for the second family , one has to break the permutational symmetry @xmath64 down to @xmath113 .
this may be done by adding to @xmath114 a term @xmath115 which is invariant under @xmath116 but breaks @xmath117 .
this can be done with two different matrices , with well defined symmetry properties . in the first case
we can assume that @xmath118 transforms as the tensorial representation of the vector @xmath104 , @xmath119 in the second case , we can assume that @xmath118 transforms as the complex tensorial symmetry term that mixes the singlet @xmath105 with the doublet vector @xmath104 .
then , in the weak basis , @xmath118 is given by @xmath120.\ ] ] we may now turn our attention to the question of breaking the @xmath121 symmetry . in order to give mass to the first family , we add another term @xmath122 to the mass matrix .
this breaking can be obtained in three different ways : first assume that @xmath122 transforms as the tensorial representation of the vector @xmath103 ; @xmath123 putting the first family in a complex representation will allow us to have a cp violating phase in the mixing matrix .
this breaking can be done in two different ways , we can assume that @xmath122 transforms as the mixed symmetry term of the doublet complex tensorial representation of the @xmath124 diagonal subgroup of @xmath64 . then , in the weak basis , @xmath122 is given by @xmath125 \end{array}\ ] ] we can assume that @xmath122 transforms as the complex tensorial symmetry term that mixes the singlet @xmath105 with the doublet vector @xmath103 .
then , in the weak basis , @xmath122 is given by @xmath126 \end{array}\ ] ] finally , adding the quark mass matrices given in eqs .
( [ eq:3.27]-[eq:3.37 ] ) , we get the mass matrix @xmath98 in the weak basis . to make explicit the assignment of particles to irreducible representations of @xmath64 , it will be convenient to make a change of basis from the weak basis to a hierarchical basis . in this basis , the quark fields are @xmath127 @xmath128 @xmath129 the subindex @xmath130 denotes the hierarchical basis . in the hierarchical basis ,
the third family quarks , @xmath131 or @xmath132 , are assigned to the invariant singlet irreducible representation @xmath133 , the other two families are assigned to @xmath134 and @xmath135 , the two components of the doublet irreducible representation of @xmath136 .
the mass matrix @xmath137 in the hierarchical basis is related to the mass matrix in the weak basis by the unitary transformation @xmath138 where @xmath139 then , in the hierarchical basis , the mass matrix @xmath98 takes the form given in eq .
( [ eq:1.7 ] ) with , @xmath140 and the phases are given by the following relations @xmath141 which are also given in eq .
( [ eq:1.14 ] ) .
+ at this stage in our derivation of hermitian mass matrices a question comes naturally to mind : which flavor permutational symmetry breaking patterns are realized in nature ? or phrased differently : which of the hermitian mass matrix textures are phenomenologically allowed ? .
to answer this question , we will analyze the phenomenological implications of maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath41 .
we are now in a position to consider the realistic @xmath142 hermitian mass matrices which allow maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath0 . in order to have maximal cp violation , we demand that the first derivative of the jarlskog invariant @xmath2 with respect to the cp violating phase @xmath0 vanishes for @xmath41 , that is ; @xmath143 where @xmath144[f_d(c_u - e_u)+b_u a_d ] \cr & + & f_d a_u[f_u ( c_d - e_d)+a_u b_d ] [ a_d(d_u - e_u)+b_u f_d]\cr&+&2a_ua_d[f_u(c_d- e_d)+a_u b_d][f_d(c_u - e_u)+a_d b_u ] \cr&+ & 2 f_uf_d[a_u(d_d - e_d)+f_ub_d][a_d(d_u - e_u)+f_d b_u]\cr&+&(a_uf_d - a_df_u)(a_ua_d+f_uf_d)[b_d(c_u - d_u)+b_u(d_d - c_d)]\end{aligned}\ ] ] we can satisfy eq .
( [ eq:5.3 ] ) by taking particular values for the hermitian quark mass matrix entries and look for the phenomenological implications of these exact zeros .
first of all notice that the quark mass matrices @xmath145 and @xmath146 commute when @xmath147 this implies that the jarlskog invariant @xmath2 is zero for these textures and cp violation is related to the condition @xmath148 thus , cp violation is related to the breaking of the @xmath121 by the term @xmath149 that transforms as the mixed symmetry term of the doublet complex tensorial representation of the @xmath124 diagonal subgroup of @xmath64 or to the breaking of the @xmath121 by the complex tensorial symmetry term @xmath150 that mixes the singlet @xmath105 with the doublet vector @xmath103 . + mass matrices with textures are over - constrained leading to the prediction of relations between mixing angles and quark masses and it is our task to find which , if any , of these textures is viable and compatible with maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath0 . for the assumptions made above
, we found 14 texture patterns that satisfy eq .
( [ eq:5.3 ] ) , 1 .
if we take @xmath151 , this gives two solutions @xmath152 or @xmath153 .
for the first solution we have the textures @xmath154 and the second solution is obtained if we change @xmath155 in eq .
( [ eq:5.9 ] ) .
if we take @xmath156 , this gives two solutions @xmath152 or @xmath153 .
for the first solution we have the textures @xmath157 and the second solution is obtained if we change @xmath155 in eq .
( [ eq:5.13 ] ) .
if we take @xmath158 , this gives two solutions @xmath152 or @xmath153 .
for the first solution we have the textures @xmath159 for the second solution we have @xmath160 4 .
if we take @xmath161 , this gives two solutions @xmath152 or @xmath153 .
these textures can be obtained by changing @xmath155 in the textures given in eqs .
( [ eq:5.19 ] ) and ( [ eq:5.17 ] ) respectively .
taking @xmath162 , we obtain six solutions , @xmath163 or @xmath164 or @xmath165 where @xmath166 .
the textures for @xmath167 are the following ; + @xmath168 + for the second solution we have @xmath169 for the third solution we obtain , @xmath170 the solutions for @xmath171 are obtained by changing @xmath155 in the textures given in eqs .
( [ eq:5.15])-([eq:5.23 ] ) .
maximal jarlskog invariant @xmath2 and democratic cp violation with maximal cp violating phase @xmath0 is the underlying theoretical assumption which leads to the structure of the quark mass textures .
this means that the basis in which the jarlskog invariant @xmath2 is maximal , and the cp violating phase @xmath0 is maximal , is a basis in the space of flavors , in which the mass matrices exhibit exact texture zeros for both up and down type quark sectors .
it is clear from eqs .
( [ eq:5.5]-[eq:5.23 ] ) that , the zeros of the quark mass textures @xmath172 and @xmath173 derived from the condition of maximal jarlskog invariant and maximal cp violating phase @xmath0 do not give up the parallelism between the structures of @xmath172 and @xmath173 . + ramond roberts and ross ( rrr ) @xcite have found that there exist five phenomenologically allowed patterns of hermitian quark mass matrices , which have five texture zeros .
the five allowed patterns are shown in table [ tab:1 ] . the question raised by rrr @xcite : `` what underlying theory can lead to such structure ? ''
has an answer in terms of the maximal jarlskog invariant and the maximal democratic cp violating phase @xmath0 .
the rrr patterns i and iii - v are obtained when we take @xmath174 in our textures . the mass pattern ii can not be obtained from the above assumptions , for this reason this texture pattern could not be a good candidate for the quark mass matrices in an underlying theory of fermion mass generation . +
the appearance of the texture zeros in the mass matrices is generally assumed to arise at some high scale q where new physics connected with mass generation comes into play , and the comparison of a mass matrix pattern which holds at a large scale q with experiment is complicated by the evolution of the yukawa couplings with energy .
this evolution can change zeros in a mass matrix at a given scale into small , but non - vanishing , contributions at a different scale .
the effect of the renormalization group evolution of the couplings is to obscure possible mass matrix patterns .
furthermore , the renormalization group ( rg ) running may not preserve the hermeticity of the quark mass textures .
+ we follow the olechowski - pokorski @xcite paper for the one loop renormalization group equations for the quark mixing matrix elements @xmath175 and the rrr @xcite analysis of the renormalization group evolution of the ckm matrix and the masses . keeping the top and bottom yukawa couplings only and neglecting thresholds , olechowski and pokorski @xcite and rrr @xcite obtained that the ckm matrix elements evolve as @xmath176 here @xmath177 where the elements are evaluated at the scale @xmath178 , @xmath179 and @xmath180 are the yukawa couplings and @xmath181 is a constant determined by the couplings of the theory . from eq .
( [ eq:1.43 ] ) and ( [ eq:5.34 ] ) , we obtain the running equation for the cp violating phase @xmath0 ; @xmath182 that is , the cp violating phase @xmath0 do not run with energy .
the running of the quark mixing matrix elements @xmath183 yield the following rg equations for the matrices @xmath184 and @xmath185 introduced in eq .
( [ eq:1.45 ] ) : @xmath186 olechowski and pokorski @xcite also found that the jarlskog invariant @xmath187 evolves via the equation @xmath188 where we can choose @xmath187 independent of the parametrization given by eq .
( [ eq:1.31 ] ) , from which we find @xmath189 it follows from this equation that the zeros imposed at some high energy scale q for the constrains @xmath190 or @xmath191 are protected during the renormalization group evolution of the ckm matrix and the masses .
this result with that of eq .
( [ eq:5.35 ] ) protects the hermiticity of the mass textures and sheds some light for the comparison with low energy experiments of the mass matrix patterns derived from maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath0 which holds at a large scale q. we can take the same texture patterns at all energies , the only constraint is to express all the running quark masses at some common energy scale . + as noticed by fritzsch and xing @xcite , with the same input values of quark mass ratios only some of the textures patterns are in good agreement with current experimental data and the 14 textures patterns that we derive can not simultaneously survive .
we find three texture patterns which can be confronted with experiment , pattern i and the patterns iii and iv in the rrr classification given in table [ tab:1 ] .
these patterns can be written in terms of the quark mass ratios without introducing any new parameters .
before proceeding to give the numerical results for the mixing matrix @xmath192 , it will be convenient to stress the following points : 1 .
the cp violation phase @xmath0 is fixed to @xmath1 from the condition of maximal jarlskog invariant and democratic cp violation .
2 . the masses of the lighter quarks are less well determined , while the moduli of the entries in @xmath193 with the largest errors , namely @xmath194 and @xmath195 are sensitive to changes in the values of the ligther quarks @xmath196 and @xmath197 respectively .
the sensitivity of @xmath194 and @xmath195 to changes in @xmath196 and @xmath197 respectively , is reflected in the shape of the unitarity triangle which changes appreciably when the masses of the lighter quarks change within their uncertainty ranges .
3 . the moduli of the entries in @xmath193 with the smallest errors , namely @xmath198 , @xmath199 , @xmath200 and @xmath201 are the most sensitive to changes in the values of the lighter quarks @xmath197 and @xmath202 . hence ,
the quality of the fit of @xmath203 to @xmath193 for @xmath204 is spoiled if relatively large changes in the masses of the lighter quarks @xmath197 and @xmath202 are made .
this puts a strong constraint on the allowed values for @xmath205 .
the best simultaneous @xmath206 fits of @xmath203 for @xmath207 to the experimentally determined quantities are very sensitive to the ratio @xmath205 and allows us to constrain @xmath208 .
4 . for the purpose of calculating quark mass ratios and computing the mixing matrix
, it is convenient to give all quark masses as running masses at some common energy scale @xcite , @xcite . in the present calculation ,
following peccei @xcite , fritzsch @xcite , the babar book @xcite and rrr @xcite , we used the values of the running quark masses evaluated at @xmath209 .
the running quark masses evaluated at @xmath209 as given by fusaoka and koide @xcite are in units of @xmath210 ; @xmath211 from the strong hierarchy in the masses of the quark families , @xmath212 , we expect @xmath213 in @xmath4 to be very close to unity .
computing the invariants of @xmath4 , @xmath214 , @xmath215 and @xmath216 and by appropriately choosing the signs of the quark masses , the entries in the mass matrix may be readily expressed in terms of the quark masses . from here
, we computed @xmath217 with the exact expressions for @xmath218 and @xmath219 .
+ we use the minuit program to perform a @xmath206 fit @xcite of the quark mixing matrix moduli @xmath220 to @xmath221 , @xmath207 . in the fit , we fixed @xmath1 and vary the quark mass ratios within the allowed limits computed from eqs .
( [ eq:5.40 ] ) .
once the minimum of the @xmath206 is found , we release the quark mass ratios and let them vary without limits . for three different allowed mass patterns the minimum of the @xmath206 fit of the exact expressions of the quark mixing matrix @xmath220 to @xmath221 , @xmath207 , is obtained .
the results for the one - standard - deviation values of the running quark mass ratios are given in table [ tab:2 ] .
+ for the texture pattern i , the entries @xmath222 , and @xmath223 in the mass matrix @xmath224 and @xmath225 , and @xmath226 in the mass matrix @xmath227 may be readily expressed in terms of the quark mass ratios as follows ; @xmath228 @xmath229\end{aligned}\ ] ] the quark mixing matrix for pattern i is a function of three quark mass ratios and @xmath0 , that is , @xmath230 , the range of values of @xmath231 are in the allowed limits computed from eqs .
( [ eq:5.40 ] ) and the quark mass ratio @xmath232 is close to the upper bound of the allowed limit as computed from eqs .
( [ eq:5.40 ] ) , while the high central value of @xmath233 obtained from @xmath234 agrees with the value of @xmath196 at the top scale found in @xcite and @xcite .
these quark mass ratio values are in good agreement with the phenomenological relations obtained by fritzsch@xcite @xmath235 it is clear from these relations that smaller values of @xmath234 favors smaller values of @xmath236 as it is shown in fig .
[ fig:1 ] .
+ for the texture pattern iii we get @xmath237 the parameters @xmath225 , and @xmath226 of the quark mass texture @xmath227 , are given by eq . ( [ eq:5.43 ] ) in terms of the quark mass ratios . for pattern iii , the quark mixing matrix is a function of the cp violating phase @xmath0 and three quark mass ratios , that is , @xmath238 . + as in the previous case , in the @xmath206 fit of @xmath220 to @xmath221 for @xmath207 we fixed the cp violation phase @xmath1 and varied the quark mass ratios within the allowed limits computed from eqs .
( [ eq:5.40 ] ) .
once the minimum is found , we release the quark mass ratios and let them vary without limits .
the minimum of the @xmath206 fit of the quark mixing matrix computed from the texture pattern iii is obtained for the quark mass ratio values given in table [ tab:2 ] .
the values of the quark mass ratios @xmath239 and @xmath232 obtained for pattern iii are close to the central values of the quark mass ratios as obtained from eqs .
( [ eq:5.40 ] ) . for this pattern the central values of the quark mass ratios
gives the central value of @xmath236 which is in very good agreement with its latest determination @xcite , @xcite and @xcite .
+ for the texture pattern iv , the entries @xmath240 , and @xmath223 in the mass matrix @xmath224 and @xmath241 , and @xmath226 in the mass matrix @xmath227 are expressed in terms of the quark mass ratios as follows : @xmath242,\end{aligned}\ ] ] @xmath243 for this texture the quark mixing matrix @xmath28 is a function of three quark mass ratios @xmath244 and the cp violation phase @xmath41 . as in the first case
, the value of @xmath233 obtained from @xmath245 agrees with the value of @xmath196 at the top scale found in @xcite and @xcite , the quark mass ratios @xmath246 and @xmath247 are close to the lower allowed limit as computed from eqs .
( [ eq:5.40 ] ) .
these values for the quark mass ratios are in very good agreement with the phenomenological relations @xcite given in eq .
( [ eq:5.58 ] ) . + in order to have an estimation of the sensitivity of @xmath203 to the uncertainty in the values of the quark mass ratios , we computed the allowed values of @xmath203 , corresponding to the range of values of the quark mass ratios given in table [ tab:2 ] , keeping @xmath0 fixed at the value @xmath1 .
the results of the @xmath206 fit of the theoretical expressions for @xmath248 , to the experimentally determined quantities @xmath249 gives the one - standard - deviation ranges shown in tab .
[ tab:3 ] .
the `` best fit '' values of the theoretical expressions for @xmath250 , using the central values of the quark mass ratios from table [ tab:2 ] are given in table [ tab:4 ] .
+ as is apparent from tables [ tab:3 ] and [ tab:4 ] , the agreement between computed and experimental values of all entries in the mixing matrix is very good .
in the three cases , the one sigma estimated range of variation of the moduli , the computed values in the four entries of the upper left corner of the matrix @xmath251 are in agreement with the error bands of the corresponding entries of the matrix of the experimentally determined values of the moduli @xmath252 .
the estimated range of variation in the computed values of the entries in the third column and the third row of @xmath248 is comparable with the error band of the corresponding entries in the matrix of experimentally determined values of the moduli , with the exception of the element @xmath208 in which case the one sigma estimated range of variation due to the uncertainty in the values of the quark mass ratios is significantly smaller than the error band in the experimentally determined value of @xmath253 .
the phenomenologically allowed quark mass patterns can be distinguished from the different predicted value for the quark mixing matrix element @xmath253 and hence from the different allowed values of the cp angles @xmath254 , @xmath255 and @xmath256 of the unitarity triangle .
the one - standard - deviation computed range of values of @xmath254 , @xmath255 and @xmath256 , corresponding to the range of values of the mass ratios given in table [ tab:2 ] with @xmath0 fixed at the value @xmath41 are given in table [ tab:5 ] .
the uncertainty in the values of the quark mass ratios is reflected in the shape of the unitarity triangle .
thus , different shapes of the unitarity triangle are equivalent to different values of the cp angles .
the one - sigma regions of the cp angle @xmath255 are shown in figs .
[ fig:1]-[fig:3 ] as function of the quark mass ratios @xmath257 and @xmath258 for the different patterns . in figs .
[ fig:1]-[fig:3 ] we also illustrate the correlations between the cp angles .
recent determinations of @xmath254 , @xmath255 and @xmath256 have been done in @xcite-@xcite .
it is interesting to compare our results with those of ali and london @xcite , who have studied the profile of the unitarity triangle in the standard model and in several variants of the mssm characterized by a single phase in the quark mixing matrix . according to these authors @xcite ,
an estimation of the range of values of the three inner angles of the unitarity triangle , compatible with the experimental information on the absolute values of the matrix elements @xmath259 , in the standard model is : @xmath260 , @xmath261 , and @xmath262 at @xmath263 cl .
+ we present in table [ tab:5 ] the one standard deviation computed ranges of @xmath254 , @xmath255 and @xmath256 in the three mentioned quark mass textures , together with the corresponding ranges obtained indirectly from the ckm unitarity .
the values of @xmath255 obtained in model i , iii , and iv and the values of @xmath254 given in i and iv are in the one - sigma allowed range given by these authors @xcite , while @xmath254 in model iii lies outside the one - sigma range , but is within the @xmath263 cl range given by ali and london @xcite .
one espects that at the b factories and hadron colliders the cp - violating angles @xmath254 , @xmath255 and @xmath256 will be directly and precisely measured .
we see from table [ tab:5 ] that whereas @xmath255 turns out to be quite similar in the three models listed , precise determination of @xmath254 and @xmath256 may lead to a discrimination among the three models .
the task of searching for phenomenologically viable quark mass textures derived from the condition of maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath0 is proposed . in order to conduct these searches more effectively ,
we have suggested in this paper the idea of democratic cp violation from flavor permutational symmetry breaking in the hermitian mass matrices as an organizing principle .
the conditions of maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath41 , encoded in eq .
( [ eq:1.39 ] ) are satisfied by taking particular zero values for the hermitian quark mass matrix entries .
+ when the cp violating phase takes the values @xmath41 , all of the elements of the quark mixing matrix moduli @xmath63 have an inflection point and none of them is preferred by nature , that is , democratic cp violation is realized in nature .
we have shown that the cp violating phase @xmath0 does not run with energy [ eq .
( [ eq:5.35 ] ) ] , and the running of the jarlskog invariant is described by eqs .
( [ eq:5.37 ] ) and ( [ eq:5.39 ] ) from this set of equations , it follows that the zeros imposed at some high energy scale q for the constrains @xmath190 or @xmath191 are protected during the renormalization group evolution of the ckm matrix and the masses . the zeros of the quark mass textures @xmath172 and @xmath173 in eqs .
( [ eq:5.5]-[eq:5.23 ] ) derived from the condition of maximal jarlskog invariant and maximal cp violating phase @xmath0 do not give up the parallelism between the structures of @xmath172 and @xmath173 .
for this reason , mass paterns with parallel structures of @xmath172 and @xmath173 could not be a good candidate for an underlying quark mass texture theory as it is supposed in some resent works @xcite , @xcite .
furthermore , we have argued that maximal jarlskog invariant @xmath2 and democratic cp violation with maximal cp violating phase @xmath0 is the underlying theoretical assumption which leads to the structure of phenomenologically viable rrr @xcite quark mass textures .
+ we have evaluated the quark mixing matrix @xmath192 directly from the hermitian quark mass matrices @xmath27 . for the quark
mass values , taken at the top quark mass scale given in eq .
( [ eq:5.40 ] ) , only the phenomenologically allowed quark mass patterns i , iii and iv in table [ tab:1 ] give a ckm mixing matrix @xmath192 in agreement with experiment .
as is apparent from table [ tab:3 ] , these patterns can be distinguished from the predicted value of the quark mixing matrix element @xmath195 .
we have also performed an analysis of the unitarity triangle .
the resulting allowed cp angles of the unitarity triangles in table [ tab:5 ] obtained from each mass pattern are very constrained .
hence , these patterns can also be distinguished from the different allowed values of the cp angles @xmath254 or @xmath256 . in the near future ,
cp - violating asymmetries in b decays will be measured at b - factories and hadron colliders .
such measurements will give us crucial information about the interior angles @xmath254 , @xmath255 and @xmath256 of the unitarity triangle .
if we are lucky , the predictions from one of these patterns will be confirmed pointing to the texture preferred by nature .
in conclusion , the hypothesis of maximal jarlskog invariant and democratic cp violation fixes the cp violating phase @xmath0 to @xmath39 and the fact that the phenomenologically viable quark mass textures for the quarks can be obtained from the conditions of maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath42 gives evidence in favor of this hypothesis .
all the analysis for the maximal jarlskog invariant and democratic maximal cp violating phase @xmath0 apply also to the lepton sector and the allowed mass patterns for the lepton sector can simply be obtained by using the georgi - jarlskog ansatz @xcite taking all the elements of @xmath264 as the same as for @xmath265 except the @xmath266 element which is multiplied by the factor 3 .
it is clear from this assumption , that once we know the right quark mass pattern , we can make predictions for the neutrino mass ratios and lepton mixing matrix .
i would like to specially thank prof .
ahmed ali for supervising this work and reading the manuscript .
i also thank nabil ghodbane for his invaluable help with minuit and dimitri ozerov for his invaluable help with paw and reading the manuscript .
this work was partly supported by an icsc - world laboratory bjorn wiik scholarship and by conacyt ( mxico ) under contract 000441 . c. jarlskog , _ phys .
rev . lett._*55 * , 1039 ( 1985 ) , see also c. jarlskog , in `` cp violation '' , edited by c. jarlskog , advanced series on directions in high energy pysics , vol 3 , ( world scientific , singapore , 1989 ) .
s. mele , _ phys .
rev . _ * d59 * , 113011 , ( 1999 ) , see also , s. mele , indirect determination of the vertex and angles of the unitarity triangle , hep - ph/0103040 .
m. ciuchini , g. dagostini , e. franco , v. lubicz , g. martinelli , f. parodi , p. roudeau , a. stocchi 2000 ckm - triangle analysis a critical review with updated experimental inputs and theoretical parameters , hep - ph/0012308 .
a. j. buras , `` weak hamiltonian , cp - violation and rare decays '' tum - hep-316/98 ; hep - ph/9806471 . to appear in `` probing the standard model of particle interactions '' , f. david and r. gupta , eds .
1998 elsevier . the babar collaboration , b. aubert , et al , _ phys .
lett . _ * 86 * , 2515 , ( 2001 ) .
belle collaboration , a. abashian et .
al . , _ phys .
_ * 86 * , 2509 , ( 2001 ) .
cdf collaboration , t. affolder et al .
_ * d61 * , 072005 , ( 2000 ) . | we argue here why cp violating phase @xmath0 in the quark mixing matrix is maximal , that is , @xmath1 . in the standard model cp violation
is related to the jarlskog invariant @xmath2 , which can be obtained from non commuting hermitian mass matrices . in this article
we derive the conditions to have hermitian mass matrices which give maximal jarlskog invariant @xmath2 and maximal cp violating phase @xmath0 .
we find that all squared moduli of the quark mixing elements have a singular point when the cp violation phase @xmath0 takes the value @xmath1 .
this special feature of the jarlskog invariant @xmath2 and the quark mixing matrix is a clear and precise indication that cp violating phase @xmath0 is maximal in order to let nature treat democratically all of the quark mixing matrix moduli .
epsf # 1@xmath3#1 |
the ultraviolet ( uv ) spectrum of seyfert galaxies is dominated by a quasi - thermal component ( shields 1978 , malkan & sargent 1982 , malkan 1983 ) which extends into the soft x - ray band ( walter & fink 1993 ) .
although this feature is usually interpreted as thermal emission from a geometrically thin , optically thick accretion disk around a supermassive black hole , recent observations cast doubt on this simple explanation .
a disconinuity at the lyman edge , either in emission or absorption , seems to be a generic feature of theoretical models of accretion disks ( blaes 1998 ) .
however , partial absorption features have been detected in @xmath2 of observed agns ( antonucci , kinney & ford 1989 , koratkar , kinney & bohlin 1992 ) .
recently , iwasawa , ( 1996 ) detected extremely broad iron k-@xmath3 line emission from the seyfert galaxy mcg-6 - 30 - 15 .
they interpret this emission as x - rays reflected by an accretion disk extending to 1.23 @xmath4 , where @xmath5 , of the central black hole . for a stable accretion disk to exist this close to the central black hole this black hole must be rotating close to its maximal value ( shapiro & teukolsky 1983 ) .
the appearance of an accretion disk around a maximally rotating kerr black hole is very different than that of a disk around a non - rotating hole .
the effective temperature of the disk is larger .
the material in the disk orbits the hole at higher velocities , increasing the range of doppler boosts of the disk radiation .
finally , the emitted radiation has to escape from a deeper potential well , which causes larger redshifts . in this paper
, we compute the observed spectrum of a viscously heated accretion disk around a maximally rotating supermassive black hole .
the rest frame spectrum of radiation emitted by the disk is computed using the code developed by sincell & krolik ( 1998 ) .
the observed disk spectrum is calculated using a ray - tracing program which we describe in the next section .
the photon equations of motion in a kerr metric are derived from the lagrangean @xmath6 using the geodesic equation @xmath7 where @xmath8 is an affine parameter and @xmath9 indicates the coordinate of interest : @xmath10 .
expressions for the metric components , @xmath11 are given by bardeen , ( 1972 ) for boyer - linquist coordinates .
equation [ eq : lagrangean eq motion ] expands to eight first - order coupled differential equations for the four space coordinates and the four components of the photon momentum .
this equation set can be reduced to six first - order equations by noting that the photon energy @xmath12 and angular momentum parallel to the symmetry access @xmath13 where @xmath14 , are conserved along a null geodesic .
it is possible to reduce the equation set further ( see bardeen , 1972 , cunningham 1975 , agol 1997 ) but this introduces complicated bookkeeping into the ray - tracing calculation .
the differential equations are integrated with a fourth - order runge - kutta routine ( press , 1992 ) .
there is one additional constant of the motion @xmath15\ ] ] which is needed to specify a particular photon geodesic . in this equation
, @xmath16 is the black hole spin parameter and @xmath17 in boyer - lindquist coordinates . to compute the observed spectrum , we choose a particular inclination angle and a grid in the impact parameters @xmath18 .
the parameter @xmath3 , which is measured in units of @xmath4 , is defined as the projected distance away from the black hole along an axis perpendicular to the spin axis of the black hole .
the parameter @xmath19 is the projected distance parallel to the spin axis .
the range of values for the impact parameters corresponds roughly to the inner and outermost radii of the accretion disk . for each pair of impact parameters , the constants of motion of the corresponding geodesic
are defined @xmath20 @xmath21 and @xmath22 .
the equations of motion are then integrated inward from @xmath23 , where @xmath24 is the mass of the black hole and we adopt geometrized units , and @xmath25 .
the integration is stopped when the geodesic either hits the disk ( @xmath26 ) , escapes to infinity or is captured by the black hole . for a geodesic
which intersects the disk at radius @xmath27 , we can compute two important quantities : the total redshift of a photon emitted by the disk at @xmath27 and escaping to infinity @xmath28 where @xmath29 are the frequency of the emitted and observed photon , and the angle at which the photon leaves the disk , @xmath30 , measured in the rest frame of the orbiting material .
explicit equations for these quantities can be found in cunningham ( 1975 ) .
the observed spectrum is found by integrating the emitted intensity over the impact parameters of all the geodesics which reach the observer @xmath31 where the factor @xmath32 arises from the invariance of the photon occupation number and @xmath33 is the spectral intensity of the radiation emitted by the disk .
for example , the intensity could be described by an analytic expression for the reflected iron line as a function of radius or a grid of numerical models of the continuum flux from an optically thick accretion disk .
in figure [ fig : continuum ] , we plot the uv continuum of an optically thick accretion disk around a maximally rotating black hole . the black hole is assumed to have a mass of @xmath34 and to accrete matter at @xmath35 , where @xmath36 is the eddington critical accretion rate .
the three curves represent three different values of the disk inclination angle , @xmath37 .
we find that the lyman edge discontinuity is unobservable when @xmath38 , but that a significant absorption feature remains at smaller inclinations .
if this model is typical of all accretion disks , then we would predict that lyman edge features should be seen in @xmath39 of all agns .
this is obviously only a very rough estimate , but it does indicate that black hole rotation is a plausible means of erasing the lyman discontinuity in a large fraction of agns . | we calculate the amplitude of the lyman discontinuity for an accretion disk around a maximally rotating black hole .
the locally emitted disk spectrum is computed with a previously developed numerical model ( sincell & krolik 1998 ) and then a new general relativistic ray - tracing code is used to determine the observed spectra as a function of inclination angle , @xmath0 .
we find that lyman discontinuities are undetectable for disks viewed at inclinations @xmath1 , but a significant feature remains for disks seen face - on .
# 1#2#3#4#4 19#3 phys .
rev .
d , # 1 , # 2 # 1#2#3#4#4 19#3 phys .
lett . , # 1 , # 2 # 1#2#3#4#4 19#3 phys .
rev .
lett . , # 1 , # 2 # 1#2#3#4#4 19#3 phys .
rev . ,
# 1 , # 2 # 1#2#3#4#4 19#3 phys .
rep . , # 1 , # 2 # 1#2#3#4#4 19#3 phys .
fluids , # 1 , # 2 # 1#2#3#4#4 19#3 proc . phys .
soc . , # 1 , # 2 # 1#2#3#4#4 19#3 nucl .
phys . , # 1 , # 2 # 1#2#3#4#4 19#3 mod .
phys .
lett . , # 1 , # 2 # 1#2#3#4#4 19#3 ap .
j. , # 1 , # 2 # 1#2#3#4#4 19#3 astr .
j. , # 1 , # 2 # 1#2#3#4#4 19#3 acta astr . ,
# 1 , # 2 # 1#2#3#4#4 19#3 rev .
mod .
phys . , # 1 , # 2 # 1#2#3#4#4 19#3 nuovo cimento c , # 1 , # 2 # 1#2#3#4#4 19#3 sov .
phys .
jetp , # 1 , # 2 # 1#2#3#4#4 19#3 sov .
ast .
aj , # 1 , # 2 # 1#2#3#4#4 19#3 pub . ast . soc .
japan , # 1 , # 2 # 1#2#3#4#4 19#3 pub .
ast .
soc .
pacific , # 1 , # 2 # 1#2#3#4#4 19#3 ann .
phys .
( ny ) , # 1 , # 2 # 1#2#3#4#4 19#3 yad . fiz .
, # 1 , # 2 # 1#2#3#4#4 19#3 sov .
j. nucl .
phys . , # 1 , # 2 # 1#2#3#4#4 19#3 ast .
ap .
, # 1 , # 2 # 1#2#3#4#4 19#3 ann .
rev .
astr .
ap .
, # 1 , # 2 # 1#2#3#4#4 19#3 m.n.r.a.s . ,
# 1 , # 2 # 1#2#3#4#4 19#3 j. de physique , # 1,#2 # 1#2#3#4#4 19#3 j. quant . spec .
rad .
transfer , # 1,#2 # 1#2#3#4#4 19#3 j.e.t.p .
lett . , # 1,#2 # 1#2#3#4#4 19#3 ap .
j. ( letters ) .
, # 1,#2 # 1#2#3#4#4 19#3 ap . j. ( supp . ) .
, # 1,#2 # 1#2#3#4#4 19#3 ap . lett . , # 1,#2 # 1#2#3#4#4 19#3 ap .
sp . sci .
, # 1,#2 # 1#2#3#4#4 19#3 nature , # 1,#2 # 1#2#3#4#4 19#3 sp .
sci .
rev .
, # 1,#2 # 1#2#3#4#4 19#3 adv .
sp .
res . , # 1,#2 |
the kaon - nucleon ( @xmath4 ) scattering processes have aroused particular interest in the past @xcite and many works have been devoted to this issue based on the constituent quark model .
but up to now , most of them can not accurately reproduce the @xmath4 phase shifts up to the orbit angular momentum @xmath7 in a sufficient consistent way . in this work
we perform a resonating group method ( rgm ) calculation of @xmath0 , @xmath1 , @xmath2 , @xmath3 wave @xmath4 phase shifts of isospin @xmath8 and @xmath9 in the chiral su(3 ) quark model , which has been successful in reproducing the energies of the baryon ground states , the binding energy of deuteron , the nucleon - nucleon ( @xmath10 ) scattering phase shifts , and the hyperon - nucleon ( @xmath11 ) cross sections by the rgm calculations @xcite .
the model is extended to include an antiquark @xmath12 and the mixing of @xmath13 and @xmath14 is considered .
a satisfactory description of the @xmath4 phase shifts for different partial waves is obtained except for the cases of @xmath5 and @xmath6 , of which the calculated phase shifts are too repulsive and a little bit too attractive respectively when the laboratory momentum of the kaon meson is greater than 400 mev in this present investigation .
in the chiral su(3 ) quark model , the potential between the @xmath15th and @xmath16th constituent quarks can be written as @xmath17 where the confinement potential @xmath18 describes the long - range nonperturbative qcd effect and the one - gluon - exchange potential @xmath19 depicts the short - range perturbative qcd effect .
the chiral - field - induced quark - quark potential is in the form of @xmath20 and mainly signifies the medium - range nonperturbative qcd effect . in this expressions ,
@xmath21 are the scalar nonet fields , and @xmath22 the pseudoscalar nonet fields . in order to study the @xmath4 system ,
we extend our model to include an antiquark @xmath12 .
the interaction between @xmath23 and @xmath12 includes two parts : direct interaction and annihilation parts , @xmath24 where @xmath25 @xmath26 now , the total hamiltonian of @xmath4 system is written as @xmath27 where @xmath28 is the kinetic energy operator of the center of mass motion , and the explicit expressions of the potentials can be found in the literature @xcite . in our calculation
, the mixing of @xmath13 and @xmath14 is considered , and the mixing angle @xmath29 is taken to be two possible values .
one is @xmath30 ( ideal mixing ) and the other is @xmath31 ( provided by dai _ et al . _
the model parameters are fixed by some special constraints @xcite and their values are tabulated in table 1 .
a rgm dynamical calculation is made to study the @xmath4 scattering process , and the calculated phase shifts are shown in figs .
[ s0s1 ] and [ pdf ] .
experimental values are taken from the analysis of hyslop _ et al .
_ @xcite and hashimoto @xcite .
-1.6 cm -1.4 cm for the @xmath0-wave we obtain the correct sign of the @xmath32 channel phase shifts comparing with the recent rgm calculation in which @xmath33 and @xmath34 boson exchanges are considered @xcite , and our results are in agreement with the experimental data for both isospin @xmath8 and @xmath9 channels , though for @xmath35 they are a little repulsive . for the higher angular momentum results , comparing with the study of lemaire _
@xcite , we now get correct signs of @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 and @xmath41 waves , and a considerable improvement on the theoretical phase shifts in the magnitude for @xmath42 , @xmath43 and @xmath6 channels .
we also compare our results with those of the previous work of black @xcite .
although our calculation achieves a considerable improvement for all partial waves , the results of @xmath5 wave are too repulsive in both black s work and our present one .
maybe the effects of the coupling to the inelastic channels and hidden color channels should be considered in future work . from figs . 1 and 2 one can see the results are very similar for the cases of @xmath44 and @xmath45 . it is comprehensible because in both of these two cases the attraction of @xmath33 is reduced , just in different approaches . when @xmath44 the interaction between @xmath23 and @xmath46 quarks vanishes , while @xmath45 the attraction of @xmath33 between @xmath47 and @xmath48 quarks is strongly reduced .
the @xmath4 scattering process is studied in the chiral su(3 ) quark model , which has been extended to include an antiquark , by solving a rgm equation .
the numerical results of different partial waves are in agreement with the experimental data except for the cases of @xmath5 and @xmath6 , which are less well described when the laboratory momentum of the kaon meson is greater than 400 mev in the present investigation .
it turns out that our model is successful to be extended to study the @xmath4 system , and some useful information of the quark - quark and quark - antiquark interactions are obtained from this study .
this work was supported in part by the national natural science foundation of china no .
90103020 . 99 r. buettgen
_ et al . _ , _ nucl .
phys . _ * a506 * , 586 ( 1990 ) .
n. black , _ j. phys . _ * g28 * , 1953 ( 2002 ) .
d. hadjimichef , j. haidenbauer , g. krein , _ phys .
rev . _ * c66 * , 055214 ( 2002 ) .
s. lemaire , j. labarsouque , and b. silvestre - brac , _ nucl .
phys . _ * a714 * , 265 ( 2003 ) .
wang , h. yang , and j.c .
su , _ phys .
* c68 * , 055204 ( 2003 ) .
z.y . zhang _
et al . _ , _ nucl .
phys . _ * a625 * , 59 ( 1997 ) .
dai , z.y .
zhang , y.w .
yu , and p. wang , _ nucl .
phys . _ * a727 * , 321 ( 2003 ) . f. huang , z.y .
zhang , y.w .
yu , and b.s .
zou , _ phys .
lett . _ * b586 * , 69 ( 2004 ) . f. huang , z.y .
zhang , and y.w .
yu , nucl - th/0406046 .
f. huang and z.y .
zhang , nucl - th/0409029 .
y.b . dai and y.l .
wu , hep - ph/0304075 .
hyslop , r.a .
arndt , l.d .
roper , and r.l .
workman , _ phys .
rev . _ * d46 * , 961 ( 1992 ) .
k. hashimoto , _ phys .
rev . _ * c29 * , 1377 ( 1984 ) . | the @xmath0 , @xmath1 , @xmath2 , @xmath3 wave @xmath4 phase shifts have been studied in the chiral su(3 ) quark model by solving a resonating group method equation .
the numerical results of different partial waves are in agreement with the experimental data except for the cases of @xmath5 and @xmath6 , which are less well described when the laboratory momentum of the kaon meson is greater than 400 mev . |
the possibility that dark matter ( dm ) , which constitutes most of the matter in the universe , consists of a particle that actually decays on a very long time scale has attracted much attention lately .
this is because the decay time scale @xmath12 can be taken to be ` short ' enough that the decay products give signals in current high energy cosmic ray experiments .
namely , if @xmath13 sec , decaying dm can be invoked to explain the excesses in the fluxes of positrons and electrons measured by , and @xcite . on the other hand , this value of @xmath12 is so much longer than the age of the universe that the slow decay does not make a dent in the overall cosmological dm abundance and does not spoil the agreement with a number of astrophysical and cosmological observations @xcite .
irrespectively of this recent activity spurred by the charged cr anomalies , decaying dm is an interesting subject by itself . a long time studied case is the one of supersymmetric gravitino dm @xcite , which is unstable due to r - parity violation .
more generically , in several particle physics models , high - scale suppressed operators may naturally mediate the decay of dm ( for a miscellaneous list of references see @xcite ) . from the phenomenological point of view
, the main feature of decaying dm with respect to the ` more traditional ' annihilating dm is that it is less constrained by neutral messenger probes ( essentially gamma rays , but also neutrinos ) originating from dense dm concentrations such as the galactic center , the galactic halo or nearby galaxies .
the reason is simple and well - known : while the signal originating from annihilating dm is proportional to the square of the dm density , for decaying dm the dependence is on the first power ; as a consequence , dense dm concentrations shine above the astrophysical backgrounds if annihilation is at play , but remain comparatively dim if dm is decaying .
decaying dm ` wins ' instead , generally speaking , when large volumes are considered .
this is why in the following we will focus on targets as large as galaxy clusters or , essentially , the whole universe . on the observational side
, the and telescopes are making unprecedented progress in the field of gamma - ray astronomy , producing measurements of many different targets including those of interest for decaying dm .
it is therefore a good time to assess the current status of the latter . to this aim
, we will compute the predicted signal from decaying dm , for a variety of decay channels ( limiting ourselves to 2-body channels , in order to remain as model - independent as possible ) , and compare it to the gamma - ray measurements , deriving constraints on the decay half - life .
in particular , in this paper , we make use of two distinct probes : a. [ dataexg ] the isotropic residual gamma ray flux recently measured by @xcite , which now extends from about 200 mev up to 580 gev .
the high - energy portion of this measurement is labelled as ` preliminary ' by the collaboration , so it should be used with care . however , we note that the data are based on solid premises , since they are obtained with the same procedure already published in @xcite with an enlargement of the dataset
. we will anyway show later the effect of considering or not the preliminary portion of the data .
b. [ datafornax ] the recent observation in gamma rays of the fornax galaxy cluster by @xcite .
we choose these probes for a number of reasons , besides the obvious fact that they are among the most recent ones . for what concerns the isotropic flux ( [ dataexg ] ) , it is known since a long time that it represents a powerful testbed for decaying dm ( see for instance @xcite ) . for what concerns ( [ datafornax ] ) ,
we are again motivated by the fact that large virialized objects such as galaxy clusters are a promising target for decaying dm ( as briefly mentioned above ) and , more in particular , by the fact that indeed stringent constraints have been recently derived using data @xcite : on the basis of this we explore the constraining power of a complemetary observatory such as .
this is particularly interesting because upcoming erenkov telescopes offer very promising prospects of improvement in the near future ( as opposed to a space based gamma ray observatory such as , which will at most increase its statistic by a factor of order 2 ) .
indeed , we will also study the sensitivity of , the upcoming large erenkov telescope array @xcite .
incremental improvements to the data and the tools relevant for our analysis also comes from other aspects : a. [ dataepem ] new data on charged cosmic rays ( cr ) , which have been recently presented . @xcite has measured the @xmath14 fraction , confirming the notorious rise exposed by in 2008 @xcite and extending the measured spectrum to larger energies ( about 200 gev ) ; has also presented updated measurements of the total flux of @xmath15 @xcite as well as the measurement of the separated fluxes of @xmath16 and of @xmath17 @xcite ; has published @xcite the measurement of the total flux of @xmath15 too ; finally , has presented results on the flux of pure @xmath17 @xcite , of @xmath18 @xcite and of @xmath19 @xcite .
b. [ ewcorrections ] electroweak ( ew ) corrections to the dm generated fluxes .
as discussed in a number of recent works @xcite , the emission of the ew gauge bosons ( @xmath20 , @xmath21 ) from the sm particles emerging from the dm decay process can significantly modify the phenomenology .
( i ) the corrections are particularly relevant for large dm masses ( above a tev ) ; ( ii ) they can alter significantly the fluxes of charged cosmic rays and gamma rays , both in their spectral shape and in their amplitude , affecting especially the low energies portion @xcite .
we include ew corrections in all our computations . c. [ improvedprop ] an improved propagation scheme for @xmath2 in the galaxy . in semi - analytic treatments of charged cr propagation in the galactic environment , usually employed in dm studies ,
the energy losses of @xmath16 and @xmath17 are generally considered as space - indepedent and their behaviour with energy @xmath22 is approximated as @xmath23 . however
, a space dependence obviously arises from the different conditions ( of magnetic field and of ambient light ) in different points of the galaxy and an energy dependence different from @xmath23 occurs in the full klein - nishina regime of inverse compton scattering .
we make use of the fluxes of @xmath2 as provided in @xcite , which include these proper space- and energy - dependences .
this impacts both the analysis of the @xmath2 fits ( mildly ) and the fluxes of ics gamma rays from the galactic halo ( more significantly ) .
some recent studies have performed work related to ours @xcite . the most closely related , as far as the isotropic gamma ray flux is concerned , is @xcite , in which a subset of us derived constraints on decaying dm with the data and the tools available at that time .
we improve , with respect to that work , by including the updated data in [ dataexg ] and by the points ( [ dataepem ] ) , ( [ ewcorrections ] ) and ( [ improvedprop ] ) discussed above .
there are also some other minor analysis differences on which we will comment in the text . among other works , ref .
@xcite made a particularly detailed analysis : we consider the more recent data in [ dataexg ] and [ datafornax ] above and we include the inverse compton scattering contribution to the gamma ray flux in the galactic component of [ dataexg ] ; the points ( [ dataepem ] ) , ( [ ewcorrections ] ) and ( [ improvedprop ] ) are new in our paper ; we also consider a few more decay channels .
for the cases which overlap with this and other studies , we will present a comparison in sec .
[ comparison ] .
the rest of this paper is organized as follows . in sec .
[ charged fits ] we update charged cr fits by including ( [ dataepem ] ) , ( [ ewcorrections ] ) and ( [ improvedprop ] ) . in sec .
[ isotropic ] we discuss the calculation of the constraints from the isotropic residual @xmath24-ray flux , while in sec .
[ fornax ] we discuss those from the fornax cluster . in sec .
[ results ] we present the combined results . in sec .
[ conclusions ] we present our conclusions .
as mentioned in sec . [ introduction ] , the anomalous , and data in @xmath16 and @xmath15 have been interpreted in terms of dm decay .
we recall here briefly the main features of the experimental data and of their dm interpretations , without entering in the details of any specific particle physics model .
we use the following data sets : * positron fraction @xcite , selecting only points with @xmath25 gev , in order both to avoid the low energy region affected by the uncertainty of solar modulation and to have a consistent overlap with the positron fraction data points ( the low energy data have very small error bars that would overconstrain the fit ) .
* positron fraction @xcite .
* @xmath15 total flux @xcite , provided in the low energy ( le ) and high energy ( he ) samples . * @xmath15 total flux @xcite , also provided in a lower energy portion and a higher energy one . * @xmath15 total flux @xcite , which however consists of only 6 data points , with error bars larger than those of and at the same energy and therefore has no effect on the global fit . * @xmath19 flux @xcite .
we perform the fit to these data using the dm generated fluxes as provided in @xcite , which include the features discussed in points ( [ ewcorrections ] ) and ( [ improvedprop ] ) above . in looking for the best fitting regions , we scan over the propagation parameters of charged cosmic rays and over the uncertainties on the slope and normalization of the astrophysical electron , positron and antiproton background .
note that a simple power law is not expected to necessarily provide a good background model , especially given that the high energy part of the electron spectrum should be dominated by the contribution from local sources . on the other hand , due to the limited number of points available at high - energy and their relatively large statistical error ,
the background power - law flux is significantly constrained by the lower energy data and the assumption is sufficient in providing a good fit , when combined with the bumpy spectrum from dark matter .
we have assumed a nfw profile for the galactic dm halo ( the specific parameters of which are discussed in sec .
[ isotropic ] ) , but other choices would have given almost indistinguishable results .
we refer to @xcite and @xcite for more details .
[ fig : example ] presents the spectra results for one specific case : the best fit point for the decay mode dm@xmath26 .
the allowed regions on the plane @xmath27@xmath12 , for the same channel , are instead shown in fig .
[ fig : exclusionexample ] .
we show the 95.45 % and 99.999 % c.l .
regions for the fit to positron and antiproton data only ( green and yellow bands ) and for the whole datasets , i.e. including ( @xmath28 ) ( red and orange blobs ) . in fig .
[ fig : exclusion ] , which we will discuss in more detail later , we show the global allowed regions for all the channels that we consider .
an example of the signals in different channels from a decaying dm candidate that fits the charged cr anomalies : we show the positron fraction ( upper left ) , all leptons ( upper right ) , antiproton ( lower left ) and isotropic gamma ray ( lower right ) spectra for the best fit candidate dm @xmath29 , namely @xmath30 gev , @xmath31 sec ( marked by a white cross in fig.s [ fig : exclusionexample ] and [ fig : exclusion ] ) .
the ` 2008 ' datapoints are reported for completeness , but we use only the more recent ` 2010 ' in the fits ( see text for details ) . in each panel , the dm contributions are dashed and the astrophysical background is shaded gray . in the lower right panel ,
the dotted ( dashed ) lines refer to the flux neglecting ( including ) extragalactic absorption , see sec .
[ isotropic ] .
_ , title="fig:",scaledwidth=50.0% ] an example of the signals in different channels from a decaying dm candidate that fits the charged cr anomalies : we show the positron fraction ( upper left ) , all leptons ( upper right ) , antiproton ( lower left ) and isotropic gamma ray ( lower right ) spectra for the best fit candidate dm @xmath29 , namely @xmath30 gev , @xmath31 sec ( marked by a white cross in fig.s [ fig : exclusionexample ] and [ fig : exclusion ] ) .
the ` 2008 ' datapoints are reported for completeness , but we use only the more recent ` 2010 ' in the fits ( see text for details ) . in each panel ,
the dm contributions are dashed and the astrophysical background is shaded gray . in the lower right panel ,
the dotted ( dashed ) lines refer to the flux neglecting ( including ) extragalactic absorption , see sec .
[ isotropic ] .
_ , title="fig:",scaledwidth=50.0% ] + an example of the signals in different channels from a decaying dm candidate that fits the charged cr anomalies : we show the positron fraction ( upper left ) , all leptons ( upper right ) , antiproton ( lower left ) and isotropic gamma ray ( lower right ) spectra for the best fit candidate dm @xmath29 , namely @xmath30 gev , @xmath31 sec ( marked by a white cross in fig.s [ fig : exclusionexample ] and [ fig : exclusion ] ) . the ` 2008 ' datapoints are reported for completeness , but we use only the more recent ` 2010 ' in the fits ( see text for details ) . in each panel , the dm contributions are dashed and the astrophysical background is shaded gray . in the lower right panel , the dotted ( dashed ) lines refer to the flux neglecting ( including ) extragalactic absorption , see sec .
[ isotropic ] .
_ , title="fig:",scaledwidth=50.0% ] an example of the signals in different channels from a decaying dm candidate that fits the charged cr anomalies : we show the positron fraction ( upper left ) , all leptons ( upper right ) , antiproton ( lower left ) and isotropic gamma ray ( lower right ) spectra for the best fit candidate dm @xmath29 , namely @xmath30 gev , @xmath31 sec ( marked by a white cross in fig.s [ fig : exclusionexample ] and [ fig : exclusion ] ) .
the ` 2008 ' datapoints are reported for completeness , but we use only the more recent ` 2010 ' in the fits ( see text for details ) . in each panel , the dm contributions are dashed and the astrophysical background is shaded gray . in the lower right panel ,
the dotted ( dashed ) lines refer to the flux neglecting ( including ) extragalactic absorption , see sec .
[ isotropic ] .
_ , title="fig:",scaledwidth=50.0% ] + illustrative example ( for the channel dm@xmath29 ) of the impact on the fit and constrained regions following from different assumptions and choices , as discussed in the text . in fig .
[ fig : exclusion ] , on the other hand , we will show only the final regions and our fiducial constraints .
_ , title="fig:",scaledwidth=50.0% ] illustrative example ( for the channel dm@xmath29 ) of the impact on the fit and constrained regions following from different assumptions and choices , as discussed in the text . in fig .
[ fig : exclusion ] , on the other hand , we will show only the final regions and our fiducial constraints .
_ , title="fig:",scaledwidth=50.0% ] the typical decay time scales that are required for the global fit are of the order of @xmath32 to @xmath33 seconds .
moreover , as well known @xcite , only ` leptophilic ' channels allow a global fit : for the quark and gauge boson channels , the few tev decaying dm needed by ( @xmath28 ) is in conflict with @xmath19 data .
the impact of the new data and of the improved analysis tools on the identification of the best fit dm properties is overall not big .
for instance , in fig .
[ fig : exclusionexample ] we show the dotted contours of the formerly preferred regions ( taken from @xcite ) , which are not very different .
however , some features can readily be identified : * the fit region for the positron and antiproton signals ( yellow and green , see fig .
[ fig : exclusionexample ] ) starts now at larger masses , due to the fact that @xmath16 data reach higher energies than the former ones from . *
an allowed fit region now appears also for the channel dm@xmath34 ( see fig .
[ fig : exclusion ] ) ; it was not present in previous analyses such as @xcite .
this is due to a number of concurring reasons , among which the fact that the inclusion of the ew corrections ( [ ewcorrections ] ) and the refined propagation ( [ improvedprop ] ) smoothens out the @xmath28 spectrum and allow a decent fit to the @xmath35 1 tev hump , the fact that we now remove from the fit the low energy @xmath16 datapoints and , also , that we are now employing for data the full error bars ( slightly larger than before ) .
* ew corrections imply the presence of a non - zero @xmath19 flux even for leptonic decay channels ( see e.g. the third panel of fig .
[ fig : example ] ) . however , the yield is not enough to appreciably affect the fit regions . a comment on the selection of the datasets and their compatibility is also in order .
we use the full set of data listed at the beginning of this section , but we also checked that dropping the data by on the positron fraction , i.e. using only data from and ( on the positron fraction and the pure @xmath16 and @xmath17 fluxes ) , does not significantly change the fit regions .
similarly , the regions are not significantly modified if we adopt for the positron astrophysical background the recent determination in @xcite ( instead of the one conventionally used , from @xcite ) .
this is expected , since the predictions in @xcite and in @xcite are very similar in the range of energies in which we are interested .
the measurements in ( [ dataexg ] ) by the satellite correspond to the ( maximal ) residual , isotropic gamma - ray flux present in their data . its origin can be in a variety of different phenomena , both in the form of unresolved sources and in the form of truly diffuse processes ( see @xcite and reference therein ) .
dm decays can also contribute to this isotropic flux , with two terms : 1 ) an extragalactic cosmological flux , due to the decays at all past redshifts ; 2 ) the residual emission from the dm halo of our galaxy .
the former is of course truly isotropic , at least as long as one neglects possible nearby dm overdensities .
the latter is not , but its minimum constitutes an irreducible contribution to the isotropic flux . in formul , the predicted differential dm flux that we compare with isotropic diffuse @xmath24-ray data is therefore given by @xmath36 for typical dm decay channels and for any dm halo profile , we find that the two contributions are of comparable amplitude , as it can be seen e.g. in the lower right panel of fig .
[ fig : example ] .
we now move to discuss the calculations of the two contributions separately , reproducing and updating the discussion of @xcite .
the extragalactic flux is given , in terms of the earth - measured photon energy @xmath37 , by @xmath38 where @xmath39 is the decay rate . here the hubble function @xmath40 , where @xmath41 is the present hubble expansion rate .
@xmath42 , @xmath43 and @xmath44 are respectively the dark matter , matter and cosmological constant energy density in units of the critical density , @xmath45 .
the gamma ray spectrum @xmath46 , at any redshift @xmath47 , is the sum of two components : ( i ) the high energy contribution due to the prompt @xmath24-ray emission from dm decays and ( ii ) the lower energy contribution due to inverse compton scatterings ( ics ) on cmb photons of the @xmath16 and @xmath17 from those same decays . using eq .
( [ smoothedmap ] ) , the extragalactic flux can therefore be computed in terms of known quantities for any specified dm mass @xmath27 and decay channel . in particular , since the flux essentially consists of the integral over @xmath48 , at the first power , the result is not affected by the formation of dm halos , the history and the properties of which are highly uncertain .
this is in contrast to the case of annihilating dm , where instead these issues affect the predictions by orders of magnitude ( see e.g. @xcite ) .
we take the resulting fluxes from @xcite , but we improve the treatment of the effect of the finite optical depth of the universe , as we now move to discuss . the factor @xmath49 in eq .
( [ smoothedmap ] ) accounts for the absorption of high energy gamma rays due to scattering with the extragalactic uv background light : essentially , a high energy photon from dm hits a lower energy uv photon and produces an energetic electron ( or positron ) which , in turn , makes inverse compton scattering on the cmb and therefore reinjects a continuum of lower energy @xmath24-rays . the process amount to a subtraction of @xmath24-ray flux from the higher energies and a redistribution towards the lower part of the spectrum . as an element of novelty in this analysis
, we take this process fully into account , using the energy and redshift dependent optical depth of the universe ( denoted @xmath50 ) of @xcite .
the effect is sizable at energies @xmath51gev : it can reduce the high energy flux by about one order of magnitude . indeed , in fig .
[ fig : example ] , lower right panel , we show with dotted lines the fluxes computed neglecting absorption and with dashed lines the effect of including it
. summing over the galactic component ( which is of course not affected by absorption ) reduces the impact of the effect , which can however remain sizable
. this will affect the constraints derived below , especially for channels which feature a large prompt contribution .
the flux from the galactic halo , coming from a generic direction @xmath52 , is given by the well known expression @xmath53 \
, \frac{d n}{d e_\gamma}\ , , \label{fluxdec}\ ] ] i.e. as the integral of the decaying dm density piling up along the line of sight individuated by the direction @xmath52 . here
@xmath54 is the dm distribution in the milky way , for which we will always take a standard navarro - frenk - white @xcite profile @xmath55 with parameters @xmath56 kpc and @xmath57 gev/@xmath58 @xcite . in principle
one could consider other choices of profiles and parameters , but , for the case of decaying dm and considering that we will not be interested in the regions of the galactic center ( gc ) where profiles differ most ( quite the opposite : we will focus on the anti - gc ) , these choices make a negligible difference .
the coordinate @xmath59 , centered on the gc , reads @xmath60 , where @xmath61 kpc is the most likely distance of the sun to the gc and @xmath62 is the angle between the direction of observation in the sky and the gc .
the spectrum @xmath46 consists again of two components : the prompt one and the ics one .
we compute both using the tools in @xcite .
in particular , for the ics flux we use the generalized halo functions for the ic radiative process provided there , which take into account the full energy losses discussed in ( [ improvedprop ] ) .
this is another refinement with respect to @xcite , which had computed the @xmath2 energy losses towards the anti - gc including cmb only : it was thus missing synchrotron losses ( which can account for up to roughly a third of the total losses , in the range 8.33 kpc @xmath63 15 kpc ) and ics losses on local ambient light other than the cmb ( which are instead less important ) .
the treatment that we now adopt represents the most refined one we can afford , at the current state - of - the - art of these semi - analytical computations . as indicated in eq .
( [ eq : main ] ) and in the discussion above it , we need to determine the minimum of the flux in eq .
( [ fluxdec ] ) . for the prompt contribution ,
the minimum is obviously located where the line - of - sight density of dm is the lowest , just because the @xmath24-ray emission density traces the dm density by definition .
this coincides with the direction of the anti galactic center ( anti - gc ) .
for what concerns the ics contribution , on the other hand , the situation is more complicated . in this case
the source of @xmath24-rays is the population of dm - originated @xmath2 , and therefore the minimum depends on what is assumed for the propagation and final distribution of the latter ones in the galactic halo .
for instance , assuming a thin diffusive halo with @xmath64 kpc ( see e.g. @xcite for a discussion and an overview of standard values ) implies that the layer of ics emitting @xmath2 is particularly shallow in the zenith ( and nadir ) direction and therefore the ics @xmath24-ray flux is minimal along lines of sight pointing directly above ( and below ) the solar system . on the other hand , for a thick @xmath65 kpc diffusive halo , the ics emission from the directions orthogonal to the galactic plane is greatly enhanced and the minimum is located again at the anti - gc . in view of this complicated morphology , we choose to locate the minimum always at the anti - gc : @xmath66 this is supported , along the lines of the discussion above , by a number of considerations , among which : ( i ) thin diffusive halos are anyhow disfavored by other arguments ( see e.g. @xcite ) ; ( ii ) a number of other uncertain astrophysical variables would enter in a detailed determination of a true minimum : for instance if the _ radial _ size of the diffusive halo is smaller than usually assumed ( @xmath3520 kpc ) , then the ics flux at the anti - gc is reduced ; ( iii ) this discussion only affects the galactic ics contribution ( not the galactic prompt and the extragalactic contributions ) , thus limiting the impact on the derived total flux and having an importance ( for the bounds we will derive below ) only for those configurations in which the ics saturates the constraints .
we thus compute @xmath67 as discussed above and we compare it with the data of ( [ dataexg ] ) .
we show an example of such a comparison in fig .
[ fig : example ] , lower right panel .
it is pretty clear ( even by inspection , in the case of this example ) that the dm signal does not agree in shape with the data , which are instead well fit by a simple power law @xcite .
for this reason , we are driven to derive constraints on the maximum dm signal , and therefore the minimum @xmath12 , admitted by the data .
there are however several possible ways to compute such constraints .
we discuss a few of them in the following and we illustrate them in the specific example of fig .
[ fig : exclusionexample ] .
* _ dm signal only . _
the most conservative option consists in demanding that the dm signal alone ( eq . ( [ eq : main ] ) ) does not exceed any single one of the data points by more than a chosen significance , which we set at 3@xmath68 .
this is overly conservative for a number of reasons .
first of all , since it supposes a vanishingly small astrophysical background . a simple inspection of the dm prediction proves that this assumption is physically untenable ; the bumpy shape of the signal is so different from the featurless observations that the ( bulk of the ) latter must be explained by astrophysical processes , leaving only a subleading role for signals from dm .
second , given the smooth nature of the dm signal and of the data , it is clear that an excess of 3@xmath68 in one point is often accompanyed by similar excesses in neighboring points , so that the ` global significance ' of the exclusion is actually higher .
for instance , the best - fit model example in fig .
[ fig : example ] is only barely excluded by such a procedure , despite the clear tension with several data points . *
_ dm signal + power - law background . _ a more realistic option consists in assuming that the astrophysical background consists of a power - law and demand that the sum of astrophysical background and dm signal does not exceed a chosen global significance .
note that one expects several potential sources to contribute to this flux , like unresolved blazars , star - forming galaxies or electromagnetic cascades from ultra - high energy cosmic ray losses .
in general , a combination of them is not expected to produce ( a priori ) an exact power - law , which may be inaccurate also for some of the different contributions taken alone .
nonetheless , within the current precision we find that the data of @xmath69 alone are very well described by a power law with index -0.41 ( very similar to the one found in resolved blazars ) and normalization 1.02 @xmath70 gev/(@xmath71 sec sr ) .
next , for a given @xmath27 we add the dm signal , whose normalization is controlled by @xmath72 .
we let the normalization of the power law background vary within a factor of 2 from the central value specified above and its index within 0 and -1 ( this choice is astrophysically plausible , although varying the parameters in a broader range would not change the results shown ) .
we then compute the @xmath73 to the data , marginalizing over these parameters .
we compute in this way 95% c.l .
limits on @xmath72 . *
_ dm signal + unconstrained background . _ a variation of the procedure described in the previous point consists in avoiding to commit on the functional form of the background and assume that it consists of an arbitrary function which describes well the data ( namely : it is consistent with each data point within a small significance , e.g. 1@xmath68 ) .
we then demand that , adding the signal , the @xmath74 does not exceed a given value , which in this case we fix very conservatively at 25 . for the practical case at hand , however , this procedure and the previous one yield similar results , as one can see on the example in fig . [ fig : exclusionexample ] .
this is not surprising , since the isotropic residual datapoints are indeed very well described by a power law .
the procedure _ ` dm signal + power - law background ' _ is the one that we will adopt to obtain the fiducial constraints shown in our results ( see fig .
[ fig : exclusion ] ) .
this procedure is ` fiducial ' also in the sense that it matches the analysis we do for charged cr anomalies ( see sec . [
charged fits ] ) and therefore fit regions and constraints are essentially consistent with each other in fig .
[ fig : exclusion ] . before moving on ,
we discuss how the constraints would be modified if we removed from the analysis the datapoints above 100 gev , that are labelled as ` preliminary ' by the collaboration , i.e. if we limited the data to those published during 2010 in @xcite only . in fig .
[ fig : exclusionexample ] the dotted gray line corresponds to how the fiducial constraints would be modified .
we see that , as expected , the limits become looser by a factor of a few .
galaxy clusters are the largest gravitationally bound structures in the universe , 80% of their total mass being in the form of dark matter . although they are located at much larger distances than other popular targets such as dwarf satellites of the milky way , they turn out to be attractive environments to search for dm due to their predicted high dm luminosity . in practice ,
which cluster is the most promising depends on a number of factors , and most notably the astrophysical background . indeed , standard astrophysical gamma - ray emission is expected , both from high energy cosmic rays interacting in the intra cluster medium ( the dominant contribution ) or from electrons in the ambient radiation field ( a subdominant contribution , see @xcite for a review ) . despite these predictions ,
no gamma - ray from these processes has been detected so far @xcite .
radio galaxies lying at the gravity center of clusters such as m87 in virgo and ngc1275 in perseus are also @xmath24-ray emitters : these have actually been directly observed .
a few galaxy clusters have been observed by : the most attractive of them for dm searches are virgo , coma and fornax . although virgo is closeby , the high energy gamma - ray emission from m87 prevents searches in the inner region .
coma is among the most massive galaxy clusters but it has been shown that it is not a privileged environment because of a relatively high cr - induced gamma - ray emission @xcite .
fornax is located at 19 mpc @xcite and shows favorably - low expected astrophysical background ( see , for instance , @xcite ) .
given its location near the tropic of capricorn , the instrument is best - suited to observe fornax with respect to other currently - operating imaging atmospheric erenkov telescopes ( iacts ) . the predicted dm @xmath24-ray flux from fornax can be easily obtained by integrating eq.([fluxdec ] ) over the observational solid angle @xmath75 , and of course replacing @xmath54 with @xmath76 , the dm distribution in the cluster .
there are a few different methods to determine @xmath76 ( and in general the dm content of galaxy clusters ) .
a widely used approach is based on x - ray measurements on the gravitationally bound hot intracluster gas . assuming a nfw profile @xcite , the latter is entirely defined from the virial mass extracted form the hiflugcs catalog catalog @xcite and the virial mass and the concentration relationship found in ref .
@xcite , in @xmath77cdm cosmology .
other choices for the profile are possible , e.g. the cored burkert profile @xcite , which may be particularly motivated by the fact that baryon physics in the inner part of clusters may significantly alter the predictions of @xmath77cdm simulations : several processes invoking baryons in galaxy formation such as dynamical friction , agn feedback or gas outflows , may flatten the dm cusps in dm cores @xcite .
for the case of fornax , various dark matter halo models have been considered in @xcite ( and references therein ) . however , as opposed to the annihilation case , the gamma - ray flux from decaying dm is less dependent on the dm distribution inside the target due to the simple linear dependence of the astrophysical factor with dm density .
varying over the profiles discussed in @xcite , the difference in the flux factor is less than a factor of 3 for a given opening integration angle .
another approach to determine the dm content of clusters consists in using dynamical tracers ( see again @xcite and references therein ) . in the following , we take as ` fiducial ' the x - ray based determination of the dm profile , hereafter referred to as the rb02 profile @xcite , which also well agrees with tracer dynamics at large distances . in this case , the dm distribution is inferred from a x - ray concentration - virial mass relationship is usually defined as the enclosed mass within a radius where the density reaches @xmath78 , with @xmath79 the critical density of the universe and @xmath80 taking values between 100 and 500 . although @xmath81 can be used to derive specific galaxy cluster properties , the x - ray concentration - mass relation is particularly well verified for @xmath80 = 200 @xcite . ] and the uncertainty on the mass determination is about 10% @xcite . as an alternative choice , we also consider the dw01 profile of @xcite , which is fully based on the dynamical tracers method .
this generates a predicted dm @xmath24-ray flux @xmath353 times smaller than rb02 .
next , we discuss the choice of @xmath82 .
in contrast to annihilating dark matter for which most often the smallest opening angle provide the most sensitive searches , at least for cuspy profiles , decaying dark matter searches require optimization of the opening angle to guarantee the highest signal - to - noise ratio . as it is straightforward from eq.([fluxdec ] ) , the luminosity scales with the size of the solid integration angle . on the other hand , background is increasing as well .
we find that the optimization of the signal - to - noise ratio versus the opening integration angle for the dark matter halo profile rb02 implies the best integration region to be 0.5@xmath83 , which corresponds to a solid angle of @xmath84 sr .
we also mention that an additional component to the decaying dark matter spectrum needs in principle to be considered , particularly for leptonic final states : that of inverse compton emission ( indeed , as opposed to dwarf galaxies , in clusters the electrons lose energy primarily through ics on the ambient radiation field and produce additional gamma rays in the final state ) .
we include it in the computations for the dm @xmath29 channel , but , due to the energy working range of , this component becomes important only for very large dm masses ( above @xmath8530 tev ) and its effect is therefore not visible in our results . in order to extract exclusion limits on the dark matter lifetime from gamma - ray astronomical observation with iacts
, the background needs to be determined to constrain the decaying dm luminosity .
the background is calculated in a region referred to as the off region , and the signal region as the on region . both on and off
regions depend on the observation mode and are specific to the iact instrument .
as for the background level , it is taken simultaneously , _
i.e. _ in the same data - taking observing conditions , to the signal events in order to allow for the most accurate estimate . in the ` standard ' on - off method
, one has to take into account the background flux in the astrophysical factor and subsequently in the upper limit calculation .
for the on region of 0.5@xmath83 and considering a off region in an annulus around the on region with an inner and outer radii of 0.5@xmath83 and 1.0@xmath83 respectively , this reduces the decaying dm flux in the on region of about 40% .
an ` improved ' background estimate procedure called the _ template _ method allows to avoid this reduction . in such method
the background events are determined in the on region but from selecting hadron - like events ( see @xcite for details on the analysis procedure ) .
we indeed make use of it in this work .
has observed fornax for a total of 14.5 hours at low zenith angle to allow for best sensitivity to low dark matter masses .
figure [ fig : exclusion ] presents the exclusion plots and therefore summarizes our main results .
one can see that the constraints from the isotropic @xmath24-ray data exclude decaying dm with a lifetime shorter than @xmath32 to few @xmath86 seconds , depending on its mass and the precise channel . therefore , in particular , they rule out the charged cr fit regions , for all the channels . as illustrated in the example in fig .
[ fig : exclusionexample ] , and inspecting fig .
[ fig : example ] , adopting the more conservative constraint procedure may marginally reallow a portion of the fit regions , for the dm @xmath29 , but leaving a clear tension . on the other hand , removing the ` preliminary ' data of @xcite and keeping only those published in 2010 in @xcite still allows to exclude the cr fit regions ( as illustrated in fig . [ fig : exclusionexample ] for the @xmath87 ; the other channels are less critical and remain safely excluded ) . the constraints from rise gently as a function of the mass , essentially as a consequence of the fact that the measured flux rapidly decreases with energy .
they also depend ( mildly , a factor of a few at most ) on the decay channel , as a consequence of the different @xmath24-ray yield and the different shape of the dm signal in each channel .
the constraints from fornax remain subdominant , roughly one order of magnitude below the ones .
however , for the case of the dm @xmath88 channel , the bound also reaches the cr fit region and essentially confirms the exclusion .
the constraints are cut at low masses by the dying out of the sensitivity in to low energy photons .
they do not look competitive with respect to even for larger masses .
the regions on the parameter space @xmath27@xmath12 that are excluded by the and constraints and that can be explored by , together with the regions of the global fit to the charged cr data , for different decay channels._,title="fig:",scaledwidth=45.0% ] the regions on the parameter space @xmath27@xmath12 that are excluded by the and constraints and that can be explored by , together with the regions of the global fit to the charged cr data , for different decay channels._,title="fig:",scaledwidth=45.0% ] + the regions on the parameter space @xmath27@xmath12 that are excluded by the and constraints and that can be explored by , together with the regions of the global fit to the charged cr data , for different decay channels._,title="fig:",scaledwidth=45.0% ] the regions on the parameter space @xmath27@xmath12 that are excluded by the and constraints and that can be explored by , together with the regions of the global fit to the charged cr data , for different decay channels._,title="fig:",scaledwidth=45.0% ] + the regions on the parameter space @xmath27@xmath12 that are excluded by the and constraints and that can be explored by , together with the regions of the global fit to the charged cr data , for different decay channels._,title="fig:",scaledwidth=45.0% ] the regions on the parameter space @xmath27@xmath12 that are excluded by the and constraints and that can be explored by , together with the regions of the global fit to the charged cr data , for different decay channels._,title="fig:",scaledwidth=45.0% ] in this study we have focussed on 2-body particle - antiparticle decay modes ( dm @xmath89 , where @xmath90 is a charged lepton ) , typical of scalar dm .
we do not address 3-body decays such as dm @xmath91 , since those are model dependent .
for fermionic dm , decay channels such as dm @xmath92 are possible : these ( in first approximation ) can be analysed in our framework with a trivial combination of the dm @xmath93 and dm @xmath94 channels .
another possibility which has been recently considered is mixed ( ` democratic ' ) leptonic channels , such as dm @xmath95 33% @xmath96 + 33% @xmath87 + 33% @xmath97 : these can of course be approximately derived on the basis of our individual bounds .
we here comment on the relative strength of the constraints in fig .
[ fig : exclusion ] with respect to bounds from other analyses or other targets .
with respect to the isotropic @xmath24-ray constraints of @xcite , the bounds derived here are stronger by a factor of 2 to 3 .
the reasons of this are discussed at length in the text and essentially amount to : updated datasets , more refined dm analyses and the adoption of a more realistic constraint procedure ( some of these effects pull towards a weakening while others for a strengthening )
. with respect to the work in @xcite , our constraints from the isotropic background are somewhat stronger than their corresponding ones ( up to about a factor of 5 for the dm @xmath26 channel ) , likely due to the several differences in the analysis discussed in sec .
[ introduction ] .
in addition , @xcite presents bounds from the observation of the the fornax cluster by : these are less stringent than our isotropic background constraints but more powerful than our -based constraints at moderate masses . at the largest masses ,
our -based constraints pick up and match theirs ( for the @xmath98 channel ) , as expected from the different operational energy range of versus . the bounds from clusters other than fornax are less powerful , according to the analysis in @xcite . ref .
@xcite has derived constraints from several clusters using data too , but their procedure is questioned in @xcite , and in any case it is now superseded . on the other hand ,
the preliminary constraints shown ( for the @xmath99 channel only ) in @xcite , obtained with a combination of several clusters in , exceed our bounds by a factor of 2 .
bounds from probes other than the isotropic flux and clusters do not generally achieve the same constraining power .
e.g. recently @xcite finds less stringent limits for the leptophilic channels , focussing on the milky way halo . within the context of observations performed by iact
, we note that the decay lifetime constraints obtained with galaxy clusters are stronger than those from dwarf galaxies . even for the ultra - faint dwark galaxy segue 1 , which is believed to be the most promising dwarf in the northern hemisphere ,
the constraints are 2 orders of magnitude higher for a dark matter particle mass of 1 tev @xcite .
we also estimate that the constraints that we derive are stronger than those that can come from neutrino observation of the galactic center ( see for instance @xcite ) . precisely computing those ones for all channels and in the same analysis framework of this study , however , is beyond the scope of this work , and would also probably benefit from a better knowledge of the data ( especially the reconstructed neutrino energy ) , which is not currently available . in summary
: with the possible exception of the preliminary bounds from a combination of clusters by for the @xmath98 channel , the constraints that we derive from the isotropic @xmath24-ray flux are the most stringent to date .
as we have seen , currently the constraints on decaying dm from the satellite are dominant with respect to those from .
however , while the former may increase its statistics by at most a factor of a few , for the latter there are prospects of developments in the mid - term future
. the next - generation iact will be a large array composed of a few tens to a hundred telescopes located on two sites , one in each hemisphere @xcite .
the goal is to improve the overall performances of the present generation : one order - of - magnitude increase in sensitivity and enlarge the accessible energy range both towards the lower and higher energies allowing for an energy threshold down to a few tens of gev . from the actual design study of , the effective area will be increased by a factor 10 and a factor 2 better in the hadron rejection is expected .
the calculation of sensitivity is performed with a background - only hypothesis , where the non - detection of a signal by would imply that the signal in the on region consists of only misidentified hadron showers .
the background level is estimated by integrating eq .
( 4 ) of @xcite after multiplying it by the effective area .
the upper limit on the number of gamma - ray events is calculated at 95% c.l . according to the method from @xcite assuming five off regions .
this calculation follows the methodology described in @xcite , from which the effective area is extracted .
the 95% c.l .
sensitivity on the decay lifetime is then given by @xmath100 } \times \frac{m_{\rm dm } \
, n_{\gamma}^{\rm 95\% c.l.}}{t_{\rm obs}\ , \int_{0}^{m_{\rm dm}/2 } a_{\rm cta}(e_{\gamma } ) \
, \frac{dn}{d e_{\gamma}}(e_\gamma ) \ , d e_{\gamma } } \ , , \ ] ] where @xmath101 is the limit on the number of gamma - ray events , @xmath102 is the cta effective area and @xmath103 the observation time .
figure [ fig : exclusion ] shows the 95% c.l .
sensitivity of on the decay lifetime for the rb02 halo profile for 50h observation time and @xmath104 sr . to conclude
, we also mention that a technique which could allow for significant improvements in the exploration of the parameter space of decaying dm using clusters is the one of stacking the observation of a large number of different clusters , as recently discussed in @xcite .
the authors find that improvements of up to 100 can be theoretically achieved , albeit this factor is @xmath355 for more realistic background - limited instruments .
decaying dark matter has come to the front stage recently as an explanation , alternative to annihilating dm , for the anomalies in cr cosmic rays in , and .
but , more generally , decaying dm is a viable possibility that is or can naturally be embedded in many dm models .
it is therefore interesting to explore its parameter space in the light of the recent observational results .
we discussed the constraints which originate from the measurement of the isotropic @xmath24-ray background by and of the fornax cluster by , for a number of decaying channels and over a range of dm masses from 100 gev to 30 tev .
we improved the analysis over previous work by using more recent data and updated computational tools .
we found ( see fig .
[ fig : exclusion ] ) that the constraints by rule out decaying half - lives of the order of @xmath32 to few @xmath33 seconds .
these therefore exclude the decaying dm interpretation of the charged cr anomalies , ( at least ) for all 2-body channels , at least adopting our fiducial constraint procedure .
the constraints by are generally subdominant . for the dm @xmath105 channel
, they can however also probe the cr fit regions and essentially confirm the exclusion .
we thank daniela borla tridon for communications on data , carlos de los heros for communications on data and alejandro ibarra for useful discussions .
this work is supported by the french national research agency anr under contract anr 2010 blanc 041301 and by the eu itn network unilhc .
ps thanks cern s theory division for hospitality during the development of this work .
chen and f. takahashi , jcap * 0902 * ( 2009 ) 004 [ arxiv : [ hep - ph ] ] .
yin , q. yuan , j. liu , j. zhang , x .-
j . bi and s .- h .
zhu , phys . rev .
d * 79 * ( 2009 ) 023512 [ arxiv : [ hep - ph ] ] .
k. ishiwata , s. matsumoto and t. moroi , phys .
b * 675 * ( 2009 ) 446 [ arxiv : [ hep - ph ] ] .
a. ibarra , d. tran , jcap * 0902 * ( 2009 ) 021 .
[ arxiv : [ hep - ph ] ] .
f. ferrer , c. nipoti and s. ettori , phys .
d * 80 * ( 2009 ) 061303 [ arxiv : [ astro-ph.co ] ]
. w. buchmuller , l. covi , k. hamaguchi , a. ibarra and t. yanagida , jhep * 0703 * ( 2007 ) 037 [ [ hep - ph ] ] .
d. eichler , phys .
* 63 * ( 1989 ) 2440 .
a. arvanitaki , s. dimopoulos , s. dubovsky , p. w. graham , r. harnik , s. rajendran , phys .
* d79 * ( 2009 ) 105022 [ arxiv : [ hep - ph ] ] . c. arina , t. hambye , a. ibarra and c. weniger , jcap * 1003 * ( 2010 ) 024 [ arxiv : [ hep - ph ] ] .
talk by m. ackermann at the http://agenda.albanova.se/conferencedisplay.py?confid=2600[tevpa 2011 conference ] ( 1 - 5 aug 2011 , stockholm , sweden ) .
a. a. abdo _ et al .
_ [ fermi - lat collaboration ] , phys . rev .
lett . *
104 * ( 2010 ) 101101 [ arxiv : [ astro-ph.he ] ] .
a. abramowski _ et al .
_ [ hess collaboration ] , astrophys . j. * 750 * ( 2012 ) 123 [ arxiv : [ astro-ph.he ] ] .
f. takayama and m. yamaguchi , phys .
b * 485 * ( 2000 ) 388 [ ] .
j. m. overduin and p. s. wesson , phys .
* 402 * ( 2004 ) 267 [ ] .
g. bertone , w. buchmuller , l. covi and a. ibarra , jcap * 0711 * ( 2007 ) 003 [ arxiv : [ astro - ph ] ] . by s.
zimmer and j. conrad at the .
s. zimmer , j. conrad , for the fermi - lat collaboration and a. pinzke , arxiv : [ astro-ph.he ] .
x. huang , g. vertongen and c. weniger , jcap * 1201 * ( 2012 ) 042 [ arxiv : [ hep - ph ] ] .
see http://www.cta-observatory.org[www.cta-observatory.org ] .
m. ackermann _ et al .
_ [ the fermi lat collaboration ] , phys . rev .
* 108 * ( 2012 ) 011103 [ arxiv : [ astro-ph.he ] ] .
o. adriani _ et al . _
[ pamela collaboration ] , nature 458 , 607 - 609 , 2009 , arxiv : .
see also o. adriani , g. c. barbarino , g. a. bazilevskaya , r. bellotti , m. boezio , e. a. bogomolov , l. bonechi and m. bongi _ et al .
_ , astropart .
* 34 * ( 2010 ) 1 [ arxiv : [ astro-ph.he ] ] .
m. ackermann _ et al .
_ [ fermi lat collaboration ] , phys .
d * 82 * ( 2010 ) 092004 [ arxiv : [ astro-ph.he ] ] .
d. b. tridon , p. colin , l. cossio , m. doro , v. scalzotto and f. t. m. collaboration , arxiv : [ astro-ph.he ] .
o. adriani _ et al . _
[ pamela collaboration ] , phys .
* 106 * ( 2011 ) 201101 [ arxiv : [ astro-ph.he ] ] .
o. adriani _ et al . _
[ pamela collaboration ] , science * 332 * ( 2011 ) 69 [ arxiv : [ astro-ph.he ] ] .
o. adriani _ et al . _
[ pamela collaboration ] , phys .
* 105 * ( 2010 ) 121101 [ arxiv : [ astro-ph.he ] ] .
see also : n. f. bell , j. b. dent , t. d. jacques and t. j. weiler , phys .
d * 78 * ( 2008 ) 083540 [ arxiv : [ hep - ph ] ] .
more recent analyses and a more complete appraisal of the phenomenological importance of the effect have been performed in : m. kachelriess , p. d. serpico and m. a. solberg , phys .
d * 80 * ( 2009 ) 123533 [ arxiv : [ hep - ph ] ] ; p. ciafaloni and a. urbano , phys .
d * 82 * ( 2010 ) 043512 [ arxiv : [ hep - ph ] ] .
the antiproton constraints ( especially for leptophilic channels ) have been considered in m. garny , a. ibarra and s. vogl , jcap * 1107 * ( 2011 ) 028 [ arxiv : [ hep - ph ] ] and m. garny , a. ibarra and s. vogl , arxiv : [ hep - ph ] .
the related phenomenon of gluon emissions is also addressed in m. asano , t. bringmann and c. weniger , arxiv : [ hep - ph ] .
p.ciafaloni , d.comelli , a.riotto , f.sala , a.strumia , a.urbano , jcap * 1103 * ( 2011 ) 019 , arxiv : .
m. cirelli , g. corcella , a. hektor , g. hutsi , m. kadastik , p. panci , m. raidal and f. sala _ et al .
_ , jcap * 1103 * ( 2011 ) 051 [ arxiv : [ hep - ph ] ] ; the numerical results are available from http://www.marcocirelli.net/pppc4dmid.html[this website ] . c. -r .
chen , s. k. mandal and f. takahashi , jcap * 1001 * ( 2010 ) 023 [ arxiv : [ hep - ph ] ] .
a. ibarra , d. tran and c. weniger , phys .
d * 81 * ( 2010 ) 023529 [ arxiv : [ hep - ph ] ] .
m. cirelli , p. panci , p. d. serpico , nucl .
. * b840 * ( 2010 ) 284 - 303 .
[ arxiv : [ astro-ph.co ] ] .
j. ke , m. luo , l. wang and g. zhu , phys .
b * 698 * ( 2011 ) 44 [ arxiv : [ hep - ph ] ] .
m. luo , l. wang and g. zhu , arxiv : [ hep - ph ] .
f. aharonian _ et al .
_ [ h.e.s.s .
collaboration ] , phys .
lett . 101 ( 2008 ) 261104 [ arxiv : ] .
f. aharonian et al .
[ h.e.s.s .
collaboration ] , arxiv : [ astro-ph.he ] .
m. cirelli , m. kadastik , m. raidal and a. strumia , nucl .
b * 813 * ( 2009 ) 1 [ arxiv : [ hep - ph ] ] .
j. lavalle , mon . not .
* 414 * ( 2011 ) 985l [ arxiv : [ astro-ph.he ] ] .
i. v. moskalenko and a. w. strong , astrophys .
j. * 493 * ( 1998 ) 694 [ ] .
e. a. baltz and j. edsjo , phys .
d * 59 * ( 1998 ) 023511 [ ] .
a. a. abdo _ et al .
_ [ fermi - lat collaboration ] , jcap * 1004 * ( 2010 ) 014 [ arxiv : [ astro-ph.co ] ] .
p. d. serpico , e. sefusatti , m. gustafsson and g. zaharijas , mon . not .
soc . * 421 * , 87l ( 2012 ) arxiv : [ astro-ph.co ] . j. f. navarro , c. s. frenk and s. d. m. white , astrophys .
j. * 462 * ( 1996 ) 563 [ arxiv : ] .
t. bringmann , f. donato and r. a. lineros , jcap * 1201 * ( 2012 ) 049 [ arxiv : [ astro-ph.ga ] ] .
p. blasi , s. gabici and g. brunetti , int .
j. mod .
a * 22 * ( 2007 ) 681 [ ] .
f. aharonian [ hess collaboration ] , a&a , * 495 * , 27 , arxiv : [ astro - ph ] . f. a. aharonian [ the hess collaboration ] , a&a , * 502 * , 437 arxiv : [ astro-ph.co ] . j. aleksic _ et al . _
[ the magic collaboration ] , astrophys .
j. * 710 * ( 2010 ) 634 [ arxiv : [ astro-ph.he ] ] . m. ackermann , m. ajello , a. allafort , l. baldini , j. ballet , g. barbiellini , d. bastieri and k. bechtol _ et al .
_ , jcap * 1005 * ( 2010 ) 025 [ arxiv : [ astro-ph.co ] ] .
m. ackermann _ et al .
_ [ fermi - lat collaboration ] , astrophys . j. , * 717 * , l71 arxiv : [ astro-ph.he ]
. t. h. reiprich and h. boehringer , astrophys .
j. * 567 * ( 2002 ) 716 [ ] .
t. e. jeltema , j. kehayias and s. profumo , phys .
d * 80 * ( 2009 ) 023005 [ [ astro - ph ] ] .
a. pinzke , c. pfrommer and l. bergstrom , phys .
* 103 * ( 2009 ) 181302 [ arxiv : [ astro-ph.he ] ] .
a. pinzke , c. pfrommer and l. bergstrom , phys .
d * 84 * ( 2011 ) 123509 [ arxiv : [ astro-ph.he ] ] .
d. a. buote , f. gastaldello , p. j. humphrey , l. zappacosta , j. s. bullock , f. brighenti and w. g. mathews , astrophys .
j. * 664 * ( 2007 ) 123 [ ] .
a. burkert , iau symp .
* 171 * ( 1996 ) 175 [ astrophys .
j. * 447 * ( 1995 ) l25 ] [ ] .
see also : p. salucci and a. burkert , arxiv : .
g. gentile , p. salucci , u. klein , d. vergani and p. kalberla , mon . not .
soc . 351 ( 2004 ) 903
[ arxiv : ] and p. salucci , a. lapi , c. tonini , g. gentile , i. yegorova and u. klein , mon . not .
378 ( 2007 ) 41 [ arxiv : ] .
a. el - zant , i. shlosman and y. hoffman , astrophys .
j. , * 560 * , 636 , .
d. martizzi , r. teyssier , b. moore and t. wentz , arxiv : [ astro-ph.co ] .
g. m. voit , rev .
* 77 * ( 2005 ) 207 [ ] .
l. dugger , t. e. jeltema and s. profumo , jcap * 1012 * ( 2010 ) 015 [ arxiv : [ astro-ph.he ] ] . r. abbasi
_ et al . _
[ icecube collaboration ] , phys .
d * 84 * ( 2011 ) 022004 [ arxiv : [ astro-ph.he ] ] . the cta consortium ( m. actis et al . ) , exper .
astron .
* 32 * ( 2011 ) 193 [ arxiv : [ astro-ph.im ] ] .
a. viana , c. m. medina , j. penarrubia , p. brun , j. f. glicenstein , k. kosack , e. moulin and m. naumann - godo _ et al .
_ , astrophys . j. * 746 * ( 2012 ) 77 [ arxiv : [ astro-ph.he ] ] .
w. a. rolke , a. m. lopez and j. conrad , nucl .
instrum .
a * 551 * ( 2005 ) 493 [ physics/0403059 ] . c. combet , d. maurin , e. nezri , e. pointecouteau , j. a. hinton and r. white , phys .
d * 85 * , 063517 ( 2012 ) [ arxiv : [ astro-ph.co ] ] . | we derive new bounds on decaying dark matter from the gamma ray measurements of ( i ) the isotropic residual ( extragalactic ) background by and ( ii ) the fornax galaxy cluster by .
we find that those from ( i ) are among the most stringent constraints currently available , for a large range of dm masses and a variety of decay modes , excluding half - lives up to @xmath0 to few @xmath1 seconds .
in particular , they rule out the interpretation in terms of decaying dm of the @xmath2 spectral features in , and , unless very conservative choices are adopted .
we also discuss future prospects for bounds from fornax which , contrary to the present constraints of ( ii ) , may allow for an interesting improvement and may become better than those from the current or future extragalactic data . *
* * marco cirelli*@xmath3 , * emmanuel moulin*@xmath4 , * paolo panci*@xmath5 , + * pasquale d. serpico*@xmath6 , * aion viana*@xmath4 _ @xmath7http://ph - dep - th.web.cern.ch / ph - dep - th/[cern theory division ] , ch-1211 genve , switzerland _ + _ @xmath8http://ipht.cea.fr / en / index.php[institut de physique thorique ] , cnrs , ura 2306 & cea / saclay , f-91191 gif - sur - yvette , france _
+ _ @xmath9http://irfu.cea.fr / en/[irfu / dsm ] , cea saclay , f-91191 gif - sur - yvette , france _ + _ @xmath10http://cp3-origins.dk/[cp3-origins ] & the danish institute for advanced study dias , + university of southern denmark , campusvej 55 , dk-5230 odense m , denmark _
+ _ @xmath11http://lapth.in2p3.fr/[lapth ] , univ .
de savoie , cnrs , b.p.110 , annecy - le - vieux f-74941 , france _ |
the _ total curvature _ of a @xmath2 path @xmath3 parameterized by arc length @xmath4 in @xmath5 , @xmath6 , is defined as @xmath7 .
w. fenchel proved in 1929 for @xmath0 and k. borsuk in 1947 for any @xmath5 that the total curvature of a closed curve is bounded from below by @xmath1 , with the equality holding only for convex simple closed curves in @xmath8 .
the _ total curvature _ of a polygonal path @xmath9 $ ] is defined as @xmath10 let @xmath11 be the set of all compact convex polyhedra in @xmath12 .
let @xmath13 , where @xmath14 is a shortest path joining two points on the boundary of a polyhedron @xmath15 .
it has been asked in @xcite whether the set @xmath16 is bounded .
we shall prove here , in theorem [ rb ] below , that the conjecture holds for polytopes @xmath17 for which the ratio @xmath18 is bounded from above : here @xmath19 and @xmath20 , respectively , are the radii of the circumscribed and inscribed ball to @xmath17 .
we define the _ spiralling number _
@xmath21 for the path @xmath14 from @xmath22 to @xmath23 on the polytopal surface by considering a variable point @xmath24 , writing this point in cylindrical coordinates as @xmath25 where the @xmath26-axis is the line through @xmath22 and @xmath23 , and @xmath27 is a continuous function .
set now @xmath28 this measures how many times the path spirals around the line through @xmath22 and @xmath23 .
the proof method of theorem [ rb ] would work for all polytopes if the function @xmath27 had bounded variation .
this , however , will be shown to be false . in section [ spiral ]
, we will construct needle - like polytopes @xmath17 with shortest path @xmath14 such that @xmath21 is arbitrarily large .
even more surprisingly , the example can be modified so that @xmath14 spirals around the line through @xmath22 and @xmath23 a hundred times in one direction , then two hundred times in the opposite direction , then thousand times in the first direction , etc .
the total curvature @xmath29 of a planar path @xmath14 is bounded sharply by @xmath1 and this bound is the lowest possible .
a triangle with one of the angles very close to @xmath30 and two points on the two sides adjacent to the wide angle but close to the vertices at the acute angles provides a simple example . in section [ total ]
, we construct an example of a shortest path on the boundary of a convex polyhedron in @xmath0 for which the @xmath1 bound does not hold . on some level
the example resembles the planar example involving a triangle with one of the angles close to @xmath30 .
on the boundary of a polytope consider a shortest path @xmath14 with non - smooth points @xmath31 .
put @xmath32 @xmath33 .
let @xmath34 be the outernormal ( of the interior ) of the facet of @xmath35 and @xmath36 , and let @xmath37 be the angle between @xmath38 and @xmath39 @xmath40 .
then the total curvature of @xmath14 is @xmath41 .
this can be easily checked .
we remark that @xmath42 with @xmath43 [ unit ] if there is a unit vector v such that @xmath44 for all i , then @xmath45 first remark that @xmath46 if @xmath47 . then @xmath48 @xmath49 since @xmath50 and @xmath51 , we have @xmath52 hence @xmath53 denote by @xmath54 the closed unit ball in @xmath0 .
[ rb ] let a polytope @xmath55 satisfy @xmath56
. then the total curvature of any shortest path on the boundary of q is less than @xmath57 let @xmath14 be a shortest path between points @xmath22 and @xmath23 on the boundary bd@xmath55 of @xmath55 .
the length of @xmath14 is less than @xmath30 .
indeed any plane through @xmath22 and @xmath23 intersects the boundary of @xmath54 along a circle of length at most @xmath1 , and the boundary of @xmath55 along a polygon of even smaller length .
so one of the two broken lines into which @xmath22 and @xmath23 divide the polygon must have length less than @xmath30 , and @xmath14 can not be longer . for a non - zero vector @xmath58 define the @xmath59-shadow @xmath60 of @xmath55 as @xmath61 assume that @xmath62 is the outernormal at an interior point @xmath63 of a facet @xmath64 of @xmath55 with @xmath65 .
we claim that @xmath66 indeed , the plane @xmath67 of @xmath64 does not meet the interior of @xmath68 .
let @xmath69 be the angle between @xmath62 and @xmath59 .
of course , @xmath70 . the distance from @xmath63 to the line through the origin
@xmath71 and @xmath59 is at most @xmath72 because @xmath73 .
further , the distance from @xmath71 to @xmath67 is at least @xmath20 .
therefore @xmath74 which proves the claim .
we are going to define points @xmath75 on @xmath14 by recursion .
set @xmath76 and assume that @xmath77 has been defined .
let @xmath78 be the first point on @xmath14 , going from @xmath77 to @xmath23 , which is not in the interior of @xmath79 .
write @xmath80 for the part of the path @xmath14 between @xmath77 and @xmath78 . since the length of @xmath80 is at least @xmath81 and
the length of @xmath14 is less than @xmath30 , there are only @xmath82 paths @xmath80 , and we put @xmath83 .
note that lemma [ unit ] applies to @xmath80 with @xmath84 by the above claim .
hence the total curvature of @xmath80 is less than @xmath85 .
summing up , the total curvature of @xmath14 is less than @xmath86 now we formulate theorem [ rb ] differently .
let @xmath87 be the ellipsoid of largest volume included in the polytope @xmath55 .
as the total curvature is invariant under angle preserving linear transformation , we may assume that @xmath87 has half - axes @xmath88 with @xmath89 call @xmath55 _ needle - like _ if @xmath23 is small , and _ pancake - like _ if @xmath22 is small compared with @xmath23 .
theorem [ rb ] on the boundedness of total curvature holds for convex bodies that are not pancake - like .
only moderate effort is needed to prove the following : if the conjecture holds for needle - like convex bodies , then it holds for pancake - like ones as well .
so it would be enough to prove the conjecture for needle - like convex bodies .
this would follow if the total variation of @xmath27 , the function defined in the first paragraph , were bounded .
but this is not true : an example showing this is in section [ spiral ] .
we construct a convex body @xmath91 with two points on the boundary such that the total curvature of the shortest path joining the points exceeds @xmath1 .
@xmath91 is constructed in four steps , with certain unbounded convex bodies @xmath92 , @xmath93 , and @xmath94 described in steps 1 - 3 .
the cartesian coordinates of a point @xmath95 are denoted by @xmath96 . for @xmath97 and @xmath98 ,
define @xmath99 to be the plane @xmath100 .
let @xmath101 be the parabola in @xmath102 and @xmath103 the parabola in @xmath104 .
denote by @xmath77 the common vertex of @xmath101 and @xmath103 , by @xmath105 the focus of @xmath101 , and by @xmath106 the focus of @xmath103 .
thus @xmath107 , @xmath108 , and @xmath109 .
denote by @xmath110 the convex region in the plane @xmath99 bounded by @xmath101 .
put @xmath111 and @xmath112 thus @xmath113 is a quarter - circle joining the vertices @xmath114 and @xmath115 and @xmath116 is a surface obtained by sliding @xmath113 along the segment @xmath117 . note that @xmath118 .
let @xmath119 be the convex hull of the union @xmath120 suppose that @xmath121 is a line in @xmath122 , crossing @xmath123 , and parallel to the directrix of @xmath123 .
denote by @xmath124 the distance between @xmath121 and the directrix of @xmath125 , and by @xmath126 the distance between @xmath121 and the directrix of @xmath123 .
note that if a shortest path @xmath14 in bd@xmath92 joining @xmath121 with @xmath127 passes through @xmath114 and @xmath115 , then the length @xmath128 of @xmath14 equals @xmath129 .
[ path ] the shortest path @xmath14 on the boundary of @xmath92 joining @xmath121 with @xmath127 is unique and passes through the vertices @xmath114 and @xmath115 .
suppose that @xmath130 is a shortest path in bd@xmath92 joining @xmath127 with @xmath121 , different from @xmath14 , crossing @xmath131 at @xmath132 and @xmath123 at @xmath133 , see fig .
[ paradist ] .
note that @xmath134 .
@xmath121 @xmath103 @xmath101 @xmath106 @xmath130@xmath135 @xmath135@xmath136 @xmath137 to show that @xmath138 , consider the two cases : @xmath139 and @xmath140 .
we have 1 . if @xmath139 , then @xmath141 and @xmath142 .
2 . if @xmath140 , then @xmath130 projects onto @xmath14 decreasing its length . [ monotone ] if @xmath143 denotes the shortest path in bd@xmath92 joining @xmath127 with @xmath121 and passing through @xmath144 , then the length @xmath145 is a monotone function of @xmath146 for @xmath147 ( and for @xmath148 ) . in the next step of the construction , we modify @xmath92 to obtain a convex unbounded slab @xmath93 .
let @xmath149 be a convex curve in the @xmath150-plane close to the segment @xmath151 , @xmath152 , and @xmath153 a positively curved surface in @xmath92 , close to @xmath116 , obtained by sliding @xmath113 along @xmath149 , keeping @xmath113 parallel to the @xmath154-plane , and the midpoint of @xmath113 and @xmath149 .
denote by @xmath93 the convex hull @xmath155 .
we require of @xmath149 ( and consequently of @xmath153 ) that 1 .
@xmath153 contains the quarter - circle @xmath113 ; 2 .
the path @xmath14 is the shortest path joining @xmath121 with @xmath127 on the boundary of @xmath93 . by remark [ monotone ] , such an @xmath93 exists .
note that @xmath93 is a thin convex unbounded slab whose top and bottom are ( horizontal ) planar regions @xmath156 s containing @xmath110 s and with positively curved side surface close to the vertices @xmath114 and @xmath115 , see fig .
[ slab ] . [ slab ] for @xmath157 ,
let @xmath94 be a ` slant slab ' obtained from @xmath93 by ` @xmath158-slanting ' the plane @xmath159 at the vertex @xmath114 in the direction perpendicular to the @xmath160-axis as follows : let @xmath161 and let @xmath162 .
note that the portion of @xmath94 above the @xmath150-plane is identical to that of @xmath93 .
[ tilted ] the shortest path @xmath163 joining @xmath121 with @xmath164 on the boundary of @xmath94 passes through the vertices @xmath114 and @xmath115 .
a path @xmath130 joining @xmath121 with @xmath165 in bd@xmath94 can be mapped ( keeping the first coordinate and the distance to the line @xmath166 unchanged ) onto the boundary of @xmath93 to a path joining @xmath121 with @xmath127 .
such a projection does not increase the length of the path .
hence , if @xmath167 , @xmath168 . for @xmath163
as defined in lemma [ tilted ] , the total curvature of @xmath163 , @xmath169 , equals @xmath170 .
@xmath115 @xmath17 @xmath171 denote by @xmath172 the side boundary of @xmath94 , _
i.e. _ , @xmath173 . for a point
@xmath174 , conv@xmath175 denotes the convex hull of @xmath94 and @xmath174 . for a point @xmath176 ,
define an _ attached cone _ with vertex @xmath174 , con@xmath177 , as the closure of conv@xmath178 provided that it does not intersect @xmath179 .
denote by @xmath180 the shortest path in @xmath172 , a quarter - circle , joining @xmath115 with @xmath181 .
let @xmath182 , @xmath183 , and @xmath184 be points close to @xmath180 for which attached cones con@xmath185 are defined and are pairwise disjoint .
note that @xmath186 is convex .
first choose points @xmath182 , @xmath183 , and @xmath184 so that 1 .
each con@xmath185 , @xmath187 , intersects @xmath180 at one point @xmath188 with @xmath189 ; 2 .
con@xmath190 and con@xmath191 are on the same side @xmath180 , and con@xmath192 on the other side of @xmath180 .
then move the points @xmath182 , @xmath183 , and @xmath184 slightly towards @xmath180 to obtain points @xmath193 , @xmath194 , and @xmath195 such that the attached cones @xmath196 intersect @xmath180 and the shortest path @xmath17 joining @xmath121 and @xmath165 in @xmath197 crosses each @xmath198 , but passes through @xmath114 , @xmath115 , and @xmath171 , see fig .
[ threecones ] .
note that @xmath17 is not planar and as it follows from the work of w. fenchel @xcite ( alternatively , see k. borsuk @xcite or j. milnor @xcite ) its total curvature is greater than that of @xmath180 .
we have [ curvature ] the total curvature @xmath199 , where @xmath200 .
finally , take @xmath201 . the line @xmath121 consists of points @xmath202 , where @xmath203 .
let @xmath204 be the part of @xmath205 cut off by the plane @xmath206 .
let @xmath207 be a symmetrical copy of @xmath204 and @xmath208 with @xmath204 and @xmath207 glued along the side @xmath206 ( see fig .
[ doubleslab ] ) .
let @xmath209 and @xmath210 be the path and the point in @xmath207 corresponding to @xmath211 and @xmath165 .
for sufficiently large @xmath212 , the path @xmath213 is the shortest path joining @xmath165 and @xmath210 in @xmath91 .
we have @xmath214 .
clearly @xmath91 is convex .
a polyhedral example can be obtained by a suitable approximation of @xmath91 .
[ delta ] there exist a convex polyhedron @xmath215 with two points @xmath216 and @xmath217 on the boundary of @xmath218 such that the total curvature of the ( unique ) shortest path joining @xmath216 and @xmath217 exceeds @xmath1 .
here we construct polytopal surfaces possessing shortest paths of arbitrarily large spiralling number .
the intrinsic metric on these surfaces will be denoted by @xmath219 .
our example is the boundary of the convex hull of a family of equilateral triangles each two of which have pairwise parallel edges .
suppose the construction performed up to the equilateral triangle @xmath222 , so that , from a fixed point @xmath223 of the ( already constructed ) surface , the intrinsic distance @xmath224 to @xmath225 ( which is @xmath226 ) is smaller than the distance to any other side of @xmath222 .
let @xmath227 be the smaller of the two differences .
we now construct the next triangle @xmath228 .
we do so that @xmath229 and any shortest path from @xmath223 to any point @xmath174 of @xmath230 necessarily crosses @xmath225 .
let @xmath231 be such that conv@xmath232 is a regular pyramid containing the already constructed surface .
then we can find a point @xmath26 behind @xmath231 ( see fig .
[ pyramid ] ) but close to the pyramid axis such that @xmath233 and conv@xmath234 includes conv@xmath232 .
we shall choose @xmath235 and put @xmath236 , @xmath237 and @xmath238 the angles @xmath239 and @xmath240 satisfy @xmath241 if @xmath26 is close enough to the pyramid axis .
consider an unfolding of the surface on the plane of @xmath246 without cutting along @xmath247 , but cutting along @xmath248 , and keep the notation .
let @xmath4 be the orthogonal projection of @xmath242 on the line @xmath113 through @xmath22 and @xmath23 .
if @xmath249 is small enough , which is the case if @xmath26 is far enough , then @xmath250 is as small as desired , in particular smaller than @xmath251 .
let @xmath252 , see fig .
[ unfold ] .
it is easily verified that @xmath253 increases when @xmath174 moves from @xmath245 to @xmath244 .
thus @xmath254 if @xmath249 is small enough and @xmath255 large enough , then @xmath256 since @xmath257 and @xmath258 , @xmath259 hence @xmath260 this ensures that any shortest path from @xmath223 to some point @xmath261 crosses @xmath225 .
indeed , if the above path crosses @xmath262 at @xmath59 , then @xmath263 @xmath264 @xmath265 @xmath62 being a point of @xmath225 closest to @xmath223 , and @xmath266 the orthogonal projection of @xmath62 on @xmath113 ; we got a contradiction .
if the path crosses @xmath267 at @xmath268 , say , then @xmath269 and a contradiction is obtained as above .
now , we have @xmath270 by choosing @xmath255 large enough , the above difference can be made as large as wished ; we make it larger than @xmath227 .
suppose now a ( rectifiable ) path from @xmath223 to @xmath271 crosses @xmath222 at @xmath217
. then its length is at least @xmath272 @xmath273 so it can not be a shortest path from @xmath223 to @xmath228 .
analogously , no path from @xmath223 to @xmath274 is a shortest path from @xmath223 to @xmath228 .
this completes the proof of all desired properties for @xmath14 .
it is clear that , iterating this procedure , the shortest path from @xmath223 to the last constructed triangle has a steadily increasing spiralling number ( by a rate close to 1/3 for each new triangle ) .
thus @xmath21 can be made as large as we wish .
the theorem is proved . in @xcite , jnos pach gives a neat example of a polytope @xmath275 and a shortest path @xmath14 on the boundary of @xmath17 with non - smooth points @xmath276 such that @xmath277 is not bounded . here
@xmath278 is the angle between the outer normals of the facets containing the segments @xmath279 $ ] and @xmath280 $ ] .
we mention that our construction has the same property . | this paper gives a partial confirmation of a conjecture of p. agarwal , s. har - peled , m. sharir , and k. varadarajan that the total curvature of a shortest path on the boundary of a convex polyhedron in @xmath0 can not be arbitrarily large .
it is shown here that the conjecture holds for a class of polytopes for which the ratio of the radii of the circumscribed and inscribed ball is bounded .
on the other hand , an example is constructed to show that the total curvature of a shortest path on the boundary of a convex polyhedron in @xmath0 can exceed @xmath1 .
another example shows that the spiraling number of a shortest path on the boundary of a convex polyhedron can be arbitrarily large . |
one of the important questions to be answered in identifying the microscopic origin of superconductivity in cuprates is the nature of the superconducting gap . a great deal of attention
@xcite has been focused on identifying the symmetry of the pairing state since it is believed that this information will provide constraints that will lead to the identification of the pairing mechanism .
for example , it has been suggested @xcite that pairing in high temperature superconductors ( htsc ) is induced by the exchange of antiferromagnetic spin fluctuations and in this case the pairing state would necessarily have d@xmath10 symmetry . from the many experiments carried out in this context
, a consensus view has emerged @xcite that associates d@xmath10 symmetry with the pairing state in optimally hole - doped cuprates . however , this consensus is lacking ( see ref .
@xcite for a review ) in the results of angle resolved photoemission @xcite , tunneling @xcite , infrared @xcite and raman spectroscopy @xcite experiments carried out on overdoped and underdoped compounds , where it appears that the material properties are very different from those of the optimally doped versions .
since an appropriate model of the normal state must be capable of predicting the effects of doping , it is clear that additional experiments are required if one is to gain a complete understanding of superconductivity in the high - t@xmath9 cuprates .
raman scattering has been widely used @xcite to investigate electronic excitations in superconductors . in electronic raman scattering experiments ,
one may investigate various parts of the fermi surface by a simple choice of the incident and scattered polarization vectors .
the b@xmath5 spectra sample @xcite regions of the fermi surface located near the brillouin zone ( bz ) principal axes [ i.e. @xmath14 , while b@xmath4 spectra sample regions located near the diagonal directions in the bz .
when raman results are interpreted @xcite in terms of the conventional model , @xcite the gap is found to have maxima along the ( @xmath15 ) and ( @xmath16 ) directions and nodes ( or minima ) along the diagonal directions .
the results are consistent with d@xmath10 symmetry and are thus in agreement with the outcome of several other experiments [ see scalapino @xcite for a review ] .
the maximum values of the gap are found to scale with t@xmath9 ( @xmath17 ) between optimally doped bi@xmath0sr@xmath0cacu@xmath0o@xmath1 ( bi2212 ) @xcite , yba@xmath0cu@xmath18o@xmath19 ( y123 ) @xcite , and la@xmath20sr@xmath21cuo@xmath22 ( la214 ) @xcite . in bi2212 ,
the value derived from raman experiments ( @xmath23 ) is also in good agreement with the results of photoemission experiments ( @xmath24 ) .
the first raman investigation of the influence of doping was carried out on y123 @xcite and it was found that , although the symmetry appeared to be independent of doping , the b@xmath5 pair - breaking frequency ( @xmath25 ) decreased much more rapidly than t@xmath9 as the doping level was increased beyond optimal .
later , kendziora _ et al .
_ @xcite carried out measurements on both heavily overdoped and underdoped bi2212 . in the heavily overdoped region
they also found that @xmath26 decreased and that the b@xmath5 and a@xmath5 peak frequencies approached equality , and suggested that this implied the existence of an isotropic gap .
the low energy frequency dependence of their measured b@xmath5 and a@xmath5 spectra also appeared to be compatible with this interpretation .
however , it was pointed out by hewitt _ et al .
_ @xcite that the gap can not be isotropic if the _ bose - factor corrected _ spectra from all scattering geometries are considered .
_ @xcite also carried out experiments on a heavily overdoped crystal of bi2212 and interpreted their results in terms of a disordered d - wave model proposed by devereaux @xcite . in an attempt to gain additional insight into the influence of doping on the superconducting order parameter in htsc we have carried out raman scattering experiments on overdoped ( od ) and underdoped ( ud ) bi2212 crystals with critical temperatures , t@xmath27 = 82k , 85k , 86k , 90.5k ( ud ) and 89k , 87k , 83k ,
70k , 55k ( od ) .
spectra have been obtained in the b@xmath5 and b@xmath4 scattering geometries in both the superconducting and normal state .
the b@xmath4 spectra of overdoped compounds , which were not carefully investigated in previous studies @xcite , have provided important information regarding the symmetry of @xmath3 .
these spectra sample regions of the fermi surface that are located near the diagonal directions in the brillouin zone and hence can be used @xcite to directly probe for the existence of d - wave type gap nodes .
our results suggest that the superconducting gap possesses d@xmath10 symmetry , for the doping range ( 0.12@xmath28 0.23 ) investigated slightly underdoped to overdoped .
in cuprates , the addition ( removal ) of oxygen increases ( decreases ) the cu valence @xcite and therefore the hole concentration . thus single crystals of bi2212 were subjected to various oxygen - rich or oxygen - poor environments in order to vary the hole concentrations , spanning the doping range 0.07 @xmath30 0.23 .
the superconducting transition temperatures ( measured by the onset of the drop in magnetization or the zero of resistance ) and hole concentrations are listed in table [ bi2212samples ] .
hole concentrations for each crystal are estimated using the empirical relation of tallon and presland @xcite , assuming t@xmath31 .
raman vibrational spectra in the frequency range 20 - 1000 @xmath32 were obtained using the 514.5 nm line of an argon ion ( ar@xmath33 ) laser as the excitation source .
the raman measurements were carried out in a quasi - backscattering geometry , with the incident laser beam directed nearly perpendicular to the freshly cleaved ( * a*,*b * ) face of the crystal .
bi2212 has an orthorhombic ( d@xmath34 ) structure with the * a * and * b * axes oriented at 45@xmath35 to the cu - o bonds . in raman experiments it is customary ( since a@xmath36 b ) to assume a tetragonal structure ( d@xmath37 ) with axes parallel to * a * and * b*. to facilitate comparison with other cuprates , however , we will consider a tetragonal cell with * x * and * y * axes parallel to the cu - o bonds .
another set of axes , * x* and * y* are rotated by 45@xmath35 with respect to the cu - o bonds .
considered within the tetragonal point group ( d@xmath37 ) , the * x**y* ( b@xmath5 ) scattering geometry means that the incident(scattered ) light is polarized along * x*(*y* ) and selection of this scattering channel enables coupling to excitations having b@xmath5 symmetry .
similarly , the * xy * geometry allows coupling to b@xmath4 excitations , which transforms as d@xmath38 .
finally , the diagonal scattering geometry * x**x* allows coupling to a@xmath5 + b@xmath4 and * x**x * to a@xmath5 + b@xmath5 excitations .
thus , by choosing the polarization of the incident and scattered light appropriately , one may select different components of the raman tensor and thus various symmetry properties of the excitations . the b@xmath5 and b@xmath4 symmetries probe @xcite complementary regions of the fermi surface ( fs ) , while the a@xmath5 spectra correspond to an approximate average over the entire fs . in particular , the b@xmath5 raman spectrum samples the fs along the ( 0,@xmath39 ) and ( @xmath39,0 ) directions whereas the b@xmath4 spectum samples the ( @xmath40,@xmath40 ) directions .
that is , the b@xmath5 spectra sample the bz principal axes while the b@xmath4 spectra sample the bz diagonals , and are therefore used as complementary probes to obtain information about the * k * dependence of the gap . the spectra presented in this section have been corrected for the bose thermal factor .
therefore , the spectra shown are proportional to the imaginary part of the raman response function ( @xmath6 ) .
spectra obtained from the cuprates at room temperature are characterized @xcite by a rather featureless continuum , extending from @xmath41 0 to 2 ev . when the samples are cooled below t@xmath9 the spectral weight at low energies ( @xmath42 )
may be redistributed to higher energy and gap - like peaks appear in the spectra . in bi2212 the superconductivity induced renormalization observed in the b@xmath5 spectra ( see fig .
[ b1 g ] ) produces a peak whose energy remains relatively constant ( for p @xmath8 0.16 ) , but decreases drastically with increasing hole concentration in the overdoped regime ( p @xmath43 0.16 ) . when the peak energy is compared ( fig .
[ gap - b1g - arpes ] ) with the maximum sc gap ( @xmath44 ) values reported by arpes @xcite and tunneling @xcite , the results are striking .
the doping dependence and energy of the b@xmath5 peak position corresponds very well with the values of @xmath25 obtained from arpes measurements on similarly doped samples ( see fig .
[ gap - b1g - arpes ] ) .
this supports the previous association of the peak frequency in the b@xmath5 spectra with the maximum ( @xmath45 ) in a d@xmath10 gap function .
this association is also supported by the low frequency dependence of @xmath6 at low temperatures , as will be discussed in section [ fdchi ] .
the ratio @xmath46 has also been plotted as a function of p in fig .
[ kbtc - vs - p ] , where the results reveal that this ratio decreases monotonically with increasing hole concentration .
a linear least - squares fit to the data is well described by the empirical relation , @xmath47 for @xmath48 .
it can be seen in fig .
[ b1 g ] that , as p is reduced , there is a reduction in the fraction of spectral weight redistributed at low frequencies .
quantitatively , this redistribution was estimated by taking the ratio of the spectral weight above and below t@xmath9 .
the frequency range of interest is selected to envelope the region where increased scattering in the superconducting state is observed . within this region , the points at which the normal and superconducting state spectra cross defines the upper and lower bounds on the integration .
in this range , the ratio of the integrated intensity of the superconducting ( a@xmath0 ) vs. normal state ( a@xmath49 ) spectra are calculated , and the results are presented in fig .
[ spectralweight ] .
the redistributed spectral weight is reduced ( fig .
[ spectralweight ] ) with decreasing hole concentration , indicating that fewer quasiparticles are participating in superconductivity .
it may be that the quasiparticle pairs are being destroyed by some collective excitation of the crystal whose strength increases with decreasing doping
. the missing spectral weight may be most likely redistributed to higher energy , as recently suggested by the work of naeini _ et al _ @xcite . in their study of la214
it was found that the b@xmath5 spectral weight that is lost at low - energies ( @xmath50 800 @xmath32 ) is transferred to the higher energy region normally occupied by multi - magnon scattering .
taken together , the results suggest that the quasiparticle density near @xmath51 is reduced by magnetic excitations of some type .
of course , this suggestion requires further quantitative analysis .
a redistribution of spectral weight also occurs below t@xmath9 in the b@xmath4 channel ( fig .
[ b2 g ] ) but it is difficult ( except in the case of the 83 k overdoped sample ) to identify any peak formation in the low temperature b@xmath4 spectra of fig .
[ b2 g ] . in the depleted regions of the spectra
the intensity rises linearly at low frequencies then levels off at a frequency we will designate as @xmath52 ( a linear - to - constant crossover frequency ) . a linear fit is made to the low frequency data and @xmath52 is defined by the point at which the line departs from the experimental data points .
this frequency can be taken as a measure of the extent of the depleted region and thus provides an estimate of the b@xmath4 `` gap '' energy @xcite .
it is interesting to note that @xmath52 has a parabolic dependence on hole concentration ( fig .
[ b2gpeak ] ) and in contrast to 2@xmath44 ( b@xmath5)actually scales with t@xmath9 throughout the doping range , consistent with the results obtained in raman experiments by opel _
_ @xcite and sugai & hosokawa @xcite .
thus , in the underdoped regime , the results obtained seem to be incompatible with the existence of a single , simple d - wave gap . on the other hand sugai and hosagawa @xcite
have suggested that the values of 2@xmath44 obtained from the b@xmath5 spectra scale with the frequency of the 2-magnon peak .
this observation suggests that b@xmath5 excitations are magnetic in origin .
although not the topic of this paper , it remains to be seen whether the holon - spinon phase diagram model of anderson @xcite and lee @xcite may offer an explanation of these results . in the overdoped regime
one can observe superconductivity induced peaks in the b@xmath5 channel . in this region , ( p@xmath53 p@xmath54 both the values of @xmath55 and @xmath56 decrease with increasing doping .
however , @xmath55 decreases much more rapidly than does @xmath56 and as a result the peak frequencies become almost equal for p @xmath36 0.25 .
this behaviour , which is similar to that observed in doping studies of la214@xcite , appears to be inconsistent @xcite with the existence of a gap with d@xmath10 symmetry .
thus , although the b@xmath5 and b@xmath4 spectra of optimally doped compounds @xcite imply the existence of a d - wave gap , there are some puzzling discrepancies in this interpretation in both the underdoped and overdoped regions . in an attempt to gain additional insight we will now turn to an investigation of the dependence of @xmath6 on @xmath57 in the low energy region of the spectrum .
the ( @xmath58 ) frequency dependence of scattering in the low energy regime , ( i.e. below @xmath59 ) provides a further test @xcite of the cooper pair symmetry . in the b@xmath4 geometry ,
the observed frequency dependence of the response function remains linear in @xmath57 , regardless of the sample @xmath60 ( fig .
[ b2 g ] ) .
the only exception is found in examining the low energy region of the b@xmath4 spectra ( fig .
[ 87kb2 g ] ) for the 87k ( od ) sample .
one observes a change in slope of the raman response function at an energy of @xmath61 or @xmath62 . below this energy
the scattering is constant and rises linearly at frequencies above 90 @xmath32 .
these results have been repeated and are consistent with that described here .
thus we are confident that the observations are intrinsic and not an experimental artifact . in an arpes investigation of a similarly overdoped ( t@xmath63= 87k )
bi2212 crystal , ding _
_ @xcite found that the sc gap function exhibited an extended node in the ( @xmath64 ) , ( @xmath65 ) directions in the bz .
the modulation direction corresponds to the ( 0,0)@xmath66 ( @xmath67 ) region of the fermi surface of bi2212 .
_ @xcite have proposed that the extended node feature ( p @xmath68 ) is an artifact of umklapp scattering , resulting from scattering of the outgoing electron by the wavevector of the modulation @xmath69 ( @xmath70 ) .
the extended node artifact near ( @xmath67 ) should be manifest in the density of excited states by an increased scattering at low energies .
therefore , we contend that the increased , and constant , scattering found at low energies in the b@xmath4 channel ( for this particular crystal ) is consistent with umklapp scattering from the modulation .
umklapp processes scatter a quasiparticle out of the first bz . in this case
, a quasiparticle on the fermi surface can be scattered by an integral multiple of @xmath69 .
therefore , in the extended zone scheme , the quasiparticle may thereby find itself on another portion of the fermi surface for which the fermi velocity is opposite to that which it possessed before being scattered .
thus , this scattering process may reverse the fermi velocity of quasiparticles with wavevectors along the direction of the modulation wavevector .
this scattering process may destroy the coherence that is required for pairing , leading to a destruction or depression of the gap along the ( @xmath71 ) direction in k - space .
the effect would therefore be seen in the b@xmath4 scattering geometry .
it is true that the raman intensity is proportional to the structure function @xmath72 .
if the scattering process is instantaneous , i.e. time reversal symmetry holds , then the structure function is related by the fluctuation - dissipation theorem to the susceptibility @xmath73 .
time reversal symmetry implies that the @xmath74 , or equivalently @xmath75 .
consequently , the value of @xmath76 for @xmath77 ......... in the b@xmath5 spectra of fig . [ b1 g ]
the renormalization is quite strong in the range of hole concentrations covering the t@xmath9 s 82 k ( ud ) to 70 k ( od ) samples . in all but one spectra ( t@xmath9 = 89 k od ) a cubic low frequency dependence is observed , below the @xmath45 pair - breaking peak .
it is also evident that there is poorer agreement with a cubic fit in the underdoped regime , where also the peak in the superconducting state appears at a relatively constant value around 565 @xmath78 .
recent tunneling spectroscopy studies of underdoped bi2212 , which probe the gap at several points on crystals show that there exists local variation in the gap size .
the results are interpreted in terms of an inhomogeneous distribution of oxygen . the poorer agreement ( fig .
[ b1 g ] ) with a cubic fit is thought to be due to this intrinsic disorder , giving rise to a linear component to the scattering at low frequencies , as described by devereaux et al @xcite and seen in overdoped bi2212 .
in addition , in other cuprate @xcite supercondcutors there is no redistribution of spectral weight below t@xmath9 in the underdoped regime , in contrast to the results presented here .
this contradiction is resolved by the assumption that the underdoped crystals contains locally varying hole concentrations - the fraction of the crystal with a local value of @xmath79 is reduced as the hole concentration is also reduced ! consequently , the peak would remain near its value at optimal doping and be reduced in intensity , as observed .
pairing functions such as an anisotropic s - wave or a d - wave gap are consistent with states lying below some maximum gap value . for a d - wave gap of the form d@xmath80 , electronic raman spectra in the b@xmath4 and b@xmath5 channels
are predicted @xcite to display a linear ( in b@xmath4 ) and cubic ( in b@xmath5 ) power law behaviour below @xmath81 .
conversely , the raman response function in the low energy regime for an isotropic s - wave gap should be @xcite characterized by the absence of states below the gap , expressed in experimental terms by exponentially activated behavior . down to the lowest frequency measured ,
the b@xmath5 and b@xmath4 spectra from all samples show scattering below the gap peaks which appear in the spectra below t@xmath9 .
in particular , at all doping levels the b@xmath4 response function shows a linear rise in scattering at low frequencies for t @xmath8 t@xmath9 .
therefore , the results are consistent with any gap function that posesses nodes ( as a line or point ) on the fermi surface near ( @xmath82 ) . in the hole concentration range @xmath83 , the * linear * and * cubic * frequency dependences found for the b@xmath4 ( fig .
[ b2 g ] ) and b@xmath5 ( fig .
[ b1 g ] ) spectra , respectively , are consistent with the behaviour predicted for a d@xmath80 gap , and thus completely inconsistent with an isotropic s - wave gap .
the cubic power law for a d@xmath10 pairing state was derived in the limit @xmath84 . however , even at the temperatures used in these studies , i.e 0.20 @xmath8 t / t@xmath85 0.28 , the power laws are still consistent with a d - wave gap .
this suggests that the gap opens quite quickly below t@xmath9 .
despite the presence of cation disorder , oxygen intercalation , b - axis modulation , and orthorhombicity that are common to bi2212 , the low - frequency power laws appear to persist throughout the doping range . the low - frequency scattering in the b@xmath5 channel is found to be consistent with a cubic frequency dependence . in b@xmath4 spectra , a linear - in-@xmath57 scattering is found for all samples studied , which indicates that there are nodes in the gap function near ( @xmath86 ) .
therefore , the combined ( for @xmath87 ) power - law behavior ( @xmath57 in b@xmath4 and @xmath88 in b@xmath5 ) points to the existence of a d@xmath10 gap function for the doping range studied .
in the b@xmath5 scattering geometry , the superconductivity induced renormalization peak frequency changes with doping . below a critical doping level ( p @xmath89 )
the renormalization is not present ( see @xcite ) . with increasing hole concentration
the peak energy remains relatively constant to p @xmath90 and then drops rather abruptly , by approximately 50 % over the hole concentration range @xmath91 ( i.e. overdoped regime ) . in the latter range , the b@xmath5 peak energy and doping dependence can be plotted on a curve that matches the @xmath25 feature observed in arpes and tunneling spectroscopy measurements .
provided that the arpes results do indeed measure 2@xmath92 , these considerations provide substantial support for the assignment of the b@xmath5 renormalization peak to @xmath25 .
one reason for the lack of a pair - breaking peak relies on the fact that the low - energy b@xmath5 spectral weight is reduced @xcite substantially with underdoping .
if the quasiparticles are scattered away from the ( @xmath15 ) regions of the fermi surface , then the resulting b@xmath5 spectra would show a reduced intensity .
a mechanism for such anisotropic scattering lies in the description of raman scattering in the nearly antiferromagnetic fermi liquid model @xcite . in the model ,
the presence of antiferromagnetic spin fluctuations , which are strongly peaked near the antiferromagnetic ordering vector @xmath93 , give rise to a enhanced scattering of quasiparticles on regions of the fermi surface that are located near the bz axes .
for example , in such a process a quasiparticle can be scattered from ( @xmath94 ) to ( @xmath95 ) by emitting a ( @xmath64 ) magnon and back to ( @xmath94 ) by emitting a ( @xmath96 ) magnon ( fig . [ hotspots ] ) . since the regions of the fs located near the diagonals are far from resonance with q , the quasiparticles in this region are much more weakly scattered .
this gives rise @xcite to the terms `` hot - spots '' for regions of the fs near the bz axes and `` cold - spots '' for those near the diagonals .
as the doping level decreases and the antiferromagnetic fluctuations grow and become more strongly peaked near q , the scattering will increase and hence the b@xmath5 spectral weight will decrease further . correspondingly , one would expect that cooper pairing of the highly damped `` hot '' quasiparticles to become increasingly unlikely in the underdoped regime . on this basis , the absence of a pair - breaking peak in the b@xmath5 channel for low doping levels ( [email protected] ) is thus to be expected .
this picture also accounts for the rapid reduction , with decreasing hole concentration , of the b@xmath5 spectral weight that is involved in superconductivity .
it is also interesting to note that the b@xmath5 spectra obtained from underdoped bi2212 appear to differ from that obtained from underdoped la214 and y123 . in the latter two compounds
the b@xmath5 spectra change much more abruptly when the doping level is reduced below optimum . in these compounds
a strong 2@xmath44 peak is observed in the b@xmath5 spectra obtained from optimally doped materials , but such a peak is completely absent @xcite in the underdoped compounds . in bi2212 however , a relatively strong b@xmath5 peak is observed still observed in the underdoped compounds ( see @xcite and fig .
[ gap - b1g - arpes ] ) .
in addition , the b@xmath5 spectral weight drops rapidly as one enters the underdoped regime @xcite of la214 or y123 but decreases more slowly with underdoping in bi2212 .
it thus appears that the optimally doped state is more sharply defined in both la214 and y123 than it is in bi2212 .
this might be attributed to the possibility that , because of cation disorder and the superstructural modulation , bi2212 is more inhomogeneous than la214 .
the hole concentration in single crystals of bi2212 has been varied from 0.10 to 0.22 by varying the oxygen concentration and by y substitution .
it has been found that the 2@xmath44 peak in the b@xmath5 spectrum increases in energy as the hole concentration is reduced from p=0.22 to p=0.12 according to the empirical relation , @xmath97 .
both the magnitude of @xmath44 and its doping dependence , as determined from the raman spectra , are in good agreement with the results of arpes measurements @xcite of the leading edge shift that occurs below t@xmath9 .
these results provide support for the association of the 2@xmath92 peak of the b@xmath5 spectra with the superconducting gap energy as determined by arpes measurements .
it is also found that the strength of the b@xmath5 renormalization , that occurs below t@xmath9 , becomes weaker as the hole concentration is reduced from p=0.22 to p=0.12 and can not be observed in crystals with hole concentrations less than [email protected] .
this observation is consistent with the selective scattering of quasiparticles on regions of the fermi surface that are located near the k@xmath21 and k@xmath98 axes of the brillouin zone .
a possible mechanism for this scattering is provided by antiferromagnetic spin fluctuations @xcite which grow in strength as the hole concentration of the crystals is decreased .
the b@xmath4 spectra show a very weak renormalization at t@xmath9 for two of the four crystals studied in this work . the 82k
the cross - over frequency ( @xmath52 ) exhibits a parabolic dependence on hole concentration - similar to that found for t@xmath9 .
in fact @xmath99 appears to scale with t@xmath9 according to the relation @xmath100
@xmath36 5k@xmath101t@xmath9 , as has been suggested by others @xcite .
the b@xmath4 response function , at t = 20k , increases linearly at low frequencies in most of the samples studied .
this observation is consistent @xcite with the presence of nodes in the superconducting gap , in the direction @xmath102 .
the only exception to this linear behaviour at low temperature was observed in the b@xmath4 spectrum of an overdoped crystal with t@xmath9 = 87k . in this case
the scattered intensity remained constant for all frequencies @xmath103 @xmath32 .
it is interesting that in arpes experiments carried out on an overdoped crystal with the same t@xmath63 it was found that gap was zero over an extended region of the fs , which could be consistent with @xmath104 constant below a certain frequency .
this constant is believed to be an artifact associated with umklapp scattering from the superstructural modulation . finally , it should be noted that results obtained here for the doping dependence of the spectra are , in most respects , similar to those obtained in other cuprates @xcite .
one exception appears to be the fact that in the underdoped regime the spectral weight depletion , and the strength of the b@xmath5 renormalization in bi2212 , decrease less abruptly as p is decreased below p@xmath36 p@xmath54 .
the financial support of the natural sciences and engineering research council of canada is gratefully acknowledged .
we are indebted to d. m. pooke of industrial research ltd .
( lower hutt , new zealand ) for some of the single crystals used in these studies .
j. mesot , m. r. norman , h. ding , m. randeria , j. c. campuzano , a. paramekanti , h. m. fretwell , a. kaminski , t. takeuchi , t. yokoya , t. sato , t. takahashi , t. mochiku , and k. kadowaki , phys .
lett . * 83 * , 840 ( 1999 ) .
h. ding , a. bellman , j. c. campuzano , m. randeria , m. r. norman , t. yokoya , t. takahashi , h. katayama - yoshida , t. mochiku , k. kawdowaki , g. jennings , and g. p. brivio , phys .
lett . * 76 * , 1533 ( 1996 ) .
spectra as a function of doping in bi2212 .
the corresponding values of @xmath25 derived from the arpes results of mesot _ et al . _
@xcite ( filled squares ) , and the tunneling data of miyakawa _ et al . _
@xcite ( filled triangles ) and ozyuzer _ et al . _ @xcite ( open triangles ) are shown for comparison . in all cases
tallon s relation was used to determine the hole concentration , assuming t@xmath31 . ]
_ can _ exchange an afm wavevector ( @xmath64 ) while those near @xmath107 _ cannot_. hence the former are highly scattered `` hot '' quasiparticles , and the latter are referred to as `` cold '' quasiparticles . ] | bi@xmath0sr@xmath0cacu@xmath0o@xmath1 crystals with varying hole concentrations ( 0.12 @xmath2 0.23 ) were studied to investigate the effects of doping on the _ symmetry _ and _ magnitude _ of the superconducting gap ( @xmath3 ) .
electronic raman scattering experiments that sample regions of the fermi surface near the diagonal ( b@xmath4 ) and principal axes ( b@xmath5 ) of the brillouin zone ( bz ) have been utilized .
the frequency dependence of the raman response function , @xmath6 , at low energies ( @xmath7 ) is found to be linear for b@xmath4 and cubic for b@xmath5 ( t@xmath8 t@xmath9 ) .
the latter observations have led us to conclude that the doping dependence of the superconducting gap is consistent with d@xmath10 symmetry , for slightly underdoped and overdoped crystals .
studies of the pair - breaking peak found in the b@xmath5 spectra demonstrate that the magnitude of the maximum gap ( @xmath11 ) decreases monotonically with increasing hole doping , for @xmath12 . based on the magnitude of the b@xmath5 renormalization
, it is found that the number of quasiparticles participating in pairing increases monotonically with increased doping . on the other hand ,
the b@xmath4 spectra show a weak `` pair - breaking peak '' that follows a parabolic - like dependence on hole concentration , for 0.12 @xmath13 0.23 . |
the broad family of complex correlated oxides are the next frontier in novel materials , providing a rich variety of tunable properties that are both functionally useful and valuable in understanding fundamental condensed matter physics .
for example , interfaces between perovskite - type oxides are relatively easily achieved at an atomically precise scale , yielding new and exciting properties.@xcite as these materials become more important , achieving a firm understanding of their surfaces , including the role and degree that electron correlations play , is crucial . in strongly correlated materials ,
the on - site coulomb energy , @xmath6 , leads to dynamical correlations between electrons , tending towards their localization at atomic sites .
incoherent electron states , the hubbard subbands , form either side of the coherent , one electron - like quasiparticle peak ( qp ) , and are referred to as lower and upper hubbard bands ( lhb and uhb respectively ) .
transfer of spectral weight away from the qp into the lhb and uhb signals the effects of stronger electron correlations .
such strong correlations can have profound effects on the electronic structure and material properties , most famously leading to insulating , rather than metallic , ground states.@xcite sr@xmath0ca@xmath1vo@xmath2 are prototypical strongly - correlated perovskites , with similar spectral weight in both the incoherent ( strongly correlated ) hubbard subbands and the coherent ( one electron - like ) quasiparticle states . although the spectral function of these materials has been well studied experimentally @xcite and theoretically,@xcite there is not a well established consensus on the form of the surface and bulk components , or as the sr content ( @xmath7 ) is varied .
both endmembers are correlated metals and have been reported to exhibit metal - insulator transitions in ultrathin films due to a reduction in the dimensionality and corresponding narrowing of the v @xmath5 bandwidth.@xcite generally , the focus of photoemission spectroscopy ( pes ) studies has been on obtaining bulk - representative spectra . for example , laser - pes has been used at very low photon energies , below the minimum in the photoelectron inelastic mean free path ( imfp ) , to extract bulk qp spectra,@xcite although the lhb is not accessible at these low energies .
conversely , soft x - rays have also been used to increase the depth sensitivity , and pes spectra were obtained which were either independent of @xmath7@xcite or which revealed an insulating surface.@xcite more recently , some progress has been made in understanding the behavior of the spectra with @xmath7 , with several reports agreeing that the lhb of cavo@xmath2 is more intense and closer to the fermi level , @xmath8 , than that of srvo@xmath2,@xcite indicative of the effects of stronger electron correlations .
additionally , band dispersions and fermi surfaces have been clearly observed in both end - members.@xcite however , the behavior of the spectra at the surface compared with the bulk has received little experimental attention , despite the role of the surface being questioned and studied in detail theoretically.@xcite moreover , our knowledge of the `` bulk '' spectra from pes has so far been limited to analytical differences between spectra containing both bulk and surface components.@xcite recently , resonant soft x - ray emission spectroscopy ( rxes ) measurements have reported truly bulk - representative information on the strongly - correlated v @xmath5 bands , illustrating that the effects of electron correlations are more pronounced in cavo@xmath2 than in sr@xmath9ca@xmath9vo@xmath2 , and that the surfaces of both materials are `` more correlated '' than their bulk.@xcite although valuable , comparisons between rxes and pes spectra are complicated by the quite different scattering processes and comparatively poorer resolution of rxes ( in part , owing to the broadening due to the lifetime of the core hole ) . in order to elicit quantitative information on the evolution of electron correlations from the surface to the bulk , depth sensitive measurements using the same probe are clearly desired . hard x - ray photoemission spectroscopy ( haxpes ) extends conventional pes measurements at ultraviolet and soft x - ray energies into the hard x - ray regime ( @xmath10 kev ) , yielding much greater depth sensitivity ( @xmath11 nm compared with 0.5 nm at ultraviolet energies).@xcite however , at such hard energies , there are several additional considerations .
most importantly , the cross section for photoionization drops off at higher energies , particularly for shallow binding energies at the valence band , @xcite and practical instrument resolutions are poorer than at low energies .
additionally , the electron momentum distribution is focused into much tighter angles , meaning the coverage of the brillouin zone of a typical haxpes measurement is much broader , although recent technical advances have facilitated the momentum - resolved spectral function at hard x - ray energies.@xcite we report here a haxpes study of the electronic structure of sr@xmath0ca@xmath1vo@xmath2 using photon energies of 4.2 kev and 2.2 kev .
the valence band of sr@xmath0ca@xmath1vo@xmath2 measured using 4.2 kev photons is quantitatively compared with _ ab initio _ theoretical calculations of the bulk , and we show that the spectra can be well described by band theory with some modest adjustments to the orbital - dependent photoionization cross - sections . additionally , we find that the evolution of the valence band electronic structure with @xmath7 is dominated by different @xmath4-site hybridization . secondly , the electronic structure of the correlated v @xmath5 states is investigated as a function of distance into the sample from the surface .
this was achieved by varying the photon energy of the haxpes measurements , and by comparison with previous surface - sensitive ultraviolet photoemission spectroscopy ( ups ) and unambiguously bulk - sensitive rxes measurements .
particular attention was paid to the _ energetics _ of the hubbard subbands , which are one of the key properties in identifying the evolution of correlated behavior .
we find that spectra recorded at 2.2 kev excitation energy still contain a large surface contribution of 30 40% , despite photoelectrons at this energy often being considered bulk probes of correlated materials . however , we show that photoelectrons at 4.2 kev are found to essentially probe the bulk electronic structure .
significantly , we demonstrate that the correlated localized hubbard bands evolve in both energy and spectral weight from the surface ( `` more correlated '' ) of the material to the bulk ( `` less correlated '' ) , in remarkable agreement with dynamical mean - field theory predictions .
( color online ) core level haxpes spectra at 4.2 kev of cavo@xmath2 and sr@xmath9ca@xmath9vo@xmath2 . ]
large high quality single crystals of cavo@xmath2 ( cvo ) and sr@xmath9ca@xmath9vo@xmath2 ( scvo ) were grown by the floating zone technique in a four mirror optical furnace , employing growth rates of 7 to 10 mm / h in an atmosphere of 1 bar of ar + 3% h@xmath12 gas.@xcite haxpes measurements were performed at beamline x24a of the national synchrotron light source , brookhaven national laboratory , with total instrument resolution ( @xmath13 ) of 180 mev at 2.2 kev and 220 mev at 4.2 kev .
photoelectrons with kinetic energies of 2.2 kev and 4.2 kev are estimated to have an imfp , @xmath14 , of 3.6 nm and 6.1 nm respectively in sr@xmath0ca@xmath1vo@xmath2 within the tpp-2 m model,@xcite compared with 0.5 nm at ultraviolet energies ( @xmath15 ev).@xcite owing to the different densities of cavo@xmath2 and srvo@xmath2 , the imfp within this model is anticipated to be @xmath16% smaller in srvo@xmath2 than in cavo@xmath2 ( mean values , roughly corresponding to sr@xmath9ca@xmath9vo@xmath2 , are quoted above ) .
samples were cleaved _
ex - situ _ , and immediately loaded into the ultra - high vacuum chamber .
the photoelectron energy was referenced to polycrystalline ag in electrical contact with the sample .
haxpes measurements at 4.2 kev of the o @xmath17 and v @xmath3 core levels are shown in fig .
[ f : shcore ] , and are in good agreement with previous reports at 1.8 kev.@xcite the two compounds exhibit essentially identical core level spectra . the v @xmath18 level exhibits two structures , originating from the ligand - screened and coherently - screened final states of the photoemission process , and the @xmath18 charge - transfer satellite
is observed as a shoulder to the o @xmath17 peak .
these spectra are free of surface contaminant species and disorder effects , such as reports of v@xmath19/v@xmath20 segregation at the surface of scraped samples.@xcite we emphasize that the enhanced electron correlations at the surface that we discuss in section [ s : nearef ] are an _
intrinsic _ property of the surface , arising from the broken translational symmetry and reduced coordination,@xcite rather than an extrinsic surface problem ( due to contaminant species and/or surface disorder ) . _
ab initio _ electronic structure calculations have been performed using the all - electron full - potential linearized augmented plane - wave ( flapw ) implemented by the elk code,@xcite within the local density approximation ( lda ) to the exchange - correlation functional .
experimental lattice parameters of 3.84 was used for cubic srvo@xmath2@xcite and @xmath21 , @xmath22 and @xmath23 were used for orthorhombic cavo@xmath2.@xcite convergence was achieved over a @xmath24 mesh in the full brillouin zone .
( color online ) ( a ) haxpes measurements at 4.2 kev of cavo@xmath2 and sr@xmath9ca@xmath9vo@xmath2 , illustrating the shirley - type background contribution ( dashed line ) .
the broadened total dos ( tdos ) is shown at the bottom of the figure for cavo@xmath2 and srvo@xmath2 .
( b ) broadened partial densities of states of cavo@xmath2 and srvo@xmath2 , showing the contribution from each orbital component , and their respective weights .
each curve has been offset vertically by 2 ev@xmath25 f.u.@xmath25 for clarity .
( c ) haxpes measurements after subtraction of the shirley - type background , compared with the results of fitting the pdos to the data ( fit ) . also shown
is the theoretical cross - section corrected pdos ( c - pdos ) . ]
haxpes measurements at 4.2 kev of the o @xmath3 and v @xmath5 valence band are shown for cvo and scvo in fig .
[ f : vb2p](a ) .
the o @xmath3 states are characterized by a strong peak at @xmath26 ev , with a prominent shoulder at @xmath27 ev , above which the v @xmath5 states are weakly visible as a double - peaked structure ( and to which we return in more detail below ) .
these spectra are in approximate agreement with previous pes measurements at low@xcite and high@xcite photon energies . for comparison , the total ( broadened ) density of states ( dos ) from the flapw calculation ,
neglecting the energy - dependent photoionization cross - section , is shown for the pure compounds at the bottom of fig .
[ f : vb2p](a ) . in fig .
[ f : vb2p](b ) , the broadened occupied partial densities of states ( pdos ) of cavo@xmath2 and srvo@xmath2 are shown , scaled ( by a factor shown in the figure ) to have approximately the same area . whereas the total dos are dominated by o @xmath3 and v @xmath5 character , the orbital - dependent photoionization cross - section at @xmath28 kev favors photoemission from the more localized @xmath29 and @xmath30 states .
for example , the strength of the peak at @xmath26 ev can not be explained without a dramatic enhancement in the v @xmath31 and ca / sr @xmath29 states , compared with the o @xmath3 pdos .
indeed , this is reflected in theoretical tabulations of these cross - sections , which predict the ca @xmath32 , sr @xmath33 and v @xmath32 cross - sections are between 13 and 21 times larger than the o @xmath3 cross - section at 4 kev.@xcite in order to more accurately assess the origin of the differences in the haxpes spectra between cvo and scvo , we turn to a more quantitative analysis , in which the relative photoionization cross - sections for each orbital are fitted to the experimental data , yielding approximate _ empirical _ relative cross - sections .
the data are first corrected for inelastic scattering processes , leading to the shirley - type background shown by the dashed line in fig .
[ f : vb2p](a ) .
the theoretical spectrum is then computed as a cross - section weighted sum of the orbital pdos components shown in fig .
[ f : vb2p](b ) , neglecting the interstitial dos ( which , being delocalized away from the ion centers , are unlikely to contribute significantly at hard x - ray energies ) , and compared to the experimental spectrum . for scvo , we take the rather crude approximation of 50% cavo@xmath2 pdos and 50% srvo@xmath2 pdos , although in practice this has only a very weak influence on the results ; for example , comparing pure srvo@xmath2 with the scvo data yields almost identical results .
the agreement between experiment and theory is optimized for both scvo and cvo simultaneously using a single set of parameters .
additionally , the o @xmath3 manifold is allowed to be rigidly shifted in energy during the fit , accounting for the inadequate description by the lda of the v @xmath5 bands . in the lda flapw calculations ,
the lhb feature ( the deeper of the two v @xmath5 peaks ) is absent , and so this rigid shift physically represents the experimental separation between the o @xmath3 and _ coherent _ v @xmath5 manifolds . finally , the experimental broadening of the dos is fitted through the convolution of the theoretical dos with a voigt function , @xmath34 , where @xmath35 is a gaussian of width @xmath13 and @xmath36 is a lorentzian of width @xmath37 .
the results of the fit yield @xmath38 ev and @xmath39 ev , accounting for the total broadening due to both intrinsic and experimental resolution effects .
all theoretical spectra in fig .
[ f : vb2p ] have been convoluted with this function , and multiplied by a fermi function of @xmath40 ev [ from a separate fit to the fermi edge of polycrystalline ag at 4.2 kev shown in fig .
[ f : vb2p](c ) ] .
although the solution from this fitting procedure is found to be relatively robust and stable , the extent to which the fitted parameters represent real photoionization cross - sections must be considered with caution .
for example , inaccuracies in the background subtraction and energy efficiency of the measurement may contribute non - trivially to the fit .
secondly , inadequacies of the lda , which may under- or overestimate hybridization between orbital characters , may be strongly exaggerated , and the lhb feature is completely absent from the theoretical curve . nevertheless , as we discuss below
, the results do reflect physical expectations , and are interpreted as providing a reasonable indication of the trends of the orbital cross - sections .
.[t : fit ] orbital - dependent photoionization cross - sections ( cs ) relative to the o @xmath3 states according to the theoretical tabulations of ref . at 4 kev and from the results of fitting haxpes
spectra at 4.2 kev of cvo and scvo to the flapw pdos .
the theoretical cross - sections marked by * correspond to nominally empty states that are not tabulated in ref . . instead , for the @xmath4 and v @xmath30 states we adopt the corresponding @xmath29 cross - section , and for the ca @xmath5 and sr @xmath41 states we use the sc @xmath5 and y @xmath41 cross - sections respectively
[ cols="<,<,<,>,<,<,>,<,^,^ , < , < , > , < , < , > , < , < " , ] in fig .
[ f : vb2p](c ) , the results of the fit are shown alongside the background - corrected haxpes spectra . a shift in the o @xmath3 manifold downwards in energy by @xmath42 ev
is required to align the energies of the peaks in the theoretical and experimental spectra ( the same shift is applied to both cvo and scvo spectra ) .
the fitted spectrum satisfyingly reproduces the experimental data , exhibiting the strong peak at @xmath26 ev and its shoulder at @xmath27 ev , as well as the rather weak relative intensity of the coherent v @xmath5 states .
moreover , the weak upwards shift in energy of the lower peak at @xmath26 ev of scvo is captured , along with the increase in the relative intensity of the shoulder . for comparison ,
the spectrum due to the theoretical cross - sections@xcite ( labeled c - pdos ) is also shown ( including the rigid shift of the o @xmath3 manifold ) in fig .
[ f : vb2p](c ) .
although this spectrum roughly reproduces the relative intensities ( and energies ) of the o @xmath3 manifold , it strongly overestimates the contribution from the v @xmath5 electrons .
the fitted , empirical relative cross - sections from the fit are shown in table [ t : fit ] , and exhibit the expected sensitivity of the haxpes experiment to the more localized @xmath29 and @xmath30 states .
surprisingly , the relative o character is strongly suppressed , including the o @xmath43 states , in contrast to other studies at similar energies ( which do not attempt to optimize agreement between experiment and theory).@xcite however , it is clear from fig . [
f : vb2p](b ) that the shape of the o @xmath43 pdos is incompatible with the data . finally , we note that incorporating an asymmetric doniach - unji lineshape,@xcite which has recently been used to explain the haxpes spectrum of v@xmath12o@xmath2,@xcite does not improve agreement between experiment and theory .
overall , these results establish that the experimental haxpes spectra of cavo@xmath2 and sr@xmath9ca@xmath9vo@xmath2 can be described well by the theoretical pdos , with some modest adjustments to the theoretical orbital cross - sections .
secondly , they demonstrate that the differences between the experimental spectra , in both relative intensity and energy , are mostly a consequence of the different @xmath4-site hybridization . the importance of @xmath4-site hybridization in @xmath44 perovskites has been pointed out in detail by ref . .
finally , the effects of the @xmath4-site hybridization are more pronounced in the experimental spectra than predicted by theory .
we now address the near-@xmath8 electronic structure , which is dominated by the v @xmath5 band .
high - statistics spectra of the v @xmath5 manifolds of cvo and scvo at 4.2 kev and 2.2 kev are shown in fig .
[ f : vb3d ] , referenced to the fermi level of polycrystalline ag foil in electrical contact with the samples . as illustrated by the relatively large density of states at @xmath8 , cvo and scvo
are both metallic in the haxpes measurement , and consist of the o @xmath3 tails below @xmath45 ev , followed by the lhb near @xmath46 ev and the qp at @xmath8 .
the dashed lines between @xmath47 and 0 ev represent the same spectra after dividing by the `` effective '' fermi function ( @xmath48 mev at 2.2 kev and @xmath49 mev at 4.2 kev ) , reflecting approximately flat qp dos up to @xmath8 .
the `` effective '' fermi function is determined by fitting a fermi function to the polycrystalline ag reference spectra ( displayed below in fig .
[ f : depth ] ) , and represents the total thermal and instrument resolution broadening .
similar observations of relatively flat dos near @xmath8 have been made using bulk - sensitive low - energy laser - pes at 7 ev.@xcite in those measurements , performed with much higher resolution than is achievable at hard x - ray energies , a distinct drop - off of the qp dos within 100 mev of @xmath8 is observed in the extracted bulk component .
although we are unable to resolve such fine details , our observation of flat dos near @xmath8 , coupled with the metallic nature of these spectra , demonstrate that recoil effects during the photoemission process at these energies are minor , and do not significantly shift the spectra , unlike other materials.@xcite in addition to the broader resolution function , the integration over several brillouin zones in the current haxpes measurements also likely contributes to the flattening of the observed spectral weight near @xmath8 .
( color online ) high - statistics haxpes spectra of the v @xmath5 states of ( a ) cavo@xmath2 and ( b ) sr@xmath9ca@xmath9vo@xmath2 at 4.2 kev and 2.2 kev incident photon energies ( vertically offset for clarity ) .
the dashed ( green ) lines indicate the contributions from the tails of the o @xmath3 band , and the black solid lines represent fits of the experimental spectra ( see text ) .
finally , the dotted lines near @xmath8 are the experimental spectra divided by the fermi function ( determined from ag foil ) . ] in order to obtain quantitative information on the spectral weight and energetics of the lhb , the spectra in fig .
[ f : vb3d ] have been fitted to a linear combination of three gaussian components , with the results multiplied by the effective fermi function .
the results of this fit , which describe the data very well , are shown by the black solid lines in fig .
[ f : vb3d ] , and the dashed line indicates the component due to the tails of the o @xmath3 states .
the other two components represent photoelectrons from the lhb and qp respectively .
( color online ) results of fitting the haxpes spectra to gaussian components .
( a ) fitted spectra ( excluding the o @xmath3 tails ) shown alongside the ups spectra of ref . .
all spectra have been broadened to have the same effective instrument resolution function , corresponding to @xmath50 mev , and normalised to the intensity of the lhb .
( b ) binding energy of the lhb for the two compounds , shown as a function of the estimated imfp , @xmath14 .
the horizontal dashed lines indicate the separation between the qp and lhb from bulk - sensitive rxes.@xcite the error bars represent estimated statistical errors associated with the fitting . ] at both excitation energies shown in fig .
[ f : vb3d ] , the relative intensity of the qp is found to be smaller in cvo than it is in scvo , which is consistent with the common notion that cvo is `` more correlated '' . in cvo ,
the narrower bandwidth , @xmath51 , of the v @xmath5 states means the relative importance of electron correlations , often parameterized through @xmath52 , is greater ( the coulomb energy , @xmath6 , is not expected to evolve between srvo@xmath2 and cavo@xmath2@xcite ) . in order to investigate how the spectral weight varies with haxpes excitation energy , we show the fitted spectra normalized to the intensity of the lhb in fig .
[ f : lhb](a ) , alongside the ups results of ref . .
to aid comparison , all spectra in fig .
[ f : lhb](a ) have been broadened to have the same instrument resolution function of @xmath53 mev ( i.e. the resolution of the 4.2 kev haxpes measurement ) .
figure [ f : lhb](a ) illustrates that the qp spectral weight is suppressed in the more surface - sensitive 2.2 kev haxpes measurement . since both the qp and lhb features derive from the same v @xmath5 states , their photoionization cross - section is the same , and therefore this evolution can be directly associated with the different depth sensitivities of the two measurements .
the ups spectra do not follow this trend ; however , these spectra were recorded on oriented samples near the @xmath54-point of the brillouin zone , and therefore represent a quite different ( and specific ) subset of momentum space . similar suppression in the qp weight with increasing surface sensitivity
has previously been reported at lower photon energies.@xcite this observation suggests that there is an _ intrinsic _ surface component of the electronic structure that is significantly `` more correlated '' than the bulk in both materials .
moreover , it indicates that an excitation energy of 2.2 kev is not sufficient to represent the bulk electronic structure of these materials .
previous efforts to extract bulk information from sr@xmath0ca@xmath1vo@xmath2 have focused on pes at soft x - ray energies up to 900 ev@xcite , 275 ev@xcite or 1486 ev@xcite ( although bulk - sensitive laser - pes has also been used@xcite the lhb lies outside the experimental range ) .
these have yielded conflicting pictures of the bulk electronic structure , for example whether it is independent of @xmath7 or not .
our measurements illustrate that hard x - rays above @xmath55 kev are required to truly examine the bulk using photoemission . whereas changes in the spectral weight of the qp and lhb are useful indications of correlated electron behavior , extracting such quantitative information from experiments can be complicated by overlapping extrinsic surface species and the tails of the o @xmath3 states , as well as the details of the brillouin zone coverage .
we now turn to the _ energetics _ of the lhb , which are experimentally much more robust , particularly since the lhb is not dispersive in momentum.@xcite as can be seen in figs .
[ f : vb3d ] and [ f : lhb ] , the energy of the lhb in both compounds lies at deeper energies in the 4.2 kev spectrum than at 2.2 kev , consistent with stronger electron correlations at the surface . at 4.2 kev ,
the lhb is found at @xmath56 and @xmath57 ev for cvo and scvo respectively ( fig .
[ f : lhb ] ) , which compares very well with the separation between the qp and lhb observed in truly bulk - sensitive rxes measurements.@xcite qualitatively , these results are in agreement with dynamical mean - field theory ( dmft ) calculations of cavo@xmath2 and srvo@xmath2 using the same coulomb parameter , @xmath6.@xcite haxpes at this energy can be considered representative of the bulk correlated electronic structure . at 2.2 kev
the lhbs of both compounds are located at @xmath58 ev , which is in rough agreement with the @xmath7-independent soft x - ray results.@xcite we discuss these spectra in more detail below , including the similarity in the lhb energy of both compounds . at the surface ,
ups suggests a markedly shallower energy for the lhb of @xmath59 and @xmath60 ev respectively , which is @xmath61% closer to @xmath8 than in the bulk .
the evolution in correlated electron behavior between the surface and the bulk has been investigated by ishida _
_ through lda+dmft calculations of the vo@xmath12- and sro - terminated surfaces of srvo@xmath2.@xcite at the vo@xmath12-terminated surface , the overall effect on the energy of the lhb is only weak , partly due to the emptying of the in - plane @xmath62 surface orbital . on the other hand , in good agreement with our results , these calculations find that the lhb is located at @xmath63 ev in the bulk and @xmath64 ev at the sro - terminated surface , and that the lhb has substantially greater spectral weight at the surface .
this reduction in the binding energy of the correlated hubbard bands , which is reproduced for both unrelaxed and relaxed sro - terminated surfaces , arises intrinsically due to the reduced coordination of the v ions at the surface , and the subsequent narrowing of the bandwidth of the out - of - plane @xmath65 orbitals.@xcite having established the variation in the energy and spectral weight of the lhb with depth , we now briefly discuss some of the implications of the spectra recorded at 2.2 kev .
in particular , the energies of the lhbs of both compounds are significantly shallower than in the bulk , and are very similar for the two different compounds , despite their differences at both the surface and the bulk . if we consider that the 2.2 kev spectra contain distinct surface and bulk components , it is possible to adequately fit the data using lhb and qp energies representative of the surface ( ups ) and bulk ( 4.2 kev haxpes ) . under this assumption , 29% and 45% of the 2.2 kev haxpes signal of cvo and scvo ,
respectively , is found to correspond to the surface signal .
indeed , this simple model reflects the larger surface sensitivity of identical photoemission measurements on srvo@xmath2 compared with cavo@xmath2 , which is due to the higher density of srvo@xmath2 and therefore shorter imfp .
however , the results at 2.2 kev do not agree well with simple calculations of the expected surface contribution . based on the imfp , @xmath14
, the surface contribution can be expressed as @xmath66 , where @xmath29 is the thickness of the surface layer . assuming a surface layer of @xmath67 ,@xcite the surface contribution is 18% and 11% at 2.2 kev and 4.2 kev respectively , with only a very minor difference ( @xmath68% ) between compounds .
this is not only at odds with the results presented above , but it is also impossible to find a reasonable value of the surface thickness at which 2.2 kev photoelectrons are significantly ( e.g. more than two times ) more sensitive to than 4.2 kev photoelectrons .
it therefore appears that either the distribution of detected photoelectrons from the sample is not simply an exponential decay , or that the notion of a definite , discrete correlated surface layer is incorrect .
the second possibility seems less likely based on lda and dmft calculations , in which changes to the electronic ( and indeed , crystal ) structure due to the surface were found to decay very rapidly from the surface.@xcite in either case , it calls into question one of the common methods of extracting bulk contributions from pes , at least applied to studying correlated electron behavior,@xcite in which a spectrum at a single photon energy is assumed to be composed of discrete surface and bulk contributions weighted by the exponential decay due to the imfp . in our data , it is not possible to reproduce either surface or bulk spectra by such a method .
( color online ) comparison between various probes of different depth sensitivities of the correlated electronic structure of cavo@xmath2 and sr@xmath9ca@xmath9vo@xmath2 .
the approximate depth sensitivity of each technique is illustrated on the left , with reference to the sro - terminated surface of cubic srvo@xmath2(100 ) . the haxpes spectra are those shown in fig .
[ f : vb3d ] after subtraction of the o @xmath3 tails
. corresponding ag reference spectra recorded at the same energies are also shown .
the ups and rxes results are from ref . .
] in fig .
[ f : depth ] , the o @xmath3 tails have been subtracted from the spectra , and the results have been compiled with the ups and rxes measurements ref . , which were recorded on the same samples as used in this study . in this figure , the rxes spectra , which represent transitions from the valence band into the v @xmath18 core level ,
have been shifted in energy by 515.2 ev , corresponding to the coherently - screened v @xmath18 haxpes core level in fig .
[ f : shcore ] .
the light gray lines show spectra of polycrystalline ag at 2.2 and 4.2 kev for comparison , underlying the metallicity of the cavo@xmath2 and sr@xmath9ca@xmath9vo@xmath2 measurements . in fig .
[ f : depth ] , the evolution in energy of the lhb from the surface ( ups ) to the bulk ( 4.2 kev haxpes ) can be clearly seen , as well as the agreement in energy of the lhb in the bulk haxpes and the rxes spectra . in summary ,
our results illustrate the evolution in correlated electron behavior from the surface of cavo@xmath2 and sr@xmath9ca@xmath9vo@xmath2 to the bulk .
spectra at 4.2 kev agree very well with truly bulk sensitive rxes measurements , and are considered representative of the bulk correlated electronic structure . on the other hand , spectra at 2.2 kev
are in good agreement with those at soft x - ray energies,@xcite but are not bulk - like and still contain a surface signal of 30 40% .
we also find that reliable surface and bulk components can not be extracted from intermediate spectra , indicating that either the surface layer is not a discrete overlayer to the bulk , or that detected photoelectrons do not originate from an exponential distribution within the sample .
signatures of evolving electron correlations are observed in both the spectral weight of the qp and lhb and in the binding energy of the lhb . at both the surface and in the bulk
, cavo@xmath2 is found to be `` more correlated '' than sr@xmath9ca@xmath9vo@xmath2 : the correlated lhb of cavo@xmath2 exhibits greater spectral weight and is located closer to @xmath8 than in sr@xmath9ca@xmath9vo@xmath2 , in good agreement with predictions from dmft of cavo@xmath2 and srvo@xmath2.@xcite at the surface , the spectral weight of the lhb of both compounds is enhanced , and is @xmath61% closer to @xmath8 than in the bulk , indicating significantly stronger electron correlations at the surface .
this result is in very good agreement with dmft calculations of the sro - terminated srvo@xmath2(001 ) surface , which indicate a 25 40% decrease in binding energy of the lhb at the surface.@xcite
we have presented a detailed haxpes study of the electronic structure of sr@xmath0ca@xmath1vo@xmath2 .
measurements of the valence band have been quantitatively compared with _ ab initio _ calculations within the lda , and are found to be in good agreement after some modest empirical adjustments to the orbital - dependent photoionization cross - sections , with the @xmath4-site hybridization in the o @xmath3 manifold mostly responsible for the differences between compounds .
secondly , our results on the correlated v @xmath5 bands support cavo@xmath2 as being a `` more correlated '' metal than srvo@xmath2 , and illustrate the significant enhancement in electron correlations at the surface compared with the bulk .
the results are found to be in good agreement with dmft predictions of the evolution in the spectral function with @xmath7 and between the surface and the bulk .
finally , we resolve the long - standing controversy of efforts to extract bulk information from pes by showing that photoelectrons of @xmath10 kev are required to obtain truly bulk - representative spectra .
the boston university program is supported in part by the department of energy under grant no .
de - fg02 - 98er45680 . supported also in part by the boston university / university of warwick collaboration fund .
the national synchrotron light source , brookhaven , is supported by the us department of energy under contract no .
de - ac02 - 98ch10886 .
additional support was provided by the national institute of standards and technology . g.b .
gratefully acknowledges financial support from epsrc grant i007210/1 .
j. chakhalian , a. j. millis and j. rondinelli , http://dx.doi.org/10.1038/nmat3225[nat .
mater . * 11 * , 92 ( 2012 ) ] ; p. zubko , s. gariglio , m. gabay , p. ghosez and j .- m .
triscone , http://dx.doi.org/10.1146/annurev-conmatphys-062910-140445[annu .
. condens .
matter phys .
* 2 * , 141 ( 2011 ) ] .
j. laverock , b. chen , k. e. smith , r. p. singh , g. balakrishnan , m. gu , j. w. lu , s. a. wolf , r. m. qiao , w. yang and j. adell , http://dx.doi.org/10.1103/physrevlett.111.047402[phys .
lett . * 111 * , 047402 ( 2013 ) ] .
a. sekiyama , h. fujiwara , s. imada , s. suga , h. eisaki , s. i. uchida , k. takegahara , h. harima , y. saitoh , i. a. nekrasov , g. keller , d. e. kondakov , a. v. kozhevnikov , th .
pruschke , k. held , d. vollhardt and v. i. anisimov , http://dx.doi.org/10.1103/physrevlett.93.156402[phys .
93 * , 156402 ( 2004 ) ] .
t. yoshida , k. tanaka , h. yagi , a. ino , h. eisaki , a. fujimori and z .- x .
shen , http://dx.doi.org/10.1103/physrevlett.95.146404[phys .
95 * , 146404 ( 2005 ) ] ; s. aizaki , t. yoshida , k. yoshimatsu , m. takizawa , m. minohara , s. ideta , a. fujimori , k. gupta , p. mahadevan , k. horiba , h. kumigashira and m. oshima , http://dx.doi.org/10.1103/physrevlett.109.056401[phys .
lett . * 109 * , 056401 ( 2012 ) ] .
r. eguchi , t. kiss , s. tsuda , t. shimojima , t. mizokami , t. yokoya , a. chainani , s. shin , i. h. inoue , t. togashi , s. watanabe , c. q. zhang , c. t. chen , m. arita , k. shimada , h. namatame and m. taniguchi , http://dx.doi.org/10.1103/physrevlett.96.076402[phys . rev .
* 96 * , 076402 ( 2006 ) ] .
h. f. pen , m. abbate , a. fujimori , y. tokura , h. eisaki , s. uchida and g. a. sawatzky , http://dx.doi.org/10.1103/physrevb.59.7422[phys .
b * 59 * , 7422 ( 1999 ) ] ; r. j. o. mossanek , m. abbate , t. yoshida , a. fujimori , y. yoshida , n. shirakawa , h. eisaki , s. kohno , p. t. fonseca and f. c. vicentin , http://dx.doi.org/10.1088/0953-8984/22/9/095601[j .
phys . : condens .
matter * 22 * , 095601 ( 2010 ) ] .
m. takizawa , m. minohara , h. kumigashira , d. toyota , m. oshima , h. wadati , t. yoshida , a. fujimori , m. lippmaa , m. kawasaki , h. koinuma , g. sordi and m. rozenberg , http://dx.doi.org/10.1103/physrevb.80.235104[phys .
b * 80 * , 235104 ( 2009 ) ] .
i. a. nekrasov , k. held , g. keller , d. e. kondakov , th .
pruschke , m. kollar , o. k. andersen , v. i. anisimov and d. vollhardt , http://dx.doi.org/10.1103/physrevb.73.155112[phys .
b * 73 * , 155112 ( 2006 ) ] .
k. byczuk , m. kollar , k. held , y .- f .
yang , i. a. nekrasov , th .
pruschke and d. vollhardt , http://dx.doi.org/10.1038/nphys538[nature phys . * 3 * , 168 ( 2007 ) ] ; m. gatti and m. guzzo , http://dx.doi.org/10.1103/physrevb.87.155147[phys .
b * 87 * , 155147 ( 2013 ) ] ; r. sakuma , ph .
werner and f. aryasetiawan , http://dx.doi.org/10.1103/physrevb.88.235110[phys .
b * 88 * , 235110 ( 2013 ) ] .
k. yoshimatsu , t. okabe , h. kumigashira , s. okamoto , s. aizaki , a. fujimori and m. oshima , http://dx.doi.org/10.1103/physrevlett.104.147601[phys . rev
. lett . * 104 * , 147601 ( 2010 ) ] ; m. gu , j. laverock , b. chen , k. e. smith , s. a. wolf and j. lu , http://dx.doi.org/10.1063/1.4798963[j .
phys . * 113 * , 113704 ( 2013 ) ] .
a. x. gray , c. papp , s. ueda , b. balke , y. yamashita , l. plucinski , j. minr , j. braun , e. r. ylvisaker , c. m. schneider , w. e. pickett , h. ebert , k. kobayashi and c. s. fadley , http://dx.doi.org/10.1038/nmat3089[nat . mater .
* 10 * , 759 ( 2011 ) ] .
s. suga , a. sekiyama , h. fujiwara , y. nakatsu , t. miyamachi , s. imada , p. baltzer , s. niitaka , h. takagi , k. yoshimura , m. yabashi , k. tamasaku , a. higashiya and t. ishikawa , http://dx.doi.org/10.1088/1367-2630/11/7/073025[new j. phys .
* 11 * , 073025 ( 2009 ) ] . | we report hard x - ray photoemission spectroscopy measurements of the electronic structure of the prototypical correlated oxide sr@xmath0ca@xmath1vo@xmath2 . by comparing spectra recorded at different excitation energies ,
we show that 2.2 kev photoelectrons contain a substantial surface component , whereas 4.2 kev photoelectrons originate essentially from the bulk of the sample .
bulk sensitive measurements of the o @xmath3 valence band are found to be in good agreement with _
ab initio _ calculations of the electronic structure with some modest adjustments to the orbital - dependent photoionization cross - sections .
the evolution of the o @xmath3 electronic structure as a function of the sr content is dominated by @xmath4-site hybridization . near the fermi level
, the correlated v @xmath5 hubbard bands are found to evolve in both binding energy and spectral weight as a function of distance from the vacuum interface , revealing higher correlation at the surface than in the bulk . |
while the milky way galaxy is the most explored galaxy , its global three dimensional ( 3d ) structure , such as the spiral arms , has not yet been clarified , because we are embedded in it .
one of the best tracers of the global structure of the galaxy is neutral hydrogen ( h i ) gas , because h i gas is one of the major components of the interstellar matter ( ism ) in spiral galaxies , and the ism is almost optically thin against the 21-cm line of the h i gas .
the earliest study of the global structure of the milky way using h i data was done by @xcite , who presented a face - on h i density map of the whole galaxy .
they reported several hi arms in the milky way : the so - called sagittarius arm , the orion arm , and the perseus arm . later , as more h i surveys were achieved , @xcite , @xcite , and @xcite traced larger - radius h i arm , which is the so - called outer arm , on a longitude velocity diagram .
@xcite analyzed the h i data to make an h i density map for the outer galaxy , and traced two arms having a pitch angle of @xmath5@xmath6 .
burton and te lintel hekkert ( 1986 ) showed three
dimensional h i maps of the outer galaxy for all galactic longitudes , and reported a warped h i disk .
@xcite also presented an h i density distribution of the outer galaxy , and discussed the warped disk and the flaring h i disk .
more recently , hartmann and burton ( 1997 ) archived a deeper and more complete survey with a much higher sensitivity . in the present work
, we tried to convert this latest data cube into an h i distribution . combining with the southern survey data cubes available
, we aimed to derive a 3d map of the whole galaxy .
this paper is the first of a series presenting our analysis of the 3d structure of the ism . in the second paper
, we will present the result for molecular gas . in this paper
we describe the data and our method for deriving the h i distribution of the whole galaxy in sections [ data ] and [ method ] , and report on the resultant h i maps and some properties of h i disk of the galaxy in section [ result ] .
three available h i survey data are used in this paper : the leiden / dwingeloo survey @xcite , parkes survey @xcite , and nrao survey @xcite , which all are archived in the fits format .
we took data within a latitude range of @xmath7 from these data cubes to study the h i distribution .
the leiden / dwingeloo survey @xcite is used for the northern hemisphere , which was obtained with the 25 m dwingeloo radio telescope in the netherlands with a half - power beam width ( hpbw ) of @xmath8 .
the data consists of 21 cm line spectra over the entire sky above declinations of @xmath9 with a grid spacing of @xmath10 . the radial velocity ( @xmath11 ) coverage is from @xmath12 to @xmath13 km s@xmath14 with a velocity resolution of @xmath15 km s@xmath14 .
the r.m.s .
noise level is 0.07 k. these spectra are averaged so that the bin width of latitude becomes @xmath16 in this paper .
the parkes survey @xcite is used for the southern hemisphere , which was obtained with the cisro 18 m telescope in australia with an hpbw of @xmath17 .
it covers galactic longitudes from @xmath18 to @xmath19 with a grid spacing of @xmath20 , and galactic latitude @xmath21 to @xmath22 , with a grid spacing @xmath23 .
the velocity range is 300 km s@xmath14 with a velocity resolution of @xmath24 km s@xmath14 .
the average r.m.s . is 0.27 k. these spectra are also averaged so that the bin width of latitude becomes @xmath16 .
the galactic center h i survey of national radio astronomy observatory is used for the southern hemisphere between @xmath25 and @xmath26 @xcite .
this data cube was obtained with the 140-feet telescope of the nrao , whose hpbw is @xmath27 .
the survey covers the region @xmath28 , @xmath29 , and a grid spacing of @xmath10 in both @xmath30 and @xmath31 .
the velocity range is @xmath32 km s@xmath14 and the velocity resolution 5.5 km s@xmath14 .
the r.m.s .
level is 0.03 k. these spectra are also averaged so that the bin width of latitude becomes @xmath16 . since the 21-cm spectra have only information about the radial velocity of the h i gas , we need to transform the radial velocity , @xmath33 , of the ism in the galaxy into the distance , @xmath34 , in the line of sight .
the relation between the radial velocity , @xmath11 , and the distance , @xmath34 , in the line of sight is obtained using the rotation curve of the galaxy , assuming that it rotates circularly . for the inner rotation curve ,
we use clemens s rotation curve @xcite , which was derived from the massachusetts
stoney brook galactic equator co survey .
clemens s rotation curve was fitted by a polynomial of the form @xmath35 though clemens listed the fitted coefficients , @xmath36 , for the cases of @xmath37 and @xmath38 , we modified them for the case of @xmath39 in order to connect smoothly to the outer rotation curve in this paper
. modified coefficients are listed in table [ coefficient ] .
.modified coefficients of the inner rotation curve . [ cols="^,^,^,^",options="header " , ] for the outer rotation curve , we use a model rotation curve of @xcite , which was determined by fitting to observational constraints ever obtained .
however , because of the small number of data points for the outer rotation curve , the mass model has not been settled yet , and four types of models are presented by @xcite . of the four models , model 2 is adopted in this work , since it has the smallest deviation from the observation among the four . in using this rotation curve
, we adopted @xmath40 kpc , @xmath41 km
s@xmath14 , which are used for the galactic constant in @xcite . the shape of the whole rotation curve used in this paper is shown in figure 1 .
in this section we describe a method for transforming the 21-cm spectra into an h i density distribution . due to a geometrical reason ,
the h i distribution should be calculated separately in several cases : the outer galaxy , the inner galaxy , and near the tangential points .
the radial velocity and the distance in a line of sight can be uniquely determined in the outer galaxy , but the distance can not be determined uniquely from the radial velocity in the inner galaxy because of the near - far ambiguity . since the h i gas has a velocity dispersion , the h i density at the tangential point can not be calculated without considering its effects .
the distance of the h i gas having a radial velocity @xmath11 can be calculated as follows .
assuming that the motion of the galactic ism is purely circular and that the gas away from the galactic plane rotates at the same speed as the gas in the underlying disk , the radial velocity @xmath11 of the ism at galactic longitude @xmath30 , galactic latitude @xmath31 , and galactocentric distance @xmath42 is expressed by @xmath43 \sin{l } \cos{b}.\ ] ] the heliocentric distance @xmath34 of the ism at @xmath44 is obtained by solving the equation @xmath45 in the outer part of the solar orbit ( @xmath46 ) , we have one solution , @xmath47 , because the other solution has a negative value , which is inadequate . in the inner galaxy ( @xmath48 )
we have two adequate solutions , @xmath49 ; that is , we have the two - fold of kinematic distance in the inner galaxy .
however , at a tangential point where @xmath50 , equation ( [ costheory ] ) has only one solution despite the inner galaxy .
hence , we have no ambiguity of distance at this point , and the kinematic distance is uniquely determined for a given radial velocity .
for example , we plot the radial velocity against the heliocentric distance ( @xmath34 ) in the case of @xmath51 in figure 2 .
the radial velocity in the longitude range @xmath52 becomes positive in the inner galaxy , and increases toward the tangential point where it gets maximum . beyond the tangential point it decreases monotonically and becomes negative in the outer galaxy .
the radial velocity for the whole galactic equator is plotted in figure 3 .
the h i density can be estimated as follows .
if the 21-cm line is assumed to be optically thin , the h i column density @xmath53 [ @xmath54 is given by integrating brightness temperature , @xmath55 [ k ] , over radial velocity , @xmath11 [ km s@xmath14 ] , @xmath57 } = 1.82 \times 10^{18 } \int^{v_{{\rm r}1}+ \delta v_{\rm r}}_{v_{{\rm r}1 } } t_{\rm b } dv_{\rm r}.\ ] ] in this work , the velocity width , @xmath58 , is taken to be @xmath59 km s@xmath14 in the case of the leiden / dwingeloo survey , @xmath60 km s@xmath14 in the case of the parkes survey , and 5.5 km s@xmath14 in the case of the nrao survey based on the velocity resolution of the southern survey .
the average volume density at an individual point is given by dividing the column density by the path length , @xmath61 , which contributes emission in the velocity range @xmath62 : @xmath63 = { n_{\rm hi } \over \delta r } = 1.82 \times 10^{18 } t_{\rm b } { \delta v_{\rm r } \over \delta r}.\ ] ] thus , we obtain the h i densities , @xmath64 , in the outer galaxy from the 21-cm spectra @xmath65
. however , the local gas with a high velocity dispersion contributes to the high - altitude emissions .
therefore , the obtained h i distribution includes the local gas in the form of @xmath66 . in this paper
, we assumed that the local gas can be expressed by a linear function of the galactic latitude , @xmath31 , and subtracted it from @xmath67 .
because of this procedure , the diffuse h i gas , which extends to a high galactic altitude @xcite , is subtracted , and can not be detected in the maps presented in this paper . at a tangential point
@xmath68 the kinematic distance can be uniquely determined even for the inner galaxy .
since the h i gas has an intrinsic velocity dispersion , the h i spectra have a higher velocity component than the terminal velocity , @xmath69 . velocity dispersion , @xmath70 ,
is assumed to be 10 km s@xmath14 ; @xcite estimate it to be @xmath71 km s@xmath14 for the first quadrant ( @xmath72 ) , @xmath73 km s@xmath14 for the fourth quadrant ( @xmath74 ) . in a given line of sight , the h i gas having a radial velocity of the range @xmath75 contributes to a higher velocity component than the terminal velocity due the velocity dispersion .
also , equation ( [ tbtonhi ] ) at the tangential points does not work because @xmath76 goes to zero there .
therefore , the volume density of the h i gas giving the radial velocity @xmath75 around the tangential points is calculated by integrating the emission within the velocity range of @xmath77 , as follows : @xmath78 @xmath79 where @xmath80 and @xmath81 denote the heliocentric distances of the points giving radial velocities of @xmath82 .
subscripts 1 and 2 indicate the near and far points , respectively . in the inner galaxy
there are two points giving the same radial velocity in a given line of sight , and thus the observed spectra give the total emission at the near and far points .
we solved this problem of the distance - degeneracy by considering the h i layer thickness .
first of all , we determined the h i layer thickness and its variations using the galactocentric distance .
for the model of the vertical distribution of h i we use spitzer s model @xcite .
@xcite analytically solved the dynamical @xmath83-distribution of stars and ism . assuming that the stars and gas in the galactic disk are isothermal and self - gravitating
, the number density @xmath64 can be expressed as a function of @xmath83 by solving poisson s equation in hydro - dynamical equilibrium .
we then obtain @xmath84 with @xmath85 being @xmath86 where @xmath83 is the height from the galactic equator , @xmath87 is the height where @xmath88 becomes the maximum value , @xmath89 , and @xmath90 is the height where @xmath64 becomes half the value of the maximum , which is half the full width of the half maximum ( fwhm ) . in equation ( [ spi_eq ] )
we assume that @xmath89 and @xmath87 vary with its position , but the scale - height , @xmath90 , depends on the galactocentric distance alone .
we determined the h i layer thickness @xmath90 as a function of the galactocentric distance by fitting the model function to the observed data .
the thickness of the inner h i layer was determined by fitting the model function to the vertical distribution at the tangential point obtained in the previous section ; we substituted the values into three parameters , ( @xmath89 , @xmath87 , and @xmath90 ) in equation ( [ spi_eq ] ) , and chose the best set of parameters to fit to the observation by the least - squares method .
the steps to search for the parameters were @xmath91 of the peak for @xmath89 , 1 pc for @xmath87 and @xmath90 .
on the other hand , the thickness of the outer disk was determined by fitting the model to all points , since there is no distance - ambiguity in the outer galaxy .
the resultant thickness of the h i layer is presented in figure 4 .
for the next stage , the total emission was divided into two h i densities at the near and far points in the line of sight using the thickness of the h i layer .
first , consider an h i gas having a radial velocity @xmath92 for a given longitude , @xmath30 .
this gas locates at corresponding kinematic distances , @xmath93 , and @xmath94 . assuming that the h i gas follows spitzer s model , the h i column density , @xmath95 , is represented as a function of latitude with four parameters : @xmath96 , @xmath97 , @xmath98 , and @xmath99 .
that is , @xmath100 where @xmath101 @xmath102 the four parameters were determined by fitting the observed data by a model using the least - squares method .
finally , the h i density at the near and far points were calculated using @xmath103 and @xmath104 , respectively .
we thus obtain an h i density cube , @xmath105 , from the 21-cm spectra @xmath106 .
finally , the @xmath64 data cube in the @xmath107 space was transformed into @xmath64 in @xmath108 space where @xmath109 are the cartesian coordinates in the galaxy .
the position @xmath110 indicates the galactic center ( g.c . ) , and the sun is situated at @xmath111 .
the space of @xmath107 was also transformed into cylindrical coordinates @xmath112 , where @xmath42 and @xmath113 are the distance from the rotation axis and the azimuth angle around it , respectively .
the angle @xmath113 was taken so that @xmath114 points to the sun and @xmath115 is parallel to @xmath116 .
the derived h i density distribution is presented in the forms of ( 1 ) a face - on h i column density map , ( 2 ) face - on hi volume density maps sliced at several altitudes , and ( 3 ) edge - on h i volume density maps sliced at several azimuthal angles around the galactic center .
_ face - on column density map_ figure 5 shows a face - on column density map of h i gas in the galaxy . according to the @xcite s model , the h i column density , @xmath117 ,
is calculated as @xmath118 \2 . _
face - on volume density maps_ figure 6 shows face - on volume density maps .
each map shows the h i volume densities in a sheet sliced parallel to the galactic equator .
the interval of each map is 500 pc ; @xmath119 pc , @xmath120 pc , @xmath121 pc , @xmath122 pc , and @xmath123 pc .
_ edge - on volume density maps_ figure 7 shows edge - on volume density maps .
each map shows the h i volume densities in a sheet through the galactic center perpendicular to the galactic plane .
individual panels show sections through @xmath124
@xmath125 , @xmath126
@xmath127 , @xmath128
@xmath129 , and @xmath130 @xmath131 .
some h i spiral arms can be traced in the resultant face - on maps , figure 5 and figure 6 . in order to investigate the properties of the spiral arms , the h i column density
is shown in polar coordinates in figure 8 .
the horizontal and vertical axes represent the azimuthal angle @xmath113 and the distance @xmath42 from the galactic center in logarithmic scale , respectively .
each arm on the polar diagram is found to be aligned in a straight line , indicating that the arms are logarithmic spiral arms .
schematic tracers of h i arms are presented in figure 9 .
the largest arm is the so - called outer arm which is traced from @xmath132 to @xmath133 .
the pitch angle is estimated to be @xmath134 .
another prominent arm is traced from @xmath135 to @xmath136 , which is the so - called sagittarius
carina arm .
its pitch angle is also estimated to be @xmath137 .
another arm can be traced from @xmath138 to @xmath139 , which corresponds to the so - called perseus arm .
the pitch angle is @xmath140 .
this arm may apparently connect to the outer arm ( figure 5 ) , but has a clearly different pitch angle .
hence , they may be distinct arms . the sun is located at a relatively h i - rich region between the sagittarius
carina arm and perseus arm , which is called the local arm .
the galactic spiral structure has been studied by many researchers using various methods .
@xcite found three h i spiral arms : the sagittarius , orion , and perseus arms .
although the orion arm may correspond to the local arm , the others are in good agreement with our traced arms .
@xcite found the outer arm , which also agrees with our result .
they estimated the pitch angle of the sagittarius
carina arm to be @xmath141 , which is consistent with our result .
@xcite reported three spiral arms , and estimated the pitch angles to be @xmath142
@xmath6 , which are pretty large compared to our values .
@xcite presented four spiral arms using the hii regions .
they traced the sagittarius
carina arm with a pitch angle @xmath143 @xmath144 , and the perseus arm with s pitch angle of @xmath0 , which are consistent with our results .
however , their scutum crux and norma arms could not be confirmed in our h i map .
@xcite fitted the far and near infrared data form cobe / dirbe by a 2- and 4-armed logarithmic spiral model .
our sagittarius and perseus arms are consistent with their spiral structure .
also , our result that the galactic arms are logarithmic spiral arms confirms their logarithmic spiral model . the radial distribution of the surface density of the h i gas is shown in figure 10 by averaging it azimuthally .
the surface density is low near the galactic center , and is lower than 1.9 @xmath145 within a radius of 6 kpc .
the h i surface density attains the maximum values at radii 7 12 kpc , where it becomes larger than 4 @xmath146 .
then , the h i density decreases toward the outer region .
it amounts to @xmath147 @xmath148 at radius 17 kpc , and therefore the size of the h i disk is 17 kpc according to the definition of @xcite .
this radius of the h i disk is about 1.3times as large as the stellar disk , since the stellar disk is estimated to be 13 kpc assuming @xmath149 kpc @xcite .
this ratio is typical for general spiral galaxies .
the h i mass within a radius of 17 kpc was calculated to be @xmath150 @xmath151 , which corresponds to 1.5% of the total mass predicted by the rotation curve .
figure 7 shows that the h i disk is strongly warping at the outer region , which is consistent with former studies @xcite .
the amplitude of warping is large in the directions @xmath152 and @xmath153 . in the direction @xmath152
, warping starts near @xmath154 kpc , and the h i disk is bending up to @xmath155 kpc height from the galactic plane at @xmath156 kpc . also , in the direction of @xmath153 , warping also starts at @xmath157 kpc .
then , the h i disk is bending down to @xmath158 kpc height from the galactic plane at @xmath156 kpc , and is flopping back to @xmath159 kpc height at @xmath160 kpc .
thus , the galactic warping is found to be asymmetric .
since the size of the stellar disk is thought to be about 13 kpc , the warping starts at the edge of the stellar disk , which follows one of the rules of warping @xcite .
hence , the warping is asymmetric .
this work displays the tilting of the inner disk @xcite . in figure 11 ,
the centroid of the h i disk @xmath87 at @xmath161 kpc is plotted versus azimuthal angle @xmath113 for an instance , which shows that the h i disk is tilted against the @xmath162 plane .
the amplitude of the tilting is @xmath163 pc .
the h i disk is bending up in the direction @xmath164 , and is bending down in the direction @xmath165 , which is the same trend as the warping of the outer disk .
figure 5 shows that the galactic h i disk is not symmetric .
the h i disk is swelled in the direction @xmath166 in the third quadrant .
the radius of the outer h i disk ( defined by the contours at @xmath167 h @xmath168 in the second quadrant ) is @xmath169 kpc , whereas it is @xmath170 kpc in the fourth quadrant .
hence , the galaxy is significantly lopsided in h i. we define the scale - height of the h i disk as the value of @xmath90 ( see subsection [ inner ] ) . the azimuthally averaged scale - height is plotted in figure 4 versus the radius .
figure 4 shows that the scale - height increases with radius , from @xmath163 pc at @xmath171 kpc to @xmath172 pc at @xmath173 kpc .
according to the @xcite s model , the scale - height is calculated as @xmath174{{2 \sigma^2 \over 3 \pi g \rho}}.\ ] ] assuming that velocity dispersion is constant at any radius , the increase in the scale - height implies a decrease in the surface mass density at the galactic plane . in subsection
[ outer ] , the volume density of h i gas ( @xmath175 ) and scale - height ( @xmath90 ) are obtained for each point of the outer galaxy . in figure 12 , we plot the scale - height , @xmath90 , versus h i volume density , @xmath175 , in an annulus of a constant radius of @xmath176 kpc .
figure 12 shows that there is a negative correlation between the h i volume density , @xmath89 , and the scale - height , @xmath90 , which were determined independently of each other in our analysis .
the h i volume density is approximately expressed by @xmath177 by a least - squares fitting . in this work ,
resultant maps were obtained based on some assumptions , which might potentially cause systematic errors in our results .
possible systematic problems are described below .
_ pure circular rotation_ though the distances of h i clouds were calculated based on an assumption that the galaxy rotates in pure circular rotation , @xcite mentioned that the streaming motion predicted by the density wave theory is not negligible .
however , because we are interested in the global distribution of the h i gas , the effects due to the streaming motion were not taken into account . since the streaming amplitudes were estimated to be between 3 and 8 km s@xmath14 , the typical errors in the calculated distances of h i clouds were on the order of 1 kpc .
particularly , in the central region , the non - circular motion may be large , because the galaxy is believed to have a bar whose size is estimated to be about 3 kpc .
hence , the h i distribution in the central 3 kpc may be incredible .
_ dependence on rotation curves_ our resultant h i map is dependent on the rotation curves and there occur two possible systematic errors .
first is the h i disk size .
if the rotation velocity becomes larger values than the one we used , the h i disk will be stretched outwards .
second is the h i contrast .
if a rotation curve has multiple peaks such as those found in rotation curves of @xcite , h i is concentrated in rings at corresponding radii , because the h i density depends on @xmath178 in equation ( [ tbtonhi ] ) .
we used a smooth rotation curve in order to avoid such artificial rings .
_ cylindrical rotation_ we assumed that the gas away from the galactic plane rotates at the same speed as the gas in the underlying disk .
this assumption was often used in former studies @xcite . however
, this assumption may not be adopted for the high - altitude h i gas , because it is known that the rotation velocity at a large distance from the galactic plane becomes slower than that of the gas at underlying disk @xcite based on observations of an external edge - on galaxy , such as ngc 891 in h i line .
_ fitting method_ we determined the h i distribution in the inner galaxy by a fitting method , as mentioned in subsection [ inner ] .
however , this fitting method can not always work precisely . largely deviated h i clouds from the surrounding gas , often seen in resultant maps , are thus doubtful , while a global feature is plausible .
_ dependence of adopted functions for @xmath83-distribution_ while the function @xmath179 was adopted in our analysis , the real gas distribution can not be perfectly represented by it , because the gas is not isothermal @xcite .
we present h i column density maps derived by using other functions , such as a gaussian function and exponential function in figure 12 .
though the resultant maps depend on the adopted functions , the global structure is not strongly affected by the difference of the chosen functions , but the resultant densities varies by a factor of @xmath180 inside the solar circle .
we obtained a three - dimensional h i distribution in the milky way galaxy using the leiden
dwingeloo h i survey @xcite , parkes survey @xcite , and the nrao survey @xcite by combining with @xcite and @xcite s rotation curves .
the distance and density of h i clouds were determined by assuming that the galaxy rotates circularly .
we solved the near far problem in the inner galaxy using a fitting method based on a model of the vertical distribution @xcite .
* acknowledgement * this work was achieved by using the large h i survey data of leiden - dwingeloo northern hemisphere survey , parkes southern hemisphere survey , and nrao galactic center survey .
we would like to thank dr .
t. sawada , dr .
j. koda , and dr .
t. handa for suggestions about writing the manuscript and for fruitful discussions .
binney , j. 1992 , , 30 , 51 broeils , a. h. , and van woerden , h. 1994 , , 107 , 129 burton , w. b. 1972 , , 19 , 51 burton , w. b. and liszt , h. s. 1978 , , 225 , 815 burton , w. b. and liszt , h. s. 1983 , , 52 , 63 burton , w. b. and te lintel hekkert , p. 1986
, , 65 , 427 clemens , d. p. 1985
, , 295 , 422 dehnen , w. and binney , j. 1998 , , 294 , 429 diplas , a. and savage , b. d. 1991 , , 377 , 126 drimmel , r. , spergel , d. n. 2001 , , 556 , 181 georgelin , y. m. , and georgelin , y. p. 1976
, , 49 , 57 hartmann , d. , and burton , w. b. 1997 , atlas of galactic neutral hydrogen , cambridge university press , cambridge .
honma , m. , and sofue , y. 1997 , , 49 , 453 kerr , f. j. 1969 , , 7 , 39 kerr , f. j. , bowers , p. f. , jackson , p. d. & kerr , m. 1986 , , 66 , 373 kulkarni , s. r. , heiles , c. , & blitz , l. 1982 , , 259 , l63 lockman , f. j. 1984 , , 283 , 90 malhotra , s. 1995 , , 448 , 138 oort , j. h. , kerr , f. j. , and westerhout , g. 1958 , , 118 , 379 robin , a. c. , creze , m. , and mohan , v. 1992 , , 400 , l25 sanders , d. b. , solomon , p. m. , & scoville , n. z. 1984 , , 276 , 182 spitzer , l. jr .
1942 , , 95 , 329 swaters , r. a. , sancisi , r. , and van der hulst , j. m. 1997 , , 491 , 140 verschuur , g. l. 1973 , , 22 , 139 weaver , h. 1970 , iau symp .
38 : the spiral structure of our galaxy , 38 , 126 fig.6
. horizontal cross section maps .
each map shows the h i volume densities in a sheet sliced parallel to the galactic equator .
the interval of each map is 500 pc ; @xmath119 pc , @xmath120 pc , @xmath121 pc , @xmath122 pc , and @xmath123 pc .
the contours are drawn at levels of 0.01 , 0.02 , 0.04 , 0.08 , 0.16 , 0.32 , 0.64 atom @xmath183 .
vertical cross section maps .
each map shows the h i volume densities in a sheet through the galactic center perpendicular to the galactic plane .
individual panel shows sections through @xmath124 @xmath125 , @xmath126
@xmath127 , @xmath128
@xmath129 , and @xmath130 @xmath131 . the contours are drawn at levels of 0.01 , 0.02 , 0.04 , 0.08 , 0.16 , 0.32 , 0.64 atom @xmath183 .
h i column density map in polar coordinates . the vertical axis denotes the galactocentric radius on the logarithmic scale , and the horizontal axis denotes the azimuthal angle around the galactic center .
the contours are drawn at levels of 1.0e+20 , 2.0e+20 , 4.0e+20,8.0e+20 , 1.6e+21 atom @xmath168 .
changes in the scale - height at an annulus of constant radius @xmath184 kpc .
the horizontal axis denotes the h i density at the centroid of the h i layer ( @xmath89 ) and the vertical axis denotes the scale - height ( @xmath90 ) .
( a ) h i column density map derived by using a gaussian function to determine the h i distribution in the inner galaxy .
( b ) h i column density map derived with an exponential function .
the contours are drawn at levels of 8.0e+19 , 1.6e+20 , 3.2e+20 , 6.4e+20 , 1.3e+21 atom @xmath168 . | we derived the three - dimensional distribution of h i gas in the milky way galaxy using the latest h i survey data cubes and rotation curves .
the distance of the h i gas was determined by the kinematic distance using a rotation curve .
we solved the near far problem in the inner galaxy by a fitting method which involves introducing a model of vertical h i distribution . in our resultant maps
we could trace three prominent arms : the sagittarius
carina arm , the perseus arm , and the outer arm .
these three arms were found to be logarithmic spiral arms .
the pitch angles of the sagittarius
carina , perseus , and outer arms were estimated to be about @xmath0 , @xmath1 , and @xmath2 , respectively .
the sun is located in a region rich in h i gas between the sagittarius
carina arm and the perseus arm .
the h i disk shows large and asymmetric warping in the outer disk : the h i disk goes up to about 1.5 kpc above the galactic plane in the northern hemisphere , and down to about 1 kpc in the southern hemisphere , which means asymmetric warping .
the inner h i disk is also found to be tilting .
the radius of the h i disk is about 17 kpc and the h i mass within this radius is estimated to be @xmath3 m@xmath4 , which corresponds to 1.5% of the dynamical mass predicted from the rotation curve .
we also found that the h i outskirt is largely swelling in the fourth quadrant , and hence the galaxy is significantly lopsided .
the scale - height of the h i layer increases with the radius , and is correlated with the h i volume density at the centroid of the h i layer . |
this paper concerns the search for a recursive formula for how denominator vectors ( @xmath0-vectors ) change under mutation of _ initial _ seeds .
it is widely expected ( see e.g. ( * ? ? ? * remark 7.7 ) ) that no satisfactory initial - seed - mutation recursion holds in general , and indeed we do not produce one . however , a very nice initial - seed - mutation recursion holds in a varied collection of cluster algebras ( including the case considered in ( * ? ? ?
* remark 7.7 ) ) .
this recursion turns out to be equivalent to a beautiful duality formula in the style of the @xmath1-vectors/@xmath2-vectors dualities of nakanishi and zelevinsky @xcite .
the first thing one notices when looking for such a recursion is that , to understand how denominators change when the initial seed is mutated , one must know something about a related family of integer vectors .
specifically , if @xmath3 is the initial cluster , then the _ negation _ of the @xmath0-vector of a cluster variable @xmath4 is the vector of lowest powers of the @xmath5 occurring in the expression for @xmath4 as a laurent polynomial in @xmath6 .
we define the * _ @xmath7-vector _ * of @xmath4 to be the vector of _ highest _ powers of the @xmath5 occurring in @xmath4 .
our initial - seed - mutation recursion for @xmath0-vectors is equivalent to a description of the @xmath7-vectors in a given cluster in terms of the @xmath0-vectors in the same cluster . in many cases
, one can establish all three formulas by reading off the duality directly from expressions for denominator vectors found in the literature @xcite .
in particular , all of them hold in finite type , in rank two , and more intriguingly , in nontrivial examples arising from marked surfaces .
we conjecture that the initial - seed - mutation recursion holds in the case of source - sink moves in arbitrary cluster algebras .
we prove this conjecture for cluster algebras arising from surfaces .
besides their usefulness in understanding denominator vectors , the @xmath7-vectors may be of independent interest .
a major goal in the study of cluster algebras is to give explicit formulas for the cluster variables .
work in this direction includes realizing cluster variables as `` lambda lengths '' in the surfaces case @xcite , combinatorial formulas in rank two @xcite , in some finite types @xcite , and for some surfaces @xcite , and interpretations in terms of the representation theory of quivers , beginning with @xcite .
short of a complete description of a cluster variable , one might instead describe its newton polytope ( the convex hull of the exponent vectors of the laurent monomials occurring in its laurent expansion ) .
however , as far as the authors are aware , there are no results describing newton polytopes in cases where general formulas for the cluster variables are not known .
( the newton polytope may also be interesting when an explicit formula for the cluster variable is available , as in @xcite . ) together , the @xmath0-vectors and @xmath7-vectors amount to coarse information about newton polytopes , namely their `` bounding boxes . ''
given a polytope @xmath8 in @xmath9 , define the * _ tight bounding box _ * of @xmath8 to be the smallest box @xmath10\times\cdots\times[a_n , b_n]$ ] containing @xmath8 .
( readers who pay attention to bounding boxes of graphics files will find the notion familiar . )
equivalently , for each @xmath11 , the values @xmath12 and @xmath13 are respectively the minimum and maximum of the @xmath14 coordinates of points in @xmath8 .
it is convenient to describe the tight bounding box by specifying the vectors @xmath15 and @xmath16 .
the tight bounding box of the newton polytope of a laurent polynomial @xmath17 in @xmath6 is @xmath10\times\cdots\times[a_n , b_n]$ ] such that @xmath12 is the lowest power of @xmath5 occurring in any laurent monomial of @xmath17 , and @xmath13 is the highest power of @xmath5 occurring .
thus when @xmath4 is a cluster variable written as a laurent polynomial in the initial cluster @xmath3 , the tight bounding box of the newton polytope of @xmath4 is given by the negation of the @xmath0-vector and by the @xmath7-vector .
our notation is in the spirit of @xcite and @xcite .
the notation @xmath18_+$ ] means @xmath19 .
we will apply the operators @xmath20 , @xmath21 , and @xmath22_+$ ] entry - wise to vectors and matrices .
we write @xmath23 for the @xmath24-regular tree with edges properly labeled @xmath25 .
symbols like @xmath26 , @xmath27 , @xmath28 , etc . will stand for vertices of @xmath23 .
we will write @xmath29 to indicate that @xmath26 and @xmath28 are connected by an edge labeled @xmath30 .
the notation @xmath31 stands for the exchange matrix at @xmath26 , where @xmath32 is the exchange matrix at @xmath27 .
similarly , @xmath33 stands for the ( coefficient - free ) cluster variable indexed by @xmath34 in the ( labeled ) seed at @xmath26 , and @xmath35 is the denominator vector of @xmath33 with respect to the cluster at @xmath27 . given a matrix @xmath36 , let @xmath37 be the matrix obtained from @xmath36 by replacing all entries outside the @xmath38 column with zeros .
similarly , @xmath39 is obtained by replacing entries outside the @xmath38 row with zeros .
let @xmath40 be the matrix obtained from the identity matrix by replacing the @xmath41-entry by @xmath42 .
the superscript @xmath43 stands for transpose .
we fix @xmath3 to be the initial cluster ( the cluster at @xmath27 ) .
we write @xmath44 for the matrix whose @xmath45 _ column _ is @xmath35 and @xmath46 for the @xmath47-entry of that matrix .
each @xmath33 is a laurent polynomial in @xmath6 .
( this is the laurent phenomenon , ( * ? ? ?
* theorem 3.1 ) . )
let @xmath48 be the matrix whose @xmath47-entry @xmath49 is the maximum , over all of the ( laurent ) monomials in @xmath33 , of the power of @xmath5 occurring in the monomial .
write @xmath50 for the @xmath45 _ column _ of @xmath48 and call this the @xmath45 * _ @xmath7-vector _ * at @xmath26 .
we now present a duality property for denominator vectors that holds in some cluster algebras , as well as two equivalent properties : an initial - seed - mutation recursion for denominator vectors and a formula for the @xmath51-matrix at a given seed in terms of the @xmath52-matrix at the same seed .
* property d. * ( @xmath52-matrix duality ) . for vertices
@xmath53 , writing @xmath54 as shorthand for @xmath31 , @xmath55 * property r. * ( initial - seed - mutation recursion for @xmath52-matrices ) .
suppose @xmath56 is an edge in @xmath23 and write @xmath57 for @xmath58 .
then @xmath59_+d_t^{b_0;t_0},\,\ , [ -b_0^{k\bullet}]_+d_t^{b_0;t_0 } \right)\\\ ] ] the recursion in property r is not on individual denominator vectors , but rather on an entire cluster of denominator vectors . for @xmath60 ,
the @xmath14 entry of each denominator vector is unchanged , while row @xmath30 of the @xmath52-matrix ( the vector of @xmath38 entries in denominator vectors ) transforms by a recursion similar to the usual recursion ( equation , below ) for how denominator vectors change under mutation .
* property m. * ( @xmath51-matrices in terms of @xmath52-matrices ) . for vertices @xmath53 , @xmath61_+d_t^{b_0;t_0},\,\ , [ -b_0]_+d_t^{b_0;t_0 } \right)\\\ ] ] when property m holds , in particular , the entire tight bounding box of a cluster variable @xmath4 can be determined directly from the denominator vectors of any cluster containing @xmath4 .
our first main result is the following theorem , which we prove in section [ general sec ] .
[ drm ] fix a ( coefficient - free ) cluster pattern @xmath62 .
the following are equivalent : 1 .
property d holds for all @xmath27 and @xmath26 .
property r holds for all @xmath27 , @xmath26 , and @xmath30 .
property m holds for all @xmath27 and @xmath26 .
a natural question is to characterize the cluster algebras in which properties d , r , and m hold . as a start towards answering this question , we prove the following three theorems in section [ special sec ] . in every case ,
the proof is to read off property d using a known formula for the denominator vectors .
[ d rk 2 ] properties d , r , and m holds in any cluster pattern whose exchange matrices are @xmath63 .
[ d finite ] properties d , r , and m hold in any cluster pattern of finite type .
[ d surface ] properties d , r , and m hold for a cluster algebra arising from a marked surface if and only if the marked surface is one of the following . 1 .
[ disk good ] a disk with at most one puncture ( finite types a and d ) .
2 . [ small atilde good ] an annulus with no punctures and one or two marked points on each boundary component ( affine types @xmath64 , @xmath65 , and @xmath66 ) .
3 . [ small dtilde good ] a disk with two punctures and one or two marked points on the boundary component ( affine types @xmath67 and @xmath68 ) .
4 . [ sphere 4 good ] a sphere with four punctures and no boundary components .
[ torus 1 good ] a torus with exactly one marked point ( either one puncture or one boundary component containing one marked point ) . in section [ general sec ]
, we also prove some easier relations on @xmath52-matrices and @xmath51-matrices that hold in general .
the first of these shows that , to understand how @xmath52-matrices transform under mutation of the initial seed , one must understand @xmath51-matrices .
[ md init ] suppose @xmath56 is an edge in @xmath23
. then @xmath69 is obtained by replacing the @xmath38 row of @xmath44 with the @xmath38 row of @xmath48 .
that is , @xmath70 the final - seed mutation recursion on denominator vectors ( * ? ? ?
* ( 7.6)(7.7 ) ) is given in matrix form as follows .
the initial @xmath52-matrix @xmath71 is the negative of the identity matrix , and for each edge @xmath72 in @xmath23 , @xmath73_+,\,\ , d_t^{b_0;t_0}[(-b_t^{b_0;t_0})^{\bullet k}]_+ \right).\ ] ] note that neither matrix inside the max in has any nonzero entry outside the @xmath38 column .
it turns out that @xmath7-vectors satisfy the same recursion , but with different initial conditions . [
usual m ] the initial @xmath51-matrix @xmath74 is the identity matrix . given an edge @xmath72 in @xmath23 , @xmath75_+,\,\ , m_t^{b_0;t_0}[(-b_t^{b_0;t_0})^{\bullet k}]_+ \right).\ ] ] finally , we present some conjectures and results on property r in the context of source - sink moves .
suppose that in the exchange matrix @xmath32 , all entries in row @xmath30 weakly agree in sign .
that is , either all entries in row @xmath30 are nonnegative ( and equivalently all entries in column @xmath30 are nonpositive ) or all entries in row @xmath30 are nonpositive ( and equivalently all entries in column @xmath30 are nonnegative ) . in this case ,
mutation of @xmath32 in direction @xmath30 is often called a * _ source - sink move _ * , referring to the operation on quivers of reversing all arrows at a source or a sink .
we conjecture that property r holds when mutation at @xmath30 is a source - sink move . in this case
, equation has a particularly simple form .
[ source - sink conj ] suppose @xmath56 is an edge in @xmath23 and @xmath57 is @xmath58 .
if all entries in row @xmath30 of @xmath32 weakly agree in sign , then @xmath76_+\\\ ] ] we also make two other closely related conjectures .
let @xmath36 be the * _ cartan companion _ * of @xmath32 , defined by setting @xmath77 for all @xmath78 and @xmath79 for @xmath80 .
then @xmath36 is a ( generalized ) cartan matrix and thus defines a root system and a root lattice in the usual way .
it also defines a ( generalized ) weyl group @xmath81 , generated by simple reflections @xmath82 given by @xmath83 , where the @xmath84 are the simple roots .
if @xmath85 is in the root lattice , then write @xmath86 $ ] for the coefficient of @xmath84 in the simple root coordinates of @xmath85 . then @xmath87=[\beta:\alpha_i]$ ] if @xmath60 and @xmath88=-[\beta:\alpha_k]+\sum_{\ell=1}^n|(b_0)_{k\ell}|[\beta:\alpha_\ell]$ ] . following ( * ? ? ?
* section 2 ) , we define a piecewise linear modification @xmath89 of @xmath90 by setting @xmath91=[\beta:\alpha_i]$ ] if @xmath60 and @xmath92=-[\beta:\alpha_k]+\sum_{\ell=1}^n|(b_0)_{k\ell}|\bigl[[\beta:\alpha_\ell]\bigr]_+$ ] .
we think of @xmath89 as a map on ( certain ) integer vectors by interpreting them as simple root coordinates of vectors in the root lattice .
we also think of @xmath89 as a map on integer matrices by applying it to each _
column_. [ source - sink sigma ]
suppose @xmath56 is an edge in @xmath23 and @xmath57 is @xmath58 . if all entries in row @xmath30 of @xmath32 weakly agree in sign , then @xmath93 . to relate conjecture [ source - sink sigma ] to conjecture [ source - sink conj
] , we quote the following conjecture , which is a significant weakening of ( * ? ? ?
* conjecture 7.4 ) .
we will say a matrix @xmath52 * _ has signed columns _ * if every column of @xmath52 either has all nonnegative entries or all nonpositive entries .
similarly , @xmath52 * _ has signed rows _ * if every row of @xmath52 either has all nonnegative entries or all nonpositive entries .
[ signed columns ] for all @xmath94 , the matrix @xmath44 has signed columns .
conjecture [ signed columns ] is not the same as another weakening of ( * ? ? ?
* conjecture 7.4 ) , namely `` sign - coherence of @xmath0-vectors , '' which asserts that for all @xmath94 , the matrix @xmath44 has signed _
rows_. we prove the following easy proposition in section [ general sec ] .
[ equiv conj ] if conjecture [ signed columns ] holds , then conjectures [ source - sink conj ] and [ source - sink sigma ] are equivalent .
theorems [ d rk 2 ] and [ d finite ] imply conjecture [ source - sink conj ] in the rank - two and finite - type cases , and theorem [ d surface ] implies it for certain surfaces . as evidence in support of the conjecture in general , we prove the following theorem in section [ source - sink surf sec ] .
[ source - sink surf ] conjectures [ source - sink conj ] and [ source - sink sigma ] hold in cluster algebras arising from marked surfaces .
we begin with the proof of proposition [ usual m ] , followed by the proof of proposition [ md init ] . to make the proof of proposition [ usual m ] completely clear , we point out two lemmas about highest powers in multivariate ( laurent ) polynomials . both are completely obvious when looked at in the right way , but otherwise one might convince oneself to worry . given a laurent polynomial @xmath95 , we write @xmath96 for the highest power of @xmath5 occurring in a term of @xmath95 . [ highest poly ] given laurent polynomials @xmath17 and @xmath97 in @xmath6 , we have @xmath98 .
write @xmath99 and @xmath100 such that the @xmath101 and @xmath102 are polynomials in the variables besides @xmath5 and @xmath103 and @xmath104 are nonzero
. then the highest power of @xmath5 in @xmath105 is @xmath106 .
( otherwise @xmath103 and @xmath104 are zero divisors . )
[ highest rat ] suppose @xmath95 is a laurent polynomial over @xmath107 in @xmath6 and @xmath17 and @xmath97 are polynomials in @xmath108 $ ] such that @xmath109 . then @xmath110 . since @xmath95 is a laurent polynomial , we can factor @xmath17 as @xmath111 and @xmath97 as @xmath112 such that @xmath113 is a monomial .
it is immediate that @xmath114 . applying lemma [ highest poly ]
, we have @xmath115 . throughout this proof ,
we omit superscripts @xmath116 .
the first assertion of the proposition is trivial . to establish the second assertion
, we compute @xmath117 , the highest power of @xmath5 occurring in @xmath118 , in terms of @xmath119 . if @xmath120 , then @xmath121 , so @xmath122 as given in the proposition .
if @xmath123 , then the exchange relation ( * ? ? ?
* ( 2.8 ) ) , with trivial coefficients , is @xmath124_+}+\prod_\ell \left(x_{\ell , t}\right)^{[-b_{\ell k;t}]_+}\biggr).\ ] ] write @xmath125 for the expression @xmath126_+}+\prod_\ell \left(x_{\ell , t}\right)^{[-b_{\ell k;t}]_+}$ ] .
each factor @xmath127 in @xmath125 has a subtraction - free expression : an expression as a ratio of two polynomials in @xmath6 with nonnegative coefficients .
therefore each term in @xmath125 has a subtraction - free expression .
write the first term as @xmath128 and the second term as @xmath129 , where @xmath130 , @xmath113 , @xmath131 , and @xmath132 are polynomials with nonnegative coefficients .
the sum @xmath125 is then @xmath133 .
since all of these expressions are subtraction - free , there is no cancellation , so @xmath134 , which equals @xmath135 which in turn equals @xmath136_+}\biggr),m_i\biggl(\prod_\ell ( x_{\ell , t})^{[-b_{\ell k;t}]_+}\biggr)\biggr).\ ] ] returning now to expressions for the @xmath127 as laurent polynomials , lemma [ highest poly ] lets us conclude that @xmath137_+,\sum_\ell m_{i\ell;t}[-b_{\ell k;t}]_+)$ ] .
now , writing @xmath138 as a rational function @xmath139 with @xmath140 and writing @xmath125 as a rational function @xmath141 with @xmath142 , equation lets us write @xmath143 as @xmath144 , so lemmas [ highest poly ] and [ highest rat ] imply that @xmath145_+,\sum_\ell m_{i\ell;t}[-b_{\ell k;t}]_+)$ ] as desired .
the cluster at @xmath146 is obtained from @xmath3 by removing @xmath147 and replacing it with a new cluster variable @xmath148 .
the two are related by @xmath149_+}+\prod_\ell x_\ell^{[-b_{\ell k}]_+}\biggr),\ ] ] where the @xmath150 are entries of @xmath32 . to show that the @xmath38 row of @xmath69 equals the @xmath38 row of @xmath48
, we appeal to the laurent phenomenon to write the cluster variable @xmath33 in the form @xmath151 for some polynomial @xmath152 not divisible by any of the @xmath5 .
we write @xmath153 , where the @xmath154 are polynomials not involving @xmath147 , with @xmath155 .
then lets us write @xmath33 as @xmath156_+}+\prod_\ell x_\ell^{[-b_{\ell k}]_+}\right)}{x'_k}+\cdots+n_p\frac{\left(\prod_\ell x_\ell^{[b_{\ell k}]_+}+\prod_\ell x_\ell^{[-b_{\ell k}]_+}\right)^p}{(x'_k)^p } } { x_1^{d_{1j;t}^{b_0;t_0}}\cdots\left(\frac{\left(\prod_\ell x_\ell^{[b_{\ell k}]_+}+\prod_\ell x_\ell^{[-b_{\ell k}]_+}\right)}{x'_k}\right)^{d_{kj;t}^{b_0;t_0}}\cdots x_n^{d_{nj;t}^{b_0;t_0}}}\ ] ] the numerator of can be factored as @xmath157 times a polynomial not divisible by @xmath148 .
the denominator can be factored as @xmath158 times a polynomial not involving @xmath148 .
we conclude that @xmath159 is @xmath160 .
the latter equals @xmath161 . to show that @xmath69 agrees with @xmath44 outside of row @xmath30 , we fix @xmath60 and consider a subtraction - free expression for @xmath33 .
the laurent phenomenon implies that this expression can be simplified to a laurent polynomial .
the simplification can , if one wishes , be done in two stages , by first factoring out all powers of @xmath5 from the rational expression and then canceling the other factors .
after the first stage , we have written @xmath33 as @xmath162 where @xmath17 and @xmath97 are subtraction - free polynomials not divisible by @xmath5 .
replacing @xmath147 in this expression by the right side of , we find that no additional powers of @xmath5 can be extracted .
( since the right side of is also subtraction - free , we obtain a new subtraction - free expression .
in particular , there can be no cancellation , so a power of @xmath5 can be extracted if and only if it is a factor in every term of the numerator or a factor in every term of the denominator .
but the right side of is not divisible by any nonzero power of @xmath5 . )
we conclude that @xmath163 .
we next prove theorem [ drm ] .
specifically , the theorem follows from the next three propositions , which more carefully specify the relations among the three properties .
[ d implies r ] for a fixed choice of @xmath32 , @xmath27 , @xmath26 and @xmath30 , let @xmath146 be the vertex of @xmath23 such that @xmath56 and write @xmath57 for @xmath58 .
suppose equation holds at @xmath164 and also at @xmath165 .
then equation holds for the same @xmath166 .
we apply at @xmath165 , then , then at @xmath164 .
@xmath167_+,\,\ , d_{t_0}^{(-b_t)^t;t}[(-b_0^t)^{\bullet k}]_+
\right)\right)^t\\ & = j_k\left(d_{t_0}^{(-b_t)^t;t}\right)^t+\ ! \max\left([b_0^{k\bullet}]_+\left(d_{t_0}^{(-b_t)^t;t}\right)^t,\,\ , [ -b_0^{k\bullet}]_+\left(d_{t_0}^{(-b_t)^t;t}\right)^t \right)\\ & = j_kd_t^{b_0;t_0}+ \max\left([b_0^{k\bullet}]_+d_t^{b_0;t_0},\,\ , [ -b_0^{k\bullet}]_+d_t^{b_0;t_0 } \right ) \end{aligned}\ ] ] in the second line , we use the fact that @xmath168 . [ r implies d ] fix a ( coefficient - free ) cluster pattern @xmath169 and vertices @xmath27 and @xmath26 of @xmath23 , connected by edges @xmath170 suppose that , for all @xmath171 , equation holds for the edge @xmath172
. then @xmath173 .
we argue by induction on @xmath174 . for @xmath175 ( i.e. @xmath176 ) , equation
says that the negative of the identity matrix is symmetric .
equation is symmetric in switching @xmath27 and @xmath146 , because @xmath177 .
thus for @xmath178 , we can use for the edge @xmath179 to write @xmath180_+d_t^{b_1;t_1},\,\ , [ -b_1^{k\bullet}]_+d_t^{b_1;t_1 } \bigr)\ ] ] by induction , we rewrite the right side of as @xmath181_+\left(d_{t_1}^{(-b_t)^t;t}\right)^t,\,\ , [ -b_1^{k\bullet}]_+\left(d_{t_1}^{(-b_t)^t;t}\right)^t \right ) \\ & \qquad= \left(d_{t_1}^{(-b_t)^t;t}j_k+ \max\left(d_{t_1}^{(-b_t)^t;t}[(b_1^t)^{\bullet k}]_+,\,\ , d_{t_1}^{(-b_t)^t;t}[(-b_1^t)^{\bullet k}]_+\right)\right)^t .
\end{aligned}\ ] ] by , this is @xmath182 .
[ r iff m ] for a fixed choice of @xmath32 , @xmath27 , @xmath26 and @xmath30 , equation holds if and only if equation holds _ in the @xmath38 row_. equation holds in the @xmath38 row if and only if @xmath183_+d_t^{b_0;t_0},\,\ , [ -b_0^{k\bullet}]_+d_t^{b_0;t_0 } \right).\\\ ] ] this equation is equivalent to in light of proposition [ md init ] .
this completes the proof of theorem [ drm ] . to conclude this section ,
we establish proposition [ equiv conj ] by proving a more detailed statement . recall that a matrix @xmath52 has signed columns if every column of @xmath52 either has all nonnegative entries or all nonpositive entries . [ equiv conj detailed ] suppose @xmath56 is an edge in @xmath23 and @xmath57 is @xmath58 . if @xmath44 has signed columns , then the right side of equals @xmath184 .
let @xmath85 be the vector in the root lattice with simple root coordinates @xmath35 .
for @xmath60 , the @xmath47-entry of the right side of is @xmath86 $ ] .
the @xmath185-entry of the right side of is @xmath186+\bigl[\sum_{\ell=1}^n|(b_0)_{k\ell}|[\beta:\alpha_\ell]\bigr]_+.$ ] by hypothesis , all of the simple root coordinates of @xmath85 weakly agree in sign , so @xmath187\bigr]_+$ ] is @xmath188\bigr]_+$ ] .
thus the right side of is @xmath189 .
we now prove theorems [ d rk 2 ] , [ d finite ] , and [ d surface ] .
the proof of theorem [ d rk 2 ] uses a formula for rank - two denominator vectors due to lee , li , and zelevinsky ( * ? ? ? * ( 1.13 ) ) .
the @xmath190-regular tree @xmath23 is an infinite path .
we label its vertices @xmath191 for @xmath192 , and abbreviate @xmath193 by @xmath194 .
as the situation is very symmetric , it is enough to take @xmath195 with @xmath113 and @xmath131 nonnegative and establish for @xmath196 with @xmath197 . when @xmath198 , the cluster pattern is of finite type and can be checked easily ( if a bit tediously ) by hand .
alternately , one can appeal to theorem [ d finite ] , which we prove below . for @xmath199 ,
the denominator vectors are given by ( * ? ? ?
* ( 1.13 ) ) .
equation is easy when @xmath200 , so we assume @xmath201 .
the labeled cluster associated to the vertex @xmath191 is @xmath202 if @xmath30 is even and @xmath203 if @xmath30 is odd .
if @xmath30 is even , we use ( * ? ? ?
* ( 1.13 ) ) to write @xmath204 where @xmath205 and the @xmath206 are chebyshev polynomials of the second kind .
( in fact , here we do not need to know anything about the @xmath206 except that they are functions of @xmath207 . )
we can similarly use ( * ? ? ?
* ( 1.13 ) ) to write an expression for @xmath208 . since @xmath30 is even @xmath209 , and thus @xmath210 . to apply ( * ? ?
* ( 1.13 ) ) in this case , we must switch the role of @xmath113 and @xmath131 .
when we do so , keeping in mind that we move now in the negative direction , we obtain exactly the transpose of the right side of .
if @xmath30 is odd , we obtain @xmath211 in this case , @xmath212 , so @xmath213 .
noticing that @xmath214 is obtained from @xmath32 by simultaneously swapping the rows and the columns , when we use ( * ? ? ?
* ( 1.13 ) ) to write an expression for @xmath208 , we also swap the rows and columns .
the result is exactly the transpose of the right side of .
the proof of theorem [ d finite ] uses a result of ceballos and pilaud @xcite giving denominator vectors in finite type , with respect to any initial seed , in terms of the compatibility degrees defined at any acyclic seed . in @xcite
, it is shown that in every cluster pattern of finite type , there exists an exchange matrix @xmath32 that is bipartite and whose cartan companion @xmath36 is of finite type .
the cluster variables appearing in the cluster pattern are in bijection with the almost positive roots in the root system for @xmath36 .
given an almost positive root @xmath85 , we will write @xmath215 for the corresponding cluster variable .
there is a * _ compatibility degree _
* @xmath216 defined on almost positive roots encoding some of the combinatorial properties of the cluster algebra .
in particular two cluster variables @xmath217 and @xmath215 belong to the same cluster if and only if the roots @xmath218 and @xmath85 are * _ compatible _ * ( i.e. if their compatibility degree is zero ) . maximal sets of compatible roots
are called * _ ( combinatorial ) clusters _ * and they correspond to the * _ ( algebraic ) clusters _ * in the cluster algebra . in the same paper fomin and
zelevinsky also showed that compatibility degrees encode denominator vectors with respect to the bipartite initial seed .
ceballos and pilaud extended this result dramatically in the following result , which is ( * ? ? ?
* corollary 3.2 ) .
( we follow ceballos and pilaud in modifying the definition of compatibility degree in an inconsequential way in order to make it easier to state the theorem . specifically , we take @xmath219 rather than @xmath220 . )
let @xmath221 be a cluster and let @xmath222 be an almost positive root .
then the @xmath0-vector of @xmath223 with respect to the cluster @xmath224 is given by @xmath225 $ ] .
passing from @xmath32 to @xmath226 has the effect of preserving the signs of entries while transposing the cartan companion @xmath36 .
the almost positive roots for @xmath227 are the almost positive co - roots associated to @xmath36 .
the following is ( * ? ? ?
* proposition 3.3(1 ) ) .
[ compat ck ] if @xmath218 and @xmath85 are almost positive roots and @xmath228 and @xmath229 are the corresponding co - roots , then @xmath230 .
the cluster pattern assigns some algebraic cluster to @xmath27 and some algebraic cluster to @xmath26 , and each of the algebraic clusters is encoded by some combinatorial cluster .
let @xmath221 be the combinatorial cluster at @xmath27 and let @xmath231 be the combinatorial cluster at @xmath26 .
now theorem [ cp cor ] and proposition [ compat ck ] are exactly property d at @xmath27 and @xmath26 .
the proof of theorem [ d surface ] relies on a result of fomin , shapiro , and thurston ( * ? ? ?
* theorem 8.6 ) giving denominator vectors in terms of tagged arcs .
we will assume familiarity with the basic definitions of cluster algebras arising from marked surfaces .
recall that tagged arcs are in bijection with cluster variables and tagged triangulations are in bijection with clusters , except in the case of surfaces with no boundary components , where plain - tagged arcs are in bijection with cluster variables and plain - tagged triangulations are in bijection with clusters .
we write @xmath232 for this bijection .
given tagged arcs @xmath218 and @xmath85 , there is an * _ intersection number _ * @xmath233 such that the following theorem ( * ? ? ?
* theorem 8.6 ) holds .
[ fst denom ] given tagged arcs @xmath218 and @xmath85 and a cluster @xmath3 with @xmath234 , the @xmath14 component of the denominator vector of @xmath215 with respect to the cluster @xmath3 is @xmath233 . in an exchange pattern arising from a marked surface ,
every exchange matrix @xmath54 is skew - symmetric , so @xmath235 .
thus we have the following corollary to theorem [ fst denom ] .
[ d sym ] in an exchange pattern arising from a marked surface , property d holds if and only if the intersection number is symmetric ( i.e. @xmath236 on all tagged arcs @xmath218 and @xmath85 that correspond to cluster variables ) .
the intersection number @xmath233 is defined in ( * ? ? ?
* definition 8.4 ) to be the sum of four quantities @xmath36 , @xmath237 , @xmath238 , and @xmath52
. to define these , we choose @xmath239 and @xmath240 to be non - self - intersecting curves homotopic ( relative to the set of marked points ) to @xmath218 and @xmath85 , and intersecting with each other the minimum possible number of times , transversally each time .
the quantity @xmath36 is the number of intersection points of @xmath239 and @xmath240 ( excluding intersections at their endpoints ) .
the quantity @xmath237 is zero unless @xmath239 is a loop ( i.e. unless the two endpoints of @xmath239 coincide ) .
if @xmath239 is a loop , let @xmath130 be its endpoint .
we number the intersections as @xmath241 in the order they are encountered when following @xmath240 in some direction . for each @xmath242
, there is a unique segment @xmath243 $ ] of @xmath239 having endpoints @xmath130 and @xmath13 and not containing @xmath244 .
there is also a unique segment @xmath245 $ ] of @xmath239 having endpoints @xmath130 and @xmath244 and not containing @xmath13 .
let @xmath246 $ ] be the segment of @xmath240 connecting @xmath13 to @xmath244 .
the quantity @xmath237 is @xmath247 , where @xmath248 is @xmath42 if the segments @xmath243 $ ] , @xmath245 $ ] , and @xmath246 $ ] define a triangle that is contractible and @xmath249 otherwise .
the quantity @xmath238 is zero unless @xmath239 and @xmath240 are equal up to isotopy relative to the set of marked points , in which case @xmath250 .
the quantity @xmath52 is the number of ends of @xmath85 that are incident to an endpoint of @xmath218 and carry , at that endpoint , a different tag from the tag of @xmath218 at that endpoint . the quantities @xmath36 and @xmath238 are patently symmetric in @xmath218 and @xmath85 , so we need not consider them in this section .
it is pointed out in ( * ? ? ?
* example 8.5 ) that @xmath52 can fail to be symmetric .
the quantity @xmath237 can also fail to be symmetric .
examples will occur below .
some immediate observations will be helpful .
[ no loops ] in a surface having no tagged arcs that are loops , @xmath237 is always @xmath251 and @xmath52 is also always symmetric .
[ no punctures ] in a surface having no punctures , @xmath52 is always zero .
[ once - punctured ] in a surface having exactly one puncture and no boundary components , @xmath52 is always zero on tagged arcs corresponding to cluster variables . for the second observation ,
recall that notched tagging may occur only at punctures .
for the third observation , recall that in a surface with exactly one puncture and no boundary components , tagged arcs correspond to cluster variables if only if they are tagged plain . to prove one direction of theorem [ d surface ] , we show that @xmath252 is symmetric in the cases listed in the theorem .
first , recall that a tagged arc may not bound a once - punctured monogon and may not be homotopic to a segment of the boundary between two adjacent marked points .
in particular , there are no loops in a disc with at most one puncture , in the unpunctured annulus with 2 marked points , in the twice - punctured disk with one marked point on its boundary , or in the four - times - punctured sphere .
thus observation [ no loops ] shows that @xmath252 is symmetric in the cases described in and , and in the simplest cases described in and . in the remaining cases described in
, observation [ no punctures ] shows that @xmath52 is always zero .
we are interested in pairs of arcs containing at least one loop ( otherwise @xmath237 is zero in both directions ) .
because an arc may not be homotopic to a boundary segment , there are no loops at a marked point if it is the only marked point on its boundary component . a marked point that is not the only marked point on its boundary component supports exactly one loop
. if there are two marked points on one component and one marked point on the other , then we are in the situation of figure [ two and one ] . in this case , numbering the points as in the figure , the only two loops in the surface are based one at @xmath253 and one at @xmath190 .
the remaining arcs start at @xmath254 , spiral around some number of times and then reach either @xmath253 or @xmath190 .
the only arc that intersects one of the loops more than once is the other loop .
these have @xmath255 in both directions , so @xmath237 is symmetric in this case .
if there are two marked points on each boundary component , the argument is similar and only slightly more complicated .
there are four loops , as illustrated in figure [ two and two ] .
each of the remaining arcs connects a point of one boundary to a point of the other boundary , with some number of spirals .
again , for any of the four loops there is only one other arc intersecting it more than once ; it is the loop based at the other marked point on the same boundary component . for each pair of intersecting loops we calculate @xmath255 in both directions .
we have finished case .
[ annuli ] the remaining case ( a disk with two punctures and two boundary points ) in is similar to the cases in .
there is a loop at each marked point on the boundary but no other loop , as illustrated in figure [ two and punctures ] . in particular , the tagging at these loops is plain , and we see that @xmath52 is symmetric .
there is exactly one arc connecting the two boundary points and four tagged arcs ( all with the same underlying arc ) connecting the two punctures .
the remaining arcs have a boundary point at one endpoint and spiral around the punctures some number of times before ending at one of the punctures , with either tagging there . once again , the only arc that intersects more than once one of the loops is the other loop , and we again have @xmath255 in both directions .
the remaining two cases are described in .
we first consider the once - punctured torus . in this cases ,
@xmath256 by observation [ once - punctured ] , so it remains to show that @xmath237 is symmetric . we will show that in fact @xmath237 is zero on all pairs of arcs .
arcs in the once - punctured torus are well - known to be in bijection with rational slopes , including the infinite slope .
( see , for example , ( * ? ? ?
* section 4 ) . )
each such slope can be written uniquely as a reduced fraction @xmath257 such that @xmath258 and that @xmath259 whenever @xmath260 if we take the universal cover ( the plane @xmath261 ) of the torus mapping each integer point to the puncture , the arc indexed by a slope @xmath257 lifts to a straight line segment connecting the origin to the point @xmath262 .
( the same arc also lifts to all integer translates of that line segment . ) it is now easy to see that @xmath263 for arcs in the once - punctured torus .
for any two arcs @xmath218 and @xmath85 , let @xmath239 and @xmath240 be the curves on the torus obtained by projecting the associated straight line segments in the plane .
this choice of representatives minimizes the number of intersections as can be seen by looking at the universal cover .
let @xmath130 and @xmath241 be the points as in the definition of @xmath237 .
given some @xmath78 between @xmath253 and @xmath264 , concatenate the curves @xmath243 $ ] , @xmath246 $ ] , and @xmath265 $ ] , and consider the lift of the concatenated curve to the plane .
this lifted curve consists of two parallel line segments and one line segment not parallel to the other two .
in particular , it is impossible for the lifted curve to start and end at the same point .
the situation is illustrated in figure [ one pun ] , where a lift of @xmath240 is shown as a solid line , several lifts of @xmath239 are shown as dotted lines and a lift of the three concatenated curves is highlighted . by the standard argument on fundamental groups and universal covers ,
we see that the concatenation of @xmath243 $ ] , @xmath246 $ ] , and @xmath265 $ ] is not a contractible triangle , and we conclude that @xmath263 on @xmath218 and @xmath85 . + the final case for this direction of the proof is the torus with one boundary component and one marked point . in this case , @xmath52 is again zero , this time by observation [ no punctures ] , so we will show that @xmath237 is symmetric .
we think of the boundary component as a `` fat point '' on the torus . with this trick ,
we can again consider lifts of arcs to the plane .
each arc lifts to a curve connecting the origin to an integer point @xmath262 with @xmath130 and @xmath113 satisfying the same conditions as above for the once - punctured torus .
however , for each such @xmath262 , there is a countable collection of arcs connecting the origin to @xmath262 .
specifically , for each integer @xmath30 , the arc may wind @xmath30 times clockwise about the fat origin point before going to @xmath262 .
( negative values of @xmath30 specify counterclockwise spirals . ) since @xmath262 and the origin both project to the same fat point on the torus , the number and direction of spirals at @xmath262 is determined almost uniquely by @xmath30 .
there are two possibilities for each @xmath30 , illustrated in figure [ two possible ] for the case where @xmath266 . for each arc @xmath218
, choosing the right change of basis of the integer lattice , we may as well assume that the lift of @xmath218 connects the origin to the point @xmath267 .
furthermore , there is a homeomorphism from the torus to itself that rotates the fat point and changes the number of spirals of @xmath218 at the origin and at @xmath267 .
rotating a half - integer number of full turns , we can assume @xmath218 lifts to a straight horizontal line segment from @xmath268 to @xmath267 .
possibly reflecting the plane through the horizontal line containing the origin ( to offset the effect of a half - turn ) , we can assume that @xmath218 looks like the solid arc shown in figure [ spirals ] , with the boundary component above the origin in the picture .
now take another arc @xmath85 and consider a lift of @xmath85 connecting the origin to @xmath262 .
since another lift connects @xmath269 to the origin , we may as well take @xmath270 .
up to a reflection in a vertical line , we can assume that the lift of @xmath85 spirals clockwise ( if it spirals at all ) as it leaves the origin .
fixing one possible number of spirals of @xmath85 at @xmath268 and fixing some @xmath262 with @xmath271 , the two possibilities for the lift of @xmath85 are shown as dashed arcs in figure [ spirals ] .
nonzero contributions to @xmath237 can arise only from segments that remain close to the fat point : by the same argument as for the once - punctured torus , the segments that do not stay near the fat point contribute nothing .
therefore it is enough to analyze the intersections of @xmath218 and @xmath85 near the origin . in each of the two possibilities we highlight in figure
[ spirals ] the segments @xmath272 $ ] of @xmath218 and @xmath273 $ ] of @xmath85 giving nonzero contributions . in the pictured examples ,
@xmath237 is symmetric in @xmath218 and @xmath85 .
it is easy to see that the symmetry survives when the number of spirals changes .
the case @xmath274 looks slightly different , but @xmath237 is still symmetric for essentially the same reasons .
( look back , for example at figure [ two possible ] . )
we have proved one direction of theorem [ d surface ] . to prove the other direction , we need to show that @xmath252 fails to be symmetric in certain cases . in each case , the failure of symmetry can be illustrated in a figure .
here , we list the cases and indicate , for each case , the corresponding figure . in some cases , we also include some comments in italics . in each case
, @xmath218 is the solid arc and @xmath85 is the dashed arc ; they intersect in at most two points .
we omit the labeling @xmath275 not to clutter the pictures .
this will complete the proof of theorem [ d surface ] . 1 .
a surface with genus greater than 1 ( figure [ genus ] ) .
+ + + _ we show the genus-2 case .
pairs of edges in the octahedron are identified as indicated by the numbering and the arrows . since all taggings are plain , @xmath256 .
however , @xmath237 is asymmetric ( @xmath42 in one direction and @xmath251 in the other ) .
the marked point shown in the figure is a puncture , but the same example works with the marked point on a boundary component . for higher genus or to have additional punctures , one can start with the surface shown and perform a connected sum , cutting a disk from the interior of the octagon shown .
_ 2 . a torus with 2 or more marked points ( figure [ torus2 ] ) .
_ opposite pairs of edges in the square are identified .
if the marked point at the corners of the square is on a boundary component , then the arcs shown in the left picture of the figure have @xmath256 but @xmath237 is asymmetric ( taking values @xmath251 and @xmath42 ) .
additional punctures and/or boundary components may exist , but the arcs @xmath218 and @xmath85 can always be chosen so that the triangle @xmath276$],@xmath277 $ ] , @xmath278 $ ] is contractible while the triangle @xmath279 $ ] , @xmath280 $ ] , @xmath281 $ ] is not . if the marked point at the corners is a puncture , then the right picture applies . in this case ,
@xmath263 but @xmath52 is asymmetric because one of the arcs is a loop and the other is not .
_ 3 . a sphere with 3 or more boundary components and possibly some punctures ( figure [ sphere3 ] )
+ _ we show a disk with 2 additional boundary components . for the arcs shown , @xmath256 but @xmath237 is asymmetric . again , additional punctures and/or boundary components may exist , but the triangle @xmath279 $ ] , @xmath280 $ ] , @xmath281 $ ] is contractible .
an annulus with one or more punctures ( figure [ annulus1 ] ) . _
@xmath263 and @xmath52 is asymmetric on the arcs shown . _ 5 .
an unpunctured annulus with 3 or more marked points on one of its boundary components ( figure [ annulus0 ] ) .
_ @xmath237 is asymmetric and @xmath256 .
_ 6 . a disk with 3 or more punctures ( figure [ disk3 ] ) . + _
@xmath263 and @xmath52 is asymmetric .
_ 7 . a disk with 2 punctures and 3 or more marked points on the boundary ( figure [ disk2 ] ) .
_ @xmath237 is asymmetric and @xmath256 .
_ 8 . a sphere with 5 or more punctures ( figure [ sphere5 ] ) .
+ + _ we show a local patch of the sphere containing all of the punctures . @xmath263 and
@xmath52 is asymmetric . _
in this section , we prove theorem [ source - sink surf ] , the assertion that conjectures [ source - sink conj ] and [ source - sink sigma ] hold for marked surfaces .
conjecture [ signed columns ] holds for surfaces because the stronger conjecture ( * ? ? ?
* conjecture 7.4 ) for surfaces is an easy consequence of ( * ? ? ?
* theorem 8.6 ) .
thus by proposition [ equiv conj ] , we need only to prove the assertion about conjecture [ source - sink conj ] . in light of theorem [ fst denom ] ,
the task is to prove a certain identity on intersection numbers .
this identity is already known ( as a special case of property r ) for the surfaces listed in theorem [ d surface ] , and it will be convenient in what follows that we need not consider those surfaces .
suppose @xmath218 is a tagged arc in a tagged triangulation @xmath43 and suppose @xmath282 is the arc obtained by flipping @xmath218 in @xmath43 .
we may as well take @xmath43 to be obtained from an ideal triangulation @xmath283 by applying the map @xmath284 of ( * ? ? ?
* definition 7.2 ) to each arc .
( any other tagged triangulation could be obtained from such a triangulation by changing tags , which by definition ( * ? ? ?
* definition 9.6 ) does not affect the associated @xmath237-matrix . ) in particular , @xmath285 .
we will abuse notation and denote by the same greek letters both ideal arcs and their corresponding tagged arcs .
suppose all of the entries in the row of @xmath286 indexed by @xmath218 weakly agree in sign .
because of the symmetry between @xmath44 and @xmath69 in , we may as well assume that all entries in the row of @xmath286 indexed by @xmath218 are nonnegative ; in this case we will say that `` @xmath218 is a source '' alluding to the usual encoding of skew - symmetric exchange matrices by quivers .
let @xmath85 be any other arc .
keeping in mind that equation , like equation before it , is true outside of row @xmath30 by proposition [ md init ] , the task is to prove the following identity : @xmath287 where @xmath288 is the entry of @xmath286 in the row indexed by @xmath218 and column indexed by @xmath222 .
the key observation in our proof is that the entries @xmath288 depends only on how @xmath43 looks locally near @xmath218 .
therefore we begin our analysis by constructing a short list of possible local configurations . to do this we build the surface and the ideal triangulation @xmath43 simultaneously by adjoining * _ puzzle pieces _ * as in ( * ?
* section 4 ) .
in fact , in ( * ? ? ? * section 4 ) , the _ ideal _ triangulation @xmath283 is built from puzzle pieces , but to save a step , we apply the map @xmath284 to the puzzle pieces _ before _ assembling , rather than _ after_. the resulting * _ tagged puzzle pieces _ * are shown in figure [ tagged pieces ] .
we will refer to them ( from left to right in the figure ) as triangle pieces , digon pieces , and monogon pieces .
the external edges of digon pieces are distinguishable ( up to reversing the orientation of the surface ) and we will call them the left edge and the right edge according to how they are pictured in figure [ tagged pieces ] . similarly , the two pairs of internal arcs in a monogon piece are distinguishable , and we will call them the left pair and right pair according to figure [ tagged pieces ] .
puzzle pieces are joined by gluing along their outer edges .
unjoined outer edges become part of the boundary of the surface . in (
* section 4 ) , one specific triangulation is mentioned that can not be obtained from these puzzle pieces , but it is a triangulation of the 4-times punctured sphere , so by theorem [ d surface ] , we need not consider it .
the list of possible local configurations around @xmath218 , given @xmath218 is a source , appears in figure [ configurations ] .
( we leave out the cases where theorem [ d surface ] applies . ) in the figure , areas just outside the boundary are marked in gray .
the curve @xmath218 is labeled , or if two curves might be a source , both of them are labeled @xmath218 . to obtain this list ,
recall that the entries in the row indexed by @xmath218 are determined by the triangles of @xmath283 containing @xmath218 or , if @xmath218 is the folded side of a self - folded triangle , by the triangles containing the other side of that self - folded triangle .
( see ( * ? ? ?
* definition 4.1 ) . ) in particular , if @xmath218 is an internal arc in a digon or monogon piece , the entries in the row indexed by @xmath218 are determined completely within the piece .
both internal arcs in the digon piece are sources if and only if the right external edge of the digon is on the boundary , as shown in the first ( i.e.leftmost ) picture of figure [ configurations ] .
we need not consider the case where both external edges of the digon piece are on the boundary , because theorem [ d surface ] applies to a once - punctured digon . in the monogon piece , both the arcs in the left pair are never sources , and the arcs of the right pair are sources if and only if the external edge of the monogon is on the boundary .
however , we do nt need to consider that case because the surface is a twice - punctured monogon , and theorem [ d surface ] applies . if @xmath218 is the external edge of a monogon piece , then each of the two left internal arcs @xmath222 has @xmath289 , so @xmath218 is not a source .
it remains , then , to consider how external edges of triangle and digon pieces can be sources .
we need to consider two cases .
suppose @xmath218 is an edge in a triangle piece and suppose @xmath222 is the edge reached from @xmath218 by traversing the boundary of the triangle in a counter - clockwise direction .
if @xmath222 is not on the boundary , then the triangle contributes @xmath42 to @xmath288 , so @xmath218 can not be a source unless either @xmath222 is on the boundary or @xmath218 and @xmath222 are also in a second triangle that contributes @xmath253 to @xmath288 .
next suppose @xmath218 is an external edge in a digon piece .
if @xmath218 is the left edge , then each of the two internal arcs @xmath222 has @xmath289 , so @xmath218 is not a source .
if @xmath218 is the right edge , let @xmath222 be the left edge .
as in the triangle case , @xmath218 can not be a source unless either @xmath222 is on the boundary or @xmath218 and @xmath222 are also in a second triangle that contributes @xmath253 to @xmath288 .
putting all of these observations together , we see that we must consider three more possibilities obtained by gluing a triangle or digon piece to another triangle or digon piece
. we can glue two triangle pieces together along one edge with opposite edges of the resulting quadrilateral on the boundary as shown in the second picture in figure [ configurations ] .
conceivably the top and bottom arcs shown in the picture are identified , but we need not consider this case because then the surface is an annulus with two marked points on each boundary component , and theorem [ d surface ] applies .
we can glue two triangle pieces together along two edges , with one of the remaining edges on the boundary as shown in the third picture in figure [ configurations ] .
we can glue a triangle piece along one of its edges to the right edge of a digon piece , with both the left digon edge and the triangle edge counter - clockwise from the glued edge on the boundary , as shown in the fourth and last picture in figure [ configurations ] .
we can glue a triangle piece along two of its edges to the two edges of a digon piece , with the remaining edge of the triangle on the boundary , but we need not consider this case , because the surface is a twice - punctured monogon , and theorem [ d surface ] applies .
we can glue two digon pieces , right edge to right edge , with the remaining two edges on the boundary .
however , we need not consider this case either , because the surface is a twice - punctured digon and theorem [ d surface ] applies . finally , we can glue both edges of a digon piece to both edges of another digon piece , but in this case , we obtain a 3-times punctured sphere , which is explicitly disallowed in the definition of marked surfaces ( * ? ? ?
* definition 2.1 ) .
thus the four configurations in figure [ configurations ] are the only local configurations near arcs that are sources , except in surfaces to which theorem [ d surface ] applies .
we will see that the first and third configurations shown are essentially equivalent for our purposes .
recall from section [ d surface sec ] that @xmath233 is the sum of four quantities @xmath36 , @xmath237 , @xmath238 , and @xmath52 .
as before , @xmath239 and @xmath240 are non - self - intersecting curves homotopic ( relative to the set of marked points ) to @xmath218 and @xmath85 respectively , intersecting with each other the minimum possible number of times , transversally each time .
recall that @xmath263 unless @xmath239 is a loop . in the configurations of figure [ configurations ]
, @xmath239 is never a loop .
furthermore , the quantity @xmath288 is nonzero only if @xmath222 is in a triangle with @xmath218 , and none of the arcs making triangles with @xmath218 is a loop in the configurations of figure [ configurations ] .
therefore , we can ignore @xmath237 in all the calculations of intersection numbers in this section .
recall also that @xmath36 is the number of intersection points of @xmath239 and @xmath240 ( excluding intersections at their endpoints ) , that @xmath290 unless @xmath239 and @xmath240 coincide , in which case @xmath250 , and that @xmath52 is the number of ends of @xmath85 that are incident to an endpoint of @xmath218 and carry , at that endpoint , a different tag from the tag of @xmath218 at that endpoint .
we observe that @xmath233 is invariant under changing all taggings of @xmath218 and of @xmath85 at some puncture .
thus for the first ( leftmost ) picture in figure [ configurations ] , we may as well take @xmath218 to be the arc tagged notched at the puncture .
figure [ labeled configurations ] shows the configurations of figure [ configurations ] with some additional information .
first , the arc @xmath282 , obtained by flipping @xmath218 , is shown and labeled . also , the arcs @xmath222 such that @xmath291 are labeled
there is either one arc @xmath292 , two arcs @xmath292 and @xmath293 , or three arcs @xmath292 , @xmath293 and @xmath294 .
the pictures in figure [ labeled configurations ] are re - ordered in the order we will consider them . we have also redrawn the last configuration more symmetrically .
our task is simplified by several symmetries .
we have already used the symmetry of changing taggings at a puncture . also , any symmetry of a configuration that fixed @xmath218 and @xmath282 or switches @xmath218 and @xmath282 preserves equation .
if the symmetry is orientation - reversing , the absolute value operation in is crucial to the symmetry .
( note however that this absolute value has been omitted in because we took @xmath218 to be a source , not a sink . )
we first consider the left picture in figure [ labeled configurations ] . since each marked point is on the boundary , there are no relevant taggings .
contributions to @xmath233 , @xmath295 , @xmath296 , and @xmath297 occur only when @xmath85 intersects the interior of the quadrilateral .
while @xmath85 may intersect the interior of the quadrilateral a number of times , each intersection can be treated separately .
in such an intersection , @xmath85 may either pass through the quadrilateral , terminate at a vertex of the quadrilateral , or connect two vertices of the quadrilateral .
up to symmetry , as discussed above , there are only three possibilities .
( the relevant symmetry group is the order-@xmath298 dihedral symmetry group of the rectangle shown . )
figure [ squarecase ] shows the possible intersections of @xmath85 ( shown as a dotted line ) with the quadrilateral , along with the contributions to @xmath233 , @xmath295 , @xmath296 , and @xmath297 .
each of these is given in the form @xmath299 .
the quantities @xmath300 and @xmath301 are both @xmath253 . in every case , we see that @xmath302 , and therefore holds .
notice that the second and third pictures of figure [ labeled configurations ] are related by a reflection that switches @xmath218 with @xmath282 . by the symmetry discussed above , we need only consider one of these configurations ; we will work with the third picture .
contributions to @xmath233 , @xmath295 , and @xmath296 only occur when @xmath85 intersects the digon , and again , we can treat each intersection separately .
figure [ twotricase ] shows many of the possible intersections of @xmath85 with the configuration , and shows @xmath233 , @xmath295 , and @xmath296 , again in the form @xmath299 . since @xmath305 ,
the desired relation is @xmath306 , and we see that this relation holds in every case .
not pictured in figure [ twotricase ] are four cases , where @xmath240 coincides with @xmath239 or @xmath307 , with two possible taggings at the point in the center of the digon . in each of these cases , @xmath308 and the @xmath36 terms of @xmath233 and @xmath295 are both zero .
the other terms are also zero , except that one of @xmath233 and @xmath295 has @xmath250 and one of @xmath233 and @xmath295 has @xmath309 . finally , we consider the last picture in figure [ labeled configurations ] .
the two cases where @xmath240 coincides with @xmath239 or @xmath310 are handled analogously to the last case of the quadrilateral condition .
up to symmetry , there are six remaining cases , pictured in figure [ digontricase ] .
once again , @xmath311 for @xmath312 , and the desired relation @xmath313 holds in every case . | cluster variables in a cluster algebra can be parametrized by two families of integer vectors : @xmath0-vectors and @xmath1-vectors .
while @xmath1-vectors satisfy two recursive formulas ( one for initial - seed - mutations and one for final - seed - mutations ) , @xmath0-vectors admit only a final - seed - mutation recursion .
we present an initial - seed - mutation formula for @xmath0-vectors and give two rephrasings of this recursion : one as a duality formula for @xmath0-vectors in the style of the @xmath1-vectors/@xmath2-vectors dualities of nakanishi and zelevinsky , and one as a formula expressing the highest powers in the laurent expansion of a cluster variable in terms of the @xmath0-vectors of any cluster containing it .
we prove that the initial - seed - mutation recursion holds in a varied collection of cluster algebras , but not in general .
we conjecture further that the formula holds _ for source - sink moves on the initial seed _ in an arbitrary cluster algebra , and we prove this conjecture in the case of surfaces . |
nonlinear one - dimensional dissipative systems at their critical points ( like chaos threshold , bifurcation points , etc ) have become one of the vivid areas of research since 1997 , after pioneering work of tsallis et al .
one of the relevant reason for this , in our opinion , was that they were able to constitute one of the first methods of inferring the entropic index @xmath0 of nonextensive entropy @xcite s_q= , and consequently determining the proper thermostatistics which the system under consideration obeys , from the microscopic dynamics of the system .
this method depends on the measure of the divergence of initially nearby orbits .
more precisely , the effect of any uncertainty on initial conditions exhibits , for periodic and chaotic orbits , an exponential temporal evolution as @xmath1 , where @xmath2 is the standard lyapunov exponent , @xmath3 and @xmath4 are the uncertainties at times 0 and @xmath5 ( @xmath6 for periodic cases , and @xmath7 for chaotic ones ) ; whereas at critical points ( e.g. , chaos threshold ) this feature is related to a power - law type such as ( t)= ^1/(1-q_mix ) where @xmath8 is a generalized version of the lyapunov exponent ( here , @xmath9 stands for _ mixing _ and the aim of this notation will be transparent soon ) .
this equation corresponds to power - law growth of the upper bounds of a complex time dependence of @xmath10 and these upper bounds allow us to infer the proper value of @xmath11 for the system under consideration .
let us call this method i of finding the proper @xmath11 of a given dynamical system .
the second method ( method ii ) is based on the geometrical aspects of the critical attractor of the dynamical system at the chaos threshold .
scaling arguments have shown that the appropriate @xmath11 is related to the multifractal structure of the critical dynamical attractor by @xcite = - ( q_mix < 1 ) where @xmath12 and @xmath13 are the end points of the multifractal singularity spectrum of the critical attractor and characterize the most concentrated and most rarefied regions respectively on the attractor .
this fascinating relation relates power - law sensitivity or mixing ( left - hand side ) with purely geometrical quantities ( right - hand side ) and constitutes another independent method for inferring the proper @xmath11 of the dynamical system under study .
finally , a third method ( method iii ) has been proposed for obtaining @xmath11 , dealing with the entropy increase rates @xcite .
this method is the one where nonextensive entropy @xmath14 enters directly and consequently it provides also a justification for the use of notation @xmath0 within the two previous methods .
the procedure for method iii is the following : we first partition the phase space into @xmath15 cells , then we choose one of these cells ( randomly or not ) and locate @xmath16 initial conditions all inside this cell . at any time @xmath5 , we have @xmath17 points in each cell ( naturally @xmath18 ) and thus we can define a set of probabilities through @xmath19 ( @xmath20 ) , which enable us to calculate the entropy @xmath21 . using this procedure , one can define a generalized version of the kolmogorov - sinai entropy as @xmath22 , and the proper @xmath11 value of a given system is the one for which the entropy production @xmath23 is finite . if @xmath24 ( @xmath25 ) , @xmath23 diverges ( vanishes ) .
these three independent methods ( from now on referred to as _ mixing - based approach _ ) have already been tested and verified with numerical calculations for a number of one - dimensional dissipative systems such as the logistic map @xcite , @xmath26-logistic map family @xcite , the circle map @xcite , @xmath26-circular map family @xcite , asymmetric logistic map family @xcite and the single site map @xcite .
quite remarkably , it is found that all three methods give one and the same value for @xmath11 of a given system .
it is worth mentioning here that all these numerical findings have been supported very recently by an analytical work @xcite .
we can now turn our attention to the second approach , which will be referred to as _ relaxation - based approach_. this approach has been developed by one of us and his collaborators @xcite and it is based on investigating numerically the critical temporal evolution of the volume of the phase space occupied by an ensemble of initial conditions spread over the _ entire _ phase space . in that sense , this approach differs from the mixing - based approach , where a set of initial conditions spread in the vicinity of the inflection point is used . in the beginning of time , all cells are occupied and @xmath14 is bounded by the equiprobability as @xmath27 . as time evolves , the number of occupied cells ( say @xmath28 ) starts contracting as @xmath29^{1/(1-q_{rel})}$ ] , here @xmath30 , @xmath31 stands for relaxation and @xmath32 is a characteristic relaxation time . in @xcite ,
@xmath26-logistic map family has been analyzed and @xmath33 values for @xmath26 parameter has been obtained .
the existence of two different classes of @xmath0 values coming from these two different approaches ( namely , mixing - based and relaxation - based approaches ) is also present in the results recently obtained for fully developed turbulence @xcite .
therefore , all these findings naturally force us to ask an intriguing question : are these classes connected to each other or not ?
the affirmative answer to this question came very recently by borges at al .
their procedure goes as follows : beginning with @xmath16 points located inside a single cell ( similar to method iii explained before ) , the dynamics of the system ( for @xmath34 ) leads it to a state with a stationary value of @xmath35 this is the attractor of the system for the particular conditions of @xmath15 , @xmath16 and the initial cell chosen .
the path that the system evolves towards its attractor changes as the initial cell is changed .
some of the initial cells spread the points very fast , increasing the entropy to values much higher than that of the attractor ( there is an overshooting in the time evolution of @xmath35 ) .
consequently , for these specific cases , the attractor is achieved from above .
the procedure consists in observing the rate in which the system goes to its attractor , for those initial cells which yield large overshoots .
specifically , the time evolution of @xmath36 is followed .
the procedure is repeated for increasing values of @xmath15 ( and correspondingly increasing values of @xmath16 ) .
a power - law regime appears and becomes more pronounced for increasing @xmath15 .
the slope ( in a log - log plot ) of the range of the @xmath37 curve that follows a power - law is identified with @xmath38 ( with @xmath39 ) .
it is worth to stress that @xmath40 , i.e. , the rate in which the system reaches its final state depends on the coarse graining ( @xmath15 ) adopted .
the connection between the mixing - based and the relaxation - based approaches is established by the scaling law [ scaling - law ] q_rel()-q_rel(w ) w^-|q_mix| . in this work ,
our aim is to provide further support to this connection by analyzing , for the first time , a two - dimensional dissipative system within this context . to accomplish this task ,
we will focus on the henon map @xcite @xmath41 where @xmath42 and @xmath43 are map parameters .
this map reduces to the standard logistic map when @xmath44 , whereas it becomes conservative when @xmath45 ; in between these two cases , it is a two - dimensional dissipative map .
specifically , we focus on small values of the @xmath43 parameter such as @xmath46 , @xmath47 , @xmath48 , etc .
this map has been studied within the mixing - based approach using the three different methods and it is found that the @xmath11 value for all values of @xmath43 parameter coincides with that of the logistic case , as expected @xcite . here , we first investigate the henon map within the relaxation - based approach and then establish the connection between these two different approaches following the same lines of @xcite .
as introduced in the previous section , for the relaxation approach , we follow the dynamical evolution of an ensemble of initial conditions uniformly distributed over the phase space . in practice ,
a partition of the phase space on @xmath15 cells of equal size is performed and a set of @xmath16 initial copies of system is followed whose initial conditions are uniformly distributed over the entire phase space . our results for the temporal evolution of the volume occupied by the ensemble are given in fig . 1 for various @xmath43 values .
as it is evident , after the power - law contraction of the volume occupied by the ensemble sets up , the slope [ @xmath49 is independent from the @xmath43 values and coincides with that of the logistic map ( namely , @xmath50 hence @xmath51 ) .
now we are at the position to establish the connection between this result and the one obtained in @xcite .
2 shows the temporal evolution of @xmath35 ( @xmath52 ) for different values of @xmath15 , at the edge of chaos for @xmath53 .
( it would be recommended that @xmath54 , but we adopted here @xmath55 , due to restricted computer capabilities .
of course , we compared this case with @xmath56 , and we found that the differences are negligible . ) it becomes evident that the limits @xmath57 and @xmath58 are non - commutable .
if the former is taken first , the system will never relax to a stationary state ( equilibrium is but a particular case ) , which leads to the conclusion that reaching equilibrium depends on the coarse graining adopted , and ultimately , depends on the observer !
equilibrium is only possible with incomplete knowledge ( @xmath59 ) . to arrive to the scaling law stated by eq .
( [ scaling - law ] ) , we must take first the temporal limit , and then @xmath60 .
when we consider the downhill side of these curves and take the slopes of @xmath61 versus time ( as explained earlier ) , then we end up with fig . 3 , where the verification of the scaling ( 4 ) is evident .
on the other hand , the extrapolation value of @xmath62 is computed as @xmath63 , which is different from the logistic case ( @xmath64 ) .
this discrepancy may be originated , most probably , from the fact that @xmath53 value is too large for analyzing , as it is also observed from the deterioration seen in fig . 1 of @xcite ( method i ) for increasing values of the @xmath43 parameter
. a careful investigation of the effect of increasing @xmath43 values to @xmath62 would be a subject of an extensive computational effort .
after the analysis of two - dimensional dissipative henon map at the chaos threshold using three distinct methods ( methods i , ii and iii ) @xcite , it became evident that the scenario , which we called here _ mixing - based approach _
, is valid not only for one - dimensional dissipative map families , but also for two - dimensional cases . at this point , what missing in the literature were ( i ) to study the _ relaxation - based approach _ for a two - dimensional dissipative system making use of the same procedure applied so far to one - dimensional cases @xcite , and ( ii ) to verify the connection ( given by the scaling relation ( 4 ) ) between the mixing - based and relaxation - based approaches for a two - dimensional dissipative system .
this work constitutes the first step along this line and provides some further elements that reinforce the idea of weak mixing .
this work has been supported by the turkish academy of sciences , in the framework of the young scientist award program ( ut / tuba - gebip/2001 - 2 - 20 ) .
* figure 1 * - time evolution of the volume occupied by the ensemble for various values of @xmath43 parameter .
thick line corresponds to a straight line with @xmath50 . + * figure 2 * - time evolution of @xmath35 , with @xmath52 ( log - log scale ) .
the henon map is at its chaos threshold for @xmath53 , that is , @xmath65 .
points correspond to @xmath66 , @xmath67 , @xmath68 , @xmath69 and @xmath70 . +
* figure 3 * - scaling law for the henon map ( eq . ( [ scaling - law ] ) ) , at the chaos threshold with @xmath53 .
correlation coefficient is @xmath71 . to improve the correlation
, it becomes necessary to use greater values of @xmath15 and @xmath16 , that means better computer machinery .
dashed line indicates extrapolated value of @xmath72 ( that is different from the value of the logistic ( @xmath44 ) map : @xmath73 ) . | the mixing properties ( or sensitivity to initial conditions ) and relaxation dynamics of the henon map , together with the connection between these concepts , have been explored numerically at the edge of chaos .
it is found that the results are consistent with those coming from one - dimensional dissipative maps .
this constitutes the first verification of the scenario in two - dimensional cases and obviously reinforces the idea of weak mixing and weak chaos .
+ nonextensive thermostatistics , henon map , dynamical systems |
this appendix contains the proof of theorem 1 : _ every reversible transformation on a system comprising @xmath0 subsystems in boxworld , with @xmath116 measurements at every subsystem each having @xmath2 outcomes , is a permutation of subsystems , followed by local operations . _
the proof idea is the same as in the case @xmath21 : find a particular representation of the vectors @xmath135 and @xmath100 ( corresponding to the previous vectors @xmath26 , @xmath38 , @xmath115 and @xmath32 and @xmath100 ) such that reversible transformations are orthogonal , and such that the scalar products of those vectors yield useful invariants .
we recall that the dual of reversible transformations preserve the cone of effects and so permute extremal effects ( this is lemma 2 applied to this case ) . rather than giving the vectors explicitly ,
we construct them implicitly from the standard @xmath0-simplex in @xmath143 : let @xmath144 be the @xmath11th standard unit vector in @xmath143 , and @xmath145 the center of those vectors , that is @xmath146 . define @xmath147 ,
so that the angles between those vectors ( @xmath140 ) are @xmath148 . by construction , we have @xmath149 , so the vectors are linearly dependent .
the set @xmath137 are then the vectors resulting from embedding normalized versions of the @xmath150s isometrically into @xmath138 .
the first two claimed equalities follow immediately .
the third can be confirmed by computing . ]
@xmath151 which involves only scalar products of the form @xmath152 .
we choose these in a particular way in order to simplify the subsequent argument : the single - site effects will be vectors in @xmath157 .
let @xmath158 denote the @xmath118th standard unit vector in @xmath159 , and @xmath100 be the unit vector on the direct sum space , @xmath160 .
then , define @xmath161 for @xmath162 and @xmath163 .
useful properties of these vectors are given in the following lemma .
since @xmath112 and @xmath113 factorize , we can compute the inner product @xmath173 termwise .
let @xmath11 be the site where @xmath112 and @xmath113 differ , and let @xmath174 and @xmath175 be the corresponding factors .
first , consider the case that @xmath174 and @xmath175 represent different outcomes of the _ same _ measurement .
then , the inner product is the negative value @xmath176 which is the smallest value that can possibly be attained .
hence @xmath177 has the same value , such that @xmath171 and @xmath172 also differ at a single site only ( where they refer to different outcomes of the same measurement ) .
the alternative case is where @xmath174 and @xmath175 represent outcomes of _ different _ measurements .
note that @xmath178 and hence @xmath179 . since @xmath78 preserves @xmath180 and maps the cone bijectively to itself ( cf .
lemma 2 ) , we have @xmath181 , from which it follows that @xmath171 and @xmath172 correspond to outcomes of different measurements on at least one factor .
furthermore , @xmath182 is preserved .
this is the largest value that can be attained subject to the constraint that they represent outcomes of different measurements on at least one factor .
it follows that @xmath171 and @xmath183 are identical in all but one tensor factor .
the proof of theorem 1 is now completed using some properties of the hamming distance .
the list of local effects ( listing the measurement - outcome pairs at the successive sites ) can be used to form a string in @xmath184 , where @xmath185 .
the hamming distance between two strings , @xmath112 and @xmath113 , is defined by @xmath186 lemma [ lem:6 ] shows that if @xmath170 are two arbitrary extremal effects with @xmath187 , then the transformed effects satisfy @xmath188 .
in fact , all reversible operations that preserve hamming distance 1 preserve the hamming distance between all effects .
furthermore , the set of hamming distance preserving transformations can be expressed as combinations of permutations of subsystems and local permutations ( see for example theorem 3.54 of @xcite ) .
we give a proof of this for completeness .
[ lem:8 ] let @xmath189 be a finite alphabet , with @xmath190 the set of length-@xmath0 words .
further , let @xmath191 be the set of permutations of letters and @xmath192 be the group of local transformations of @xmath190 , which act independently at each position . for @xmath202 ,
consider the set , @xmath203 of words of the form @xmath204 i.e. @xmath51 and strings that differ from @xmath51 only at position @xmath11 . because the elements of @xmath203 all have mutual hamming distance equal to one , there must be a function @xmath205 such that @xmath206 .
since @xmath78 is invertible , @xmath205 is a permutation , which may be thought of as an element of @xmath196 . because @xmath207 for all @xmath11 , there is no loss of generality in assuming that @xmath78 takes @xmath203 to itself . employing yet another local operation if necessary , we may even assume that @xmath71 acts like identity on all elements of @xmath203 , and hence on all strings with hamming distance 1 to @xmath51 .
define the _ weight _ of an element @xmath208 to be @xmath209 .
what we have shown so far amounts to the fact that @xmath78 fixes all words of weight zero and one .
next , we prove by induction that @xmath78 fixes the words of any weight @xmath210 ( and hence all of @xmath211 .
assume the claim has been established for weights up to @xmath212 .
if @xmath213 has weight @xmath214 , it is uniquely specified by the @xmath210 words @xmath215 which have weight @xmath216 and hamming distance @xmath217 to @xmath213 ( in fact , any two words @xmath218 from this set are sufficient to specify @xmath213 ) .
but since the weights of all of the @xmath215 , and @xmath219 are preserved by @xmath78 by the induction hypothesis , @xmath78 must fix @xmath213 .
it follows that all transformations in boxworld can be formed by composing subsystem permutations and local permutations .
however , the set of allowed local permutations is further restricted ( hamming distance preservation is a necessary but not sufficient condition on @xmath78 ) : [ lem:9 ] the only local reversible operations allowed in boxworld are relabellings of measurements and their outcomes ( separately for each measurement ) . furthermore , all possible local relabellings are allowed transformations , regardless of the total number of subsystems , @xmath0 .
recall and note that these are the only combinations of extremal effects that sum to @xmath100 : otherwise the identity @xmath100 could be decomposed into a sum involving two effects @xmath135 and @xmath220 with @xmath221 , so there would be a state @xmath51 with @xmath222 , for which @xmath223 , a contradiction .
consider a measurement @xmath118 , then @xmath224 so each member of @xmath225 must correspond to the same measurement .
hence all local reversible adjoint transformations permute the measurements , and , for each measurement separately , permute the outcomes . to see that all permutations are allowed if the number of outcomes @xmath4 is the same for every measurement @xmath118 , note that the representation of @xmath135 as in ( [ eq : repxm ] ) on @xmath226 permits that all those permutations are implemented as allowed linear ( hence orthogonal ) transformations : relabelling the measurements corresponds to permuting the standard unit vectors @xmath158 of @xmath159 ( constituting an @xmath3-dimensional irreducible representation of the symmetric group @xmath227 ) , while relabelling the outcomes corresponds to symmetry transformations of the @xmath228-simplex in @xmath229 with @xmath4 vertices @xmath230 ( a @xmath228-dimensional irreducible representation of @xmath231 ) .
in the case that the number of outcomes @xmath232 depends on the measurement @xmath118 , the vector space carrying the local effects will analogously be @xmath233 .
this allows us to represent the permutation of outcomes linearly , as before , while the permutations of measurements @xmath118 and @xmath234 with @xmath235 correspond to permutations of direct summands .
we have thus proven that every local relabelling transformation @xmath78 is an allowed transformation in boxworld .
it remains to show that @xmath78 is allowed if the single system is coupled to others ( i.e. that @xmath236 is an allowed transformation ) .
( the analogue in quantum theory is to prove complete positivity ) .
this follows from the fact that local relabellings preserve the no - signalling , positivity and normalization constraints , such that @xmath237 maps the no - signalling polytope ( that is , the state space ) onto itself . in the final part of the appendix
, we prove that a state @xmath51 is a pure product state if and only if all the probabilities @xmath238 are either @xmath239 or @xmath28 with respect to extremal effects , @xmath96 ( in the most general case that the number of measurements and outcomes varies from site to site ) .
this has been used in the proof of theorem 2 .
+ [ lem : extremal01 ] let @xmath240 be a normalized state on @xmath0 arbitrary boxworld subsystems ( some of which may be classical ) .
then , @xmath51 is a pure product state if and only if @xmath126 for all extremal effects @xmath241 . if @xmath124 is a product of pure states , then @xmath242 for every @xmath11 , such that @xmath243 is either @xmath239 or @xmath28 .
it remains to prove the converse .
suppose that @xmath51 is any state with @xmath126 for all @xmath88 .
the idea is to construct a pure product state @xmath244 with @xmath245 for all @xmath96 , which proves that @xmath246 . to this end , note that the decomposition of the identity given in has the following generalization : if the @xmath247 are arbitrary local measurement devices , then @xmath248 where the sum is over all outcomes ( the number of outcomes may depend on the measurement ) .
it follows that @xmath249 so exactly one of the addends must be @xmath28 , while all others are @xmath239 .
hence , to every string @xmath250 describing local choices of measurements , there is a unique string of corresponding outcomes @xmath251 such that @xmath252 , while the inner product is @xmath239 for all other outcome combinations .
it remains to show that each @xmath253 only depends on @xmath247 ; in this case , we can construct a state @xmath254 , generating the same probabilities as @xmath51 , factor by factor .
so suppose there were strings @xmath255 and @xmath256 with @xmath257 , but @xmath258 , ( where @xmath259 ) then @xmath260 where the last inequality follows by applying the decomposition of the identity to @xmath261 factors .
this is a contradiction . | a remarkable feature of quantum theory is non - locality ( i.e. the presence of correlations which violate bell inequalities ) .
however , quantum correlations are not maximally non - local , and it is natural to ask whether there are compelling reasons for rejecting theories in which stronger violations are possible . to shed light on this question , we consider post - quantum theories in which maximally non - local states ( non - local boxes ) occur .
it has previously been conjectured that the set of dynamical transformations possible in such theories is severely limited .
we settle the question affirmatively in the case of reversible dynamics , by completely characterizing all such transformations allowed in this setting .
we find that the dynamical group is trivial , in the sense that it is generated solely by local operations and permutations of systems .
in particular , no correlations can ever be created ; non - local boxes can not be prepared from product states ( in other words , no analogues of entangling unitary operations exist ) , and classical computers can efficiently simulate all such processes . _ _
introduction.__|quantum mechanics exhibits the remarkable feature of non - local correlations , as highlighted in bell s seminal paper @xcite .
such correlations have ( up to a few remaining loopholes ) been extensively verified in experiments @xcite . aside from their theoretical importance , non - local correlations can be exploited for technological use : they are vital in entanglement - based quantum key distribution schemes @xcite , for example , where their presence can be used to guarantee security ( see also @xcite for a recent review ) .
while quantum mechanics violates bell inequalities , it does not do so in the maximal possible way .
there are conceivable devices , so - called _ non - local _ or _ popescu - rohrlich boxes _ , that permit even stronger correlations than quantum mechanics does , while respecting the no - signalling principle @xcite . such correlations are not observed in nature and the question arises as to whether other fundamental principles might be violated if they were to exist .
there has already been some progress towards answering this question .
for example , the existence of non - local boxes would lead to some communication complexity problems becoming trivial @xcite , the possibility of oblivious transfer @xcite and the lack of so - called information causality @xcite .
it has also been realized that in a theory in which maximally bell violating correlations emerge , the set of possible dynamical transformations would be severely restricted compared to those allowed in quantum theory @xcite . while a complete classification of the dynamics has remained elusive , it has been shown , for example , that entanglement swapping is impossible @xcite .
furthermore , the question of the computational power of such a theory has been raised @xcite .
we work in the framework of generalized probabilistic theories @xcite , adopting the pragmatic operational view that the physical content of a theory is in the predicted statistics of measurement outcomes given preparations and transformations .
the framework makes minimal assumptions and allows for mathematical rigour .
we consider a system composed of @xmath0 subsystems . to each subsystem one of @xmath1 measurements
may be applied , yielding one of @xmath2 outcomes ( in the following , unless otherwise stated , we assume each subsystem has the same @xmath3 and @xmath4 ) .
the state space contains all non - signalling correlations , corresponding to so - called _ generalized non - signalling theory _
@xcite or , more colloquially , _
boxworld_. our main result ( theorem 1 ) is that ( except in the case @xmath5 which corresponds to classical theory ) the set of reversible transformations in boxworld is trivial : _ all _ such operations are a combination of local operations on a single system ( which correspond to relabellings of measurements and their outcomes ) and permutations of local systems ( which correspond to relabellings of subsystems ) .
this solves the aforementioned open problem concerning the computational power of boxworld in the case of reversible dynamics @xcite .
another interesting consequence is that , in boxworld , measurements and dynamics are necessarily distinct physical processes , in the sense that a measurement can not be seen as a reversible dynamics on the system comprising the state and measurement device ( cf . quantum theory , where the measurement process can be seen as a unitary evolution from the point of view of an external observer ) .
we discuss this further in the final section .
we note that , in the case of a classical - boxworld hybrid system , theorem 1 does not hold|we give an example of a cnot operation on this system at the end of the paper .
however , for all types of system , including those where the number of measurements and outcomes differs among the subsystems , reversible dynamics map pure product states to pure product states|that is , non - local states can not be reversibly prepared from product states .
this is our second main result ( theorem 2 ) .
a geometric intuition behind this result is as follows .
the state space of the theory is a convex polytope , and reversible transformations must map it to itself .
they therefore correspond to symmetries of the polytope .
the polytope is in some way _ stellated _ , with the vertices corresponding to maximally non - local states having a different character from local ones .
they are hence not connected by symmetries of the polytope .
a two - dimensional caricature is shown in figure [ fig:1 ] .
the presentation proceeds as follows .
we begin by formally introducing boxworld , then proceed to give the mathematical framework we will work with .
this is essentially the standard generalized probabilistic framework , as used in @xcite . for clarity of exposition , in the main text
we restrict to the case of two binary measurements ( @xmath6 , @xmath7 ) and give proofs of the main theorems for this case .
the general case is deferred to the appendix , where the proofs are slightly more complicated but analogous .
_ _ boxworld.__|recall that we have a system comprising @xmath0 subsystems and , on each subsystem , one of @xmath3 possible measurements can be applied ( corresponding to different measurement devices ) , yielding one of @xmath4 possible outcomes ( in the most general case , @xmath4 depends on the measurement ) .
the local measurements are denoted @xmath8 . a measurement on the entire system made up of local measurements
can then be described by a string @xmath9 , where @xmath10 specifies the measurement applied to the @xmath11th subsystem .
similarly , the corresponding outcomes are denoted @xmath12 , with @xmath13 .
measurement - outcome pairs are called _ effects _ , e.g. a measurement of @xmath14 giving outcome @xmath15 .
a _ state _ is then a function @xmath16 $ ] , which gives the probability of the effect that @xmath9 is measured and gives outcomes @xmath12 .
more general measurements are possible : a measurement is a collection of effects for which the sum of the outcome probabilities over the collection is 1 when acting on any state .
such measurements include procedures whereby the measurement performed on a particular subsystem depends on the outcomes of previous measurements , convex combinations of such procedures and more @xcite .
however , the statistics of the local measurements @xmath9 are sufficient to uniquely determine the outcome probabilities of all measurements , and hence can be used to specify the state .
this non - trivial assumption is known as the local observability principle @xcite .
furthermore , the subsystems can be spatially separated , and hence we require that @xmath17 satisfies the _ non - signalling _ conditions , i.e.that @xmath18 is independent of @xmath19 .
this implies that the marginal distribution on some set of subsystems is independent of the choice of measurement(s ) on other subsystems .
boxworld is a physical theory whose state space consists of any @xmath17 subject to : ( _ i ) _
@xmath17 takes values in @xmath20 $ ] ; _ ( ii ) _
@xmath17 is normalized in the obvious sense ; and _ ( iii ) _
@xmath17 satisfies the non - signalling conditions ( [ eqn : nos ] )
. the constraints _ ( i ) _ _ ( iii ) _ are such that the state space is a convex polytope , which turns out to have a non - trivial structure .
we first deal with the special case @xmath21 ( the case of so - called _ gbits _
@xcite ) .
the corresponding state spaces ( defined below ) contain interesting non - local states , for example , non - local boxes with maximal bell violating correlations .
we label the measurements @xmath22 and @xmath23 .
* ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) ) .
here , states are represented as vectors embedded in a real vector space .
effects will also be represented as vectors , such that the probabilities of outcomes will be given by inner products between the relevant vectors .
we begin with the case of a single system ( @xmath24 ) .
we choose three linearly independent vectors @xmath25 .
the vector @xmath26 is identified with @xmath27 , which is the effect that the @xmath26 measurement gives outcome @xmath28 .
we define a vector @xmath29 and associate it with @xmath30 .
the prefix @xmath31 may be interpreted as a negation .
lastly , the @xmath32 effect is defined analogously as @xmath33 .
because @xmath34 are linearly independent , for every state @xmath17 , there is a unique vector @xmath35 representing @xmath17 in the sense that @xmath36 it follows that @xmath37 and likewise for @xmath38 .
we will refer to the set @xmath39 as the single - site _ extremal effects _ , for reasons that will become clear below .
( note that the quantum analogue of our effect vectors are projectors , and the inner product is analogous to the hilbert - schmidt scalar product , mapping states , @xmath40 , and projectors , @xmath41 , to probabilities , @xmath42 . )
the @xmath0-subsystem extremal effects @xmath43 are defined to be the tensor products @xmath44 , where @xmath45 ( the reason for this definition is that it recovers the full set of non - signalling distributions for the state space , as will be shown in lemma 1 ) .
we further define the identity on @xmath0 sites , @xmath46 .
a central object is the convex cone @xmath47 generated by @xmath43 @xcite .
this cone is the collection of all vectors which can be written as a linear combination of elements of @xmath43 with non - negative coefficients . for any convex cone @xmath48
, one can associate a _ dual cone _ @xmath49 .
we will identify this with the set of unnormalized states .
our interest in cones and duality stems from the following lemma , which characterizes the state space of boxworld in terms of the cone @xmath47 .
it also implies the well - known result that there are no entangled effects in boxworld . _
_ lemma 1.__|let @xmath50 be the set of vectors @xmath51 in the dual cone @xmath52 which satisfy @xmath53 .
the space of ( normalized ) states in boxworld can be represented by @xmath50 .
we use the notation @xmath54 and @xmath55 for @xmath56 . the vectors @xmath57
will henceforth be called _ states _ ; they satisfy @xmath58 and @xmath59 for all @xmath60 . to every state @xmath51
, we associate a probability distribution @xmath17 via @xmath61 for @xmath62 and @xmath63 .
first we show that every such @xmath17 is a valid non - signalling probability distribution . by definition , @xmath17 is non - negative . to see that it is normalized , note that we can decompose the identity @xmath64 for any choices of @xmath62 , so that @xmath65 to see that @xmath17 is non - signalling , consider @xmath66 which is independent of @xmath19 . to show that every non - signalling distribution has an associated state , note that there is a unique vector @xmath67 such that @xmath68 for @xmath69 ( since these effects span the space ) .
it is then easy to see that the no - signalling property enforces consistency also in the case that @xmath70 for some @xmath11 .
non - negativity and normalization follow directly from the corresponding statements for the probability distribution @xmath17 . _
transformations.__|we now consider transformations in boxworld .
first note that all allowed dynamical transformations in general probabilistic theories ( reversible or not ) are linear|this follows from the fact that they have to respect convex combinations , which correspond to probabilistic mixtures . for
a general proof of this fact see @xcite .
the _ allowed _ transformations , @xmath71 , are defined to be linear maps with the property that for all @xmath57 , @xmath72 .
a transformation is _ reversible _ if both @xmath71 and @xmath73 are allowed transformations .
it follows that a reversible transformation maps the state space @xmath50 bijectively onto itself .
furthermore , since @xmath71 is a linear map , it is also the case that @xmath71 maps extremal states to extremal states .
( more generally , one would only consider a transformation allowed if @xmath74 is also allowed|a condition analogous to complete positivity in quantum theory .
however , our result applies without this additional requirement . )
note that the states @xmath75 themselves do not have a physical meaning|only their scalar products with effects do , i.e. @xmath76 ( which are probabilities ) . since @xmath77 , the dynamics may equivalently be specified by means of the adjoint map @xmath78 .
( in quantum theory , the analogue is passing from the schrdinger to the heisenberg picture . )
reversible transformations , @xmath71 , map the state space bijectively onto itself and , likewise , the adjoint transformations @xmath78 act accordingly on the cone of effects @xmath47 .
_ _ lemma 2.__|adjoint reversible transformations @xmath78 map the cone of effects @xmath47 bijectively onto itself .
moreover , they map the set of _ extremal effects _ , @xmath43 , onto itself .
any vector @xmath79 can be written @xmath80 for some @xmath81 and some @xmath75 .
then , for any outcome @xmath60 , we have @xmath82 since @xmath83 . from the definition of the dual cone
, it follows that @xmath84 ( note that @xmath85 for every closed convex cone @xmath48 ( cf .
@xcite ) ) .
therefore , @xmath78 maps the cone of effects @xmath47 into itself .
the same argument applies to the inverse @xmath86 , hence the cone is mapped bijectively onto itself .
since it is a convex cone , @xmath47 is completely characterized by its extremal rays . by linearity , @xmath78 maps the extremal rays of @xmath47 onto themselves . from the definition of @xmath47 , we know that the cone is the convex hull of the @xmath87 rays formed by all @xmath88 .
it is elementary to check that these are indeed the extremal rays .
therefore , for every @xmath88 , there exists an @xmath89 and a non - negative number @xmath90 such that @xmath91 . to see that @xmath90 must equal @xmath28 , observe that for every @xmath92
, there exist ( product ) states @xmath93 such that @xmath94 and @xmath95 .
since this holds in particular for both @xmath96 and @xmath97 , it follows that @xmath98 . _ _ orthogonal representation of transformations.__|there are @xmath87 extremal effects , and thus @xmath99 permutations acting on @xmath43 .
we go on to show that only a tiny fraction of those is actually realizable in boxworld
. it will be convenient to use a specific representation of @xmath26 , @xmath38 and @xmath100 : @xmath101 _ _ lemma 3.__|with respect to the representation above , it holds that any reversible transformation @xmath71 is orthogonal , i.e. on @xmath0 subsystems , @xmath102 , where @xmath103 is the @xmath104-dimensional identity matrix . first observe that with this choice , @xmath105 and hence ( since @xmath43 factorizes ) @xmath106 .
then , since @xmath78 permutes the extremal effects , @xmath107 . the fact that @xmath71 ( and thus @xmath78 ) is orthogonal , gives rise to a host of invariants .
if one picks any two extremal effects @xmath108 , then clearly their inner product is a conserved quantity : @xmath109 .
however , @xmath110 , where @xmath111 is the _ hamming distance _ between @xmath112 and @xmath113 ,
i.e. the number of places at which @xmath112 and @xmath113 differ . thus the hamming distance of extremal effects is a conserved quantity : @xmath114 .
it is well - known in the theory of error correction @xcite that the set of maps on finite strings which preserve the hamming distance is highly restricted : the group of those maps is generated by local transformations and permutations of sites only ( for a proof , see the appendix ) .
thus @xmath78 acts as such an operation on @xmath43 .
moreover , since the states in @xmath43 span the entire space , the action on this set is sufficient to completely specify @xmath78 .
furthermore , it is straightforward to show that the set of allowed local operations comprises exchanging @xmath26 and @xmath38 ( relabelling measurements ) , exchanging @xmath26 and @xmath115 ( relabelling the outcome upon input @xmath26 ) , exchanging @xmath38 and @xmath32 ( relabelling the outcome upon input @xmath38 ) and combinations thereof ( see lemma [ lem:9 ] in the appendix , for a proof in the general case ) . _ _ main results.__|combining all the previous results proves the following theorem in the special case of @xmath6 measurements with @xmath7 outcomes ( a full proof for all @xmath116 and @xmath2 is given in the appendix ) : _ _ theorem 1.__|every reversible transformation on a system comprising @xmath0 subsystems in boxworld , with @xmath116 measurements at every subsystem each having @xmath2 outcomes , is a permutation of subsystems , followed by local relabellings of measurements and their outcomes .
furthermore , we show the following : _ _ theorem 2.__|in boxworld , every reversible transformation maps pure product states to pure product states . this is true even if the system is coupled to an arbitrary number of classical systems , and if the number of devices and outcomes varies from subsystem to subsystem . before giving the proof , we need to slightly extend the notion of outcome vectors to the general case .
we denote the set of extremal effects for the @xmath11th subsystem by @xmath117 , where @xmath118 labels the measurements ( the number of different @xmath118s may depend on @xmath11 ) and @xmath119 the corresponding outcomes ( the number of different @xmath119s may depend on @xmath118 and on @xmath11 ) .
these vectors satisfy @xmath120 , where @xmath121 represents the identity . except for these relations ,
no linear dependencies occur .
the identity on the full system is then @xmath122 , and the extremal effects are @xmath123 .
the convex cone @xmath47 and the state space @xmath50 are defined analogously to the binary case previously described .
the statements and proofs of lemmas 1 and 2 remain valid in this more general case , hence , in particular , adjoint reversible transformations map @xmath43 onto itself . to complete the proof of theorem 2 , note that a state @xmath75 is a pure product state ( that is , of the form @xmath124 , where all @xmath125 are pure ) if and only if @xmath126 for all extremal effects @xmath88 ( a proof is given in lemma [ lem : extremal01 ] ) .
suppose that @xmath51 is a pure product state and @xmath71 a reversible transformation , then @xmath127 which proves that @xmath128 must also be a pure product state .
note that theorem 1 does not , in general , apply to the case of site - dependent numbers of measurements .
for example , suppose that we have two sites , where the first has two binary measurements , @xmath26 and @xmath38 , and the second allows only a single binary measurement , @xmath129 .
( in other words , a gbit is coupled to a classical bit . )
it is then straightforward to construct a reversible cnot operation , where the classical bit is the control bit .
for example , there is an adjoint reversible transformation that acts as @xmath130 for all @xmath131 . in the case of a system composed of several classical subsystems ,
theorem 1 also does not hold|the dynamics in such a case is non - trivial . nevertheless , theorem 2 does apply to this case|it remains impossible to prepare entangled states from separable ones . _
_ conclusions.__|we have shown that the set of reversible operations in boxworld is trivial : the only possible operations relabel subsystems , local measurements and their outcomes .
in particular , there is no boxworld analogue of an entangling unitary in quantum theory , one can not reversibly prepare non - local states from separable ones , nor perform useful computations reversibly .
in addition , the results have consequences for the interplay between dynamics and measurements in boxworld : suppose we have a system comprising a particle , @xmath96 , and two observers , @xmath132 and @xmath133 , initially in an uncorrelated tripartite product state . in quantum theory , if @xmath132 measures @xmath96 , but @xmath133 does not take part in the interaction , then @xmath133 can model the corresponding dynamics by a unitary transformation on the @xmath134-system .
that is , @xmath133 can view the whole interaction as reversible while retaining the ability to correctly predict the outcome probabilities of any future measurements .
( theories with such a property might be called _
fundamentally reversible_. ) in boxworld , on the other hand , this is not true : @xmath132 s measurement on @xmath96 would have to create correlations between @xmath96 and @xmath132 , but this could never be achieved by a reversible transformation .
hence @xmath133 would have to model the @xmath134-measurement using irreversible dynamics , even if @xmath133 did not take part in the interaction itself .
it would be interesting to extend our result to explore which state spaces are compatible with fundamentally reversible theories in this sense , or with theories that are _ transitive _ , i.e. that every pure state can be reversibly mapped to any other .
this property has been used by hardy as an axiom for quantum theory @xcite .
both conditions seem to strongly restrict the possible geometry of the state space , and an interesting open question is how non - local such theories can be . _
_ acknowledgments.__|we are grateful to howard barnum , volkher scholz , and reinhard werner for interesting discussions and to nicholas harrigan for comments which improved the presentation .
od would like to thank matthew leifer for introducing him to generalized probabilistic theories , and mm would like to thank christopher witte for the same reason and for the quadratic cocktail model of a gbit .
rc and od acknowledge support from the swiss national science foundation ( grant no .
200021 - 119868 ) .
dg s research is supported by the eu ( corner ) .
99 j. bell , cambridge university press , cambridge ( 1987 ) .
a. aspect , nature * 398 * , 189 ( 1999 ) .
a. ekert , phys .
rev .
lett . ,
* 67 * , 661 , ( 1991 ) .
v. scarani _ et .
al . _ ,
arxiv:0802.4155 ( 2008 ) . l. a. khalfin and b. s. tsirelson , world scientific , singapore , 441 ( 1985 ) .
b. s. tsirelson , hadronic journal supp .
* 8*:4 , 329 ( 1993 ) . s. popescu , d. rohrlich , found .
phys .
* 24 * 379 ( 1994 ) .
w. van dam , quant - ph/0501159 ( 2005 ) .
g. brassard , h. buhrman , n. linden , a. mthot , a. tapp and f. unger , phys .
rev .
lett . , * 96 * , 250401 , ( 2006 ) .
h. buhrman _
et al . _ ,
proc .
royal society a * 462 * 1919 ( 2006 ) .
m. pawlowski _
et al . _ ,
arxiv:0905.2292 ( 2009 ) .
j. barrett , phys .
rev .
a , * 75 * , 032304 ( 2007 ) .
t. short , s. popescu , and n. gisin , phys .
rev .
a * 73 * , 012101 ( 2006 ) t. short and j. barrett , arxiv:0909.2601 ( 2009 ) .
h. barnum , j. barrett , m. leifer , and a. wilce , quant - ph/0611295 ( 2006 ) .
l. hardy , quant - ph/0101012 ( 2001 ) .
a. s. holevo , `` probabilistic and statistical aspects of quantum theory '' , north holland ( 1982 ) . g. m. dariano , quant - ph/0701217 ( 2007 ) . c. d. aliprantis , r. tourky , `` cones and duality '' , american mathematical society ( 2007 ) .
p. kaski , p. r. j. stergrd , `` classification algorithms for codes and designs '' , springer ( 2006 ) . |
cosmologists have noticed the @xmath8 50% shortfall of the cosmic baryons at low redshift compared to the baryon content synthesized in the big bang for a long time @xcite .
the cosmic baryon budget has been predicted from a series of cosmic microwave background ( cmb ) surveys @xcite , e.g. @xmath12 from the most recent measurement @xcite . at high redshifts ,
ly@xmath13 forest can account for all baryons expected @xcite .
however , careful census at low redshift revealed that baryons in collapsed objects such as galaxies , galaxy groups and clusters only account for about @xmath8 10% of the cosmic mean of the baryonic content , and @xmath830% of the baryons lie in the gas phase revealed by the ly@xmath13 absorption line . the rest of cosmic baryons are missing . in light of cosmological hydrodynamic simulations , majority of the missing baryons
are believed to reside in a phase called warm hot intergalactic medium ( whim ) with temperature between @xmath14 k @xcite .
the whim in the absorbers of ultraviolet ( uv ) and x - ray absorption lines signatured by ly@xmath13 , ovi , ovii and other highly ionized metals has been extensively investigated ( e.g. * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and the evidence of the whim is accumulating .
a related problem is the galactic missing baryon problem @xcite .
observationally , the baryonic fraction increases monotonically with the mass of the gravitational system , and only galaxy clusters and massive galaxies contain the cosmic mean fraction of baryons @xcite . for low mass galaxies , stars and interstellar medium ( ism ) only account for a small fraction of the baryons , leaving @xmath6 80% of the baryons undetected in observations .
the theoretical concept of a hot ( @xmath8 10@xmath15k ) diffuse corona has been discussed and modeled in numerous literatures @xcite .
these hot coronae could have been shock heated by accretion , or ejected from galaxies into the intergalactic medium ( igm ) or still detained in the galactic dark halo as cgm , depending on the feedback energy in the structure formation process .
there have been a number of tentative detections of the cgm for both early type galaxies and late type galaxies in uv and x - ray band , via both emission and absorption ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
on the other hand , for our own galaxy , the presence of hot halo can explain the lack of hi signature in local dwarf spheroidal galaxies @xcite and the head tail structure of the high velocity clouds ( hvc)@xcite . the x - ray emission measure also puts a constraint on the baryonic content in the cgm of the milky way ( mw ) .
@xcite discussed that an extended hot gaseous halo profile @xcite can agree well with the existent constraints .
although no broad conclusion of the extent and the profile of the cgm has been drawn , the search for the cgm has been a critical issue , and it s possible to detect the cgm through a careful design of observation considering the ubiquity of the cgm in galaxies .
the sunyaev - zeldovich ( sz ) effect @xcite is one of the most promising methods of detecting the missing baryons .
after the reionization epoch , our universe is almost completely ionized after z=6 , and free electrons are prevailing in galaxy clusters as intracluster medium ( icm ) and in between as igm .
they will scatter off cmb photons via inverse compton scattering , and generate secondary cmb anisotropies known as the sz effect .
the sz effect is therefore contributed by all electrons , and it can serve as a tool to probe the electrons in ionized gas and hence the missing baryons in the icm / igm / cgm . due to relatively low density and temperature of the cgm , it s difficult to detect them via the thermal sz ( tsz ) effect ( @xmath16 ) and x - ray emission ( @xmath17 ) .
the kinetic sz ( ksz ) effect , however , directly measures the electron momentum @xmath18 along the line of sight , and is thus less weighted toward hot gas ( icm ) in galaxy groups and clusters . for less massive halos with virial temperature around @xmath19 k ,
the ksz effect turns out to be larger than the tsz effect @xcite . considering the @xmath20% fraction of missing baryons ,
the prevailing igm / cgm around galaxies can become a detectable ksz effect source .
blind detection of the ksz signal from the auto angular correlation alone seems impossible because of the overwhelming contamination of cosmic infrared background ( cib ) and the degeneracy between the ksz effect and the primary cmb .
the state - of - art observations like the planck satellite , south pole telescope ( spt ) , and atacama cosmology telescope ( act ) alone are not able to measure the auto - correlation signal of the kinetic sz effect .
most of these surveys alone can only put an upper limit of the ksz effect @xcite .
taking into account the additional information of the tsz bispectrum can help improve the constraint on the ksz effect @xcite .
however , the results inevitably depend on both the thermal and kinetic sz templates .
since the ksz effect traces the peculiar momenta of the free electrons and thus leaves imprints in the cmb sky , it can be used to constrain cosmological models or cosmological parameters .
recent works have devoted to measuring the ksz effect on the cmb sky at the positions of galaxy clusters or galaxies .
works have focused on measuring the bulk flow of the local universe using galaxy clusters @xcite , however , the existence of the bulk flow remained controversial due to low significance @xcite .
the first solid detection of the ksz effect @xcite reported a 3.8 @xmath7 significance by measuring the pairwise ksz effect of galaxy clusters , taking advantage of the synergy between the boss spectroscopic galaxy group catalogue and the act survey .
alternatively , the ksz effect was proposed to probe the missing baryons @xcite .
progress has been made towards the measurement of the ksz effect using galaxies .
@xcite claimed the detection of the local bulk flow by measuring the ksz effect toward the direction of nearby galaxies , which indicates the existence of hot plasma around galaxies .
moreover , recent works @xcite found the evidence of ksz signature out to about three times the mean virial radius of the central galaxy catalogue ( cgc ) sample , in which they employed both the pairwise momentum estimator and the ksz temperature - velocity field correlation estimator .
these results suggest that the hot plasma around galaxies may be responsible for the missing baryons of galaxies , and produce detectable ksz effect . considering
the cgm may account for a significant fraction of the missing baryons of galaxies , we here present a work to probe the prevailing cgm through the measurement of the kinetic sz effect at the galaxy positions in the cmb map .
previous work by @xcite has theoretically studied the tsz / ksz effect of the cgm , and discussed the detectability of the sz effect of the cgm as well as the constraints on the gas fraction .
we here alternatively propose to measure the ksz effect of the cgm and the detectability of the signal with different profiles .
we try to design an unbiased minimum variance estimator of the ksz effect of the cgm in order to minimize the noise , and then stack the velocity weighted ksz signal of galaxies to further improve the measurement .
the minimum variance estimator involves finding out an optimal filter to filter out the large scale contamination of the primary cmb while keep the ksz signal of cgm conserved .
the design is unprejudicedly constructed for both the icm in galaxy groups and the cgm around galaxies , provided we have the knowledge of the profile of the optical depth profile of the icm / cgm ( e.g. see * ? ? ? * for the application to mock sample of clusters ) . _ in particular , we here focus on the ksz effect of the cgm of milky way(mw)-size halos . _
since the baryonic budget and the radial extent of the cgm are still under debate @xcite , we here confine the cgm to two compelling profiles : the hot extended halo profile ( * ? ? ?
* hereafter mb profile ) and the empirical @xmath0 profile @xcite . in this work , we are mostly interested in the relatively flatter mb profile , as extensively discussed in @xcite concerning the milky way , and it proves a promising profile of the missing baryons of the mw - size halos . we treat the @xmath0 profile as a comparison model , which can give some insight to the upper limit of detectability of the cgm .
the fiducial cosmology we adopt in this work is the standard flat @xmath21cdm cosmology @xcite : @xmath22 , the cosmic baryon fraction @xmath23 .
we focus on the detectability of ska - like galaxy surveys and act - like sz surveys , and discuss the possible application to other surveys in [ sec : discussion ] .
ska will survey @xmath24 hi galaxies in a wide sky area and a deep volume , and will make a fair sample of the reservoir of the cgm of late type galaxies . for mw- size galaxies to be resolved ,
we need sz surveys with the resolution like the act / spt survey .
the organization of the paper is as follows .
we first introduce the models and show the ksz effect of the cgm in [ sec : modeling_cgm_ksz ] , and then raise the formalism and methodology to measure the ksz effect from the cgm in section
[ sec : methodology ] .
we then detail the cross correlating ksz effect with recovered velocity in [ sec : isolating_ksz ] and further apply the matched filter to improve the detection and predict the detectability in
[ sec : results ] .
we present the conclusion and discussion in
[ sec : discussion ] .
when cmb photons travel along the line of sight , the high energy electrons in the intervening medium ( e.g. icm , igm or cgm ) will scatter off the cmb photons and distort the energy spectrum of the cmb via the inverse compton scattering .
this process gives rise to the tsz effect due to the thermal motion of electrons , and to the ksz effect if the scattering medium has a bulk motion with respect to the rest frame of the cmb .
for the cgm of a milky way - size galaxy , the ksz effect is larger than the tsz effect @xcite .
suppose we have a galaxy with its centre at the position @xmath25 in the 2d sky , where we have adopted the flat sky approximation as the galaxy s scale is very small @xcite .
the peculiar momenta of the electrons in the surrounding cgm will induce the kinetic sz effect @xmath26 . in this work
, we assume the ksz signal of the cgm is spherical symmetry around the galaxy centre , and we can always put the galaxy centre @xmath27 at the origin , therefore from herein we write any function of @xmath28 as a function of the radial separation @xmath29 .
the induced ksz signal is proportional to the electron momentum integrated along the line of sight @xmath30 @xmath31 is the optical depth profile of the cgm , where @xmath32 is the scale factor , @xmath33 is the comoving radial distance , @xmath34 is the electron number density profile , and @xmath35 is the thompson cross section .
since for a single galaxy , the peculiar velocity along the line of sight @xmath36 is nearly constant , the ksz induced temperature distortion is proportional to the product of the optical depth and the line of sight velocity @xmath37 here , we have defined @xmath38 as @xmath39 .
the observed optical depth profile is the convolution @xmath40 between the underlying profile @xmath41 and the beam window @xmath42 .
we assume the beam of the sz survey is a gaussian profile @xmath43 , with @xmath44 , and @xmath45 the full width half maximum ( fwhm ) of the beam profile .
accordingly , the beam window in fourier space can be written as @xmath46 . in practice , we calculate the smoothed optical depth profile in fourier space as @xmath47 .
the mass fraction and the underlying spatial distribution of the cgm are still under debate .
self - regulated galaxy simulations @xcite indicate that the baryonic fraction of galaxies can vary substantially .
a reasonable range covers 30 - 100% of the cosmic mean @xmath48 . on the other hand ,
observations have made little advance in constraining the mass fraction of the cgm , which covers an even broader range of the mass fraction @xmath8 10 - 100% @xcite .
the profile of the cgm is also an open question , and the empirical @xmath0 profile is often adopted in the analysis of the observational data , especially for elliptical and massive spiral galaxies @xcite .
however , @xcite found that the mb profile @xcite can account for the missing baryons of the milky way , and is favoured by current observational constraints . in this work , we take the mb profile as a fiducial profile of the cgm for mw - size spiral galaxies , and the @xmath0 model as a comparison profile .
we calculate the ksz effect in the context of future galaxy surveys such as ska .
such survey will detected a huge number of hi field galaxies and is less subject to the contamination of the hot icm .
we expect that most galaxies will be located in halos with mass range @xmath49@xmath50 , and assume that gas within the virial radius is in the hydrostatic equilibrium .
given the redshift distribution and mass distribution of the surveyed galaxies , we can in principle calculate the differential contribution of host halos in certain mass and redshift bin , and obtain the total stacked signal .
though , we will take the example of the typical milky way - size halo of @xmath51 m@xmath52 in this work .
the radial density profile of the mb model ( also known as extended hot halo model ) follows @xmath53 with the profile @xmath54^{3/2}\ , \ ] ] where @xmath55 is the normalized value of gas density at virial radius ( please find details in @xcite ) .
@xmath56 , and @xmath57 is the characteristic radius defined by @xmath58 , where the virial radius @xmath59 kpc and the concentration @xmath60=12 have been adopted for milky way - size host halo . assuming a baryonically closed system , the baryonic mass of the cgm within the virial radius is @xmath61 . in the following , we assume the baryon fraction @xmath62 , and the gas fraction @xmath63 within the virial radius .
we assume the missing mass of hot gas is @xmath64 as in their work , which implies @xmath65 of the cosmic mean . for comparison
, besides the fiducial profile , we also test the detectability of a steeper density profile , the @xmath0 profile , which follows @xmath66 the @xmath0 profile usually fits well to the hot gas around elliptical galaxies @xcite and galaxy groups and clusters@xcite .
a realistic @xmath0 profile would require fitting two parameters , the core radius @xmath67 and the profile index @xmath0 , to the x - ray emission data .
we here assume the index @xmath68 , and fix @xmath69 for a direct comparison between the mb profile and the @xmath0 profile .
we also require a density normalization that integrated gas mass within the virial radius to be @xmath70 .
though the @xmath0 profile is often adopted for the hot gas around elliptical galaxies and icm in galaxy clusters , it has been applied to giant spiral galaxies lately @xcite . whether @xmath0 profile is appropriate for cgm in late type galaxies is still under debate @xcite .
we here choose the @xmath0 profile to demonstrate the detectability of the cgm , which may put constraints on its upper limit .
the profile of the optical depth can be obtained by integrating the gas density alone the sightline : @xmath71 where @xmath72 is the ionization fraction , and @xmath73 is the projection length . at low redshift
, we assume the universe is fully ionized such that @xmath74 .
@xmath75 is the projected distance of the line of sight to the centre of the galaxy , and @xmath76 is the upper range of the integration which means we cut the integration at the virial radius .
@xmath77 is the electron molecular weight , and @xmath78 is the proton mass .
the observed optical depth @xmath79 is the convolution of the underlying optical depth @xmath80 and the beam profile @xmath42 , i.e. @xmath81 . for a mw - size halo with a virial mass @xmath82 , the typical virial radius is @xmath5 arcmin at @xmath83 .
it can be marginally resolved by the act - like sz survey . at higher
redshift , the galaxies are not resolved or just marginally resolved so that most of the signal is within the beam and concentrated in the central pixel regardless of spatial profile .
thus a similar optical depth profile for different cgm profiles is expected after the beam convolution at higher redshifts .
below redshift @xmath83 , the galaxy can be well resolved into several concentric rings of pixels .
the beam convolved ksz profiles of the cgm are shown in fig .
[ fig : tksz ] at several redshifts ( @xmath84 ) . at redshifts
@xmath3 , the ksz profile of the mb model is much flatter and the signal amplitude is much smaller than those of the @xmath0 profile in the central region . we expect less benefit from filtering the mb s profile , because the filter may not well separate the spatial pattern between the ksz effect of extended hot halo and the primary cmb .
profile ( lower panel ) .
the ksz signal of mb model is much flatter in the entire region , and smaller than @xmath0 profile in the central region .
[ fig : tksz],width=307 ]
the ksz effect is immersed in the overwhelming contamination constituted by the thermal sz effect @xmath85 , the primary cmb fluctuation @xmath86 , the foreground cosmic infrared background @xmath87 @xcite , and the instrument noise @xmath88 imposed by the detector .
the noisy temperature distortion profile @xmath89 of the cgm at galaxy position is given by @xmath90 where we have contracted @xmath91 as a term @xmath92 , as they are all noise to the ksz effect .
in realistic surveys , the sky map is pixelated on the receiver . for each pixel in the observed cmb map ,
both sky and the instrument noise contribute to the measured temperature .
all the components except the instrument noise @xmath88 should be smoothed by the telescope beam . to model the instrument noise
, we assume that errors are uncorrelated between pixels and have uniform variance @xmath93 , i.e. @xmath94 @xcite .
the power spectrum of the instrument noise thus can be written as @xmath95 for act - like survey , we adopt @xmath96 arcmin for the beam size , pixel size @xmath97 arcmin , and @xmath98k noise per pixel . these fiducial parameters are adopted through the paper unless specified otherwise .
the problem of probing the ksz effect of the cgm is now equivalent to the problem of optimally solving for @xmath99 in eq .
[ eq : noisy_map ] .
the basic idea is outlined as follows .
the problem requires separating the ksz signal from the noisy cmb map by constructing a correlation , and then beating the noise from all the other components , mainly from the primary cmb .
as shown in previous works , the ksz effect of the cgm is correlated with an effective velocity field @xmath100 , which we can reconstruct from spectroscopic galaxy surveys . the correlation is only contributed by ksz effect , but the noise term is contributed by all the components in the cmb sky plus the instrument noise .
we can then optimally remove the noise with the matched filter .
this would require the knowledge of the optical profile of the underlying ksz signal of the cgm .
finally we stack galaxies of the same host halo mass to boost the signal to noise @xmath101 .
the presence of the overwhelming noise , e.g. the primary cmb , the cib from dusty star forming galaxies , the tsz from the very source of ksz effect , and the potential correlation between the cib and the tsz , makes the direct detection of the ksz effect nearly impossible . in principle , one can separate the signal if there are different spectral dependences between different components of the noise term , by taking advantage of modern cmb surveys with multiple bands . however , the ksz effect itself is not only much smaller than other components , but is also degenerate with the primary cmb , as it lacks spectral dependence . the local velocity or momentum is a good tracer of the underlying ksz effect .
we can therefore weight the temperature distortion profile @xmath102 around galaxies with an effective velocity @xmath103 at the galaxy position .
galaxy density @xmath104 traces the underlying dark matter distribution @xmath105 within an accuracy of a bias factor , i.e. @xmath106 . with the aid of spectroscopic galaxy survey
, we can reconstruct an effective velocity field @xmath103 that traces the underlying peculiar velocity @xmath107 , by assuming that the matter density @xmath105 follows the linear growth theory .
such approximation was widely adopted , and the reconstructed velocity traces the underlying velocity well ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for more discussion ) .
given the galaxy distribution , we can construct the underlying dark matter density field @xmath105 , which should follow the continuity equation of dark matter in the linear approximation @xmath108 where @xmath107 is the underlying peculiar velocity . in the regime of the linear perturbation theory
, one can reconstruct the effective velocity field @xmath103 with @xmath109 where @xmath110 is the hubble parameter at redshift @xmath111 , @xmath112 and @xmath113 is the linear density growth rate .
we have employed that the density follows the linear perturbation theory , and the galaxy is linearly biased against the dark matter with @xmath106 , where @xmath114 is the linear bias of galaxies . from galaxy surveys , we can measure the galaxy bias @xmath114 beforehand .
though there are stochasticities in galaxy bias , studies have shown that the uncertainty in galaxy bias is likely to be at the level of 10% @xcite , and can be negligible at the scales of interest @xcite .
this construction cancels out the cross correlation between the velocity and all the other components except the ksz effect , since they do nt have the characteristic directional dependences and should be uncorrelated with the reconstructed velocity .
details can be found in @xcite .
therefore , we are able to isolate the ksz signal of the cgm . statistically , the line of sight component of the recovered velocity @xmath115 is correlated with underlying velocity field @xmath36 , and is thus correlated with the ksz effect @xmath116 ( momentums ) . so that we obtain the correlation of the recovered velocity and the cmb map @xmath117 \right > \\ & = \langle v_{\parallel } t_{\rm ksz}(\theta ) \rangle \\ & = \langle v_{\parallel } \tilde{u } \rangle \tau_b ( \theta ) . \end{split}\ ] ] in the second line of the equation , we have dropped out @xmath118 .
when we stack the correlation , we assume that all the galaxies share the same optical depth profile @xmath119 ( see the discussion below for the possible uncertainty of @xmath80 profile ) . now that one separates the ksz signal from the noisy map by constructing the correlation estimator between the reconstructed velocity and the underlying ksz signal of the cgm , one need to assess the performance of the estimator .
one of the most important statistics to evaluate the performance is the measurement error , which is known as the variance of the correlation @xmath120 , @xmath121 ^ 2 \rangle \\ & = \langle [ v_\parallel \delta t(\theta)]^2 \rangle- \langle
v_\parallel \delta t(\theta ) \rangle^2 \\ & \simeq
\langle v^2_\parallel \rangle \langle \delta t^2(\theta ) \rangle \ . \end{split}\ ] ] in the last line of eq .
[ eq : stacking_var_nofilter ] , we omit @xmath122 since it s much smaller than the variance @xmath123 .
we also made an assumption that @xmath115 and @xmath102 follow the distribution of gaussian fields .
we then expand the equation using wick theorem , and omit any terms concerning 3-point correlation .
the measurement error is constituted by the product of the velocity dispersion @xmath124 and the temperature variance @xmath125 of each pixel in the cmb survey .
it s clear from eq .
[ eq : stacking_var_nofilter ] that the temperature variance of each pixel is contributed by all the components in the pixel , including the predominant cmb variance .
therefore , cross correlating the cmb map with the reconstructed velocity @xmath115 merely separates the ksz effect of cgm from other components of the noisy cmb map , while the detection is still subject to large measurement error .
as shown later in [ subsec : signal_to_noise_nofilter ] , stacking @xmath8 1 million mw - size halos is necessary to achieve 1 @xmath7 detection .
thus a direct cross correlation between the ksz effect of the cgm with the reconstructed velocity picks out only the ksz effect . to improve the detection ,
we need to further reduce the measurement error by removing the contamination of all the other components .
this involves utilizing the knowledge of the covariance matrix of the noise , and the spatial distribution of the ksz signal of the cgm , which in turn lies in the optical depth profile @xmath119 as in eq .
[ eq : stacking_corr ] . given the profile of @xmath80
, we can design an unbiased estimator @xmath126 of the underlying velocity field @xmath39 with a matched filter @xmath127 to optimally remove the noise mainly from the primary cmb . then by cross correlating @xmath128 one can achieve a much higher s / n .
we now turn to the problem of finding out an optimal estimator of @xmath129 from eq .
[ eq : noisy_map ] .
the primary cmb dominates the variance in each pixel . as shown in
[ subsec : signal_to_noise_nofilter ] , the primary cmb gets the most of the contribution from large scale multipoles , while the ksz signal from the cgm is mainly from the scale of galaxy size .
thus we can take advantage of the different scale dependences and optimally subtract the primary cmb from large scales with an unbiased minimum variance estimator @xmath126 , by employing matched filter ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
an optimal matched filter @xmath130 should satisfy the following requirements : ( 1 ) @xmath126 is an unbiased estimator of the amplitude @xmath39 of the underlying ksz signal ; ( 2 ) the variance between the estimator @xmath126 and underlying signal @xmath39 is minimized . the nominal temperature distortion profile around one galaxy is @xmath131 .
we can construct an effective quantity @xmath126 as an estimate of @xmath39 , @xmath132 where @xmath29 the radial separation from the galaxy centre and @xmath130 is the matched filter .
the first condition requires an unbiased estimator , thus the expectation of @xmath126 should be @xmath39 @xmath133 which means @xmath134 in the above equation , we have used the property that the ensemble average of the noise term @xmath92 vanishes .
the second condition of the estimator requires minimizing of the variance @xmath135 between the estimator and the underlying amplitude .
we can write the variance in fourier space using parseval theorem @xmath136 ^ 2 \rangle - \langle \tilde{u } \rangle ^2 \\ & = \frac{1}{(2\pi)^4 } \langle \left [ \int\!\mathrm{d}^2 l \phi(l ) n(l)^{\ast } \right ] \left[\int \!\mathrm{d}^2 l^{\prime } \phi^{\ast}(l^{\prime } ) n(l^{\prime } ) \right ] \rangle \\ & = \frac{1}{(2\pi)^2 } \int\!\mathrm{d}^2 l \int\!\mathrm{d}^2 l^{\prime } \delta^d(l - l^{\prime } ) ( c_l b_l^2+\sigma_p^2 \theta_p^2 ) \phi(l ) \phi^{\ast } ( l^{\prime } ) \\ & = \frac{1}{(2\pi)^2}\int\!\mathrm{d}^2 l ( c_l b_l^2 + \sigma_p^2 \theta_p^2)\ , \left|\phi(l)\right|^2 \ , \end{split}\ ] ] in the following transform convention @xmath137 which has been reduced to the zero - order hankel transform due to circular symmetry .
@xmath138 is the angular power spectrum of the primary cmb , for which we have used the camb code @xcite .
we have assumed that there s no correlation between the primary cmb and the instrument noise .
therefore , solving the matched filter can be reduced to finding the local extremum of a function that is subject to an equality constraint .
specifically , here we need to solve for @xmath127 and a lagrangian multiplier @xmath139 that minimize @xmath140 , \ ] ] with the constraint @xmath141 hence the following equation is demanded @xmath142 = 0 \ .\ ] ] we thus have the solution by solving eq .
[ eq : unbiased_fourier ] and [ eq : lambda_fourier ] @xmath143 where the normalization @xmath139 is given by @xmath144^{-1 } \ .\ ] ] we then apply this filter to the noisy cmb map , and the cross correlation signal would be @xmath145 ^ 2 \rangle \\ & = \langle v_\parallel \tilde{u } \rangle \left ( \int d^2\theta \phi(\theta ) \tau_b(\theta)\right ) \\ & = \langle v_\parallel u_\parallel \rangle t_{\rm cmb}/c . \end{split}\ ] ] by constructing the correlation of the optimal estimator and the recovered velocity , we expect not only to get the ksz effect isolated from the noisy map , but also to have a greatly reduced measurement error . by stacking galaxies with the same host halo mass
, we will further beat the statistical noise and get a higher s / n .
if no filter is applied to the noisy cmb map , emission from all scales of every component in the cmb sky will be retained in the measurement error . from eq .
[ eq : stacking_corr ] and [ eq : stacking_var_nofilter ] , we can now estimate the signal to noise ratio of this correlation , @xmath146 ^ 2 } { \langle \delta t^2(\theta)\rangle } \\ & = r_{uv}^2 \frac{\langle \delta t^2_{\rm ksz}(\theta ) \rangle}{\langle \delta t^2 \rangle } , \end{split}\ ] ] where @xmath147 the correlation coefficient defined as @xmath148 .
thus the signal to noise ratio of each pixel is essentially proportional to the correlation coefficient @xmath147 and the ratio of the variance of the ksz effect to the overall variance of the cmb sky .
all the components in the cmb map constitute the pixel noise to the ksz signal , and it can be obtained from the following equation @xmath149 @xmath138 is the measured power spectrum of the cmb sky , which is the sum of the primary cmb , the sz effect , the cib , etc . . however , the primary cmb dominates on scales @xmath150 and gets most of its variance around the first several acoustic peaks .
the power spectra of other components will only exceed cmb on smaller scales , and get suppressed by the beam window of the telescope , and thus contribute negligible power to the pixel variance .
hence we calculate the pixel variance in eq .
[ eq : pixel_variance ] with only the primary cmb , which we get from the camb code .
the resulting pixel variance of the temperature fluctuation is @xmath151 k@xmath152 , which means @xmath153 k of the r.m.s .
noise in a pixel . in the context of the standard cosmology and the linear perturbation theory , both baryons and dark halos emmersed in the same gravitational potential exerted by the large scale density perturbation and shall follow the same velocity field on large scales . in linear regime and even quasilinear regime
, one can reconstruct an effective velocity field using the continuity equation , given a galaxy density field from observations .
on small scales , shell crossing will complicate the dynamics and fail the linear prediction .
nevertheless , the velocity field reconstructed from spectroscopic galaxy surveys can trace the underlying velocity flow on large scales .
hydrodynamic plus n - body simulations have examined the relation between the underlying velocity and the reconstructed velocity .
it has been shown @xcite that the recovered velocity field @xmath103 follows the underlying baryonic velocity @xmath107 fairly good . for simplification
we here denote @xmath154 and @xmath100 as the projected velocity component along any given direction of the 3d velocity @xmath107 and @xmath103 . to quantify the performance of the velocity reconstruction , one usually introduces in the fourier space the cross correlation coefficient @xmath155 , and the velocity bias @xmath156 ( see e.g. * ? ? ?
* for a different but related definition of the coefficient @xmath157 ) , where the power spectrum @xmath158 , and @xmath159 is defined as @xmath160 . for the underlying baryon velocity , @xmath161 for the recovered velocity from real space matter density , @xmath162 for the cross power spectrum between the two .
the recovered velocity power spectrum and the cross power spectrum , given matter density in redshift space , are shown as @xmath163 and @xmath164 .
the shot noise power spectrum of velocity , which has been scaled to the fiducial galaxy number density , is shown in triangles . _
the lower two panels _ : the cross correlation coefficient @xmath157 between the underlying baryon velocity @xmath154 and the recovered velocity given galaxy density in real space and in redshift space.,width=307 ] as a function of @xmath165 at z=0 ( left panel ) and z=0.5 ( right panel ) , defined as @xmath166 . the cross correlation and auto correlation are integrated up to @xmath165 in the convention of eq .
[ eq : cross_correlation ] .
thick and thin lines denote the nr run and ww run respectively . here
@xmath147 encodes the uncertainties induced by the redshift space distortion , the gas physics , and the shot noise .
, width=307 ] a perfect reconstruction of the underlying velocity means @xmath167 and @xmath168 at all scales . in realistic galaxy surveys
, there are a number of issues that will complicate the velocity reconstruction @xcite , so that @xmath157 and @xmath169 both deviate from unity . however , deterministic bias does not affect the estimation of s / n as clearly shown in eq .
[ eq : signal_to_noise_nofilter ] , leaving the dependence of s / n only on @xmath147 .
hence we focus on @xmath147 to evaluate the reconstruction .
several important issues which may degrade the velocity reconstruction are listed here : ( 1 ) eq . [ eq : vk_reconstruction ] only holds when @xmath170 , and we have also simplified the equation by assuming the linear density evolution , and a deterministic bias between the galaxy density @xmath104 and the matter density @xmath105 ; ( 2 ) the galaxy number density in the redshift space is what is measured in galaxy survey , while the equation holds in the real space ; ( 3 ) in the reconstruction we lose the information of velocity vorticity which contributes significantly to the ksz signal on small scales ; ( 4 ) the lack of knowledge of the underlying gas dynamics will induce uncertainties in the recipes of gas dynamics employed in simulation and thus in the reconstructed velocity field ; and finally ( 5 ) the mask and geometry of galaxy surveys will reduce the effectiveness of the reconstruction .
these factors will of course degrade the reconstruction .
fortunately , the velocity correlation length is @xmath171 @xmath172mpc , and the reconstruction is fairly good on linear scale and even on quasilinear scales .
we test several issues mentioned above against a set of hydrodynamic simulations .
the simulations are run with the gadget2 code @xcite in a lcdm cosmology with parameters : @xmath173 .
the box size of the simulation is l = 300 @xmath172mpc on each side , in which @xmath174 dark matter particles and @xmath174 gas particles are initially seeded .
two runs are implemented with the same initial condition , one of which is non - radiative ( nr run ) , and the other radiative ( ww run ) with star formation and weak wind feedback ( wind speed is 480 km / s ) .
we measure the power spectra of the underlying velocity of baryons @xmath175 , the recovered velocity @xmath176 given dark matter distribution , the cross power spectrum @xmath177 , and the corresponding power spectra in the redshift space , and show them in the upper panels of fig .
[ fig : cor_uv ] .
it s shown that the recovered power spectrum of the velocity in the redshift space is larger than that in the real space on large scales , and on scales @xmath178 @xmath179mpc@xmath180the power spectrum in redshift space drops rapidly due to the finger of god effect .
also shown is the shot noise power spectrum , which has been scaled to the fiducial galaxy number density @xmath181@xmath182mpc@xmath2 , and it will dominate the velocity power spectrum at @xmath183@xmath179mpc@xmath180 . comparing the nr run ( thin lines ) and ww run ( thick lines ) , we find gas physics only modify small scale power spectrum .
we also show the correlation coefficient in fourier space @xmath157 in the lower two panels of fig .
[ fig : cor_uv ] .
the cross correlation coefficient is large than 0.9 on scale @xmath184 @xmath179mpc@xmath180 , and decreases rapidly on smaller scales when nonlinear evolution dominates , and redshift space distortion further reduces the coefficient . comparing nr run and ww run , we find gas physics only affects the reconstruction on scales @xmath185 @xmath179mpc@xmath180 .
all of above factors play a role in shaping the recovered velocity .
though there are indeed significant deviations between the power spectra of recovered velocity and the underlying velocity , what we really measure is the cross correlation of the recovered velocity and the cmb distortion at galaxy positions .
thus to quantity the performance of the reconstruction , we measure the cross correlation coefficient @xmath186 .
the cross correlation between the reconstructed velocity and the underlying velocity along the line of sight ( l.o.s ) can be calculated by integrating the cross power spectrum as @xmath187 the cross correlation converges rapidly to the auto correlation of the underlying velocity .
on one hand , linear theory predicts that peak contribution comes from @xmath188 around @xmath189@xmath179mpc@xmath180(see fig .
[ fig : cor_uv ] ) , and 70% ( 90% ) of the velocity dispersion is contributed on scales @xmath184@xmath179mpc@xmath180(@xmath190@xmath179mpc@xmath180 ) . on the other hand , observations and simulations
have also shown that only 15% of the velocity dispersion comes from @xmath191@xmath179mpc@xmath180 , where the linear approximation holds very well @xcite .
we measure @xmath192 , @xmath193 and @xmath194 in the simulation with the convention as shown in eq .
[ eq : cross_correlation ] . as shown in fig .
[ fig : cor_uv ] , most of the correlation signal comes from large scale modes , which makes sure that the correlation @xmath194 converges on large scales .
we then calculate the cross correlation coefficient in the redshift space @xmath147 as a function of @xmath165 , and show them in fig .
[ fig : coef ] . here @xmath147 encodes errors induced by the redshift space distortion , the gas physics , and the shot noise .
we find that the correlation coefficient @xmath147 is larger than 0.8 for galaxy surveys sampling a galaxy density corresponding to @xmath195@xmath179mpc@xmath180 , and then decrease to below 0.6 at smaller scales when the shot noise becomes overwhelming .
therefore in this work we choose @xmath196 as the fiducial value , and the signal to noise ratio can be scaled accordingly for other values of @xmath147 .
we have tested the correlation coefficient @xmath147 against various uncertainties in the velocity reconstruction in hydrodynamic simulations , and we assume @xmath147 applies to our fiducial cosmology . for our fiducial cosmology , the underlying zero - lag auto - correlation along the line of sight @xmath197 is calculated using the linear perturbation theory in this work @xmath198 where @xmath199 is the power spectrum of the dark matter obtained from camb .
we have adopted the assumption that the velocity is isotropic .
the linear prediction of the velocity variance is @xmath200 km / s in our fiducial cosmology . as shown in the eq .
[ eq : signal_to_noise_nofilter ] , the signal to noise ratio is proportional to the r.m.s of the underlying velocity dispersion and the correlation coefficient @xmath147 .
simple calculation shows that the central ksz signal at @xmath201 for a mw - size halo is @xmath202k ( eq . [ fig : tksz ] ) .
therefore , a direct stacking without optimal filtering requires more than one million galaxies to achieve 1 @xmath7 detection . even with more pixels of resolved galaxies at lower redshifts ,
the required galaxy number scales inversely to the number of pixels per galaxy , and is still beyond current galaxy surveys for late type galaxies . compared to the variance of the background sky ,
the instrument noise term is very small .
therefore , the measurement error of the cgm s ksz signal is almost entirely contributed by the primary cmb .
thanks to the different scale dependences between the ksz effect and the primary cmb , we can reduce the variance by filtering out the contribution of the large scale primary cmb .
for extended hot halo model ( mb ) profile and @xmath0 profile in fourier space .
the first peaks of the filters are almost at @xmath203 5000- 6000 .
the filters are smooth and they look similar for both models at high redshift , while at low redshift the filter for the flatter mb model fluctuates much more severely than that for the @xmath0 profile .
[ fig : phil],width=307 ] for act - like cmb survey for both mb profile ( upper panel ) and empirical @xmath0 density profile ( lower panel ) for cgm .
[ fig : phit],width=307 ]
as shown in eq . [ eq : phi_fourier ] , the filter profile is the most sensitive to the scales where the signal becomes relatively dominant with respect to the primary cmb plus the instrument noise . on large scales ( low @xmath204 ) , @xmath205 almost vanishes since @xmath206 is much smaller than the primary cmb spectrum .
however , the primary cmb is damped exponentially , while @xmath206 decreases slowly for @xmath207 . in the case of act - like cmb survey
, the filter gets its maximum contribution from those scales where the instrument noise is comparable to the primary cmb , i.e. @xmath208 at almost all redshifts . at lower redshifts z@xmath209 0.06 ,
there are several comparable troughs and peaks for the mb model , as shown in the fourier space in fig .
[ fig : phil ] .
because of the acoustic fluctuation in the damping tail of the cmb , there s small dithering of the filter .
we can also see the behaviour of the filter @xmath130 in fig .
[ fig : phit ] in real space , in which we have adopted the normalization so that @xmath210 .
for all the selected redshifts shown here , the first trough is located at @xmath211 arcmin away from the galaxy center , which is just around the beam size ( the first trough of 0-th bessel function ) .
however , at z@xmath209 0.1 , the mb profile is very shallow and the ksz signal from scales larger than the beam size is non - negligible , and thus part of the ksz signal is compromised by the filter .
that s why one sees that at low redshifts , the filter fluctuates heavily even in the outer region of the halo .
we also show in fig .
[ fig : phit ] the filter for the @xmath0 profile which turns out to be similar for all redshifts . as shown in the figure , although the spatial extent of the @xmath0 profile ( see fig . [
fig : tksz ] ) increases by an order of magnitude from z=0.3 to the local universe , the shape of the filter at different redshifts changes only slightly .
this feature will simplify the stacking procedure , as the filter follows almost the same shape , and thus apply to all galaxies at different redshifts . also as shown in fig .
[ fig : phit ] , when galaxies became well resolved by the sz survey , the filter profiles diverge significantly between the two cgm profiles .
the different behaviours will in turn shed light on how to discriminate between the two cgm profiles , as the ksz effect of a given profile only responds to the corresponding matched filter , and gives the optimal measurement accordingly .
the matched filter gives an optimal estimator @xmath126 of the underlying ksz signal which minimizes the variance . with this estimator in hand ,
we can cross correlate the estimator @xmath126 and the constructed velocity field @xmath100 , and evaluate the signal to noise ratio for a single galaxy as follows @xmath212^{-1 } \\ & = \left [ 1+\frac{\langle v_\parallel^2 \rangle \langle \tilde{u}^2 \rangle } { \langle v_\parallel \tilde{u } \rangle^2 } + \frac{\langle v_\parallel^2\rangle}{\langle v_\parallel \tilde{u } \rangle^2 } \frac { ( 2\pi)^2\int\mathrm{d}^2 l \phi^2(l)\left [ c_l b^2(l)\!+\!\sigma^2_p \theta^2_p\right ] } { \left [ \int\!\mathrm{d}^2 l \phi(l ) \tau(l ) b(l ) \right]^2 } \right]^{-1 } .
\end{split}\ ] ] we have transformed the estimation into fourier space in the last line . as addressed in
[ subsec : uncertainties_in_r ] , the signal to noise ratio is strongly correlated with the cross correlation between the reconstructed velocity and the underlying velocity .
even if there s a deterministic bias of velocity between @xmath100 and @xmath154 , the bias will cancel out in the denominator and the numerator , leaving the dependence only on the underlying velocity dispersion and the correlation coefficient @xmath147 .
we can rewrite the signal to noise estimation in the case of matched filtering as @xmath213 } { t^2_{\rm cmb } \left [ \int\mathrm{d}^2 l \phi(l ) \tau(l ) b(l ) \right]^2 } \right]^{-1 } .
\end{split}\ ] ] in real survey , even if the contamination is filtered , the residual noise is still much larger than the ksz signal of mw - size halos .
thus the first two terms can be neglected , if the correlation coefficient is good enough .
therefore , the signal to noise ratio is proportional to the correlation coefficient @xmath147 , and that s why a good velocity tracer is the key quantity in this method .
it s clear from fig .
[ fig : phil ] that @xmath205 will remove the contribution from large scale modes of the primary cmb @xmath138 . though the large scale modes of the ksz effect @xmath206 are also removed ,
the loss of signal is negligible and the total signal is conserved by design . after applying this filter , the contribution of the primary cmb to the noise is greatly reduced , and the r.m.s .
noise is dramatically reduced from @xmath214 k to a level of @xmath215 k at @xmath201 .
assuming we stack galaxies of the same host halo mass , for example 10000 mw - size halos , in a narrow redshift bin around @xmath216 , we will reduce the statistical noise and achieve @xmath6 1 @xmath7 detection . at lower redshifts , the cgm will have higher ksz effect amplitude , and cover more extended pixels , which make the filter much more effective .
therefore , the s / n increases until @xmath217 , and with 10000 mw - size halos one can achieve @xmath218 detection . below this redshift ,
the signal to noise decreases , as the ksz effect get non - negligible contribution from large scales for spatially extended halos , and thus the filter can not effectively suppress the primary cmb on these scales .
the signal to noise ratio as a function of the redshift for 10000 galaxies in a redshift bin ( @xmath219 for @xmath220 , @xmath221 for @xmath3 ) is plotted in fig .
[ fig : snz ] .
in fact , we have simplified the estimation here , and the total signal to noise ratio will inevitably depend on the number distribution , the redshift distribution , and the mass distribution of the host halos .
careful calculation is needed when applied to the real galaxy catalogue .
also shown in the figure is the s / n for the @xmath0 profile , which is much steeper in the outer region than the fiducial mb profile .
for the @xmath0 profile , the mass is more concentrating to the centre , leading to a higher ksz effect in the central region , with little ksz signal coming from large scale wave mode . the ksz amplitude of the @xmath0 profile is much higher than the mb profile , and thus we can expect a higher signal to noise ratio at the same redshift . as shown in the lower panel of fig .
[ fig : phil ] , the filter for the @xmath0 profile effectively removes the large scale contribution from the primary cmb , and retains the ksz signals very well .
thus the s / n increases with decreasing redshift till the turnover redshift , z=0.03 for a @xmath0 profile , where we get a maximum s / n=19 and a dramatic drop - off below this redshift .
besides the intrinsic shape of the ksz profile of the cgm , the beam size of the telescope also determine the shape of the filter . for comparison
, we also predict the detectability for the two profiles in the case of a planck - like cmb survey , i.e. , @xmath222 arcmin , @xmath223k per pixel .
the results are also shown in the fig .
[ fig : snz ] .
there s similar trends except much lower signal to noise ratio .
the signal to noise ratio reaches the maximum at z@xmath9 1 ( z@xmath9 0.03 ) for the mb profile ( the @xmath0 ) profile when the size of the halo is about 5 arcmin , just around the beam size of planck . the signal to noise ratio will be @xmath8 1 and 2 for the mb profile and @xmath0 profile , respectively . for 10000 mw - size halos as a function of redshift for an act - like sz survey , for both the mb profile and the @xmath0 profile with @xmath68 . also shown
are the predictions for sz surveys like planck .
[ fig : snz],width=307 ]
in this work , we investigate the ksz effect of the cgm for mw - size halos . by weighting the cmb sky with the reconstructed velocity at the galaxy position
, one can pick out the ksz effect of the cgm .
the cross correlation is however subject to heavy contamination mainly from the primary cmb . to remove the large scale contaminations from the primary cmb
, one can take advantage of different scale dependences between the primary cmb and the ksz signal , and choose a minimum variance estimator by designing a matched filter .
one can then stack the filtered map to further suppress the statistical error . with this method
, we alternatively provide a statistical means to detect the cgm via the ksz effect .
furthermore , in its own right , ksz effect of the missing galactic baryons can be cleanly measured , given field galaxies with similar host halo mass . since current detection of the ksz effect from the missing galactic baryons is mixed with the ksz effect from icm / igm , a clean sample of galaxies , such as field galaxies from ska , should be essential . for comparison , we take the compelling mb profile as well as the empirical @xmath0 model to test the detectability . the matched filter is designed to remove the primary cmb on large scales , and keep the small scale ksz signal conserved . for the mb profile , given an act - like sz survey and 10000 milky way - size halos in a narrow redshift bin , we can achieve more than @xmath224 for the cgm of mw - size halos at @xmath216 . for galaxies occupying a single pixel , the matched filter degenerates to the aperture photometry ( ap ) filter , which can be easily implemented in the cmb map by subtracting the average of an outer ring of equal area from the central part ( see * ? ? ? * for an example ) . at lower redshifts ,
more resolved pixels of each galaxy are involved to regulate the filter , the s / n increase greatly with the same mount of galaxies , and reaches a maximum s / n about 7.5 at @xmath225 .
the cgm with the @xmath0 model is more concentrating and the s / n turns out to be much higher . though at high redshift the two profiles show similar s / n , the ksz signal of @xmath0 profile has much higher s / n at @xmath226
. it can reach up to a maximum 19 @xmath7 detection at a lower redshift ( z@xmath8 0.03 ) .
although it is unclear a steeper profile like the @xmath0 model is favoured by observational constraints for the cgm in milky way - size halos , ksz effect of such profile can give insight to its upper limit . due to the different responses of the minimum variance estimator to the filter , mainly at low redshift ( @xmath3 ) , it is possible to discriminate mb profile from the @xmath0 profile , as the underlying cgm will optimally respond to the corresponding filter profile and show the highest significance of detection among a set of filters .
we also test a planck - like survey , and with the same number of galaxies , it can only marginally detect the ksz signal of the cgm at very low redshift .
though this work targets the cgm of late type galaxies , it can also serve to probe the ksz effect for early type galaxies given the knowledge of the cgm s profile .
in this work , we have made a set of assumptions , and some of them are still under debate .
we discuss several major concerns here . 1 . _ it is still under debate the total mass fraction and the spatial distribution of the cgm . _
such information depends on the halo mass , the accretion history , the star formation , the feedback energy deposited and the merging process . on one hand , simulations of galaxy formation
use a broad baryon fraction ranging from 30 - 100% . on the other hand , there are controversial constraints on baryon fraction @xmath48 within the virial radius .
for example , @xcite found a consistent @xmath48 with the cosmic mean , while @xcite draw conclusions that only 10 - 20% missing baryons can be accounted for within the virial radius of the host halo .
the radial extent of x - ray observations , the spectral modeling , and the metallicity distribution suggests different conclusions .
therefore , in this work , we take @xmath22762.5% of the cosmic mean , which is the missing baryon fraction of the milky way , as a fiducial fraction of the hot coronae constituting the missing part of baryons .
we can easily scale the signal to noise ratio with @xmath228 with other values using eq .
[ eq : optical_depth ] and eq .
[ eq : signal_to_noise_filtered ] .
the uncertainty of the optical depth profile .
_ we can see that for a given sz survey with fixed resolution , the two profiles of cgm lead to similar filter profile and thus similar s / n at high redshifts ( @xmath229 for act , @xmath230 for planck ) when galaxies are unresolved . when galaxies become well resolved at lower redshifts , the more concentrating @xmath0 profile has steeper ksz effect , and the filter profiles diverge significantly . since either the mb profile or the @xmath0 model may not account for the true profile of the underlying cgm , or the profiles of halos with the same mass can vary due to different galaxy formation histories
, the uncertainty in the profile can induce variance in the estimator and degrade the detectability .
future application of this method requires more observational constraints on the average profile and the scatter of the profile for the cgm .
_ mass range of host halos .
_ in this work , we have shown that for mw - size halos , the method works well . in real survey ,
host halos of galaxies span a much broader mass range .
this technique can be carried out with careful design .
on one hand , as shown in eq .
[ eq : optical_depth ] that there s degeneracy between @xmath228 and m@xmath231 , high mass with low @xmath228 ( e.g. beta profile for massive ellipticals ) can lead to similar s / n for halos with low mass and high @xmath228 . on the other hand , we can always bin host halos of a broad mass range into subsamples of narrow mass bin .
we can then apply the filter to galaxies to obtain a measure of @xmath232 for each subsample , and stack them to reduce the statistical noise further .
the projection effect . _ as a temperature distortion upon the cmb , the ksz signal is an effect that involves projecting all the free electrons thus the ionized gas along the line of sight .
therefore , the ksz effect of the cgm would suffer from the contamination of the chance alignment of background galaxies , clusters or diffuse baryons around galaxies , which are within the velocity correlation length ( see * ? ? ?
* for the observational evidence of baryons beyond the virial radius of halos ) .
one can in principle model the contribution these sources and filter them out , by accounting for the two - halo term of the ksz effect , given the cosmology and the halo occupation distribution for late type galaxies or the distribution of the icm for clusters .
for example , @xcite calculated the power spectrum of the ksz from the cgm and clusters using halo model , and they found the ksz signal from cgm would only dominate the clusters on scales @xmath233 .
we can alternatively compute the cross correlation between the icm and the cgm using hydrodynamic simulations , and quantify the influence of the projection of icm along the line of sight of the cgm .
another possible contamination is the potentially prevailing igm or filamentary structures bridging galaxies beyond the virial radius . for field galaxies from ska hi survey , this problem is less significant .
the question is still open whether the missing baryons in galaxies are detained in the cgm or have been expelled to the igm , or have never been accreted into galaxies in the galaxy formation process in the first place .
the projection effect is thus to some extent degenerate to the gas fraction @xmath228 within the virial radius .
taking @xmath228 as a free parameter and assigning @xmath228 with some empirical distribution would be a reasonable choice .
the stochasticity in the velocity correlation . _
a number of issues can induce uncertainties in the velocity correlation .
we test the velocity reconstruction and the cross correlation under the influence of the redshift space distortion , the shot noise , and the gas physics using a set of hydrodynamic simulations .
we have found these effect will not fail the reconstruction .
we have also assumed a deterministic galaxy bias @xmath114 in the velocity reconstruction .
though there are stochasticities in @xmath114 , the uncertainties are under control @xcite on scales of interest . as discussed in [ subsec : velocity_dispersion ] ,
the peak contribution to the velocity correlation is from around @xmath234@xmath179mpc@xmath180 , and 70% contribution is from @xmath235@xmath179mpc@xmath180 . given a galaxy survey with number density larger than @xmath236mpc@xmath2 , the recovered velocity traces the underlying velocity very well on scales of interest , as shown in fig .
[ fig : coef ] , and we take @xmath196 as a fiducial value in the s / n estimation . the application of the velocity reconstruction to the real survey has also proved effective @xcite .
6 . _ the number density and the redshift distribution of galaxy surveys_. the realistic measurement of the ksz signal of cgm of galaxies will depend on the number density and the redshift distribution of the galaxies of the real survey .
on one hand for example , ska @xcite will provide a wide and deep survey of hi galaxies covering the act / spt sky partly , and map @xmath24 field galaxies .
it will be sufficient to recover the velocity field from the density field , and thus recover the ksz effect of the cgm .
on the other hand , 2mass redshift survey ( 2mrs ) @xcite have independently completed the peculiar velocity survey , and there are roughly 6000 edge - on spiral galaxies in local universe .
this catalogue alternatively provide a measure of the velocity field .
however , the spiral galaxy number ( @xmath8 150 ) covering the act sky ( 780 deg@xmath152 released in the southern sky ) is so small that the cgm of these galaxies may not be detectable . with even smaller beam size and data release of full 2500 @xmath237 sky in the future
, spt can serve to probe the cgm of spiral galaxies in 2mrs .
the cgm of early type galaxies . _ the method presented in this paper can in principle apply to the cgm of all galaxies with given profiles .
for example , the cgm in massive early type galaxies , even with low baryon fraction , can contain more hot gas than in late type galaxies @xcite .
if this is the prevailing case , early type galaxies should contribute a huge reservoir of cgm for the ksz effect .
6df galaxy redshift survey ( 6dfgrs ) @xcite has measured the peculiar velocity of @xmath8 9000 early type galaxies at @xmath238 in the southern sky .
a very rough estimate shows that there are @xmath8 300 galaxies in the area covering the act sky .
given the @xmath0 profile , as a usually adopted profile for the cgm in elliptical galaxies , the cgm in the halo with mass of @xmath51 m@xmath52 ( as shown in fig .
[ fig : snz ] ) can be detected with @xmath239 sigma level with a rough estimate via this stacking scheme .
massive elliptical galaxies with higher mass can show even higher s / n .
we thank pengjie zhang for helpful discussion , and weipeng lin for kindly providing the simulation data .
js is supported by the national natural science foundation of china under grant no .
tf was partially supported by the national natural science foundation of china grant no .
11273021 , by the strategic priority research program `` the emergence of cosmological structures '' of the chinese academy of sciences , grant no .
xdb09000000 , and by the national natural science funds for distinguished young scholar grant no . | we propose the use of the kinetic sunyaev - zeldovich ( ksz ) effect to probe the circumgalactic medium ( cgm ) , with the aid of a spectroscopic survey covering the same area of a sz survey .
one can design an optimal estimator of the ksz effect of the cgm with a matched filter , and construct the cross correlation between the estimator and the peculiar velocity recovered from the galaxy survey , which can be measured by stacking a number of galaxies .
we investigate two compelling profiles for the cgm , the mb profile @xcite and the @xmath0 profile , and estimate the detectability against the synergy of a fiducial galaxy survey with number density @xmath1 mpc@xmath2 and an act - like sz survey .
we show that the shape of the filter does not change much with redshift for the @xmath0 profile , while there are significant side lobes at @xmath3 for the mb profile . by stacking @xmath4 milky way - size halos around z @xmath5
, one can get @xmath6 1 @xmath7 signal to noise ( s / n ) for the both profiles .
the s / n increases with decreasing redshift before it reaches a maximum ( @xmath8 7.5 at z @xmath9 0.15 for the mb profile , @xmath10 at @xmath11 for the @xmath0 profile ) . due to the large beam size
, a planck - like cmb survey can marginally detect the ksz signal by stacking the same number of galaxies at @xmath3 .
the search for the cgm in realistic surveys will involve dividing the galaxies into subsamples with similar redshift and mass of host halos , and scaling the results presented here to obtain the s / n .
cosmology : theory cosmic background background
methods : statistical galaxies : intergalactic medium |
suppose we take samples , @xmath2 , of size @xmath3 , which are independent and identically distributed ( i.i.d . ) as a @xmath0-variate distribution . here , we consider situations where the data dimension @xmath0 is very high compared to the sample size @xmath1 .
let @xmath4 and assume @xmath5 , @xmath6 , with @xmath7 $ ] and @xmath8 .
we assume that @xmath9 has an unknown mean vector , @xmath10 , and unknown covariance matrix , @xmath11 that is , @xmath12 , @xmath13 , @xmath14 and cov@xmath15 .
let @xmath16 be the @xmath17-th diagonal element of @xmath18 for @xmath19 , and assume @xmath20 for all @xmath21 .
we denote the correlation coefficient matrix between @xmath22 and @xmath23 by @xmath24 , where @xmath25 @xmath26 . here
, @xmath27 denotes the diagonal matrix of elements , @xmath28 . in this paper
, we consider testing the correlation coefficient matrix between @xmath22 and @xmath23 by @xmath29 for high - dimensional settings .
when @xmath30 or @xmath31 , ( [ 1.1 ] ) implies the test of correlation coefficients .
aoshima and yata @xcite gave a test statistic for the test of correlation coefficients and yata and aoshima @xcite improved the test statistic by using a method called the _ extended cross - data - matrix ( ecdm ) methodology_. the test of correlation coefficient matrix is a very important tool of pathway analysis or graphical modeling for high - dimensional data .
one of the applications is to construct gene networks .
see figure 1 . ) .
on can apply the test to constructing gene networks . ]
drton and perlman @xcite and wille et al .
@xcite considered pathway analysis or graphical modeling of microarray data by testing an individual correlation coefficient .
for example , wille et al .
@xcite analyzed gene networks of microarray data with @xmath32 ( @xmath33 and @xmath34 ) and @xmath35 . on the other hand , hero and rajaratnam @xcite
considered correlation screening procedures for high - dimensional data by using a test of correlations .
lan et al .
@xcite and zhong and chen @xcite considered tests of regression coefficient vectors in linear regression models .
as for the test of independence , see fujikoshi et al .
@xcite , srivastava and reid @xcite and yang and pan @xcite .
also , one may refer to szkely and rizzo @xcite about distance correlation . in section 2 ,
we give several assumptions to construct a high - dimensional correlation test for ( [ 1.1 ] ) . in section 3
, we produce a test statistic for ( [ 1.1 ] ) by using the ecdm methodology and show the unbiasedness of ecdm estimator .
we also show that the ecdm estimator has the consistency property and the asymptotic normality when @xmath36 and @xmath37 . in section 4
, we propose a test procedure for ( [ 1.1 ] ) by the ecdm estimator and evaluate its asymptotic size and power when @xmath36 and @xmath37 theoretically and numerically . in section 5 ,
we give several applications of the ecdm estimator .
finally , we demonstrate how the test procedure performs in actual data analyses by using a microarray data set .
in this section , we give several assumptions to construct a test procedure for ( [ 1.1 ] ) .
we have the eigenvalue decomposition of @xmath38 by @xmath39 , where @xmath40diag@xmath41 having eigenvalues , @xmath42 , and @xmath43 is an orthogonal matrix of the corresponding eigenvectors .
let @xmath44 , @xmath45 , where @xmath46 and @xmath47 . here
, @xmath48 denotes the identity matrix of dimension @xmath0 .
note that if @xmath9 is gaussian , the elements of @xmath49 are i.i.d . as the standard normal distribution , @xmath50 .
we assume the following model : @xmath51 where @xmath52 is a @xmath53 matrix for some @xmath54 such that @xmath55 , and @xmath56 , are i.i.d .
random vectors having @xmath57 and @xmath58 .
let @xmath59 , where @xmath60 with @xmath61s @xmath62 for @xmath6 .
then , we have that @xmath63 for @xmath6 .
note that @xmath64 . also , note that ( [ 2.1 ] ) includes the case that @xmath65 and @xmath66 .
let @xmath67 , @xmath68 .
we assume that @xmath69 for all @xmath70 . similar to bai and saranadasa @xcite and aoshima and yata @xcite
, we assume that ( a - i ) : : @xmath71 and @xmath72 + for all @xmath73 . we assume the following assumption instead of ( a - i ) as necessary : ( a - ii ) : : @xmath74 + for all @xmath75 $ ] and @xmath76 $ ] , @xmath77 , where @xmath78 and @xmath79 .
see chen and qin @xcite and zhong and chen @xcite about ( a - ii ) .
note that ( a - ii ) implies ( a - i ) .
when @xmath9 is gaussian , it holds that @xmath65 and @xmath66 in ( [ 2.1 ] ) . note that ( a - ii ) is naturally satisfied when @xmath9 is gaussian because the elements of @xmath49 are independent and @xmath80 for all @xmath70
we assume the following assumption for @xmath18s as necessary : ( a - iii ) : : @xmath81 as @xmath82 .
we note that if @xmath83 and @xmath84 as @xmath82 , ( a - iii ) holds even when @xmath85 is fixed for @xmath86 . also , note that @xmath87 as @xmath82 " is equivalent to @xmath88 as @xmath82 " . here
, @xmath89 denotes the largest eigenvalue of @xmath18 .
let @xmath90 and @xmath91 , where @xmath92 is the frobenius norm .
we note that @xmath93 is equivalent to @xmath94 .
we assume one of the following two assumptions as necessary : ( a - iv ) : : @xmath95 as @xmath96 ; ( a - v ) : : @xmath97 .
note that ( a - v ) holds under the null hypothesis @xmath98 in ( [ 1.1 ] ) .
yata and aoshima @xcite developed the ecdm methodology that is an extension of the cdm methodology given by yata and aoshima @xcite .
one of the advantages of the ecdm methodology is to produce an unbiased estimator having small asymptotic variance at a low computational cost .
see section 2.5 of yata and aoshima @xcite for the details . in this section
, we give a test statistic for ( [ 1.1 ] ) by the ecdm methodology .
we consider an unbiased estimator of @xmath99 by the ecdm methodology .
let @xmath100 and @xmath101 , where @xmath102 denotes the smallest integer @xmath103 .
let @xmath104 for @xmath105 , where @xmath106 denotes the largest integer @xmath107 .
let @xmath108 denote the number of elements in a set @xmath109 .
note that @xmath110 , @xmath111 , @xmath112 and @xmath113 for @xmath105 .
also , note that @xmath114 let @xmath115 for @xmath105 .
let @xmath116 for all @xmath117 .
then , from ( [ 3.1 ] ) , we emphasize the following facts : @xmath118 for all @xmath117 .
let @xmath119 .
we propose an unbiased estimator of @xmath99 by @xmath120 * remark 1*. one can save the computational cost of @xmath121 by using previously calculated @xmath122 and @xmath123 . then , the computational cost of @xmath121 is of the order , @xmath124 .
+ if one considers a naive estimator of @xmath99 as @xmath125 having @xmath126 with @xmath127 , it follows that under ( a - i ) @xmath128 note that the bias term of @xmath125 becomes very large as @xmath0 increases . srivastava and reid @xcite considered an estimator of @xmath99 by @xmath129 with @xmath130s the sample covariance matrices when the underlying distribution is gaussian .
they showed that @xmath131 .
however , @xmath132 is very biased without the gaussian assumption .
contrary to that , the proposed estimator , @xmath133 , is always unbiased and one can claim that @xmath134 without any assumptions . + * remark 2*. we give the following mathematica algorithm to calculate @xmath133 : + * input * : sample size @xmath1 and @xmath135 data matrices @xmath136 $ ] , @xmath6 , such as @xmath137=({\mbox{\boldmath { $ x$}}}_{i1}, ... ,{\mbox{\boldmath { $ x$}}}_{in})^t$ ] . +
* mathematica code * : * @xmath138ceiling@xmath139;$ ] @xmath140 @xmath141 * v@xmath142 $ ] : = if [ floor@xmath143\ge n1 $ ] , take@xmath144floor@xmath143-n1 + 1 $ ] , floor@xmath143\}],$ ] + join[take@xmath145 floor@xmath143\}],$ ] take@xmath146 \{floor@xmath143+n2 + 1 , n\}]\ ] \ ] $ ] * v@xmath147 $ ] : = if [ floor@xmath143\le n1 $ ] , take@xmath144floor@xmath143 + 1 $ ] , floor@xmath143 + n2 \}]$ ] , + join@xmath148take@xmath145 floor@xmath143-n1\},$ ] take@xmath144floor@xmath143 + 1 , n\}]\ ] \ ] $ ] * do@xmath148m@xmath149=$]mean@xmath150\ ] $ ] , @xmath151 @xmath152 , @xmath153 $ ] * @xmath154sum[(part@xmath155 , i]-$]m@xmath156$]).(part@xmath155 , j]-$]m@xmath157 $ ] ) + @xmath158(part@xmath159 , i]-$]m@xmath160$]).(part@xmath159 , j]-$]m@xmath161)$ ] , @xmath162 then , one obtains @xmath163 .
we first consider the consistency property of @xmath121 in the sense that @xmath164 as @xmath96 . assume ( a - i ) .
it holds that as @xmath96 @xmath165 * remark 3 .
* when the underlying distribution is gaussian and @xmath166 , srivastava and reid @xcite showed that as @xmath96 @xmath167 under certain regularity condition which is stronger than ( a - iii ) .
however , as for @xmath121 , one can claim that var(@xmath121 ) in lemma 3.1 is asymptotically equivalent to var(@xmath132 ) under ( a - iii ) and @xmath166 .
+ note that @xmath80 for all @xmath70 when the underlying distribution is gaussian . from lemma 3.1
, we have the consistency property of @xmath121 as follows : assume ( a - i ) and ( a - iv ) .
then , it holds that as @xmath168 @xmath169 the consistency property holds under ( a - iv ) . when ( a - iv ) is not met , we consider the asymptotic normality of @xmath121 .
let @xmath170 .
we give the following result .
assume ( a - i ) , ( a - iii ) and ( a - v ) .
then , it holds that as @xmath168 @xmath171 from lemma 3.2 we have the asymptotic normality of @xmath121 as follows : assume ( a - ii ) , ( a - iii ) and ( a - v ) .
then , it holds that as @xmath168 @xmath172 where
@xmath173 " denotes the convergence in distribution and @xmath50 denotes a random variable distributed as the standard normal distribution .
since @xmath174s are unknown in @xmath175 , it is necessary to estimate @xmath174s for constructing a test for ( [ 1.1 ] ) .
yata and aoshima @xcite gave an estimator of @xmath176 by @xmath177 note that @xmath178 .
from lemma 3.1 , we have the following result .
assume ( a - i ) .
then , it holds as @xmath96 that for @xmath6 @xmath179 * remark 4*. in section 2.5 of yata and aoshima @xcite , they compared @xmath180 with other estimators of @xmath174 theoretically and computationally .
they showed that @xmath180 has small asymptotic variance at a low computational cost .
+ let @xmath181 .
then , by combining theorem 3.2 with lemma 3.3 , we have the following result .
assume ( a - ii ) , ( a - iii ) and ( a - v ) .
it holds that as @xmath168 @xmath182 now , we considered an easy example such as @xmath183 , @xmath184 , @xmath185 , @xmath186 and @xmath65 .
let @xmath187 for @xmath6 , where @xmath188 with eigenvalues , @xmath189 , and @xmath190 is an orthogonal matrix of the corresponding eigenvectors .
we considered two cases : ( a ) @xmath93 ( @xmath191 and @xmath192 ) , and ( b ) @xmath193 ( @xmath194 @xmath195 and @xmath196 ) . here
, @xmath197 were generated independently from a pseudorandom normal distribution with mean vector zero and covariance matrix @xmath38 for each case of @xmath198 , @xmath199 and @xmath200 .
note that ( a - ii ) , ( a - iii ) and ( a - v ) hold from the fact that @xmath201 . in figure 2
, we gave two histograms of 2000 independent outcomes of @xmath202 for ( a ) and ( b ) in each case of @xmath203 together with probability densities of @xmath50 and @xmath204 .
from corollary 3.1 , we expected that @xmath202 is close to @xmath50 when @xmath93 and @xmath204 when @xmath205 . when @xmath198 , the histograms appear far from the probability densities .
when @xmath206 , the histogram for ( a ) fits well the probability density of @xmath50 .
however , the histogram for ( b ) is still far from the probability density of @xmath204 .
this is because the convergence in lemma 3.2 is slow for @xmath207 compared to @xmath93 . as expected , both the histograms fit well the probability densities when @xmath208 . for other simulation settings such as @xmath209 and @xmath210 ,
see section 2 of yata and aoshima @xcite . and
the histograms of @xmath202 for cases of ( a ) @xmath93 and ( b ) @xmath205 fit the solid lines with increasing dimension and sample size : @xmath198 , @xmath211 and @xmath212.,title="fig : " ] + @xmath198@xmath206 @xmath208
in this section , we propose a test procedure for ( [ 1.1 ] ) in high - dimensional settings .
let @xmath213 be a prespecified constant .
from corollary 3.1 , we test ( [ 1.1 ] ) by @xmath214 where @xmath215 is a constant such that @xmath216 .
then , we have the following result . under ( a - ii ) and ( a - iii ) ,
the test by ( [ 4.1 ] ) has that as @xmath96 @xmath217 where @xmath218 denotes the c.d.f .
of @xmath50 and power(@xmath219 ) denotes the power when @xmath220 for given @xmath221 . when ( a - iv ) is met , we have the following result from theorem 3.1 .
assume ( a - i ) .
assume ( a - iv ) under @xmath222 .
then , the test by ( [ 4.1 ] ) has for any @xmath223 that as @xmath96 @xmath224 * remark 5*. let @xmath225 then , from lemma 3.1 , it holds that @xmath226 as @xmath96 under ( a - i ) and ( a - iii ) .
hence , from theorem 3.2 , one may write the power in theorem 4.1 as @xmath227 in order to study the performance of the test by ( [ 4.1 ] ) , we used computer simulations .
we set @xmath228 , @xmath183 , @xmath184 , @xmath229 , @xmath230 and @xmath65 , where @xmath231.\ ] ] note that @xmath232 .
we set ( a ) @xmath93 and ( b ) @xmath193 that are the same settings as in figure 2 .
we considered three distributions for @xmath233s : ( i ) @xmath234 , ( ii ) @xmath235 @xmath236 in which @xmath237s are i.i.d .
as the chi - squared distribution with @xmath238 degree of freedom and ( iii ) @xmath239s are i.i.d . as @xmath0-variate @xmath240-distribution , @xmath241 , with mean zero , covariance matrix @xmath48 and degrees of freedom @xmath242 . note that ( a - ii ) is met in ( i ) and ( ii ) .
however , ( a - i ) ( or ( a - ii ) ) is not met in ( iii ) . we set @xmath243 and @xmath244 .
we note that ( a - iii ) and ( a - v ) hold for ( a ) and ( b ) .
we compared the performance of @xmath133 with @xmath245 by srivastava and reid @xcite , where @xmath246 and @xmath247\{{\mbox{tr}}({\mbox{\boldmath { $ s$}}}_i^2)-{\mbox{tr}}({\mbox{\boldmath { $ s$}}}_i)^2/(n-1)\}$ ] , @xmath6 .
they showed that @xmath245 has the asymptotic normality as @xmath248 when the underlying distribution is gaussian and @xmath93 . also , note that @xmath131 only under the gaussian assumption .
contrary to that , from corollary 3.1 , @xmath249 has the asymptotic normality as @xmath248 even for non - gaussian situations and @xmath207 .
also , one can claim that @xmath134 without any assumptions such as ( a - i ) . in figure 3
, we summarized the findings obtained by averaging the outcomes from 4000 @xmath250 say ) replications for ( i ) to ( iii ) . here
, the first @xmath251 replications were generated for ( a ) when @xmath93 and the last @xmath251 replications were generated for ( b ) when @xmath207 .
we defined @xmath252 when @xmath98 was falsely rejected ( or not ) for @xmath253 , and @xmath222 was falsely rejected ( or not ) for @xmath254 .
we gave @xmath255 to estimate the size in the left panels and @xmath256 to estimate the power in the right panels .
their standard deviations are less than @xmath257
. let @xmath258 . from theorem 4.1 in view of remark 5
, we expected that @xmath259 and @xmath260 for ( [ 4.1 ] ) are close to @xmath261 and @xmath262 , respectively . in figure 4
, we gave the averages ( in the left panels ) and the sample variances ( in the right panels ) of @xmath263 and @xmath264 by the outcomes for ( b ) when @xmath207 in cases of ( i ) to ( iii ) .
from remark 5 , the asymptotic variance for @xmath263 was given by @xmath265 . from figures 3 and 4
, we observed that @xmath266 gives good performances for the gaussian case .
however , for non - gaussian cases such as ( ii ) and ( iii ) , @xmath132 seems not to give a preferable performance .
especially , it gave quite bad performances for ( iii ) .
that is probably because ( a - i ) ( or ( a - ii ) ) is not met in ( iii ) . on the other hand
, @xmath133 gave adequate performances for high - dimensional cases even in the non - gaussian situations .
we observed that @xmath133 is quite robust against other non - gaussian situations as well .
hence , we recommend to use @xmath133 for the test of ( [ 1.1 ] ) and for the estimation of @xmath99 .
are denoted by the dashed lines in the left panels and the values of @xmath260 are denoted by the dashed lines in the right panels for the tests by ( [ 4.1 ] ) and @xmath245 ( sr ) in cases of ( i ) to ( iii ) .
the asymptotic powers were given by @xmath258 which was denoted by the solid lines in the right panels .
, title="fig : " ] + ( i ) @xmath234
. + are denoted by the dashed lines in the left panels and the values of @xmath260 are denoted by the dashed lines in the right panels for the tests by ( [ 4.1 ] ) and @xmath245 ( sr ) in cases of ( i ) to ( iii ) .
the asymptotic powers were given by @xmath258 which was denoted by the solid lines in the right panels .
, title="fig : " ] + ( ii ) the chi - squared distribution with @xmath238 degree of freedom .
+ are denoted by the dashed lines in the left panels and the values of @xmath260 are denoted by the dashed lines in the right panels for the tests by ( [ 4.1 ] ) and @xmath245 ( sr ) in cases of ( i ) to ( iii ) .
the asymptotic powers were given by @xmath258 which was denoted by the solid lines in the right panels .
, title="fig : " ] + ( iii ) @xmath267 . and
@xmath264 are denoted by the dashed lines in the left panels and their sample variances , @xmath268 and @xmath269 , are denoted by the dashed lines in the right panels in cases of ( i ) to ( iii ) .
the asymptotic variance of @xmath263 was given by @xmath265 which was denoted by the solid lines in the right panels.,title="fig : " ] + ( i ) @xmath234 .
+ and @xmath264 are denoted by the dashed lines in the left panels and their sample variances , @xmath268 and @xmath269 , are denoted by the dashed lines in the right panels in cases of ( i ) to ( iii ) .
the asymptotic variance of @xmath263 was given by @xmath265 which was denoted by the solid lines in the right panels.,title="fig : " ] + ( ii ) the chi - squared distribution with @xmath238 degree of freedom .
+ and @xmath264 are denoted by the dashed lines in the left panels and their sample variances , @xmath268 and @xmath269 , are denoted by the dashed lines in the right panels in cases of ( i ) to ( iii ) .
the asymptotic variance of @xmath263 was given by @xmath265 which was denoted by the solid lines in the right panels.,title="fig : " ] + ( iii ) @xmath267 .
in this section , we give several applications of the results in section 3 .
we construct a confidence interval for @xmath99 by @xmath270,\end{aligned}\ ] ] where @xmath271 .
then , from corollary 3.1 , it holds that as @xmath96 @xmath272 under ( a - ii ) , ( a - iii ) and ( a - v ) .
hence , one can estimate @xmath99 by @xmath273 .
if one considers @xmath274 as a candidate of @xmath275 , one can check whether @xmath274 is a valid candidate or not according as @xmath276 or not . as discussed in section 3
, @xmath121 holds the consistency property when ( a - iv ) is met , and @xmath121 holds the asymptotic normality when ( a - v ) is met . here ,
we propose a method to check whether ( a - iv ) holds or not .
let @xmath277 .
we have the following result .
assume ( a - i ) .
it holds that as @xmath96 .
@xmath278 from proposition 5.1 , one can distinguish ( a - iv ) and ( a - v ) .
if @xmath279 is sufficiently small , one may claim ( a - iv ) , otherwise ( a - v ) .
let @xmath280 . here , @xmath281 is the ( population ) rv - coefficient which is a multivariate generalization of the squared pearson correlation coefficient .
note that @xmath282 $ ] .
see robert and escoufier @xcite for the details .
smilde et al .
@xcite considered the rv - coefficient for high - dimensional data .
let @xmath283 .
then , we have the following result .
assume ( a - i ) .
it holds that as @xmath96 @xmath284 thus , one can estimate the rv coefficient by @xmath285 for high - dimensional data .
we consider testing @xmath286 where @xmath274 is a candidate covariance structure .
let @xmath287 and @xmath288 where @xmath119 .
note that @xmath289 .
then , we consider a test statistic for ( [ 5.1 ] ) by @xmath290 note that @xmath291 .
let @xmath292 .
then , we have the following result . [ lem5.1 ]
assume ( a - i ) .
then , it holds that as @xmath96 @xmath293 from lemma [ lem5.1 ] , theorems 3.1 and 3.2 , we have the following results .
assume ( a - i ) . assume also ( a - iv ) with @xmath294 .
then , it holds that as @xmath168 @xmath295 assume ( a - ii ) , ( a - iii ) and ( a - v ) .
assume also ( a - v ) with @xmath294 .
then , it holds that as @xmath168 @xmath296 hence , one can apply @xmath297 to a test for ( [ 5.1 ] ) .
in this section , we demonstrate how the proposed test procedures perform in actual data analyses by using a microarray data set .
we analyzed gene expression data of arabidopsis thaliana given by wille et al .
@xcite in which the data set consists of @xmath298 samples having @xmath299 genes : @xmath300 isoprenoid genes and @xmath301 additional genes . all the data were logarithmic transformed .
wille et al .
@xcite considered a genetic network between the two gene sets . by using a graphical gaussian modeling , they constructed an isoprenoid gene network given in figure 2 of @xcite . in figure 5 , we gave the illustration of the isoprenoid gene network and the additional genes .
we first considered testing ( [ 1.1 ] ) by using ( [ 4.1 ] ) .
see figure 1 for the illustration .
let @xmath228 .
we calculated @xmath302 and @xmath303 , so that @xmath304 . from ( [ 4.1 ] ) and @xmath305 , we rejected @xmath98 .
thus we concluded that two networks have some connections .
in addition , we calculated @xmath306 .
thus , with the help of proposition 5.1 one may conclude that ( a - iv ) is met , so that the power of the test is @xmath238 asymptotically and @xmath307 from theorem 3.1 and corollary 4.1 . also , with the help of proposition 5.2 we obtained @xmath308 as an estimate of the rv - coefficient . ,
ppds@xmath238 and so on are names of genes .
dpps@xmath309 is connected with both mep pathway and mva pathway .
other genes of mitochondrion are not connected with either mep pathway or mva pathway . ]
next , we considered testing ( [ 1.1 ] ) between some part of the isoprenoid genes and the additional genes .
the isoprenoid genes consisted of three types as mep pathway ( 19 genes ) , mva pathway ( 15 genes ) and mitochondrion ( 5 genes ) .
see @xcite for the details . from figure 5
we expected that ( i ) the correlation between dpps@xmath309 and the additional genes is high , and ( ii ) the correlation between the genes of mitochondrion ( except dpps@xmath309 ) and the additional genes is low .
we set @xmath23 as the additional genes ( @xmath34 ) .
we considered three tests for @xmath22 : ( a ) the genes of mitochondrion ( @xmath310 ) ; ( b ) dpps@xmath309 ( @xmath210 ) ; and ( c ) upps@xmath238 , ggpps@xmath311 ( @xmath312 ) . by using the first 50 samples ( @xmath313 ) of the @xmath298 samples
, we constructed ( [ 4.1 ] ) .
then , with @xmath228 , we rejected @xmath98 for ( a ) since @xmath314 and for ( b ) since @xmath315 . on the other hand
, we accepted @xmath98 for ( c ) since @xmath316 .
similar to section 5 in yata and aoshima @xcite , we considered a high - dimensional linear regression model : @xmath317 where @xmath318 is an @xmath319 response matrix , @xmath320 is an @xmath321 fixed design matrix , and @xmath322 is a @xmath323 parameter matrix
. the @xmath1 rows of @xmath324 are independent and identically distributed as a @xmath325-variate distribution with mean vector zero .
let @xmath22 be the @xmath17th sample of the @xmath326 isoprenoid genes ( except upps@xmath238 , ggpps@xmath311 ) .
let @xmath327 , @xmath328 .
we set @xmath329^t$ ] and @xmath330^t$ ] with @xmath331 .
we noted that the standard elements of @xmath332 are path coefficients from the isoprenoid genes to the additional genes . by using the observed samples of size @xmath313 as a training data set
, we obtained the least squared estimator of @xmath322 by @xmath333 .
we investigated prediction accuracy of the regression with @xmath334 by using the remaining samples of size @xmath335 as a test data set .
we considered the prediction mean squared error ( pmse ) by @xmath336 . by using the test samples @xmath337 and @xmath338 , @xmath339 , we applied the bias - corrected and accelerated ( bca ) bootstrap by efron @xcite .
then , we constructed @xmath340% confidence interval ( ci ) of the pmse by @xmath341 $ ] from 10000 replications . on the other hand , we considered the pmse for the full isoprenoid ( @xmath342 genes ) .
then , similar to above , we constructed @xmath340% ci of the pmse by @xmath343 $ ] .
the pmse by the @xmath326 isoprenoid genes is probably smaller than that of the full isoprenoid genes .
thus we conclude that the test procedure by ( [ 4.1 ] ) effectively works for this data set .
throughout , we assume that @xmath344 and @xmath345 without loss of generality . let @xmath346 , @xmath347 and @xmath348 . note that @xmath349 from the fact that @xmath350 for @xmath6 .
then , we note that @xmath351 , where @xmath352 is given in remark 5 .
let @xmath353 and @xmath354 for all @xmath355 .
note that @xmath356 .
let @xmath357 , @xmath358 and @xmath359 for all @xmath360 . note that @xmath361 for all @xmath360 and @xmath362 for all @xmath363 .
let @xmath364 , @xmath365 and @xmath366 .
let @xmath367 and @xmath368 for all @xmath369 .
we write that @xmath370 for all @xmath369 .
note that all the terms of @xmath371 in ( [ a.2 ] ) are uncorrelated under ( a - i ) . from ( [ a.1 ] )
, it holds that under ( a - i ) @xmath372 for all @xmath360 .
similarly , under ( a - i ) , it holds that @xmath373 for all @xmath360 .
then , we have that under ( a - i ) @xmath374 then , we have that as @xmath96 @xmath375 under ( a - i ) . on the other hand , we have that as @xmath376 @xmath377 under ( a - i ) .
then , we have that as @xmath96 @xmath378=o(k^2 ) \label{a.4}\end{aligned}\ ] ] under ( a - i ) .
hence , by combining ( [ a.3 ] ) with ( [ a.4 ] ) , we have that as @xmath96 @xmath379\\ & = { \mbox{var } } ( u_{n})+{\mbox{var}}(u_n-\widehat{t}_n ) -2e[\{u_n-(\widehat{t}_n-\delta)\}u_n]\\ & = k^2\{1+o(1)\}+o(\omega^{1/2}/n^2)\end{aligned}\ ] ] under ( a - i ) from the fact that @xmath380|\le \{{\mbox{var}}(u_n-\widehat{t}_n ) { \mbox{var}}(u_n)\}^{1/2}$ ] by schwarz s inequality .
it concludes the result .
let @xmath381 .
when we consider the singular value decomposition of @xmath382 , it follows that @xmath383 , where @xmath384 denote singular values of @xmath382 , and @xmath385 ( or @xmath386 ) denotes a unit left- ( or right- ) singular vector corresponding to @xmath387 .
then , it holds that @xmath388 similarly , we claim that @xmath389 , so that @xmath390 thus under ( a - iii ) , it holds that @xmath391 as @xmath392 .
then , we claim that @xmath393 as @xmath392 , so that under ( a - iii ) and ( a - v ) @xmath394 by noting that @xmath395 and @xmath396 , from lemma 3.1 and ( [ a.6 ] ) , we have that as @xmath96 @xmath397 under ( a - i ) , ( a - iii ) and ( a - v ) from the fact that @xmath398 as @xmath392 under ( a - v ) . from ( [ a.5 ] )
, it holds that @xmath399 , so that @xmath400 .
then , from lemma 3.1 and @xmath401 , it holds that as @xmath96 @xmath402 under ( a - i ) .
thus , under ( a - iv ) , from chebyshev s inequality , we can claim the result .
we first consider the first result of lemma a.1 .
let @xmath404 for all @xmath405 .
let @xmath406 and @xmath407 .
note that @xmath408 and @xmath409 under ( a - ii ) . here , we claim that @xmath410 , so that @xmath411 then , under ( a - ii ) , we have that @xmath412 for @xmath413 , it is necessary to consider the terms of @xmath414 because it does not hold that @xmath415 unless @xmath416 or @xmath417 . here , under ( a - ii ) , we evaluate that for sufficiently large @xmath418 @xmath419 from the fact that @xmath420 for all @xmath70 .
similarly , for other terms , we can evaluate the order as @xmath421 .
hence , we can claim that @xmath422 under ( a - ii ) , so that @xmath423 from ( [ a.7 ] ) . by noting that @xmath424
, we can conclude the first result of lemma a.1 . from ( [ a.2 ] ) , we have that under ( a - i ) @xmath427 then , from ( [ a.6 ] ) , we have that as @xmath96 @xmath428 under ( a - i ) , ( a - iii ) and ( a - v ) . by combining ( [ a.8 ] ) with ( [ a.4 ] ) , from the fact that @xmath429 , we can conclude the result .
let @xmath430 for @xmath431 .
note that @xmath432 .
also , note that @xmath433 here , we have for @xmath434 , that @xmath435 .
then , we consider applying the martingale central limit theorem given by mcleish @xcite .
let @xmath436^{1/2}$ ] , @xmath431 .
note that @xmath437 and @xmath438 .
let @xmath439 denote the indicator function . from ( [ a.1 ] ) , under ( a - ii ) and ( a - iii )
, it holds that as @xmath82 @xmath440 then , by using chebyshev s inequality and schwarz s inequality , from lemma a.1 , under ( a - ii ) and ( a - iii ) , it holds for lindeberg s condition that as @xmath96 @xmath441 for any @xmath442 . here , from lemma a.1 , ( [ a.9 ] ) and ( [ a.10 ] ) , under ( a - ii ) and ( a - iii )
, we evaluate that as @xmath96 @xmath443= o\big ( \frac{\psi^2}{b^2n } + \frac{\omega}{b^2 } \big)\to 0,\end{aligned}\ ] ] so that @xmath444\to 0 .
\label{a.11}\end{aligned}\ ] ] then , by using the martingale central limit theorem , from ( [ a.10 ] ) and ( [ a.11 ] ) , under ( a - ii ) and ( a - iii ) , we obtain that as @xmath96 @xmath445 note that @xmath446^{1/2}\to 1 $ ] as @xmath96 under ( a - ii ) , ( a - iii ) and ( a - v ) .
then , by combining ( [ a.12 ] ) with lemmas 3.2 and a.2 , we have that as @xmath96 @xmath447 under ( a - ii ) , ( a - iii ) and ( a - v ) .
it concludes the result .
we first consider the proof of corollary 4.1 . from theorem 3.1 , under ( a - i ) and ( a - iv ) , we have that as @xmath96 @xmath453 from the fact that @xmath454 as @xmath96 under ( a - i ) and ( a - iv ) .
it concludes the result of corollary 4.1 .
next , we consider the proof of theorem 4.1 . from corollary 3.1 , under ( a - ii ) , ( a - iii ) and ( a - v ) , we have that as @xmath96 @xmath455 from the fact that @xmath456 as @xmath96 under ( a - ii ) .
we can conclude the results of size and power when ( a - v ) is met in theorem 4.1 .
we note that @xmath457 as @xmath96 under ( a - iv ) , so that we obtain the result of power when ( a - iv ) is met from corollary 4.1 .
hence , by considering the convergent subsequence of @xmath458 , we can conclude the result of power in theorem 4.1 .
the proofs are completed .
next , we consider the case when ( a - v ) is met . from ( [ a.5 ] ) , it holds that @xmath460 , so that @xmath461 as @xmath96 under ( a - v ) .
then , from lemma 3.1 and ( [ a.1 ] ) , under ( a - i ) and ( a - v ) , we claim that @xmath462 as @xmath96 .
note that @xmath463 as @xmath96 under ( a - v ) .
thus under ( a - i ) and ( a - v ) , it holds that @xmath464 as @xmath96 .
then , from lemma 3.3 , under ( a - i ) and ( a - v ) , we have that as @xmath96 @xmath465 it concludes the result when ( a - v ) is met .
the proof is completed .
let @xmath469 .
similarly to ( [ a.5 ] ) , it holds that @xmath470 then , by noting that @xmath471 and @xmath472 , from lemma 5.1 , we have that as @xmath96 @xmath473 under ( a - i ) .
thus , under ( a - iv ) with @xmath294 , from chebyshev s inequality , we can claim the result of corollary 5.1 .
similarly to the proof of lemma a.2 , under ( a - i ) , ( a - iii ) , ( a - v ) and ( a - v ) with @xmath294 , we can claim that @xmath474 as @xmath96 .
thus , similar to ( [ a.13 ] ) , from lemma 3.3 , we can conclude the result .
research of the first author was partially supported by grant - in - aid for young scientists ( b ) , japan society for the promotion of science ( jsps ) , under contract number 26800078 .
research of the second author was partially supported by grants - in - aid for scientific research ( b ) and challenging exploratory research , jsps , under contract numbers 22300094 and 26540010 .
m. aoshima , k. yata , asymptotic normality for inference on multisample , high - dimensional mean vectors under mild conditions , methodol .
( 2013 ) , in press .
doi : 10.1007/s11009 - 013 - 9370 - 7 .
z. bai , h. saranadasa , effect of high dimension : by an example of a two sample problem , statist .
sinica 6 ( 1996 ) 311 - 329 .
chen , y .- l .
qin , a two - sample test for high - dimensional data with applications to gene - set testing , ann .
38 ( 2010 ) 808 - 835 .
a. wille , p. zimmermann , e. vranova , a. frholz , o. laule , s. bleuler , l. hennig , a. prelic , p. von rohr , l. thiele , e. zitzler , w. gruissem , p. bhlmann , sparse graphical gaussian modeling of the isoprenoid gene network in arabidopsis thaliana , genome biol . 5 ( 2004 ) | in this paper , we consider testing the correlation coefficient matrix between two subsets of high - dimensional variables .
we produce a test statistic by using the extended cross - data - matrix ( ecdm ) methodology and show the unbiasedness of ecdm estimator .
we also show that the ecdm estimator has the consistency property and the asymptotic normality in high - dimensional settings .
we propose a test procedure by the ecdm estimator and evaluate its asymptotic size and power theoretically and numerically .
we give several applications of the ecdm estimator .
finally , we demonstrate how the test procedure performs in actual data analyses by using a microarray data set .
+ + * keywords : * correlations test ; cross - data - matrix methodology ; graphical modeling ; large @xmath0 , small @xmath1 ; pathway analysis ; rv - coefficient 0.5 cm * kazuyoshi yata and makoto aoshima * + institute of mathematics , university of tsukuba , ibaraki , japan + |
the periodic potential in the slip plane of a crystal leads to stable positions for a straight dislocation line separated by maxima in the potential energy known as peierls barriers .
when the peierls barrier is large compared to the available thermal energy , the mobility of dislocations is limited by the discreteness of the crystal structure . in that case dislocation glide takes place by the thermally activated nucleation and propagation of kinks@xcite , which are localized regions connecting dislocation segments lying in adjacent valleys of the potential in the slip plane . the existence of kinks is clearly exhibited in md simulations reported here and in many other investigations@xcite , and their movement effects the glide of the dislocations on which they lie . to understand the kink mechanism it is necessary to investigate both kink propagation and nucleation of pairs of kinks
+ when the formation energy of kink pairs is large compared to the thermal energy it becomes impossible to obtain statistically significant data on kink nucleation within md time scales without resorting to unrealistic applied stresses , typically six orders of magnitude larger than those encountered in experiments@xcite . but
dislocation core structures and peierls barriers are known to change with applied stress @xcite , making it difficult to relate such simulations to experimental reality . as a result , while the kink mechanism is well established in dislocation theory there is a sparsity of md data on the parameters controlling kink motion , without which it is impossible to predict the velocity - stress relationship of the host dislocation for realistic stresses .
+ to circumvent the problem of kink nucleation in md we use boundary conditions on the simulation supercell which enforce the existence of an isolated kink on the dislocation line .
we use the md code lammps@xcite with a recently developed potential by gordon _ _ et al.__@xcite which gives the best available representation of the screw dislocation core structure and bulk phonon dispersion . to avoid free surfaces ,
periodic boundary conditions must be imposed .
the dislocation supercell must contain defects with no net burgers vector to avoid a divergent elastic energy , and in this work we use dislocation dipoles .
thus the supercell contains two dislocations with equal and opposite burgers vectors , and each dislocation has one kink in the supercell .
+ dislocation dipoles were introduced by removing an appropriate number of atoms for an edge dislocation dipole or shearing the simulation supercell for a screw dislocation dipole@xcite , and then applying the anisotropic elastic displacement field for the dipole .
the system was relaxed by a conjugate gradients algorithm , followed by an annealing process which heated the system to 200k then back to zero temperature over 100ps ( 100,000 timesteps ) to ensure that the system was in the ground state . to heat the system ,
atomic velocities were gradually rescaled according to a maxwell - boltzmann distribution of increasing temperature . once the desired temperature was achieved the system
was evolved microcanonically and data was taken .
this has a firmer statistical basis than using a thermostat because it relies on the real atomistic heat bath of a large system rather than any particular thermostat algorithm .
however , unrealistic results can be obtained if there is significant heat generation or absorption as this may affect the probability of other activated processes@xcite .
this is likely to be the case , for example , when a high energy kink pair annihilates or is created during the simulation .
however , due to their large activation energies such processes did not occur during the simulations with the zero stress conditions investigated here , and the system temperature was observed to be constant throughout the simulation runs . +
a kink connects two straight dislocation segments both parallel to a lattice vector @xmath5 lying in the same slip plane . for the segments to be crystallographically equivalent they must be separated by a lattice vector , which may be uniquely identified , modulo @xmath6 , with a lattice vector @xmath7 which we call the ` kink vector '
some dislocations may exist with a variety of core structures and there is a corresponding variety of atomic structures of kinks @xcite .
nevertheless , the kink vector uniquely identifies any kink on a dislocation line in a given slip plane with a given burgers vector and the same core structure on either side of the kink .
we note in passing that a similar classification may also be applied to jogs , the sessile equivalent of a kink@xcite , where the ` jog vector ' will be a lattice vector with a component normal to the slip plane .
while the term kink vector has already appeared in the literature@xcite there has been no attempt to relate it to the host crystal lattice . to see the utility of our definition ,
consider a relaxed straight dislocation dipole , parallel to a lattice vector @xmath5 , in a supercell formed from lattice vectors @xmath8 , @xmath9 and @xmath10 , where the @xmath11 are all integers . to impose the boundary conditions required for a kink on each dislocation we create a new supercell from @xmath8 , @xmath12 and @xmath10 , as shown in figure [ 101 ] . in principle @xmath7
may be any lattice vector modulo @xmath6 lying in the slip plane , but for high index lattice vectors the relaxed structure will possess multiple kinks as one might expect if @xmath7 spans many minima of the potential in the slip plane .
the set of low index lattice vectors lying in the slip plane quickly provides an enumeration of the possible kinks a straight dislocation line may support , generalizing the approach taken in previous work@xcite to produce isolated kinks on dislocation lines .
we have used this procedure to generate the simulation supercell geometries employed here .
each supercell contained approximately 700,000 atoms , with the dislocation line either initially sharply kinked as in figure [ 101 ] or parallel to the @xmath12 supercell vector .
the relaxed configurations were independent of this initial preparation .
+ , @xmath12 and @xmath10 .
the broken line in the left figure highlights one of the two slip planes , which are separated by half the supercell height @xmath13.[101 ] ] ] having obtained a relaxed configuration , the atoms were grouped into atomic planes normal to the unkinked dislocation line direction @xmath5 .
the potential energy in each plane has clearly defined peaks , the centres of gravity of which identify the positions of the dislocation core , as illustrated in figure [ slice ] .
the coarse grained representation of the dislocation comprises a line threading nodes at each of these core positions .
there is one node for each atomic plane normal to the dislocation line .
the coarse grained representation is well defined , independent of dislocation character , and it yields the position of the dislocation with atomic resolution .
the unique mapping between the atomistic and coarse grained representations enables the energy of the coarse grained representation to be determined at each time step of the simulation .
the position of each node on the dislocation line moves in one dimension , normal to @xmath6 and the slip plane normal @xmath14 .
the kink position and width were determined from the center and width of the maximum in the core energy along the dislocation line , with the trajectories of the two kinks in each supercell each forming a time series @xmath15,@xmath16 .
the kink position is readily located graphically as shown in figure [ 110_kink ] , where the dislocation segments on either side of the kink are straight lines at absolute zero .
the dynamics of the kink positions are simulated with md in the following section and we aim to reproduce the dynamics with a coarse grained , many body , stochastic model .
+ the effectiveness of the coarse graining relies on the uniformity of the potential energy of atoms in the bulk ; however , at finite temperature it is expected and observed that random fluctuations in atomic positions and energies due to thermal vibrations obscure the dislocation position . to filter out this noise
it is necessary to average atomic positions and energies over a period of a few thermal oscillations .
the atomic coordinates and energies can then be processed in an identical fashion to the relaxed zero temperature system , again yielding localized dislocation core positions .
kinks appear as localized geometric and energetic regions along the coarse grained dislocation line , allowing the determination of the kink position , width and formation energy by calculating the total deviation from the core energy of a straight dislocation line .
+ other techniques to determine the dislocation position are to calculate the deviation of the atomic displacements from the anisotropic elastic field@xcite , or the bonding disregistry across the slip plane@xcite .
however , we found that the procedure employed here gave better localization at finite temperature and is applicable to many different dislocation geometries .
+ table [ kvs ] shows kink formation energies calculated for edge dislocations in bcc fe .
these values were obtained by calculating the excess energy in a cylindrical slice coaxial with the average dislocation line direction , relative to the energy of a slice containing the same number of atoms for a straight dislocation .
the slices contain one atomic plane normal to the dislocation lines .
the radius of the cylinder was enlarged , using the periodic boundary conditions if necessary to generate atomic coordinates outside the supercell , until the excess energy of the dislocation core in the slice reached a constant asymptotic value .
this excess core energy is plotted in green in figure [ 110_kink ] .
the kink formation energy is the sum of these excess core energies along the dislocation line .
convergence in the core energy per atomic plane was typically achieved for a supercell length of thirty burgers vectors for kinks on edge dislocations .
this implies that the interaction energy between a kink and its periodic images along the line is not detectable at separations of more than thirty burgers vectors .
the kink formation energies are in broad agreement with other studies@xcite .
0.48@cccccburgers & glide & tangent & kink & formation + vector ( @xmath17 ) & plane & vector ( @xmath5 ) & vector ( @xmath18 ) & energy + @xmath19 $ ] & @xmath20 & @xmath21 $ ] & @xmath19 $ ] & 0.15 ev + & @xmath22 & @xmath23 $ ] & @xmath24 $ ] & 0.03 ev + & & & @xmath25 $ ] & 0.02 ev + @xmath26 $ ] & @xmath27 & @xmath28 $ ] & @xmath26 $ ] & 0.51 ev + & @xmath29 & @xmath30 $ ] & @xmath31 $ ] & 0.25 ev + + ( 4874.00,3628.00 ) ( 528,815 ) ( 0,0)[r ] 0 ( 528,1036 ) ( 0,0)[r ] 0.2 ( 528,1258 ) ( 0,0)[r ] 0.4 ( 528,1480 ) ( 0,0)[r ] 0.6 ( 528,1701 ) ( 0,0)[r ] 0.8 ( 528,1923 ) ( 0,0)[r ] 1 ( 528,2144 ) ( 0,0)[r ] 0 ( 528,2366 ) ( 0,0)[r ] 0.2 ( 528,2587 ) ( 0,0)[r ] 0.4 ( 528,2809 ) ( 0,0)[r ] 0.6 ( 528,3031 ) ( 0,0)[r ] 0.8 ( 528,3252 ) ( 0,0)[r ] 1 ( 660,484 ) ( 0,0 ) 10 ( 1130,484 ) ( 0,0 ) 15 ( 1600,484 ) ( 0,0 ) 20 ( 2070,484 ) ( 0,0 ) 25 ( 2539,484 ) ( 0,0 ) 30 ( 3009,484 ) ( 0,0 ) 35 ( 3479,484 ) ( 0,0 ) 40 ( 3949,484 ) ( 0,0 ) 45 ( 4081,957 ) ( 0,0)[l ] 0 ( 4081,1210 ) ( 0,0)[l ] 0.03 ( 4081,1464 ) ( 0,0)[l ] 0.06 ( 4081,1717 ) ( 0,0)[l ] 0.09 ( 4081,1970 ) ( 0,0)[l ] 0.12 ( 4081,2223 ) ( 0,0)[l ] 0 ( 4081,2477 ) ( 0,0)[l ] 0.03 ( 4081,2730 ) ( 0,0)[l ] 0.06 ( 4081,2983 ) ( 0,0)[l ] 0.09 ( 4081,3236 ) ( 0,0)[l ] 0.12 ( 154,2033 ) ( 4586,2033 ) ( 2304,154 ) ( 0,0 ) @xmath32 plane index ( 0,0)(1\bar{2}1)$ ] edge dislocation line at 0k ( bottom ) and 90k ( top ) .
the kink is clearly localized as measured by the position of the core ( red ) and the core energy ( green ) .
note the narrow kink width in contrast to the screw dislocation kinks in figure [ screw_e ] .
( color online ) [ 110_kink],title="fig : " ] as @xmath26(011)$ ] edge dislocations have not been directly observed in experiment they are of little interest and we do not consider them further here .
the very low formation energy of 0.03 ev for kinks on @xmath33 edge dislocations indicates that the mobility of these dislocations is not limited by kinks except possibly at the very lowest temperatures .
therefore we investigate isolated kinks on @xmath34(1\bar{2}1)$ ] and @xmath26(010)$ ] edge dislocations , whose motion is known to be kink - limited , at temperatures up to which kinks remain isolated on md time - scales .
+ ( 010)$ ] edge dislocation in the bulk .
below : a kink on this dislocation with the bulk atoms removed .
the non - planar core gives a large kink formation energy of 0.61ev .
( color online ) [ 100_kink],title="fig:",scaledwidth=40.0% ] + ( 010)$ ] edge dislocation in the bulk .
below : a kink on this dislocation with the bulk atoms removed .
the non - planar core gives a large kink formation energy of 0.61ev .
( color online ) [ 100_kink],title="fig:",scaledwidth=35.0% ] the kink formation energy for @xmath19(1\bar{2}1)$ ] dislocations is 0.15ev . at temperatures below 300k it is possible to observe and analyze the stochastic motion of an isolated kink for md runs of several nanoseconds .
similarly , for @xmath26(010)$ ] edge dislocations no additional kinks are expected to be nucleated in md runs of several nanoseconds at temperatures up to 700k owing to their large formation energy of 0.61ev .
this large formation energy is due to the non - planar core , shown in figure [ 100_kink ] .
an isolated kink is localized geometrically and energetically as shown in figure [ 110_kink ] . + ( 4874.00,3004.00 ) ( 528,704 ) ( 0,0)[r ] -100 ( 528,958 ) ( 0,0)[r ] 0 ( 528,1213 ) ( 0,0)[r ] 100 ( 528,1467 ) ( 0,0)[r ] 200 ( 528,1722 ) ( 0,0)[r ] 300 ( 528,1976 ) ( 0,0)[r ] 400 ( 528,2230 ) ( 0,0)[r ] 500 ( 528,2485 ) ( 0,0)[r ] 600 ( 528,2739 ) ( 0,0)[r ] 700 ( 660,484 ) ( 0,0 ) 0 ( 1208,484 ) ( 0,0 ) 200 ( 1756,484 ) ( 0,0 ) 400 ( 2305,484 ) ( 0,0 ) 600 ( 2853,484 ) ( 0,0 ) 800 ( 3401,484 ) ( 0,0 ) 1000 ( 3949,484 ) ( 0,0 ) 1200 ( 22,1721 ) ( 2304,154 ) ( 0,0 ) time [ ps ] ( 2005,884 ) ( 0,0)[l ] t=90k ( 3358,2566 ) ( 0,0)[r ] kink 1 ( 3358,2346 ) ( 0,0)[r ] kink 2 ( 0,0)(1\bar{2}1)$ ] edge dislocation dipole at t=90k .
( color online)[traj],title="fig : " ] figure [ traj ] shows the trajectories of the two kinks in a simulation supercell on a @xmath35 edge dislocation dipole .
the two kinks appear to be moving independently , but we can not be sure that there is no significant interaction between them .
any correlation arising from their interaction may be eliminated by analyzing the center of mass @xmath36 , defined here to be the mean of the two kink positions , @xmath37 .
it may be shown@xcite that such a quantity is independent of any interaction , and it yields a diffusion constant one half that of a free kink , @xmath38 .
thus we construct from the two kink positions a single time series @xmath39 for the center of mass to ensure such correlation effects do not affect our results .
we look for diffusive behavior in the mean squared displacement ( msd ) @xmath40 over a range of intervals @xmath41 , defined as @xmath42 which is the variance of the displacement .
it is well known@xcite that for diffusive motion with a diffusion constant @xmath38 , @xmath43 + ( 4704.00,3004.00 ) ( 528,704 ) ( 0,0)[r ] 0 ( 528,1043 ) ( 0,0)[r ] 50 ( 528,1382 ) ( 0,0)[r ] 100 ( 528,1722 ) ( 0,0)[r ] 150 ( 528,2061 ) ( 0,0)[r ] 200 ( 528,2400 ) ( 0,0)[r ] 250 ( 528,2739 ) ( 0,0)[r ] 300 ( 660,484 ) ( 0,0 ) 5 ( 1106,484 ) ( 0,0 ) 10 ( 1551,484 ) ( 0,0 ) 15 ( 1997,484 ) ( 0,0 ) 20 ( 2442,484 ) ( 0,0 ) 25 ( 2888,484 ) ( 0,0 ) 30 ( 3333,484 ) ( 0,0 ) 35 ( 3779,484 ) ( 0,0 ) 40 ( 22,1721 ) ( 2219,154 ) ( 0,0 ) time interval @xmath44 [ ps ] ( 1251,2566 ) ( 0,0)[l ] t=100k ( 1251,2346 ) ( 0,0)[l ] t=130k ( 1251,2126 ) ( 0,0)[l ] t=170k ( 1251,1906 ) ( 0,0)[l ] linear fit ( 0,0 ) ) for the kink center of mass on a @xmath34(1\bar{2}1)$ ] edge dislocation dipole .
the linear relationship with time is in agreement with diffusive behavior , equation ( [ mdmsd_d]).[ouedge],title="fig : " ] examples of the msd are shown for the system of kinks on edge dislocations considered above in figure [ ouedge ] .
the msd clearly shows the linear time dependence characteristic of diffusive behavior , with the diffusion constant as defined in equation ( [ mdmsd_d ] ) shown in figure [ arrhenius ] for kinks on @xmath45(1\bar{2}1)$ ] edge dislocations .
kinks on @xmath46(010)$ ] edge dislocations exhibit similar behavior .
the diffusivity rises exponentially with temperature in both cases , indicating that the kink motion is thermally activated across the kink migration barrier@xcite .
we therefore conclude that the kink performs one dimensional stochastic motion in a periodic migration potential @xmath47=@xmath48 whose amplitude @xmath49=@xmath50 is large compared to the thermal energy .
while the traditional analysis for such data is to fit an arrhenius form @xmath51 for the diffusion constant , in one dimension there exists an exact solution , given by the lifson - jackson formula@xcite @xmath52 where @xmath53 is boltzmann s constant and @xmath54 is the friction , or dissipation , parameter@xcite , which measures the rate of momentum transfer from the diffusing object ( here a kink ) to the heat bath .
@xmath54 plays a key rle in the stochastic equations of motion introduced in section [ sec : le ] , defining the frictional force @xmath55 and it is the inverse of the kink mobility . to gain insight into equation ( [ lj ] )
we investigate limiting cases .
when the amplitude of the migration potential @xmath49=@xmath50 is much greater than thermal energy @xmath56 , as for the case of kinks on edge dislocations here , we may evaluate the integrals in ( [ lj ] ) by the method of steepest descents . denoting @xmath57 for the second
derivative , ( [ lj ] ) becomes @xmath58 which is precisely the arrhenius form given by kramers@xcite for thermally activated diffusive motion .
we note that the traditional temperature independence of the prefactor in ( [ kramersd ] ) requires that @xmath54 be independent of temperature . in the other limit ,
when the thermal energy @xmath56 is much larger than @xmath49 , the integrals ( [ lj ] ) are constant , giving a diffusivity @xmath59 as first described by einstein@xcite for a freely diffusing particle .
we note that a linear temperature dependence in the diffusivity ( [ freed ] ) implies @xmath54 is again independent of temperature .
we will see that kinks on screw dislocations exhibit the diffusive behavior of ( [ freed ] ) due to their negligible migration barrier and thus the relation ( [ lj ] ) is able to capture the wide range of diffusive behavior exhibited by kinks on dislocation lines .
a numerical illustration of ( [ lj ] ) is shown in figure [ th ] , where we indeed see the failure of the arrhenius law when @xmath60 . in section [ sec : le ] we show that to an excellent approximation the migration barrier for a kink is sinusoidal , @xmath61 , allowing an exact expression of ( [ lj ] ) @xmath62 where @xmath63 is the zeroth order modified bessel function@xcite .
we thus perform a two parameter fit of ( [ bessd ] ) to the kink diffusion constant with temperature to determine @xmath54 and @xmath49 for the kink systems investigated here , the results of which are shown in table [ kink_table ] .
+ ( 4874.00,3004.00 ) ( 528,704 ) ( 0,0)[r ] -3 ( 528,995 ) ( 0,0)[r ] -2 ( 528,1285 ) ( 0,0)[r ] -1 ( 528,1576 ) ( 0,0)[r ] 0 ( 528,1867 ) ( 0,0)[r ] 1 ( 528,2158 ) ( 0,0)[r ] 2 ( 528,2448 ) ( 0,0)[r ] 3 ( 528,2739 ) ( 0,0)[r ] 4 ( 910,484 ) ( 0,0 ) 0.005 ( 1460,484 ) ( 0,0 ) 0.006 ( 2010,484 ) ( 0,0 ) 0.007 ( 2560,484 ) ( 0,0 ) 0.008 ( 3110,484 ) ( 0,0 ) 0.009 ( 3660,484 ) ( 0,0 ) 0.01 ( 286,1721 ) ( 2304,154 ) ( 0,0 ) 1/t [ k@xmath64 ( 895,1275 ) ( 0,0)[l ] @xmath49 = 74mev ( 895,1072 ) ( 0,0)[l ] @xmath54 = 1.79 m@xmath65ps@xmath66 ( 3358,2566 ) ( 0,0)[r ] diffusion constant @xmath67 ( 3358,2346 ) ( 0,0)[r ] equation ( [ bessd ] ) ( 0,0 ) for the kink center of mass on a @xmath34(1\bar{2}1)$ ] edge dislocation dipole
. a two - parameter fit of equation ( [ bessd ] ) gives a migration barrier of 74mev , a large fraction of the 150mev kink formation energy .
the linear gradient implies the dissipation parameter @xmath54 is independent of temperature ( see text ) .
( color online ) [ arrhenius],title="fig : " ] there are two points of note in the md results for kinks on edge dislocations in table [ kink_table ] .
firstly , we find that the migration barrier is comparable to the formation energy , implying that the nature of the kink mechanism on edge dislocations is complex , with double kink nucleation and kink migration occurring on similar timescales .
this agrees with previous simulations on edge dislocations@xcite where the mobility was found to be independent of the dislocation segment length .
secondly , the linear gradient of the arrhenius plot in figure [ arrhenius ] also implies , by equation ( [ kramersd ] ) , that the dissipation parameter for the kink @xmath54 is temperature independent .
+ we show in section [ sec : le ] that the dissipation parameter for a kink is proportional to the dissipation parameter for the host dislocation line , which should therefore also be temperature independent .
this is in agreement with several other studies of dislocations with a large lattice resistance@xcite , self - interstitial defects@xcite and prismatic loops@xcite in bcc iron .
however , decades of theoretical work@xcite conclude that the dissipation parameter for a dislocation should increase linearly with temperature due to the increasing phonon population , as has been found by simulation for dislocations with a negligible lattice resistance , such as @xmath34(1\bar{2}1)$ ] edge dislocations@xcite in bcc iron and many dislocations in fcc metals@xcite .
we return to this important issue concerning the coupling of dislocations to the heat bath in the following section on screw dislocations , as the diffusive form ( [ freed ] ) they exhibit allows an even more direct investigation of @xmath54 . + ( 4874.00,3004.00 ) ( 660,484 ) ( 0,0 ) 0 ( 1154,484 ) ( 0,0 ) 0.2 ( 1649,484 ) ( 0,0 ) 0.4 ( 2143,484 ) ( 0,0 ) 0.6 ( 2637,484 ) ( 0,0 ) 0.8 ( 3132,484 ) ( 0,0 ) 1 ( 3626,484 ) ( 0,0 ) 1.2 ( 4120,484 ) ( 0,0 ) 1.4 ( 440,1721 ) ( 2436,154 ) ( 0,0 ) kt [ barrier height ] ( 1251,2566 ) ( 0,0)[l ] lifson - jackson form , eq .
( [ bessd ] ) ( 1251,2346 ) ( 0,0)[l ] arrhenius form ( 1251,2126 ) ( 0,0)[l ] stochastic simulation ( 0,0 ) ) ( blue ) , the appropriate arrhenius form ( green ) and numerical data ( red ) are compared across a wide temperature range . at low temperatures all
three agree but at intermediate to high temperatures a linear temperature dependence emerges in simulation and equation ( [ bessd ] ) .
( color online)[th],title="fig : " ] while edge dislocations may be thought of as an inserted half plane of atoms@xcite producing a bonding disregistry perpendicular to the dislocation line direction , screw dislocations create a bonding disregistry along the dislocation line direction , which does not require the addition or removal of material .
screw dislocations possess a non planar core structure in bcc metals , which gives a large peierls barrier@xcite .
the complex core structure is heavily influenced by the choice of interatomic potential used in classical atomistic simulations .
the vast majority of existing potentials predict a screw dislocation has multiple core structures@xcite , leading many authors to suggest that a screw dislocation may pass through a metastable core structure during the kink nucleation process@xcite . under an applied stress this can produce a new kink formation pathway leading to a discontinuity in the flow stress@xcite .
however , this discontinuity is not shown in experiment , and recent _ ab initio _
calculations@xcite rule out any metastable core structure , with the nucleation pathway seen to occur almost entirely in the @xmath68 slip planes . a recently developed potential by gordon _ _ et al.__@xcite attempts to address these issues , concluding that while the metastable core may be removed from the nucleation pathway , multiple core structures remain .
using this potential , we introduce kinks with the core structure predicted from first principles calculations , thereby minimizing unphysical effects due to the interatomic potential . +
a screw dislocation dipole requires a triclinic simulation cell to avoid spurious image stresses ; we refer the reader elsewhere for details of the simulation method@xcite which are well established .
as before , the kink vector was added to the supercell vectors to give the boundary conditions required for isolated kinks to form on each dislocation under relaxation . as there is no mirror symmetry along @xmath69 directions in the bcc lattice , the so called ` right ' and ` left ' kinks forming a kink pair on a screw dislocation will have different kink vectors and
thus are expected to be asymmetric .
previous zero temperature calculations of isolated kinks on screw dislocations in bcc iron@xcite with the mendelev _
_ potential@xcite found a noticeable difference between the formation energies of right and left kinks , which correspond to kink vectors @xmath70=@xmath71 $ ] and @xmath72=@xmath28 $ ] .
they also found the kink formation energy converged to a constant value when the supercell length was greater than the widths of two kinks .
the supercell length measures the separation between a kink and its periodic images .
these findings are inconsistent with elasticity theory of kink interactions@xcite , according to which the far field interaction between kinks should decrease with the inverse of the kink separation .
+ to investigate these discrepancies we performed similar calculations with the improved potential by gordon _
_ , extending the supercell length to 240 burgers vectors , more than double that used in @xcite . in agreement with elasticity theory we found that when the supercell length was greater than two kink widths the kink formation energy decreased with the inverse of the supercell length , by 0.011ev for both right and left kinks over a distance of 200 burgers vectors .
the formation energy for the right and left kinks was 0.604ev and 0.13ev respectively , giving a double kink formation energy of 0.734ev , in good agreement with the previous study .
+ we conclude that the long range kink interaction , while decaying inversely with separation as predicted by elasticity theory@xcite , is a small perturbation to the kink formation energy in the atomistic simulations performed here and in @xcite . however , the difference in formation energies of left and right kinks is still unexplained . to gain insight into the kink structure , figure [ screw_e ] shows the excess potential energy , relative to a straight dislocation , per atomic plane normal to the line of a screw dislocation with a right or left kink , obtained by the coarse graining procedure described above .
the kinks appear as well defined peaks of approximately the same height but also with long range tails which differ markedly between the two kinks .
these long range tails are the source of the difference in the formation energies , with the ` core ' of each kink very similar in size and energy . + the nature of these tails
can be understood by noting that the kinks may be regarded as two short segments of dislocation with edge character .
however , the edge segments are of equal and opposite sign because their line senses are reversed . for the right kink
we see a projection of the kink vector along the dislocation line of @xmath73 , which implies the insertion of two atomic planes and is often thought of as an interstitial kink@xcite , whose compressive far field locally raises the core energy of the host dislocation . for the left kink
we see a projection of the kink vector along the dislocation line of @xmath74 , which is most appropriately thought of as the removal of @xmath75=@xmath76 planes due to the stacking sequence of the dislocation core , and is often referred to as a vacancy kink@xcite .
the tensile far field then lowers the core energy locally due to a slight relaxation towards the bulk crystal structure .
this picture also implies that the kink fields should cancel at long range . to test this we average the energy per unit length for the left and right kinked screw dislocation lines , shown in figure [ screw_e ] , where we indeed see a localized peak with a width of around 20@xmath77 , much wider than the sharp kinks of width 3@xmath77 seen on edge dislocations .
it is this core energy and width which we take to define kinks on screw dislocations . + ( 4704.00,6008.00 ) ( 528,704 ) ( 0,0)[r ] -0.01 ( 528,1040 ) ( 0,0)[r ] 0 ( 528,1376 ) ( 0,0)[r ] 0.01 ( 528,1712 ) ( 0,0)[r ] 0.02 ( 528,2048 ) ( 0,0)[r ] 0.03 ( 528,2384 ) ( 0,0)[r ] -0.01 ( 528,2720 ) ( 0,0)[r ] 0 ( 528,3056 ) ( 0,0)[r ] 0.01 ( 528,3391 ) ( 0,0)[r ] 0.02 ( 528,3727 ) ( 0,0)[r ] 0.03 ( 528,4063 ) ( 0,0)[r ] -0.01 ( 528,4399 ) ( 0,0)[r ] 0 ( 528,4735 ) ( 0,0)[r ] 0.01 ( 528,5071 ) ( 0,0)[r ] 0.02 ( 528,5407 ) ( 0,0)[r ] 0.03 ( 660,484 ) ( 0,0 ) 40 ( 1106,484 ) ( 0,0 ) 60 ( 1551,484 ) ( 0,0 ) 80 ( 1997,484 ) ( 0,0 ) 100 ( 2442,484 ) ( 0,0 ) 120 ( 2888,484 ) ( 0,0 ) 140 ( 3333,484 ) ( 0,0 ) 160 ( 3779,484 ) ( 0,0 ) 180 ( 806,5236 ) ( 0,0)[l ] left kink ( 806,5001 ) ( 0,0)[l ] @xmath72=@xmath28 $ ] ( 806,3556 ) ( 0,0)[l ] right kink ( 806,3321 ) ( 0,0)[l ] @xmath70=@xmath71 $ ] ( 806,1756 ) ( 0,0)[l ] average ( -110,3223 ) ( 2219,154)(0,0)abc stacking index ( spacing @xmath78=@xmath34 $ ] ) ( 0,0)(1\bar{1}0)$ ] screw dislocation .
the kinks appear as peaks of similar height with asymmetric tails .
the tails are removed under averaging ( black ) as described in the text .
note the large kink width as compared to the edge dislocation kinks in figure [ 110_kink ] .
( color online)[screw_e],title="fig : " ] ( 4874.00,3004.00 ) ( 528,704 ) ( 0,0)[r ] 0 ( 528,930 ) ( 0,0)[r ] 20 ( 528,1156 ) ( 0,0)[r ] 40 ( 528,1382 ) ( 0,0)[r ] 60 ( 528,1608 ) ( 0,0)[r ] 80 ( 528,1835 ) ( 0,0)[r ] 100 ( 528,2061 ) ( 0,0)[r ] 120 ( 528,2287 ) ( 0,0)[r ] 140 ( 528,2513 ) ( 0,0)[r ] 160 ( 528,2739 ) ( 0,0)[r ] 180 ( 660,484 ) ( 0,0 ) 50 ( 1130,484 ) ( 0,0 ) 100 ( 1600,484 ) ( 0,0 ) 150 ( 2070,484 ) ( 0,0 ) 200 ( 2539,484 ) ( 0,0 ) 250 ( 3009,484 ) ( 0,0 ) 300 ( 3479,484 ) ( 0,0 ) 350 ( 3949,484 ) ( 0,0 ) 400 ( 22,1721 ) ( 2304,154 ) ( 0,0 ) t [ k ] ( 1251,2566 ) ( 0,0)[l ] left @xmath79 m@xmath65ps@xmath66 ( 1251,2346 ) ( 0,0)[l ] right @xmath80 m@xmath65ps@xmath66 ( 1251,2126 ) ( 0,0)[l ] linear fit ( 851,1906 ) ( 0,0)[l ] ( @xmath81 ) ( 0,0),title="fig : " ] at finite temperature , the kink trajectories were analyzed in a similar manner to that detailed above for edge dislocations .
however , the temperature dependence of the kink diffusivities , shown in figure [ screw_d ] , is markedly different .
they exhibit a linear temperature dependence , which by equations ( [ lj ] ) , ( [ freed ] ) implies a negligible migration barrier , as found in static calculations@xcite .
it also implies that the dissipation parameter @xmath54 is independent of temperature .
+ the temperature independence is significant as all theories of dislocation damping since liebfried@xcite have concluded that the dissipation parameter for a dislocation must increase linearly with temperature due to the increased phonon population .
we emphasize that the temperature independence of the dissipation parameter is exhibited in both thermally activated diffusion , where the prefactor of the arrhenius law is independent of temperature , and in essentially free diffusion , where the gradient of the diffusion constant with temperature is independent of temperature .
the prediction of a linear temperature dependence in the dissipation parameter essentially arises due to the vanishingly small phase space predicted for one phonon scattering , meaning the leading contribution to the thermal force is a two phonon scattering term@xcite .
a consequence of the fluctuation - dissipation theorem is that the variance of the thermal phonon force acting on a body is equal to the system temperature multiplied by the dissipation parameter for that body .
it may be shown that the thermal force produced by one phonon scattering has a variance which increases linearly with temperature , meaning the dissipation parameter is temperature independent as seen here , whilst the thermal force produced by two phonon scattering has a variance which increases quadratically with temperature , giving a dissipation parameter which increase linearly with temperature , as predicted by the standard theory and as observed in the drift motion of dislocations with a low lattice resistance .
we suggest that the presence of a strong lattice coupling leads to a strong enhancement of the one body phonon scattering term , giving a temperature independent dissipation of momentum , though it is clear that this discrepancy between theory and simulation remains unexplained and is an important topic for future investigation . in this section
we have reported the results from large scale molecular dynamics simulations of isolated kinks on edge and screw dislocation lines in bcc iron .
the large simulation cells required significant computational power to obtain statistically significant results ; however the total real time simulated was still of the order of nanoseconds .
we now introduce a model which aims to reproduce the coarse grained data from the full atomistic simulation at a fraction of the computational cost , allowing access to experimentally relevant time and length scales .
there are two principal methods for simulating dislocation motion that avoid an explicit treatment of atomic dynamics : dislocation dynamics and kinetic monte carlo methods . in conventional dislocation dynamics
codes the motion of dislocations is entirely deterministic @xcite ; the stochastic dislocation dynamics observed in md simulations and experimentally can not be simulated with such codes .
the traditional technique to model thermally dominated motion is a master equation approach@xcite .
this assigns probabilities from the canonical ensemble to transitions between different system states , which are then implemented in a kinetic monte carlo simulation .
however , the large state space available to even an isolated flexible dislocation quickly renders the technique extremely cumbersome .
the assignment of a canonical distribution is hard or impossible to justify in non - equilibrium environments and while the logarithmic time scale employed improves efficiency it obscures comparison to the real - time trajectories given by experiment and atomistic simulation .
+ in this section we introduce a model which aims to reproduce the results from full atomistic simulation . the model we employ is the well known discrete frenkel - kontorova - langevin ( fkl ) model@xcite which treats the dislocation line as a discrete elastic string sitting in a periodic substrate potential . the representation of a dislocation as an elastic line was first used to model pinning by trapping sites@xcite and due to its simplicity equivalent systems appear in many areas@xcite .
a discrete langevin equation approach has recently been used to model the thermal motion of @xmath82 prismatic and vacancy dislocation loops in bcc fe by derlet _ _ et al.__@xcite . however , the absence of a substrate potential rendered the model unable to capture any kink mechanisms and the discreteness had no relation to the crystallography of the corresponding atomic system .
in contrast , the spacing of nodes of the elastic string in our model is determined by the spacing of atomic planes normal to the dislocation line .
we find this is essential to reproduce the structure and dynamics of kinks seen in atomistic simulations .
while these discreteness effects have previously been investigated theoretically by jos and duesbery@xcite in covalent materials there has been no investigation , to our knowledge , of the dynamical behavior they predict .
many different shapes of the substrate potential used in the fkl model have been investigated@xcite .
however , the only qualitative change occurs in the presence of deep metastable minima , which imply the existence of a metastable core structure . as discussed above , even for the complex case of screw dislocations , recent _
ab initio _ calculations show the kink formation process to take place in the slip plane , with no metastable core structure .
this allows us to take the substrate potential in the fkl model as a simple sinusoid , and is consistent with the approximation of taking a dislocation line to be a string of constant internal structure moving only in the slip plane .
+ the main criticism of the fkl model is its inability to capture the long range kink interaction predicted from elasticity and the different formation energies of left and right kinks seen in atomistic simulation .
however , we have seen that the long range interaction is a minor perturbation on the formation energy .
additionally , the asymmetric kink formation energy was seen to result from the long range fields of the kinks . the simplest term which captures this behavior is linear in the dislocation line gradient@xcite and
consequently will not affect the equations of motion as it may be integrated out of the lagrangian @xcite .
investigation of more complicated terms is beyond the requirements of the current investigation as very little difference was found in the diffusivities of left and right kinks . in effect our model treats the localized , symmetric kink ` cores ' shown in the lower panel of figure [ screw_e ] .
+ first we obtain analytic expressions for the kink formation energy , width and migration barrier in terms of the parameters defining the fkl model .
we then introduce the stochastic equations of motion which govern the system dynamics , obtaining an analytic expression identical in form to equation ( [ lj ] ) for the kink diffusivity . by equating these analytic expressions to the values obtained for the kink formation energy , width , migration barrier and diffusivity from md
, we may solve numerically for the fkl model parameters .
the system is then stochastically integrated and compared to the output from full atomistic simulation .
the fkl model treats a dislocation line as a discrete elastic string sitting in a periodic substrate potential .
the string is constructed from a set of harmonically coupled nodes spaced by a fixed distance @xmath83 , which we set equal to the distance between atomic planes normal to the dislocation line .
the string sits in a substrate potential of period @xmath84 , which we set equal to the projection of the relevant kink vector normal to the dislocation line , often known as the kink height . as a result
the two length scales of the model , @xmath83 and @xmath84 , are determined by the crystallography of the corresponding atomistic system .
+ taking a co - ordinate system @xmath85 , where @xmath86 lies along the ( unkinked ) dislocation line direction and @xmath87 is normal to @xmath86 in the slip plane , each dislocation is represented by a discrete line of points @xmath88 , where @xmath89 and only the @xmath90 vary with time .
each node is thus constrained to move only in the @xmath87 direction .
the potential energy is as follows : @xmath91 where @xmath92 is the amplitude of the substrate potential and @xmath93 is the harmonic coupling strength , both in units of energy per unit length , with displaced periodic boundary conditions to account for the presence of the kink @xmath94 to obtain an analytic expression for the shape of the static kink we first take the continuum limit @xmath95 . in this limit , the system energy ( [ dv ] ) with boundary conditions ( [ kbc ] ) is minimized by the soliton kink @xmath96 @xmath97 where @xmath98 is the kink position and @xmath99 is the kink width , which is proportional to @xmath100 .
an illustration of ( [ fkkink ] ) is shown as the red curve in figure [ kinkprof ] .
the soliton kink ( [ fkkink ] ) shape interpolates the numerically minimized discrete structure of ( [ dv ] ) well .
however , in the continuum limit the system energy ( [ dv ] ) is independent of the kink position , while in the discrete system the energy varies periodically with the kink position as the continuous translation symmetry is broken , in direct analogy to the peierls barrier for a dislocation .
this position dependent energy produces the kink migration barrier discussed above .
it may be shown@xcite that substituting ( [ fkkink ] ) into ( [ dv ] ) gives @xmath101 @xmath102 the first few @xmath103 are shown as functions of the equilibrium kink width @xmath99 in figure [ fkharm ] .
we see that for realistic kink widths of more than @xmath104 the leading term @xmath105 dominates by a factor of at least ten , allowing us to approximate the migration potential as a sinusoid of period @xmath83 .
the kink energy ( [ kmp ] ) has a minimum when the kink center of mass lies between two nodes , @xmath106 , as in this configuration no node lies at the maximum of the substrate potential .
this minimum kink energy , which should be equated to the formation energy from the relaxed atomistic simulation , is thus @xmath107 @xmath108 the form of the migration barrier @xmath49 provides insight into the observed behavior of kinks in atomistic simulation .
as can be seen in figure [ fkharm ] , @xmath49 decreases rapidly with the equilibrium kink width , in agreement with the observation that the narrow kinks on edge dislocations have a significant migration barrier as compared to the essentially free motion of the wide kinks on screw dislocations .
these significant effects would be entirely lost in any continuum model , emphasizing the importance of atomistic resolution in modeling dislocation dynamics .
+ we now introduce the stochastic equations of motion which govern the dynamics of the fkl model ( [ dv ] ) .
we recall that the atomistic data was time - averaged .
applying the same procedure to our model is formally equivalent to taking a strong damping limit@xcite and thus permits a first order equation of motion for the node displacements .
this approximation is supported by the absence of any ballistic motion , even over short time intervals , in the simulation data .
the output from the stochastic model may thus be subjected to identical analysis as the output from atomistic simulation , allowing us to determine whether the data produced is statistically equivalent .
+ ( 4874.00,3004.00 ) ( 528,704 ) ( 0,0)[r ] 0 ( 528,1043 ) ( 0,0)[r ] 0.1 ( 528,1382 ) ( 0,0)[r ] 0.2 ( 528,1722 ) ( 0,0)[r ] 0.3 ( 528,2061 ) ( 0,0)[r ] 0.4 ( 528,2400 ) ( 0,0)[r ] 0.5 ( 528,2739 ) ( 0,0)[r ] 0.6 ( 660,484 ) ( 0,0 ) 1 ( 1104,484 ) ( 0,0 ) 1.5 ( 1548,484 ) ( 0,0 ) 2 ( 1992,484 ) ( 0,0 ) 2.5 ( 2437,484 ) ( 0,0 ) 3 ( 2881,484 ) ( 0,0 ) 3.5 ( 3325,484 ) ( 0,0 ) 4 ( 3769,484 ) ( 0,0 ) 4.5 ( 4213,484 ) ( 0,0 ) 5 ( 40,1721 ) ( 2436,154 ) ( 0,0 ) kink width @xmath99 [ @xmath83 ] ( 3622,2566 ) ( 0,0)[r ] @xmath110 ( 3622,2306 ) ( 0,0)[r ] @xmath111 ( 3622,2046 ) ( 0,0)[r ] @xmath112 ( 0,0 ) ) as a function of the equilibrium kink width @xmath113 .
we see the leading term @xmath105 dominates but all three terms decrease rapidly with increasing @xmath113 .
( color online ) [ fkharm],title="fig : " ] the thermal behavior of the system is investigated through the stochastic integration of first order langevin equations@xcite , which balance a frictional force proportional to the velocity , @xmath114 , the conservative force @xmath115 and a ` fluctuation force ' which will be detailed below .
this simplified equation of motion allows us to integrate the system on a much coarser timescale .
in addition , any notion of a dislocation mass is assigned to the dissipation parameter @xmath116 , which measures the rate of momentum transfer from the dislocation to the heat bath . in this way
we avoid the controversial concept of dislocation inertia as inertial effects were not exhibited in the atomistic simulations we wish to reproduce with this model .
the first order equation of motion for our discrete system ( [ dv ] ) , with boundary conditions ( [ kbc ] ) , is @xmath117 where @xmath116 is the dissipation parameter for the dislocation line and the @xmath118 are independent gaussian random variables@xcite representing the stochastic force from the surrounding heat bath .
they are defined under an ensemble average @xmath119 of all heat baths at a temperature @xmath120 , possessing only an average and standard deviation by the central limit theorem .
these read @xmath121 the amplitude of the fluctuations @xmath122 is uniquely determined by the fluctuation - dissipation theorem , which requires that the steady state solution to the fokker - planck equation associated with ( [ leom ] ) is the canonical distribution@xcite .
the absence of any spatial correlation in the noise forces reflects the chaotic atomic dynamics of the surrounding heat bath and does not preclude any correlation in the dislocation motion ;
however the delta function @xmath123 is strictly the limiting case of a vanishingly small correlation time in the atomic collisions which constitute the heat bath@xcite .
this limit may be taken only when we operate on a sufficiently coarse timescale much longer than an individual collision , which is indeed the case in the first order equations of motion investigated here .
+ we now have a completely specified system ; we will integrate the equations stochastically ( [ leom ] ) , using a pseudorandom number algorithm@xcite to generate the stochastic forces ( [ gwn ] ) . however , in order to have expressions for all the quantities extracted from atomistic simulation in terms of the model parameters we must also derive the kink diffusion constant .
+ ( 4874.00,3004.00 ) ( 528,844 ) ( 0,0)[r ] 0 ( 528,1195 ) ( 0,0)[r ] 0.5 ( 528,1546 ) ( 0,0)[r ] 1 ( 528,1897 ) ( 0,0)[r ] 1.5 ( 528,2248 ) ( 0,0)[r ] 2 ( 528,2599 ) ( 0,0)[r ] 2.5 ( 660,484 ) ( 0,0 ) 30 ( 1208,484 ) ( 0,0 ) 40 ( 1756,484 ) ( 0,0 ) 50 ( 2305,484 ) ( 0,0 ) 60 ( 2853,484 ) ( 0,0 ) 70 ( 3401,484 ) ( 0,0 ) 80 ( 3949,484 ) ( 0,0 ) 90 ( 22,1721 ) ( 2304,154 ) ( 0,0 ) @xmath32 plane index ( 0,0),title="fig : " ] to do this , we obtain first the kink equation of motion , assuming that the kink position @xmath124 is slowly varying in time in comparison to the fluctuations of the @xmath125 , and that the kink shape suffers only small perturbations due to these fluctuations throughout its motion .
it will be seen that these approximations still give excellent agreement with the results from stochastic integration , and allow us to equate the field center of mass to the kink position @xmath124 . to derive the kink equation of motion
we evaluate the total velocity projected along the dislocation line .
we consider segments connecting neighboring nodes of the dislocation line and take the projection of the nodal velocity along the segment normal .
we then take the projection of this normal along the line direction @xmath86 to obtain the contribution to the net projection from that segment .
a very similar calculation is used when calculating the driving force on a ship s sail@xcite and is formally equivalent to the derivation of the field momentum of ( [ dv ] ) in the continuum limit@xcite .
it can be shown that this gives a kink velocity @xmath126 using ( [ leom ] ) , ( [ kmp ] ) and ( [ fkkink ] ) we obtain the kink equation of motion @xmath127 where the kink dissipation parameter @xmath54 is given by @xmath128 @xmath129 is a one dimensional gaussian random variable with an average and standard deviation @xmath130 and @xmath49 is given in equation ( [ kink_energy ] ) .
we note that equation ( [ gamma_kink ] ) shows the dissipation parameter for the dislocation line , @xmath116 , to be directly proportional to the dissipation parameter for the kink , @xmath54 , and thus @xmath116 is also temperature independent .
equation ( [ dkel ] ) describes a point particle undergoing one dimensional stochastic motion in a sinusoidal potential . as discussed in section [ sec : md ] , under an ensemble average the mean squared displacement exhibits diffusive behavior with a diffusion constant ( [ bessd ] ) , where now @xmath49 and @xmath54 are given explicitly in terms of the free model parameters @xmath131 and the crystallographically determined @xmath132 . by inverting the relations ( [ fkkink ] ) , ( [ kink_energy ] ) and ( [ gamma_kink ] )
we may determine the model parameters for the dislocations considered here .
with these parameters , the system ( [ dv ] ) was first relaxed by a conjugate gradient algorithm , with the boundary conditions ( [ kbc ] ) , to determine the formation energy and kink width , then the equations of motion ( [ leom ] ) were stochastically integrated .
the kink formed remains well defined at finite temperature , as shown in figure [ kinkprof ] .
it can be seen that the data is identical in form to that produced from atomistic simulation , shown in figure [ 110_kink ] , with the kink trajectories extracted and analyzed in an identical manner .
kink_table ] shows the results from these simulations as compared to the results from atomistic simulation , displaying excellent agreement over a wide range of temperature .
+ it is emphasized that the discreteness of the model is essential to produce a kink migration barrier .
we also note that the discreteness , which is determined by the underlying crystallography through @xmath83 and @xmath84 , influences the kink formation energy as well as the migration barrier , as can be seen in equation ( [ kink_energy ] ) . as we simulate a line of only @xmath2@xmath133 nodes on a coarse time step of @xmath134 ps , as opposed to the entire atomistic system of @xmath135 atoms on a very fine time step of @xmath136 fs
, we may generate data sets equivalent to those produced from atomistic simulation at around @xmath2@xmath3 of the computational cost .
therefore , despite the atomistic resolution along the line , the model affords enormous computational savings as compared with a full md simulation .
this significant efficiency gain allows us to simulate dislocation motion at experimental strain rates , while retaining atomistic resolution and a statistically rigorous temperature .
0.48@cccccdislocation & simulation & @xmath137 & @xmath49 & @xmath54 + system & method & [ ev ] & [ ev ] & [ m@xmath65ps@xmath64 + @xmath138(1\bar{2}1)$ ] edge & md & 0.150 & 0.074 & 1.79 + & fkl & 0.148 & 0.072 & 1.74 + @xmath26(010)$ ] edge & md & 0.510 & 0.222 & 2.61 + & fkl & 0.505 & 0.218 & 2.58 + @xmath138(1\bar{1}0)$ ] screw & md & 0.367 & - & 1.83 + ( average ) & fkl & 0.367 & - & 1.82 + the parameters obtained from the kink diffusion simulations are now used to investigate the motion of straight dislocations at experimental stress levels .
this important regime is not accessible to atomistic simulation for dislocations which have a large kink formation energy .
therefore , this is an ideal application of the fkl model . for a discrete dislocation segment of @xmath139 nodes , we supplement the equations of motion ( [ leom ] ) with a force per node @xmath140 to induce drift of the dislocation line , @xmath141 with periodic boundary conditions @xmath142 we then extract the position of the center of mass @xmath143 at each timestep , obtaining the ensemble average center of mass velocity @xmath144 in an identical manner to that shown in equation ( [ mdmsd ] ) , @xmath145 to obtain the relationship between the force @xmath140 and an applied stress , we recall the peach - koehler formula@xcite for the force per unit length @xmath146 on a dislocation of burgers vector @xmath78 and line direction @xmath6 under an applied stress @xmath147 , @xmath148 the nodal force @xmath140 is then the projection of ( [ pk ] ) along the displacement direction of the @xmath125 , @xmath149 , multiplied by the segment separation @xmath83 .
we apply a shear stress across the slip plane of magnitude @xmath150 in the direction of the dislocation burgers vector @xmath78 , resulting in a force per node of @xmath151 we apply experimental stresses of 40 mpa , which corresponds to a very small force per node of @xmath2@xmath152ev/@xmath84 . to demonstrate phenomena
this discrete model can treat , we investigate the effect of segment length on dislocation velocity .
figures [ m_e_s],[m_e_d ] show typical center of mass trajectories of @xmath138(1\bar{1}0)$ ] screw and @xmath26(010)$ ] edge dislocation segments . extracting the center of mass velocity through equation ( [ drift_fkl ] ) over a wide range of segment lengths
gave a length independent velocity for edge segments , whereas the velocity increased linearly with segment length as @xmath153 at 300k for screw dislocations .
this is in good agreement with the linear relationship of velocity with length of around @xmath154 at 300k for screw dislocation segments that has recently been observed experimentally@xcite .
we note that this gradient depends exponentially on temperature due to the activated nature of the dislocation migration .
+ we can understand these differences in terms of the discrete structure of the fkl dislocation lines .
we have seen that the wide kinks on screw dislocations have a negligible migration barrier ; as a result , once a double kink is formed it will move quickly under the applied stress until it meets another kink . for long segments
we therefore expect the dislocation velocity to scale linearly with segment length due the increased number of possible locations for kink nucleation , as observed .
in contrast , as the narrow kinks on edge dislocations have a large kink migration barrier , comparable to the kink formation energy . as a result
we expect the mobility always to be independent of the segment length as kink migration and kink nucleation occur on similar timescales .
this behavior , which is found only in a discrete line model , is expected to have consequences in many aspects of dislocation behavior , for example the difference in the effect of impurities on the mobilities of edge and screw dislocations@xcite .
( 4874.00,3004.00 ) ( 528,704 ) ( 0,0)[r ] 0 ( 528,1017 ) ( 0,0)[r ] 5 ( 528,1330 )
( 0,0)[r ] 10 ( 528,1643 ) ( 0,0)[r ] 15 ( 528,1956 ) ( 0,0)[r ] 20 ( 528,2269 ) ( 0,0)[r ] 25 ( 528,2582 ) ( 0,0)[r ] 30 ( 660,484 ) ( 0,0 ) 0 ( 1548,484 ) ( 0,0 ) .5 ( 2437,484 ) ( 0,0 ) 1 ( 3325,484 ) ( 0,0 ) 1.5 ( 4213,484 ) ( 0,0 ) 2 ( 90,1721 ) ( 2436,154 ) ( 0,0 ) time [ s ] ( 1415,2087 ) ( 0,0)[l ] l @xmath155 200 nm , @xmath15616@xmath157 ( 2081,917 ) ( 0,0)[l ] l @xmath155 100 nm , @xmath1568@xmath157 ( 0,0)(1\bar{1}0)$ ] screw dislocation segments , under an applied stress of 40 mpa , at a temperature of 300k .
the highly stepped motion reflects directly the nucleation of kinks , which then quickly propagate along the entire line due to the negligible kink migration barrier . as a result
longer lines , which have a greater number of nucleation sites , have a velocity which increases linearly with segment length.[m_e_s],title="fig : " ] ( 4874.00,3004.00 ) ( 528,702 ) ( 0,0)[r ] 0 ( 528,1124 ) ( 0,0)[r ] 3 ( 528,1546 ) ( 0,0)[r ] 6 ( 528,1968 ) ( 0,0)[r ] 9 ( 528,2390 ) ( 0,0)[r ] 12 ( 660,484 ) ( 0,0 ) 0 ( 1548,484 ) ( 0,0 ) .5 ( 2437,484 ) ( 0,0 ) 1 ( 3325,484 ) ( 0,0 ) 1.5 ( 4213,484 ) ( 0,0 ) 2 ( 90,1721 ) ( 2436,154 ) ( 0,0 ) time [ s ] ( 1015,2007 ) ( 0,0)[l ] l @xmath155 200 nm , @xmath1567@xmath158 ( 2081,1274 ) ( 0,0)[l ] l @xmath155 100 nm , @xmath1566.5@xmath158 ( 0,0)(010)$ ] edge dislocation segments , under an applied stress of 40 mpa , at a temperature of 600k . as the kink migration barrier is comaparable to
the kink formation energy double kink nucleation occurs on the same timescale as kink migration and thus the line propagates gradually . as a result the segment velocity is almost independent of the segment length , as distinct from screw dislocation segments , a feature not captured by a continuum line model.[m_e_d],title="fig : " ]
in this paper we have presented results from large scale md simulations of edge and screw dislocation lines in bcc fe .
boundary conditions were exploited to produce isolated kinks , which were seen to remain isolated on the timescale of the md simulations provided the thermal energy was significantly less than the double kink formation energy .
the dislocation lines were coarse grained , while retaining atomistic resolution , by assigning a dislocation core position in each atomic plane normal to the line direction .
kinks were clearly identifiable , with statistical analysis allowing extraction of a kink diffusion constant .
+ kinks on edge dislocations were seen to be narrow , exhibiting thermally activated stochastic motion which was described well by an arrhenius law . in contrast , kinks on screw dislocations were broad with a diffusivity that varied linearly with temperature , implying a vanishingly small migration barrier .
the difference in the formation energies of left and right kinks on screw dislocations was seen to be attributable to the structure of the kinks outside their cores . the kink interaction energy predicted by elasticity
was observed to be much smaller than the kink core energy . + an analytic result for one dimensional stochastic motion in a periodic potential , valid for all temperatures and barrier heights ,
was seen to rationalize the wide range of diffusive behavior , leading to the conclusion that the frictional force on a kink , and hence the host dislocation , is temperature independent .
this result from direct atomistic simulation is in agreement with other studies on dislocations with a large lattice resistance , but directly opposes the textbook theory which states that the frictional force should be proportional to temperature .
this significant disagreement is an important topic for future investigation . +
the discrete one - dimensional fkl model was seen to be able to reproduce the observed motion of kinks on edge and screw dislocations over a wide range of temperatures and dislocation geometries , with all length scales fixed by the crystallography .
the discrepancy between kinks on edge and screw dislocations was explained through an analytic relationship between the kink width and migration barrier ; it was noted that the migration barrier depends sensitively on the discrete structure , with the atomistic resolution of the model being essential to reproduce the detailed behavior of thermally activated dislocation glide .
this discrete structure , which is absent in conventional dislocation dynamics codes , is expected to be significant in thermally activated dislocation phenomena such as impurity interaction and cross slip .
+ the application of the fkl model to the motion of initially straight dislocation segments under experimental applied stresses found a noticeable length dependence for screw dislocation segments , due to the negligible kink migration barrier , whereas the large migration barrier for kinks on edge dislocations suppressed a length dependence .
the consequences of this highly anisotropic mobility of discrete dislocation lines for microstructural evolution will be investigated in future work .
we are grateful to a referee for helpful comments on an earlier draft of this paper .
tds was supported through a studentship in the centre for doctoral training on theory and simulation of materials at imperial college london funded by epsrc under grant number ep / g036888/1 .
this work , partially supported by the european communities under the contract of association between euratom and ccfe , was carried out within the framework of the european fusion development agreement .
the views and opinions expressed herein do not necessarily reflect those of the european commission .
this work was also part - funded by the rcuk energy programme under grant ep / i501045 . | isolated kinks on thermally fluctuating @xmath0 screw , @xmath1 edge and @xmath0 edge dislocations in bcc iron are simulated under zero stress conditions using molecular dynamics ( md ) .
kinks are seen to perform stochastic motion in a potential landscape that depends on the dislocation character and geometry , and their motion provides fresh insight into the coupling of dislocations to a heat bath .
the kink formation energy , migration barrier and friction parameter are deduced from the simulations . a discrete frenkel - kontorova - langevin
( fkl ) model is able to reproduce the coarse grained data from md at @xmath2@xmath3 of the computational cost , without assuming an _ a priori _ temperature dependence beyond the fluctuation - dissipation theorem .
analytic results reveal that discreteness effects play an essential rle in thermally activated dislocation glide , revealing the existence of a crucial intermediate length scale between molecular and dislocation dynamics .
the model is used to investigate dislocation motion under the vanishingly small stress levels found in the evolution of dislocation microstructures in irradiated materials .
dislocation motion is limited by two general processes , the formation and migration of kinks and pinning by impurities and other defects @xcite . in this paper
we investigate the motion of kink - limited screw and edge dislocations in bcc fe , where the kink formation energy is much larger than the thermal energy . to obtain dislocation motion on the time - scales accessible to molecular dynamics ( md ) simulations
some researchers have resorted to inducing kink formation by applying stresses some six orders of magnitude greater than those pertaining experimentally @xcite . but
dislocation core structures and peierls barriers are known to be highly stress - dependent@xcite , making it difficult to relate simulation to the vanishingly low stress conditions found in thermally activated evolution of dislocation microstructures .
+ to avoid the problem of nucleating kinks in md simulations we use periodic boundary conditions which enforce the existence of an isolated kink on each dislocation line in the supercell .
this allows us to study the detailed dynamics of kink migration in isolation , a crucial and previously unexplored aspect of kink - limited dislocation motion , as seperate from the kink nucleation process . under no applied stress ,
kinks are seen to undergo stochastic motion in a potential landscape that varies with the dislocation character and burgers vector .
a coarse graining procedure is introduced , that facilitates statistical analysis yielding many properties of the kink motion , and which provides physical insight into the coupling of dislocations to the heat bath . in particular , the friction parameter for a dislocation is found to be temperature - independent , contradicting decades of theoretical work since liebfried@xcite .
but this temperature - independence is seen in many other investigations of dislocations@xcite with a large lattice resistance and in the stochastic motion of point defects@xcite , although to the best of our knowledge it has not been highlighted before .
+ our main result is that a stochastic , discrete line representation of the dislocation , the discrete frenkel - kontorova - langevin ( fkl ) model@xcite , is able to reproduce quantitatively the coarse grained data obtained from md simulations of thermally activated dislocation glide at @xmath2@xmath3 of the computational cost , with no _ a priori _ assumption of any temperature dependence beyond the fluctuation - dissipation theorem .
the computational efficiencies of the model are exploited to investigate dislocation motion under experimental stress levels inaccessible to md .
+ we find the discreteness of the model , which is determined by the underlying crystallography , is essential to produce the thermally activated dynamics of dislocations seen in atomistic simulation .
this shows that in order to model the kinetics of thermally activated dislocation motion beyond the limitations of atomistic simulation , a coarse grained model has to be sensitive to length scales smaller than those considered in traditional dislocation dynamics simulations .
+ the paper is structured as follows . in section [ sec : md ] we briefly review dislocation glide in bcc metals and the kink mechanism .
we present results from large scale md simulations of isolated kinks on @xmath0 screw , @xmath0 edge and @xmath4 edge dislocation dipoles in bcc fe .
kinks are given a crystallographic label , the kink vector , which enables a systematic enumeration of the kinks a dislocation line may support .
the observed asymmetry between left and right kinks on screw dislocations is rationalized in terms of the kink vector , with the kink ` core ' being essentially symmetric . at finite temperature , under zero applied stress , kinks are seen in the simulations to perform stochastic motion in a potential landscape that varies with the dislocation character and geometry .
we then introduce our coarse graining procedure .
+ in section [ sec : le ] we review the fkl model , using analytic expressions for kink properties to determine the parameters of the fkl model from md simulations . the fkl model is known to lack any interaction between kinks @xcite ; we show that this deficiency has a very small effect on the kink formation energy and may be neglected in the investigation of very well separated kinks , allowing us to deduce parameters of the fkl model for the host dislocation .
numerical integration of the stochastic equations of motion of the fkl model produces data which may be processed identically to that from atomistic simulation , allowing us to compare the statistical results obtained from both methods .
we find the transport properties of kinks in the md simulations and their fkl counterparts to be in excellent agreement over a wide range of temperature for different dislocations .
the parametrized fkl models are used to investigate screw and edge dislocation mobilities at applied stresses too low to induce dislocation motion in md . |
any function @xmath1 which maps a phase space @xmath2 into itself may be considered as a classical dynamical system .
an arbitrary point @xmath3 determines the trajectory uniquely , @xmath4 , where the natural index @xmath5 plays the role of time .
dynamical system defined by @xmath1 is deterministic .
there exist different ways to introduce randomness into the system .
one possibility relays on taking into account an additive noise , @xmath6 , where @xmath7 are uncorrelated random variables drawn according to a prescribed probability distribution @xmath8 .
the deterministic case is obtained if @xmath8 tends to the dirac delta distribution , @xmath9 . in this work
we analyze another class of stochastic systems , in which the choice of a system is random at each step .
such systems are called _ iterated function system _ ( ifs ) , and may be considered as a generalization of the notion of a classical dynamical system .
an ifs is defined @xcite by a set of @xmath10 functions @xmath11 , which represent dynamical systems in the classical phase space @xmath2 .
the functions @xmath12 act randomly with given probabilities @xmath13 .
they characterize the likelihood of choosing a particular map at each step of the time evolution of the system . in general
, the probabilities may be place dependent . analyzing a given iterated function system , one may ask , how an initial point @xmath14 is transformed by the random process .
one may also consider a more general question , how does a probability measure @xmath15 on @xmath2 change with respect to the frobenius - perron operator @xmath16 associated with the ifs .
if the functions @xmath12 are strongly contracting , then there exists unique invariant measure @xmath17 of @xmath16 see for instance @xcite and references therein .
interestingly , for a large class of ifs s the invariant measure @xmath18 is characterized by a fractal structure .
if the invariant measure is attracting , such ifs s may be used to generate fractal sets in the space @xmath2 . in particular , there exist iterated function systems leading to popular fractal sets : cantor set , sierpinski carpet or sierpinski gasket @xcite .
in fact by considering a slightly generalized class of ifs s it is possible to design a random system , the invariant measure of which possesses any desired properties . in this work we shall visualize the statement , by presenting an euroattractor - a system , the attractor of which has the shape resembling the contours of europe .
this paper is organized as follows . in the next section we recall the definition and basic properties of the standard ifs .
section iii is devoted to a weakly contracting ifs , which exhibit nontrivial invariant measures , although they do not fulfill the standard contractivity assumptions . in section
iv we define a generalized class of the systems and present the definition of an euroattractor .
an iterated function system ( ifs ) is defined by a set of @xmath10 functions ( dynamical systems ) @xmath12 , @xmath19 , defined on a compact space @xmath2 and the set of @xmath10 positive probabilities @xmath13 , which sum to unity . in the ifs of the first kind the probabilities are constant , while in the more general , ifs of the second kind @xcite , they are functions of the position , @xmath20 . for any @xmath21
the condition @xmath22 has to be fulfilled . in general , ifs @xmath23
is defined by a set @xmath24 \}$ ] with @xmath25 .
any ifs generates a frobenius
perron operator @xmath16 , which describes the time evolution of the probability densities on @xmath2 . in the simplest case ,
in which the space @xmath2 is an interval in @xmath26 and the maps @xmath12 are invertible , any initial probability density @xmath27 is mapped into @xmath28 by the frobenius
perron operator @xcite @xmath29 = \sum_{i=1}^k p_i \bigl(f_i^{- 1}(x)\bigr)\rho\bigl(f_i^{-1}(x)\bigr ) \left| \frac{d f_i^{-1}(x)}{dx } \right| .
\label{fp1}\ ] ] if @xmath27 is a non - singular probability density , so is @xmath28 .
however , in the limit @xmath30 the image @xmath31 $ ] needs not to be a smooth probability density , so it is advantageous to work in the larger space of probability measures on @xmath2 .
this space includes also also singular measures like the dirac delta distributions and their convex combinations .
if an ifs is sufficiently contracting , i.e. \a ) the functions @xmath12 satisfy the lipschitz condition , @xmath32 ) for any @xmath33 . here
@xmath34 denotes the distance between both points and the lipschitz constants @xmath35 are assumed to be smaller than unity , \b ) the probabilities @xmath36 are continuous functions , strictly positive for any @xmath37 , the ifs is called _ hyperbolic _ and there exists a unique _ invariant probability measure _ @xmath18 satisfying the equation @xmath38 @xcite .
interestingly , the conditions a ) and b ) are not necessary to assure the existence of a unique invariant probability measure .
some other less restrictive sufficient assumptions which regard also ifs of the second kind , with place dependent probabilities , were analyzed in @xcite .
the invariant measures of an ifs display often fractal properties .
for instance , consider an ifs of the first kind defined on an interval and consisting of @xmath39 affine transformations , @xmath40 , \
f_1(x)=x/3 , \ f_2(x)=x/3 + 2/3 ; \quad p_1=p_2=\frac{1}{2 } \}. \label{ifs1}\ ] ] since both functions are continuous contractions with lipschitz constants @xmath41 , and the probabilities @xmath13 are constant and nonzero this ifs is hyperbolic .
thus there exists a unique , attracting invariant measure @xmath18 .
it is not difficult to show @xcite that @xmath18 is concentrated uniformly on the cantor set of the fractal dimension @xmath42 .
although @xmath18 is singular , it is possible to give a general prescription , how to compute integrals with respect to this invariant measure .
this is due to the fact that the measure @xmath18 is _ attractive _ , i.e. , @xmath43 converges weakly to @xmath18 if @xmath44 .
therefore , in order to obtain the exact value of an integral @xmath45 for any continuous function @xmath46 it is sufficient to find the limit of the sequence @xmath47 for an arbitrary initial measure @xmath15 .
this method of computing integrals over the invariant measure @xmath18 is purely _
deterministic _ @xcite .
alternatively , one may employ a _ random iterated algorithm _ by generating by the ifs a random sequence @xmath48 , @xmath49 which originates from an arbitrary initial point @xmath50 .
due to the the ergodic theorem for ifs s @xcite the mean value @xmath51 $ ] converges in the limit @xmath52 to the desired integral , @xmath45 , with probability one . to show an example of an ifs of the second kind let us change probabilities in @xmath53 and define @xmath54 , \
f_1(x)=x/3 , \ f_2(x)=x/3 + 2/3 ; \quad p_1(x)=x , \
p_2(x)=1-x \}. \label{ifs2}\ ] ] the probability @xmath55 vanish at the point @xmath56 , so this ifs is not hyperbolic . in spite of
this fact there exists the unique invariant measure @xmath18 concentrated in a non - uniform way on the cantor set @xcite . in this case
@xmath18 displays multifractal properties , since the generalized dimension of the measure depends on the rnyi parameter @xcite .
contraction conditions a ) and b ) imply existence of a unique invariant measure for the ifs . however , these conditions are too strong , and there exist ifss with an attracting invariant measure which do not obey them . if the space @xmath57 and the functions @xmath12 are linear , the lipschitz condition a ) is equivalent to the statement that the moduli of all elements of the jacobi matrices @xmath58 are strictly smaller then one .
this is not the case for the following ifs defined on the square @xmath59\times [ 0,1 ] , \
k & = & 2 ; \quad p_1=p_2 = 1/2 \\ f_1(x , y ) & = & ( x y , ~ y ) \\
f_2(x , y ) & = & ( ( x-1)y+1 , ~ y+\delta ) \end{array } \right\ } , \label{ifs3}\ ] ] since the jacobi matrices @xmath60{cc } y & x\\ 0 & 1 \end{array } \right ) { \rm \quad and \quad } j_2=\left ( \begin{array } [ c]{cc } y & x-1\\ 0 & 1 \end{array } \right ) \label{jacobi}\ ] ] contain entries equal to unity . this reflects the fact that both functions @xmath61 and @xmath62 are not contracting in the @xmath63direction .
numerical experiments suggest that in spite of this fact for irrational values of @xmath64 , the associated frobenius
perron operator @xmath16 possess a unique invariant measure @xmath18 .
such an ifs , not fulfilling the condition a ) or b ) will be called _
weakly contracting_. ( [ ifs3 ] ) obtained for @xmath65 .
the cross - section of its support at @xmath66 , denoted by a dashed line , forms the cantor set.,scaledwidth=60.0%,scaledwidth=50.0% ] each random trajectory of the ifs @xmath67 exhibits a slow diffusion in the @xmath63direction , the speed of which is governed by the rotation parameter @xmath68 . due to periodic boundary conditions , @xmath69 ,
every trajectory is confined to the square @xmath2 . if the parameter @xmath68 is chosen to be irrational , there are no periodic trajectories
. if @xmath68 is sufficiently small than the contraction in the @xmath70direction will dominate the diffusion in @xmath63 , so the trajectory will be sticked to the support of the invariant measure @xmath18 .
figure 1 shows a typical trajectory consisting of @xmath71 points generated by this ifs . to present the asymptotic properties of the invariant measure
the first @xmath72 points were omitted .
note that if the parameter @xmath68 is set to zero , the variable @xmath63 would be fixed to the initial value @xmath73 .
such an ifs would have an invariant measure concentrated on the interval @xmath74 . analyzing the contraction
constant we see that for @xmath75 its support is equal to the cantor set generated by the cantor ifs ( [ ifs1 ] ) . in general , for any @xmath76 its dimension would be @xmath77 we conjecture that this formula gives also the fractal dimension of the horizontal cross - section at @xmath74 of the invariant set of the ifs ( [ ifs3 ] ) in the limit @xmath78 .
another example of an invariant measure of a nonlinear , weakly contracting ifs is presented in fig.2 .
the system is defined by @xmath79\times [ 0,1 ] , \
k & = & 2 ; \quad p_1=p_2 = 1/2 \\ f_1(x , y ) & = & ( \frac{1}{2 } x \sin^2(\pi y ) , ~ y ) \\
f_2(x , y ) & = & ( \frac{1}{2 } ( x-1)\sin^2(\pi y)+1 , ~ y+\delta ) \end{array } \right\ } .
\label{ifs4}\ ] ] as in the previous example each random realization of the ifs generates a trajectory confined in the torus , and for small values of the shift parameter @xmath68 the contraction in the horizontal direction dominates the vertical diffusion . for @xmath80
the ifs reduces to one - dimensional , and the fractal dimension of its invariant set reads @xmath81 this number is smaller than one for all initial values apart of @xmath82 .
numerical results performed for @xmath64 suggest that the same formula describes the dimension of the cross - section at @xmath74 of the invariant set of the ifs ( [ ifs4 ] ) in the limit @xmath83 . in other words , the support of the invariant measure @xmath18 of the ifs @xmath84 may be considered as a collection of cantor sets , with a continuously varying dimension @xmath85 , placed horizontally , one parallel to another . )
obtained for @xmath65 .
, scaledwidth=60.0%,scaledwidth=50.0% ]
in the original definition of the ifs , presented in section ii , all functions @xmath12 map the space @xmath2 into itself .
one may , however , relax this condition , and consider generalized ifs constructed of functions which map some fragments of @xmath2 out of this space .
such functions are called _ repellers _ , since they do not conserve the probability , and during the time evolution the probability flux leaks out of the system .
a simple @xmath86-d repeller is obtained from the standard logistic map , @xmath87 , if the control parameter @xmath88 exceeds the critical value , @xmath89 @xcite . in such systems
a randomly selected initial value leads to a trajectory , which escapes the system with probability one .
the complementary set of initial points which generate infinite trajectories is of measure zero and often forms a fractal set .
the same property is characteristic to a large class of chaotic 1-d maps with a gap - a subset of @xmath2 mapped out of @xmath2 .
an analysis of the fractal dimension of the invariant set for the tent map with a gap is presented in @xcite . in this chapter
we will consider an ifs defined on the square , @xmath90 \times [ 0,1]$ ] , consisting of @xmath10 repelling functions .
assume that each function @xmath12 restricted to a given set @xmath91 acts as an affine transformation defined by real matrices @xmath92 of size @xmath93 and two components translation vectors @xmath94 , but the complementary subset @xmath95 is mapped out of @xmath2 , @xmath96 such functions @xmath12 are repelling , so instead of analyzing single trajectories , generated by a random initial point , it is convenient to work with initially smooth probability measures
. let us take for the initial measure @xmath97 the uniform ( lebesgue ) measure on @xmath2 . to iterate the repelling ifs we find the images of the measures covering each of the subsets @xmath98 by the maps @xmath12 , @xmath19 and add all contributions with prescribed weights
probabilities @xmath13 .
more formally , we use the the dimensional analogue of the frobenius - perron operator ( [ fp1 ] ) @xmath99 = \sum_{i=1}^k p_i \bigl(f_i^{- 1}(x , y)\bigr)\rho\bigl(f_i^{-1}(x , y)\bigr ) \left| \frac{\partial f_i(x , y)}{\partial ( x , y ) } \right|^{-1 } , \label{fp2}\ ] ] where the contributions from the function @xmath12 have to be taken into account only if the analyzed point @xmath100 belongs to @xmath101 .
then the preimage @xmath102 exists , so the @xmath103-th term in the expression ( [ fp2 ] ) is meaningful . which send the polygons @xmath98 into @xmath104 , and the remaining set @xmath105
is mapped out of the system.,scaledwidth=80.0% ] as an example of such a generalized , repelling system we consider the ifs @xmath106 consisting of @xmath107 functions with equal weights , @xmath108 .
the detailed form of the affine functions @xmath12 as well as the definitions of the sets @xmath98 are provided in the appendix .
figure 4 shows time evolution of the initially uniform measure @xmath97 .
the right hand side of the figure shows the measures @xmath109 $ ] , while the left hand side shows the sets , where the measure @xmath110 is concentrated .
initially uniform measure @xmath97 converges fast to the asymptotic measure @xmath17 , and already after @xmath111 iterations can hardly be distinguished from it .
the shape of the support of @xmath17 explains why the name _ euroattractor _ has been concocted for this system . in the square .
already after @xmath111 iterations the measure can not be distinguished from the invariant measure @xmath18 . ]
the purpose of this work was to demonstrate usefulness of the concept of _ iterated function system_. this generalization of the concept of a dynamical system incorporates deterministic and stochastic behavior and may be applied in various fields of physics , and in general , of nonlinear science .
a possible extension of the concept of ifs for quantum mechanics was recently worked out in @xcite .
the ifs s belong to a larger class of random systems studied in @xcite . from a mathematical point of view
several questions concerning the rigorous theory of ifs remain open .
for instance , the sufficient conditions , under which an ifs possesses a unique , attracting invariant measure are known , but the formulation of the necessary conditions remains still an exciting mathematical challenge @xcite . in this work
we have introduced a weakly contracting ifs s @xmath67 and @xmath84 , which do not satisfy the standard contractivity conditions , but do posses attracting invariant measures . for
any given set @xmath112 in the plane one may try to find an ifs , such that its invariant measure @xmath18 has support on @xmath112 .
this versatility of ifs s allows one to apply them for image synthesis , image processing and image encoding @xcite .
for instance , the attractor of a generalized ifs @xmath106 , described in this work , has the shape resembling the contours of * europe*. it is a pleasure to thank radosawa bach and wojciech somczyski for constructive remarks and fruitful interaction .
this paper was presented at the european interdisciplinary school on nonlinear dynamics _ euroattractor 2002 _ organized in warsaw in june 2002 .
we are grateful to wodzimierz klonowski for an inspiring name of the conference as well as for the opportunity to present this work during that event .
in this appendix we provide an explicit definition of an _ euroattractor _ a generalized , repelling ifs @xmath106 , the invariant measure of which is shown in fig.4 .
the space @xmath2 is equal to an rectangle . to address each pixel in a simple way its size is taken to be @xmath113 .
the ifs consists of @xmath107 affine functions @xmath12 defined in ( [ ifs5 ] ) , and all the weights are constant , @xmath114 . the parameters defining the corners of the quadrangles @xmath98 , and the elements of the transformation matrices @xmath92 and the translation vectors @xmath115 are collected in table 1 .
observe that some entries of the jacobi matrix @xmath116 have modulus larger than one , so this ifs does not satisfy the contractivity condition . | in this work we propose a definition of an _ euroattractor _ : an attracting invariant measure of a certain iterated functions system ( ifs ) .
an ifs is defined by specifying a set of functions , defined in subsets of @xmath0 or in a classical phase space , which act randomly on the initial point , so it may be considered as a generalization of the notion of classical dynamical system .
if the functions are sufficiently contracting , there exists an invariant measure of the system , often concentrated on a fractal set .
we investigate invariant measures of a certain class of generalized or weakly contracting ifs s . |
given a 3-manifold @xmath8 in @xcite ozsvth and szab defined topological invariants of @xmath8 , indicated @xmath9 , @xmath10 , @xmath11 and @xmath12 .
these groups are the _ heegaard floer homologies of @xmath8 _ in the respective versions .
moreover ozsvth and szab in @xcite and rasmussen in @xcite proved that any homologically trivial knot @xmath0 in @xmath8 induces a `` knot filtration '' on the heegaard floer chain complexes .
the first pages of the associated spectral sequences ( in each versions ) result then to be topological invariants of @xmath0 : these are bigraded homology groups @xmath13 , @xmath14 , @xmath15 and @xmath3 called _ heegaard floer knot homologies _ ( in the respective versions ) .
these homologies are powerful invariants for the couple @xmath16 .
for instance in @xcite and @xcite , it has been proved that @xmath17 _ categorifies the alexander polynomial @xmath18 of @xmath0 _ , i.e. @xmath19 where @xmath20 means that the two sides are equal up to change sign and multiply by a monic monomial and @xmath21 denotes the _ graded euler characteristic_. this was the first categorification of the alexander polynomial ; a second one ( in seiberg - witten - floer homology ) was discovered later by kronheimer and mrowka ( @xcite ) . in @xcite ozsvth and
szab developed a similar construction for any link @xmath22 in @xmath23 and got invariants @xmath24 and @xmath25 for @xmath22 , which they called _ heegaard floer link homologies_. now these homologies come with an additional @xmath26 degree , where @xmath27 is the number of the connected components of @xmath22 .
ozsvth and szab proved moreover that _
@xmath24 categorifies the multivariable alexander polynomial of @xmath22 _ , which is a generalization of the classic alexander polynomial .
they found in particular that : @xmath28 and @xmath29 in the series of papers @xcite-@xcite , colin , ghiggini and honda prove the equivalence between heegaard floer homology and _ embedded contact homology _ for three manifolds .
the last one is another floer homology theory , first defined by hutchings , which associates to a contact manifold @xmath1 two graded modules @xmath30 and @xmath31 .
[ theorem : hf and ech in intro ] @xmath32 where @xmath33 is the manifold @xmath8 with the inverted orientation . in light of theorem [
theorem : hf and ech in intro ] , it is a natural problem to find an embedded contact counterpart of heegard floer knot homology . in analogy with the _ sutured heegaard floer theory _ developed by juhsz ( @xcite ) , in @xcite the authors define a sutured version of embedded contact homology .
this can be thought of as a version of embedded contact homology for manifolds with boundary . in particular , given a knot @xmath0 in a contact three manifold @xmath34 , using sutures they define a _ hat version @xmath2 of embedded contact knot homology_. roughly speaking , this is the hat version of @xmath35 homology for the contact manifold with boundary @xmath36 , where @xmath37 is a suitable thin tubular neighborhood of @xmath0 in @xmath8 and @xmath38 is a contact form satisfying specific compatibility conditions with @xmath0 . in @xcite the following conjecture is stated : [ conjectureintro : eck iso to hfk hat ] @xmath39 in this paper we first define a _ full version of embedded contact knot homology _
@xmath40 for knots @xmath0 in any contact three manifold @xmath34 endowed with a ( suitable ) contact form @xmath38 for @xmath41
. moreover we generalize the definitions to the case of links @xmath22 with more then one components to obtain homologies @xmath42 we state then the following : [ conjectureintro : eck iso to hfk ] for any link @xmath22 in @xmath8 , there exist contact forms for which : @xmath43 next we compute the graded euler characteristics of the @xmath6 homologies for knots and links in homology three - spheres and we prove the following : [ theoremintro : euler characteristics of eck and alex ] let @xmath22 be an @xmath27-component link in a homology three - sphere @xmath8 .
then there exists a contact form @xmath38 such that @xmath44 here @xmath45 is the _ alexander quotient _ of the complement of @xmath22 in @xmath8 .
the theorem is proved using fried s dynamic reformulation of @xmath46 ( @xcite ) .
classical relations between @xmath47 and @xmath48 imply the following result : [ theoremintro : euler characteristic of eck ] let @xmath22 be any @xmath27-component link in @xmath23 .
then there exists a contact form @xmath38 for which : @xmath49 and @xmath50 this implies that the homology _
@xmath6 is a categorification of the multivariable alexander polynomial . _ a straightforward application of last theorem
is obtained by comparing it with equations [ equationintro : euler characteristics of hfl^- ] and [ equationintro : euler characteristics of widehat hfl ] : in @xmath23 , conjectures [ conjectureintro : eck iso to hfk hat ] and [ conjectureintro : eck iso to hfk ] hold at level of euler characteristics .
i first thank my advisors paolo ghiggini and vincent colin for the patient and constant help and the trust they accorded to me during the years i spent in nantes , where most of this work has been done .
i also thank the referees for having reported my ph.d .
thesis , containing essentially all the results presented here .
i finally thank vinicius gripp , thomas guyard , christine lescop , paolo lisca , margherita sandon and vera vrtesi for all the advice , support or stimulating conversations we had .
this subsection is devoted to remind some basic notions about contact geometry , holomorphic curves , morse - bott theory and open books . a ( co - oriented ) _ contact form _ on a three dimensional oriented manifold @xmath8 is a @xmath51 such that @xmath52 is a positive volume form .
a _ contact structure _ is a smooth plane field @xmath41 on @xmath8 such that there exists a contact form @xmath38 for which @xmath53 .
the _ reeb vector field _ of @xmath38 is the ( unique ) vector field @xmath54 determined by the equations @xmath55 and @xmath56 .
simple reeb orbit _ is a closed oriented orbit of @xmath57 , i.e. it is the image @xmath58 of an embedding @xmath59 such that @xmath60 is positively tangent to @xmath58 in any @xmath61 .
reeb orbit _ is an @xmath62-fold cover of a simple reeb orbit , with @xmath63 .
+ the form @xmath38 determines an _ action _
@xmath64 on the set of its reeb orbits defined by @xmath65 . by definition @xmath66 for any non empty orbit @xmath67 .
a basic result in contact geometry asserts that the flow of the reeb vector field ( abbreviated reeb flow ) @xmath68 preserves @xmath41 , that is @xmath69 for any @xmath70 ( see ( * ? ? ? * chapter 1 ) ) . given a reeb orbit @xmath58 , there exists @xmath71 such that @xmath72 for any @xmath61 ; if @xmath73 is the smallest possible , the isomorphism @xmath74
is called the ( _ symplectic _ ) _ linearized first return map _ of @xmath75 in @xmath76 . the orbit @xmath58 is called _ non - degenerate _ if @xmath77 is not an eigenvalue @xmath78 .
there are two types of non - degenerate reeb orbits : _ elliptic _ and _
@xmath58 is elliptic if the eigenvalues of @xmath78 are on the unit circle and is hyperbolic if they are real . in the last case we can make a further distinction : @xmath58 is called _ positive _ ( _ negative _ ) hyperbolic if the eigenvalues are both positive ( resp .
negative ) .
the _ lefschetz sign _ of a non - degenerate reeb orbit @xmath58 is @xmath79 it is easy to check that @xmath80 if @xmath58 is elliptic or negative hyperbolic and @xmath81 if @xmath58 is positive hyperbolic . to any non - degenerate orbit @xmath58 and a trivialization @xmath82 of
@xmath83 we can associate also the _ conley - zehnder index _ @xmath84 of @xmath58 with respect to @xmath82 . even if we do not give a precise definition ( that can be found for example in @xcite or @xcite ) we will provide an explicit description of this index ( see ( * ? ? ?
* section 3.2 ) ) . given @xmath61 , using the basis of @xmath83 determined by @xmath82 we can regard the differentials @xmath85 of the reeb flow as a path in @xmath86 $ ] of @xmath87 symplectic matrices .
in particular @xmath88 is the identity matrix and , if @xmath73 is as above , @xmath89 is a matrix representation for @xmath78 .
if @xmath58 is elliptic , following this path for @xmath86 $ ] , @xmath90 will represent a rotation by some angle @xmath91 with @xmath92 ( since @xmath58 is non degenerate ) .
then @xmath93 , where @xmath94 is the highest integer smaller then @xmath95 . otherwise , if @xmath58 is hyperbolic , then the symplectic matrix of @xmath90 rotates the eigenvectors of @xmath78 by an angle @xmath96 with @xmath97 if @xmath58 is positive hyperbolic and @xmath98 if @xmath58 is negative hyperbolic .
then @xmath99 .
even if @xmath100 depends on @xmath82 , its parity depends only on @xmath58 . indeed ,
if @xmath58 is elliptic , then @xmath101 .
moreover suppose that @xmath58 is hyperbolic and @xmath99 ; if @xmath102 differs from @xmath82 by a twist of an angle @xmath103 with @xmath104 , the rotation by @xmath96 on the eigenvectors will be composed with a rotation by @xmath103
. then @xmath105 .
if @xmath58 is non - degenerate then for any @xmath82 @xmath106 [ definition : orbit sets ] given @xmath107 , we will indicate by @xmath108 the set of simple reeb orbits of @xmath38 contained in @xmath109
orbit set _
( or _ multiorbit _ ) in @xmath109 is a formal finite product @xmath110 , where @xmath111 and @xmath112 is the _ multiplicity _ of @xmath113 in @xmath67 , with @xmath114 whenever @xmath113 is hyperbolic .
the set of multiorbits in @xmath109 will be denoted by @xmath115 .
note that the empty set is considered as an orbit , called _ empty orbit _ and it is indicated by @xmath116 .
an orbit set @xmath110 belongs to the homology class @xmath117 = \sum_i k_i [ \gamma_i ] \in h_1(y)$ ] ( unless stated otherwise , all homology groups will be taken with integer coefficients ) .
moreover the action of @xmath67 is defined by @xmath118 .
we recall here some definitions and properties about holomorphic curves in dimension @xmath119 .
we refer the reader to @xcite and @xcite for the general theory and @xcite and @xcite-@xcite for an approach more specialized to our context .
let @xmath109 be an oriented even dimensional manifold .
an _ almost complex structure _ on @xmath109 is an isomorphism @xmath120 such that @xmath121 and @xmath122 .
if @xmath123 and @xmath124 are two even dimensional manifolds endowed with an almost complex structure , a map @xmath125 is _ pseudo - holomorphic _ if it satisfies the _ cauchy - riemann equation _
@xmath126 a _ pseudo - holomorphic curve _ in a four - dimensional manifold @xmath127 is a pseudo - holomorphic map @xmath128 , where @xmath129 is a riemann surface . note
that here we do not require that @xmath130 is connected .
in this paper we will be particularly interested in pseudo - holomorphic curves ( that sometimes we will call simply holomorphic curves ) in `` symplectizations '' of contact three manifolds .
let @xmath1 be a contact three - manifold and consider the four - manifold @xmath131 .
call @xmath132 the @xmath133-coordinate and let @xmath57 be the reeb vector field of @xmath38 .
the almost complex structure @xmath134 on @xmath131 is _ adapted to @xmath38 _ if 1 .
@xmath134 is @xmath132-invariant ; 2 . @xmath135 and @xmath136 at any point of @xmath131 ; 3 .
@xmath137 is compatible with @xmath138 , i.e. @xmath139 is a riemannian metric . for us , a holomorphic curve @xmath140 in the symplectization of @xmath1 is a holomorphic curve @xmath141 , where : 1 .
@xmath134 is adapted to @xmath38 ; 2 .
@xmath142 is a riemann surface obtained from a closed surface @xmath130 by removing a finite number of points ( called _ punctures _ ) ; 3 . for any puncture @xmath143
there exists a neighborhood @xmath144 such that @xmath145 is mapped by @xmath140 asymptotically to a cover of a cylinder @xmath146 over an orbit @xmath58 of @xmath75 in a way that @xmath147 , where @xmath148 is the projection on the @xmath133-factor of @xmath131 .
we say that @xmath143 is a _ positive puncture _ of @xmath140 if in the last condition above the limit is @xmath149 : in this case the orbit @xmath58 is a _ positive end _ of @xmath140 .
if otherwise the limit is @xmath150 then @xmath143 is a _ negative puncture _ and @xmath58 is a _ negative end _ of @xmath140 .
if @xmath58 is the reeb orbit associated to the puncture @xmath143 , then @xmath140 near @xmath143 determines a cover of @xmath58 : the number of sheets of this cover is the _
local @xmath143-multiplicity of @xmath58 in @xmath140_. the sum of the @xmath143-multiplicities over all the punctures @xmath143 associated to @xmath58 is the _ ( total ) multiplicity _ of @xmath58 in @xmath140 .
if @xmath67 ( @xmath151 ) is the orbit set determined by the set of all the positive ( negative ) ends of @xmath140 counted with multiplicity , then we say that @xmath140 is a _ holomorphic curve from @xmath67 to @xmath151_. a _ cylinder over an orbit set @xmath67 _ of @xmath8 is the holomorphic curve @xmath152 .
note that if there exists a holomorphic curve @xmath140 from @xmath67 to @xmath151 , then @xmath117 = [ \gamma ' ] \in h_1(y,{\mathbb{z}})$ ] .
we state now some result about holomorphic curves that will be useful later .
[ lemma : holomorphic curves low the actions of the orbits ] if @xmath140 is a holomorphic curve in the symplectization of @xmath1 from @xmath67 to @xmath151 , then @xmath153 with equality if and only if @xmath154 and @xmath140 is a union of covers of a cylinder over @xmath67 .
[ theorem : a holomorphic has only isolated critical points ] let @xmath155 be a non - constant holomorphic curve in @xmath127 , then the critical points of @xmath156 are isolated .
in particular , if @xmath157 denotes the projection @xmath158 , @xmath159 is transverse to @xmath54 away from a set of isolated points . from now on
if @xmath140 is a map with image in @xmath131 , we will set @xmath160 and @xmath161 .
holomorphic curves also enjoy the following property , which will be essential for us : see for example @xcite .
[ theorem : positivity of intersection in dimension 4 ] let @xmath140 and @xmath162 be two distinct holomorphic curves in a four manifold @xmath163
. then @xmath164
. moreover , if @xmath76 is an intersection point between @xmath165 and @xmath166 , then its contribution @xmath167 to the algebraic intersection number @xmath168 is strictly positive , and @xmath169 if and only if @xmath140 and @xmath162 are embeddings near @xmath76 that intersect transversely in @xmath76 .
when the almost complex structure does not play an important role or is understood it will be omitted from the notations . the morse - bott theory in contact geometry has been first developed by bourgeois in @xcite .
we present in this subsection some basic notions and applications , mostly as presented in @xcite . a _ morse - bott torus _ ( briefly m - b torus ) in a @xmath170-dimensional contact manifold @xmath1 is an embedded torus @xmath73 in @xmath8 foliated by a family @xmath171 , of reeb orbits , all in the same class in @xmath172 , that are non - degenerate in the morse - bott sense . here
this means the following .
given any @xmath173 and a positive basis @xmath174 of @xmath175 where @xmath176 ( so that @xmath177 is transverse to @xmath178 ) , then the differential of the first return map of the reeb flow on @xmath175 is of the form @xmath179 for some @xmath180 .
if @xmath181 ( resp .
@xmath182 ) then @xmath73 is a _ positive _ ( resp .
_ negative _ ) m - b torus .
we say that @xmath38 is a _ morse - bott contact form _ if all the reeb orbits of @xmath38 are either isolated and non - degenerate or come in @xmath183-families foliating m - b tori . as explained in @xcite and (
* section 4 ) it is possible to modify the reeb vector field in a small neighborhood of a m - b torus @xmath73 preserving only two orbits , say @xmath184 and @xmath185 , of the @xmath183-family of reeb orbits associated to @xmath73 .
moreover , for any fixed @xmath186 , the perturbation can be done in a way that @xmath184 and @xmath185 are the only orbits in a neighborhood of @xmath73 with action less then @xmath22 . if @xmath73 is a positive ( resp . negative ) m - b torus and @xmath82 is the trivialization of @xmath41 along the orbits given pointwise by the basis @xmath187 above , then one can make the m - b perturbation in a way that @xmath185 is positive hyperbolic with @xmath188 and @xmath184 is elliptic with @xmath189 ( resp .
@xmath190 ) .
the orbits @xmath184 and @xmath185 can be seen as the only two critical points of a morse function @xmath191 defined on the @xmath183-family of reeb orbits foliating @xmath73 and with maximum corresponding to the orbit with higher c - z index .
often m - b tori will be implicitly given with such a function .
[ observation : a m - b perturbation creates non isolated and degenerate orbits ] it is important to remark that , before the perturbation , @xmath73 is foliated by reeb orbits of @xmath38 and so these are non - isolated .
moreover the form of the differential of the first return map of the flow of @xmath41 implies that these orbits are also degenerate .
after the perturbation , @xmath73 contains only two isolated and non degenerate orbits , but other orbits are created in a neighborhood of @xmath73 and these orbits can be non - isolated and degenerate .
see figure [ figure : dynamic of the m - b perturbation near k ] later for an example of m - b perturbations . for any m - b torus @xmath73 and any @xmath192 there
exists a m - b perturbation of @xmath73 such that , with the exception of @xmath184 and @xmath185 , all the periodic orbits in a neighborhood of @xmath73 have action greater then @xmath22 . a torus @xmath73 foliated by reeb orbits all in the same class of @xmath172 ( like for example a morse - bott torus ) can be used to obtain constraints about the behaviour of a holomorphic curve near @xmath73 . following ( * ? ? ?
* section 5 ) , if @xmath67 is any of the reeb orbits in @xmath73 , we can define the _ slope of @xmath73 _ as the equivalence class @xmath193 of @xmath117 \in h_1(t,{\mathbb{r } } ) - \{0\}$ ] up to multiplication by positive real numbers .
let @xmath194 $ ] be a neighborhood of @xmath195 in @xmath8 with coordinates @xmath196 such that @xmath197 is a positive basis for @xmath198 and @xmath199 is directed as a positive normal vector to @xmath73 .
suppose that @xmath155 is a holomorphic curve in the symplectization of @xmath1 ; by theorem [ theorem : a holomorphic has only isolated critical points ] , there exist at most finitely many points in @xmath194 $ ] where @xmath200 is not transverse to @xmath54 .
then , if @xmath201 and @xmath202 intersects @xmath203 , we can associate a slope @xmath204 to @xmath205 , for any @xmath206 $ ] : this is defined exactly like @xmath193 , where @xmath205 is considered with the orientation induced by @xmath207)\right)$ ] . note that if @xmath140 has no ends in @xmath208 $ ] , then @xmath209 ) = u_y(f ) \cap t_{y ' } - u_y(f ) \cap t_{y}\ ] ] and @xmath210 .
the following lemma is a consequence of the positivity of intersection in dimension four ( see ( * ? ? ?
* lemma 5.2.3 ) ) .
[ lemma : blocking lemma ] let @xmath73 be linearly foliated by reeb trajectories with slope @xmath211 and @xmath140 a holomorphic curve be as above . 1 .
if @xmath140 is homotopic , by a compactly supported homotopy , to a map whose image is disjoint from @xmath212 , then @xmath213 .
2 . let @xmath194 $ ] be a neighborhood of @xmath214 .
suppose that , for some @xmath206 \setminus \{0\}$ ] , @xmath140 has no ends in @xmath215 $ ] if @xmath216 $ ] or in @xmath217 if @xmath218 .
if @xmath219 then @xmath140 has an end which is a reeb orbit in @xmath73 .
let now @xmath143 be a puncture of @xmath130 whose associated end is an orbit @xmath67 in @xmath73 ; if there exists a neighborhood @xmath220 of @xmath143 in @xmath130 such that @xmath221 then @xmath67 is a _ one sided end _ of @xmath140 in @xmath143 .
this is equivalent to requiring that @xmath222 is contained either in @xmath223 or in @xmath224 .
the following is proved in @xcite ( lemma 5.3.2 ) .
[ lemma : trapping lemma ] if @xmath73 is a positive ( resp .
negative ) m - b torus and @xmath225 is a one sided end of @xmath140 associated to the puncture @xmath143 , then @xmath143 is positive ( resp .
negative ) . given a surface @xmath226 and a diffeomorphism @xmath227 ,
the _ mapping torus _ of @xmath228 is the three dimensional manifold @xmath229}{(x,2)\sim ( \phi(x),0)}.\ ] ] in this paper we use the following definition of open book decomposition of a 3-manifold @xmath8 .
this is not the original definition but a more specific version based on @xcite .
an _ open book decomposition for @xmath8 _ is a triple @xmath230 such that * @xmath231 is an @xmath27-component link in @xmath8 ; * @xmath226 is a smooth , compact , connected , oriented surface with an @xmath27-components boundary ; * @xmath232 is an orientation preserving diffeomorphism such that on a small neighborhood @xmath233 \times s^1 $ ] of @xmath234 , with coordinates @xmath235 near each component , it acts by @xmath236 ( and in particular @xmath237 ) ; * for each @xmath238 there exists a tubular neighborhood @xmath239 of @xmath238 such that @xmath8 is diffeomorphic to @xmath240 where the union symbol means that for any @xmath241 , @xmath242}{0\sim 2}$ ] is glued to @xmath243 in a way that , for any @xmath244 , @xmath245}{(0\sim 2)}$ ] is identified with a meridian of @xmath238 in @xmath246 .
the link @xmath22 is called the _ binding _ , the surfaces @xmath247 are the _ pages _ and the diffeomorphism @xmath248 is the _ monodromy _ of the open book .
when we are interested mostly in the mapping torus part of an open book decomposition we will use a notation of the form @xmath228 , omitting the reference to its binding . sometimes we will call @xmath228 an _ abstract open book_. following @xcite , we will often consider each @xmath243 as a union of a copy of @xmath249}{(0\sim 2)}\times [ 1,2 ] \times s^1 $ ] , endowed with the extension of the coordinates @xmath250 , glued along @xmath251 to a smaller neighborhood @xmath252 of @xmath238 .
the gluing is done in a way that the sets @xmath253 are identified with meridians for @xmath0 and the sets @xmath254 are identified to longitudes . by the giroux s work in @xcite
there is a one to one correspondence between contact structures ( up to isotopy ) and open book decompositions ( up to _ giroux stabilizations _ ) of @xmath8 . in order to simplify the notations
, we consider here open books with connected binding
. given @xmath255 we can follow the thurston - wilkenkemper construction ( @xcite ) to associate to it an _ adapted contact form _
@xmath38 on @xmath8 as explained in ( * ? ? ?
* section 2 ) . in @xmath256
the resulting reeb vector field @xmath257 enjoys the following properties : * @xmath75 is transverse to the pages @xmath258 @xmath259 $ ] ; * the first return map of @xmath75 is isotopic to @xmath248 ; * each torus @xmath260}{(0\sim 2 ) } \times \{y\}$ ] , for @xmath261 $ ] , is linearly foliated by reeb orbits and the first return map of @xmath75 on @xmath262 is @xmath263 the last implies that when the set of orbits foliating @xmath262 comes in an @xmath183-family , @xmath73 is morse - bott . to explain the behaviour of @xmath75 on @xmath37 ,
let us extend the coordinates @xmath196 to @xmath264}{(0\sim 2)}\times [ 2,3 ) \times s^1 $ ] , where @xmath265
. for @xmath266 set @xmath267}{(0\sim 2)}\times \{y\ } \times s^1 $ ] .
given a curve @xmath268 in @xmath262 we can define the _ slope _ of @xmath67 in @xmath269 by @xmath270 in particular if a meridian has constant slope , this must be @xmath149 and @xmath271 has slope @xmath272 .
note that the slope of @xmath262 as given by @xmath273 where now @xmath67 is a parametrization of a reeb trajectory in @xmath262 and @xmath274 .
note in particular that if @xmath275 is irrational then @xmath262 does not contain reeb orbits , and if @xmath262 is foliated by meridians ( like @xmath276 ) then @xmath277 . on @xmath249}{(0\sim 2)}\times [ 1,2 ]
\times s^1 $ ] the contact form will depend on a small real constant @xmath278 : call @xmath279 the contact form on all @xmath8 .
let @xmath280 be a smooth function such that : * @xmath281 has minimum in @xmath282 of value @xmath283 ; * @xmath284 ; * @xmath285 near @xmath286 ; * @xmath287 for @xmath288 and @xmath289 for @xmath290 .
then the reeb vector field @xmath75 of @xmath279 in @xmath291 is such that : * @xmath75 is transverse to the annuli @xmath292 \times s^1 $ ] @xmath293}{0\sim 2}$ ] ; * the tori @xmath262 , @xmath294 $ ] are foliated by reeb orbits with constant slope and first return map given by @xmath295 .
finally in @xmath296 each torus @xmath262 is linearly foliated by reeb orbits whose slope vary in @xmath297 $ ] for @xmath298 going from @xmath170 ( not included ) to @xmath299 and , where @xmath300 is a positive real number .
moreover @xmath0 is also a reeb orbit .
note that for every @xmath58 , @xmath276 is a negative m - b torus foliated by orbits with constant slope @xmath301 . as explained in [ subsection : morse - bott theory ] we can perturb the associated @xmath183-family of orbits into a pair of simple reeb orbits @xmath302 , where @xmath184 is an elliptic orbit with c - z index @xmath303 and @xmath185 is positive hyperbolic with c - z index @xmath272 ( the indexes are computed with respect to the trivialization given by the torus ) .
similarly the positive m - b torus @xmath304 is also foliated by orbits with constant slope @xmath301 and a m - b perturbation gives a pair of simple reeb orbits @xmath305 in @xmath304 , where @xmath306 is elliptic of index @xmath77 and @xmath307 is hyperbolic of index @xmath272 ( in the papers @xcite-@xcite the orbits @xmath306 and @xmath307 are called @xmath308 and @xmath309 respectively ) .
[ h ] in the rest of the paper , if not stated otherwise , when we talk about contact forms and their reeb vector fields adapted to an open book we will always refer to them assuming the notations and the properties explained in this subsection . in particular the m - b tori @xmath276 and @xmath304 will be always assumed to be perturbed into the respective pairs of simple orbits . in the case of open books with
non - connected binding @xmath22 , the reeb vector field of an adapted contact form satisfies the same properties above near each component of @xmath22 .
we saw that to any open book decomposition @xmath230 of @xmath8 it is possible to associate an adapted contact form .
let us now say something about the inverse map of the giroux correspondence . [
theorem : construction of an open book adapted to a contact structure ] given a contact three - manifold @xmath34 , there exists an open book decomposition @xmath230 of @xmath8 and an adapted contact form @xmath38 such that @xmath310 . given any contact structure @xmath41 on @xmath8 , in @xcite giroux explicitly constructs an open book decomposition @xmath230 of @xmath8 for which there exists a compatible contact form @xmath38 such that @xmath310 .
following ( * ? ? ?
* section 3 ) , the proof can be carried on in three main steps .
the first step consists in providing a cellular decomposition @xmath311 of @xmath8 that is , in a precise sense , `` _ _ compatible with @xmath41 _ _ '' .
it is important to remark that , up to take a refinement ( in a way that each @xmath170-cell is contained in a darboux ball ) any cellular decomposition of @xmath8 can be isotoped to make it compatible with @xmath41 . in the second step
, @xmath311 is used to explicitly build @xmath230 .
we describe now some of the properties of @xmath226 , seen as the embedded @xmath272-page of the open book .
+ let @xmath312 be the @xmath241-skeleton of @xmath311 and let @xmath313 be a tubular neighborhood of @xmath314 .
suppose that @xmath315 is a tubular neighborhood of @xmath316 such that @xmath317 is homeomorphic to a tubular neighborhood of @xmath318 .
then : 1 .
@xmath319 , @xmath320 and @xmath321 ; 2 .
@xmath322 is a disjoint union of strips which are diffeomorphic to @xmath323 $ ] with @xmath318 corresponding to @xmath324 ; the fact that @xmath311 is compatible with @xmath41 implies that @xmath22 intersects each @xmath299-simplex exactly twice and it is possible to use this fact to prove that the complement of @xmath22 in @xmath8 fibers in circles over @xmath226 , which implies that @xmath22 is the binding of an open book with @xmath272-page the complement in @xmath226 of a small neighborhood of @xmath226 .
the third step consists finally in defining the contact form @xmath38 with the required properties .
[ theorem : giroux correspondance ] let @xmath38 and @xmath325 be contact structures on @xmath8 that are adapted to the open books @xmath230 and , respectively , @xmath326
. then @xmath38 and @xmath325 are isotopic if and only if @xmath326 can be obtained from @xmath230 by a sequence of giroux stabilizations and destabilizations .
a _ giroux stabilization of an open book _ is an operation that associates to an open book decomposition @xmath230 of @xmath8 another open book decomposition @xmath326 of @xmath8 , obtained as follows .
choose two points @xmath327 and @xmath328 in @xmath271 ( not necessarily in the same connected component ) and let @xmath67 be an oriented embedded path in @xmath226 from @xmath327 to @xmath328 .
let now @xmath329 be the oriented surface obtained by attaching a @xmath77-handle to @xmath226 along the attaching sphere @xmath330 . consider the closed oriented loop @xmath331 defined by @xmath332 , where @xmath333 is the core curve of the @xmath77-handle , oriented from @xmath328 to @xmath327 , and the gluing is done along the common boundary @xmath334 of the two paths . by the definition of monodromy of open book that we gave , the @xmath248 is the identity along @xmath271 .
so @xmath248 extends to the identity map on the handle : we keep calling @xmath248 the resulting diffeomorphism on @xmath329 .
if @xmath335 is a positive dehn twist along @xmath336 , define @xmath337 .
it results that @xmath338 embeds in @xmath8 and that @xmath339 is a disjoint union of solid tori .
then , if @xmath340 is the set of the core curves of these tori , @xmath326 is an open book decomposition of @xmath8 , which is said to be obtained by _
giroux stabilization _ of @xmath230 along @xmath67 .
there is an obvious inverse operation of the stabilization : with the notations above , we say that @xmath230 is obtained by _
giroux destabilization _ of @xmath326 along @xmath151 .
note that a giroux stabilization does not change the components of @xmath22 that do not intersect the attaching sphere .
moreover it is not difficult to see that the number of connected components of @xmath22 and @xmath340 differs by @xmath77 : if @xmath327 and @xmath328 are chosen in the same component then @xmath340 has one component more than @xmath22 ; otherwise @xmath340 has one component less then @xmath22 .
we briefly remind here the hutchings original definition of @xmath30 and @xmath31 for a closed contact three - manifold @xmath1 .
let @xmath1 be a closed contact three - manifold and assume that @xmath38 is non - degenerate ( i.e. , that any reeb orbit of @xmath38 is non - degenerate ) . for a fixed @xmath341 ,
define @xmath342 to be the free @xmath343-module generated by the orbit sets of @xmath8 ( recall definition [ definition : orbit sets ] ) in the homology class @xmath344 and pose @xmath345 this is the _
@xmath35 chain group of @xmath1_. the _ @xmath35-differential _
@xmath346 ( called simply @xmath347 when no risk of confusion occurs ) is defined in @xcite in terms of holomorphic curves in the symplectization @xmath348 of @xmath1 as follows .
given @xmath349 , let @xmath350 be the set of ( possibly disconnected ) holomorphic curves @xmath351 from @xmath67 to @xmath58 , where @xmath142 is a punctured compact rieamannian surface .
it is clear that @xmath140 determines a relative homology class @xmath352 \in h_2({\mathbb{r}}\times y;\gamma,\delta)$ ] and that if such a curve exists then @xmath117 = [ \delta ] \in h_1(y)$ ] .
if @xmath353 and a trivialization @xmath82 of @xmath354 is given , to any surface @xmath355 with @xmath356 it is possible to associate an _ @xmath35-index _
@xmath357 which depends only on the relative homology class of @xmath300
. here * @xmath358 is the _ first relative chern class _ of @xmath300 ; * @xmath359 is the _
@xmath82-relative intersection paring _ of @xmath131 applied to @xmath300 ; * @xmath360 , where @xmath361 is the conley - zehnder index defined in section [ subsection : contact geometry ] .
we refer the reader to @xcite for the details about these quantities . if @xmath140 is a holomorphic curve from @xmath67 to @xmath58 set @xmath362 ( well defined up to approximating @xmath165 with a surface in the same homology class ) .
define @xmath363 .
the @xmath35-differential is then defined on the generators of @xmath364 by @xmath365 where the fraction means that we quotient @xmath366 by the @xmath133-action on the curves given by the translation in the @xmath133-direction in @xmath131 . in (
* section 5 ) hutchings proves that @xmath367 is a compact @xmath272-dimensional manifold , so that @xmath368 is well defined .
the _ ( full ) embedded contact homology of @xmath1 _ is @xmath369 it turns out that these groups do not depend either on the choices @xmath134 in the symplectization or the contact form for @xmath41 .
it is possible to endow @xmath370 with a canonical absolute @xmath371-grading as follows . if @xmath110 set @xmath372 where @xmath373 is the lefschetz sign of the simple orbit @xmath113 .
note that @xmath374 is given by the parity of the number of positive hyperbolic simple orbits in @xmath67 . if @xmath140 is a holomorphic curve from @xmath67 to @xmath58 , by simple computations it is possible to prove the following _ index parity formula _
( see for example section 3.4 in @xcite ) : @xmath375 it follows then that the lefschetz sign endows embedded contact homology with a well defined absolute grading .
fix now a generic point @xmath376 . given two orbit sets @xmath67 and @xmath58 , let @xmath377 be the map defined on the generators by @xmath378 hutchings proves that @xmath379 is a chain map that counts only a finite number of holomorphic curves and that this count does not depend on the choice of @xmath380 .
so it makes sense to define the map @xmath381 for any @xmath380 as above .
this is called the _ u - map_. the _ hat version of embedded contact homology of @xmath1 _ is defined as the homology @xmath31 of the mapping cone of the u - map . by this
we mean that @xmath31 is defined to be the homology of the chain complex @xmath382 with differential defined by the matrix @xmath383 where the element of the complex are thought as columns .
also @xmath31 has the relative and the absolute gradings above .
[ observation : ech groups split in direct sum over h_1 y ] note that @xmath346 and @xmath384 respect the homology class of the generators of @xmath385 .
this implies that there are natural splits : @xmath386 we end this section by stating the following result ( see for example @xcite ) .
let @xmath116 be the empty orbit .
then @xmath387 \in ech(y,\xi)$ ] is an invariant of the contact structure @xmath41 . the class @xmath387 $ ] is called @xmath35 _ contact invariant _ of @xmath41 . in order to define @xmath35 for contact three - manifolds @xmath388 with nonempty boundary
, some compatibility between @xmath38 and @xmath389 should be assumed . in this paper
we are particularly interested in three - manifolds whose boundary is a collection of disjoint tori . in ( * ? ? ?
* section 7 ) colin , ghiggini and honda analyze this situation when @xmath389 is connected .
if @xmath390 is homeomorphic to a torus , then they prove that the @xmath35-complex and the differential can be defined almost as in the closed case , provided that @xmath57 is tangent to @xmath391 and that @xmath38 is non - degenerate in @xmath392 .
if the flow of @xmath393 is irrational they define @xmath394 while , if it is rational , they consider the case of @xmath391 morse - bott and do a m - b perturbation of @xmath38 near @xmath391 ; this gives two reeb orbits @xmath185 and @xmath184 on @xmath391 and , since @xmath38 is now a m - b contact form , the @xmath35-differential counts special holomorphic curves , called _ m - b buildings_. let @xmath38 be a morse - bott contact form on the three manifold @xmath8 and @xmath134 a regular almost complex structure on @xmath395 .
suppose that any m - b torus @xmath73 in @xmath1 comes with a fixed a morse function @xmath396 .
let @xmath397 be the set of simple reeb orbits in @xmath8 minus the set of the orbits which correspond to some regular point of some @xmath396 .
a _ morse - bott building in @xmath1 _ is a disjoint union of objects @xmath140 of one of the following two types : 1 .
@xmath140 is the submanifold of a m - b torus @xmath73 corresponding to a gradient flow line of @xmath396 : in this case the positive and negative end of @xmath140 are the positive and , respectively , the negative end of the flow line ; 2 .
@xmath140 is a union of curves @xmath398 of the following kind .
@xmath399 is a @xmath134-holomorphic curve in @xmath395 with @xmath27 ends @xmath400 corresponding to regular values of some @xmath401 .
then , for each @xmath241 , @xmath399 is augmented by a gradient flow trajectory @xmath402 of @xmath403 : @xmath402 goes from the maximum @xmath404 of @xmath403 to @xmath405 if @xmath405 is a positive end and goes from @xmath405 to the minimum @xmath406 of @xmath403 if @xmath405 is a negative end .
the ends of @xmath140 are obtained from the ends of @xmath399 by substituting each @xmath405 with the respective @xmath404 or @xmath406 .
a morse - bott building is _ nice _
the @xmath399 above has at most one connected component which is not a cover of a trivial cylinder .
suppose now that @xmath8 is closed and @xmath407 is a solid torus embedded in @xmath8 . if @xmath408 , under some assumption on the behaviour of @xmath38 in a neighborhood of @xmath409 , in @xcite the authors define relative versions @xmath410 and @xmath411 of embedded contact homology groups and prove that @xmath412 the notation suggests that these new homology groups are obtained by counting only orbits in @xmath256 and quotienting by orbits on @xmath389 .
let us see the definition of these homologies in more details .
as mentioned , to define these versions of embedded contact homology and prove the isomorphisms above , some compatibility between @xmath38 and @xmath409 is required .
we refer the reader to ( * ? ? ?
* section 6 ) for the details . essentially two conditions are required .
the first one fixes @xmath38 near @xmath409 in a way that @xmath75 behaves similarly to the reeb vector field defined in section [ subsection : open books ] near @xmath37 , where @xmath0 was the binding of an open book decomposition of @xmath8 .
+ briefly , this means that there exists a smaller closed solid torus @xmath413 and a neighborhood @xmath414 $ ] of @xmath415 in @xmath8 such that : 1 .
@xmath416 \subset n$ ] , @xmath417 ) \cup v$ ] and @xmath418 ; 2 .
@xmath419 is foliated by reeb trajectories for any @xmath420 $ ] ; 3 . if @xmath421 , then @xmath0 is a reeb orbit and @xmath422 is foliated by concentric tori , which in turn are linearly foliated by reeb trajectories that intersect positively a meridian disk for @xmath0 in @xmath296 .
4 . @xmath423 and @xmath424 are negative and , respectively , positive m - b tori foliated by reeb orbits which are meridians of @xmath0 . as in subsection [ subsection : open books ] , the families of reeb orbits in @xmath276 and @xmath304 are perturbed into two pairs of reeb orbits @xmath302 and , respectively , @xmath305 : here @xmath184 and @xmath306 are elliptic and @xmath185 and @xmath307 are positive hyperbolic ( see figure [ figure : dynamic of the m - b perturbation near k ] ) .
if @xmath38 satisfies the conditions above we say that _ @xmath38 is adapted to @xmath0_.
the second condition of compatibility is that there must exist a seifert surface @xmath425 for @xmath0 such that @xmath75 is positively transverse to @xmath426 . in this case
we say that @xmath38 is _ adapted to @xmath226_. ( see theorem 10.3.2 in @xcite ) [ lemma : there exists alpha copatible with k and s ] given a null - homologous knot @xmath0 and a contact structure @xmath41 on @xmath8 there exists a contact form @xmath38 for @xmath41 and a genus minimizing seifert surface @xmath226 for @xmath0 such that : 1 .
@xmath38 is adapted to @xmath0 ; 2 .
@xmath38 is adapted to @xmath226 .
we give here only the proof of 1 ) , referring the reader to @xcite for 2 ) . up to isotopy , we can assume that @xmath0 is transverse to @xmath41 and let @xmath325 be any contact form for @xmath41 . up to isotopy of @xmath325
we can suppose that @xmath0 is a reeb orbit . since the compatibility condition with @xmath0 can be arranged on a neighborhood of @xmath0 , by the darboux - weinstein neighborhood theorem ( see for example @xcite ) there exists a contact form @xmath38 which is compatible with @xmath0 and contactomorphic to @xmath325 .
if @xmath255 is an open book decomposition of @xmath8 and @xmath38 is a contact form adapted to @xmath255 , then it is adapted also to @xmath0 and to any page of @xmath255 .
in @xcite the authors prove that it is possible to define the @xmath35-chain groups without taking into account the orbits in @xmath427 and in @xmath428 , so that the only interesting orbits in @xmath37 are the four orbits above ( plus , obviously , the empty orbit ) .
moreover the only curves counted by the ( restriction of the ) @xmath35-differential @xmath347 have projection on @xmath8 as depicted in figure [ figure : dynamic near binding in w ] .
these curves give the following relations : @xmath429 note that the two holomorphic curves from @xmath185 to @xmath184 , as well as the two from @xmath306 to @xmath307 , cancel one each other since we work with coefficients in @xmath371 .
the compactification of the projection of the holomorphic curve that limits to the empty orbit is topologically a disk with boundary @xmath307 , which should be seen as a cylinder closing on some point of @xmath0 .
this curve contribute to the `` @xmath116 part '' of the third of the equations above , which gives @xmath430 = [ \emptyset]$ ] in @xmath35-homology . in the rest of this manuscript
the fact that this disk is the only @xmath35 index @xmath77 connected holomorphic curve that crosses @xmath0 will be essential .
[ h ] from now on we will use the following notation .
if @xmath1 is understood , given a submanifold @xmath431 and a set of reeb orbits @xmath432 , we will denote @xmath433 the free @xmath434-module generated by orbit sets in @xmath435 . unless stated otherwise ,
the group @xmath433 will come with the natural restriction , still denoted @xmath346 , of the @xmath35-differential of @xmath364 : if this restriction is still a differential the associated homology is @xmath436 this notation is not used in @xcite , where the authors introduced a specific notation for each relevant @xmath35-group . in particular with their notation : @xmath437 as mentioned before , even if in @xmath409 there are other reeb orbits , it is possible to define chain complexes for the @xmath35 homology of @xmath1 only taking into account the orbits @xmath438 . the blocking and trapping lemmas and the relations above imply that the restriction of the full @xmath35-differential of @xmath8 to the chain group @xmath439 is given by : @xmath440 where @xmath441 and a term in the sum is meant to be zero if it contains some elliptic orbit with negative total multiplicity or a hyperbolic orbit with total multiplicity not in @xmath442 ( see ( * ? ? ?
* section 9.5 ) ) .
we remark that the blocking lemma implies also that @xmath443 . the further restriction of the differential to @xmath444 is then given by @xmath445 combining the computations of sections 8 and 9 of @xcite the authors get the following result .
[ theorem : ech of y iso to the relative versions ] suppose that @xmath38 is adapted to @xmath0 and there exists a seifert surface @xmath226 for @xmath0 such that @xmath38 is adapted to @xmath226 .
then @xmath446 it is important to remark that the empty orbit is always taken into account as a generator of the groups above .
this implies that if orbit sets with @xmath307 are considered , @xmath346 counts also the holomorphic plane that contributes to the third of relations [ equation : ech boundary near the binding ] .
later we will give the definition of another differential , that we will call @xmath447 , which is obtained from @xmath346 by simply deleting that disk .
define now the _ relative embedded contact homology groups of @xmath448 _ by @xmath449 \sim [ \gamma]}\\ \widehat{ech}(n,\partial n,\alpha ) & = & \frac{ech(n,\alpha)}{[e \gamma ] \sim [ \gamma]}.\end{aligned}\ ] ] since @xmath307 does not belong to the complexes @xmath450 and @xmath451 , the blocking lemma implies that the @xmath35-differentials count only holomorphic curves in @xmath256 .
this `` lack '' is balanced by the quotient by the equivalence relation @xmath452 \sim [ \gamma].\ ] ] the reason behind this claim lie in the third of the relations [ equation : ech boundary near the binding ] .
indeed we can prove the following : [ lemma : how kill h+ and get the quotient e gamma sim gamma ] @xmath453 \sim [ \gamma]}.\ ] ] using the fact that @xmath307 can have multiplicity at most @xmath77 , it is not difficult to see that the long exact homology sequence associated to the pair @xmath454 is @xmath455 where : * @xmath456 is the inclusion map ; * @xmath457 is the module generated by orbit sets of the form @xmath458 with @xmath459 ; * @xmath460 is the quotient map sending to @xmath272 all generators having no contributions of @xmath307 ; * @xmath461 is the standard connecting morphism , that in this case is defined by @xmath462 ) = [ \gamma + e\gamma].\ ] ] we can then extract the short exact sequence @xmath463 where @xmath464 \sim [ \gamma]}.\ ] ] since @xmath465 , the map @xmath466 is an isomorphism . similarly , the fourth line of equation [ equation : ech boundary near the binding ] `` explains '' why we can avoid considering @xmath185 in the full @xmath30 .
in fact with similar arguments of the proof of last lemma , one can prove : @xmath467 .
observe that since @xmath468 , the differential is compatible with the equivalence relation .
so , instead of take the quotient by @xmath469 \sim [ \gamma]$ ] of the homology , we could take the homology of the quotient of the chain groups under the relation @xmath470 , and we would obtain the same homology groups .
we will use this fact later .
note moreover that for every @xmath471 , @xmath472 = [ \emptyset]$ ] .
equations [ equation : ech in terms of the relative version ] and [ equation : ech widehat in terms of the relative version ] follow then from last two lemmas and theorem [ theorem : ech of y iso to the relative versions ] .
an important example of the situation depicted above is when @xmath0 is the binding of an open book decomposition @xmath255 of a closed three manifold @xmath8 , and @xmath256 is the associated mapping torus considered in subsection [ subsection : open books ] . using the same notations , define the _ extended pages _ of @xmath228 to be the surfaces @xmath474}{0\sim 2}.\ ] ] let @xmath38 be a contact form on @xmath8 compatible with @xmath255 . in particular @xmath38
is adapted to both @xmath0 and any page of @xmath255 .
[ definition : degree of multi orbits ]
if @xmath67 is a reeb orbit in @xmath475 , define the _ degree _ of @xmath67 by @xmath476 if @xmath477 is some orbit set , we define @xmath478 . if @xmath479 , we indicate by @xmath480 ( resp .
@xmath481 ) the set of multiorbits in @xmath109 with degree equal ( resp .
less or equal ) to @xmath241 . note that @xmath482 depends only on the homology class of @xmath67 in @xmath475 . in this context
the relative embedded contact homology groups can also be defined in terms of limits as follows .
define @xmath483 to be the free @xmath343-module generated by orbit sets in @xmath484 .
similarly let @xmath485 be generated by orbit sets in @xmath486 .
define the inclusions @xmath487 given by the map @xmath488 each of these chain groups can be endowed with ( the restriction of ) the @xmath35-differential , which counts m - b buildings in @xmath256 .
let @xmath489 and @xmath490 be the associated homology groups . then the relative embedded contact homology groups above can be defined also by @xmath491 if @xmath492 , let @xmath493 be the homology of @xmath494 with the @xmath35-boundary map .
the `` stabilization '' theorem 1.0.2 of @xcite implies that for the definition of @xmath411 it is sufficient to take into account just orbit sets in @xmath495 .
then : @xmath496 \sim [ \gamma]}. \label{equation : equivalence ech n , partial n , alpha and ech_2 g n , partial n}\ ] ] let @xmath0 be a homologically trivial knot in a contact three - manifold @xmath1 .
in this subsection we recall the definition of a hat version of contact homology for the triple @xmath497 .
this was first defined in ( * ? ? ?
* section 7 ) as a particular case of _ sutured _ contact homology . on the other hand , following ( * ? ? ?
* section 10 ) , it is possible to proceed without dealing directly with sutures : we follow here this approach .
let @xmath226 be a seifert surface for @xmath0 . by standard arguments in homology , it is easy to compute that @xmath498 & \longmapsto & ( i_*[a]\ , \langle a , s \rangle ) \end{array}\ ] ] is an isomorphism . here
@xmath499 is the inclusion and @xmath500 denotes the intersection number between @xmath501 and @xmath226 : this is a homological invariant of the pair @xmath502 and is well defined up to a slight perturbation of @xmath226 ( to make it transverse to @xmath501 ) .
note that a preferred generator of @xmath503 is given by the homology class of a meridian for @xmath0 , positively oriented with respect to the orientations of @xmath226 and @xmath8 .
if @xmath8 is a homology three - sphere , the number @xmath500 depends only on @xmath501 and @xmath0 .
this is the _ linking number between @xmath501 and @xmath0 _ and it is usually denoted by @xmath504 . if @xmath505 is a finite formal product of closed curves in @xmath475 , then @xmath506 .
if @xmath255 is an open book decomposition of @xmath8 and @xmath38 is an adapted contact form , then @xmath507 for any orbit set @xmath508 , where @xmath509 is given in definition [ definition : degree of multi orbits ] .
[ proposition : holomorphic curves preserve the filtration in general ] suppose that @xmath0 is an orbit of @xmath54 and let @xmath226 be any seifert surface for @xmath0 .
if @xmath67 and @xmath58 are two orbit sets in @xmath475 and @xmath510 is a holomorphic curve from @xmath67 to @xmath58 , then @xmath511 let @xmath512 be the compactification of @xmath140 in @xmath513 \times y$ ] .
since @xmath140 has no limits in @xmath0 , then @xmath514 \times k \rangle = \langle u , \mathbb{r } \times k \rangle \geq 0,\ ] ] where the inequality follows by the positivity of intersection in dimension 4 ( since @xmath0 is a reeb orbit , @xmath515 is holomorphic ) .
consider the two surfaces @xmath516 and define the closed surface @xmath517 \times k ) \cup l_{1 } \end{array}\ ] ] where the first gluing is made along @xmath518 and the second along @xmath519 .
since @xmath520 \in h_2([-1,1 ] \times y)$ ] , then : @xmath521 \times k \rangle + \langle \widehat{u } , l_{1 } \rangle & = \\ & = & \langle \delta , s \rangle + \langle \widehat{u } , [ -1,1 ] \times k \rangle - \langle \gamma , s \rangle . & \end{array}\ ] ] the result then follows by observing that the last equation implies that @xmath522 \times k \rangle \geq 0.\ ] ] suppose that @xmath38 is adapted to @xmath0 in the sense of subsection [ subsection : ech for manifolds with boundary ] . a choice of ( a homology class for ) the seifert surface @xmath226 for the orbit @xmath0 defines a _ knot filtration _ on the chain complex @xmath523 for @xmath31 , where , recall , @xmath256 is the complement of a neighborhood @xmath37 of @xmath0 in which the only `` interesting '' orbits and holomorphic curves are the ones represented in figure [ figure : dynamic near binding in w ] .
let @xmath524 be the free sub - module of @xmath525 generated by orbit sets @xmath67 in @xmath526 such that @xmath527 .
define moreover @xmath528 the direct sum above is not in general finite . on the other hand
if @xmath38 is adapted to @xmath226 then @xmath529 for any @xmath67 and the sum is finite for any @xmath461 .
even if @xmath38 is not adapted to @xmath226 , the intersection number induces an exhaustive filtration @xmath530 on @xmath525 .
the filtration above is the _ knot filtration _ induced by @xmath0 .
if @xmath67 is a generator of @xmath524 , the integer @xmath461 is the _ filtration degree _ of @xmath67 .
the differential @xmath346 of @xmath525 respects the knot filtration .
proposition [ proposition : holomorphic curves preserve the filtration in general ] applied to the m - b buildings counted by @xmath346 implies immediately that @xmath531 for any @xmath461 and the result follows .
suppose now that @xmath38 is adapted to @xmath226 . by standard arguments in algebra , the filtration above induces a spectral sequence whose page @xmath532 is isomorphic to @xmath533 and whose page @xmath272 is the chain complex @xmath534 where @xmath524 should be seen as @xmath535 and @xmath536 is the map induced by @xmath346 on the quotient , i.e , it is the part of @xmath537 that strictly preserves the filtration degree .
[ observation : the only lowing - degree curve is the disk from h_+ in the hat case ] the proof of proposition [ proposition : holomorphic curves preserve the filtration in general ] implies that the holomorphic curves counted by @xmath346 that strictly decrease the degree are exactly the curves that intersect @xmath0 .
so we can interpret @xmath447 as the restriction of @xmath346 ( given by equation [ equation : the complete ech differential with all the orbits ] ) to the count of curves that do not cross a thin neighborhood of @xmath0 .
this is indeed the proper @xmath35-differential of the manifold @xmath538 ( and not the restriction of the @xmath35-differential of @xmath8 to the orbit sets in @xmath538 ) .
note that , by definition of @xmath525 , all the holomorphic curves contained in @xmath539 strictly preserve the filtration degree .
in fact the only holomorphic curve that contributes to @xmath540 and decreases the degree ( by @xmath77 ) is the plane from @xmath307 to @xmath116 .
equation [ equation : the hat ech differential with all the orbits ] gives then @xmath541 where @xmath441 and any term is meant to be zero if it contains some orbit with total multiplicity that is negative or not in @xmath442 if the orbit is hyperbolic the _ hat version of embedded contact ( knot ) homology _ of the triple @xmath497 is @xmath542 in @xcite @xmath2 is called @xmath543 and in theorem 10.3.2 it is proved that @xmath544 where , recall , with our notation @xmath545 . on the other hand , by using exactly the same arguments of lemma [ lemma : how kill h+ and get the quotient e gamma sim gamma ] , it is easy to see that @xmath546 [ observation : alpha compatible or not with s to define eck ] note that in order to define @xmath2 , we supposed that @xmath38 is compatible with @xmath226 .
this hypothesis is not present in the original definition ( via sutures ) in @xcite .
indeed , without this condition we can still apply all the arguments above and define the knot filtration on @xmath525 exactly in the same way .
the page @xmath77 of the spectral sequence is again the well defined homology in the definition above , and the page @xmath532 is still isomorphic to @xmath444 .
the only difference ( a priori , see lemma [ lemma : no need of alpha adapted to s ] below ) is that now we do not know that @xmath533 , since in theorem [ theorem : ech of y iso to the relative versions ] the hypothesis that @xmath38 is adapted to @xmath226 is assumed .
this homology comes naturally with a further relative degree , inherited by the filtered degree : if @xmath547 then @xmath548 sometimes , in analogy with heegaard floer , we will call this degree the _
alexander degree_. suppose that @xmath255 is an open book decomposition of @xmath8 and that @xmath38 is an adapted contact form .
since any non - empty reeb orbit in @xmath475 has strictly positive intersection number with @xmath226 , @xmath549 \rangle_{{\mathbb{z}}/2}.\ ] ] this is the @xmath35-analogue of the fact that if @xmath0 is fibered , then @xmath550\rangle_{{\mathbb{z}}/2},\ ] ] where @xmath551 is the genus of @xmath0 and @xmath333 is the associated contact element ( see @xcite ) .
[ observation : the relative alexander degree is well defined ] we remark that the alexander degree can be considered as an absolute degree only once a relative homology class in @xmath552 for @xmath226 has been fixed , since the function @xmath553 defined on @xmath554 changes if @xmath555 $ ] varies . on the other hand ,
suppose that @xmath117 = [ \delta ]
\in h_1(y \setminus k)$ ] and let @xmath556 be a surface such that @xmath557 .
computations analogue to that in the proof of proposition [ proposition : holomorphic curves preserve the filtration in general ] imply that @xmath558 and the alexander degree , considered as a relative degree , does not depend on the choice of a homology class for @xmath226 .
obviously if @xmath559 , the alexander degree can be lifted to an absolute degree .
in @xcite the authors conjectured that their sutured embedded contact homology is isomorphic to sutured heegaard - floer homology .
both the hat version of embedded contact knot homology and of heegaard floer knot homology can be defined in terms of sutures . in this case
their conjecture becomes [ conjecture : eck iso to hfk hat ] for any knot @xmath0 in @xmath8 : @xmath560 where @xmath38 is a contact form on @xmath8 adapted to @xmath0 .
let @xmath0 be a homologically trivial knot in a contact three - manifold @xmath1 . as recalled in subsection [ subsection : widehat ech for knots ] , if @xmath38 is adapted to @xmath0 , a choice of a seifert surface @xmath226 for @xmath0 induces a filtration on the chain complex @xmath561 , where @xmath392 is homeomorphic to @xmath475 . moreover if @xmath38 is also adapted to @xmath226 , the homology of @xmath561 is isomorphic to @xmath31 , and the first page of the spectral sequence associated to the filtration is the hat version of embedded contact knot homology @xmath2 . in this section
we generalise the knot filtration in two natural ways .
in subsection [ subsection : the full eck ] we extend the filtration induced by @xmath0 on the chain complex @xmath562 .
this filtration is defined in a way completely analogue to the hat case .
we define the _ full version of embedded contact knot homology _ of @xmath497 to be the first page @xmath4 of the associated spectral sequence .
moreover we remove the condition that @xmath38 must be compatible with @xmath226 , in order to consider a wider class of contact forms : the knot spectral sequence is still well defined , but at the price of renouncing to a proof of the existence of an isomorphism between @xmath30 and the page @xmath532 of the spectral sequence .
in subsection [ subsection : eck for links ] we generalise the knot filtration to @xmath27-components links @xmath22 .
the resulting homologies , defined in a way analogue to the case of knots , are the _ full and hat versions of embedded contact knot homologies of @xmath563 _ , which will be still denoted @xmath564 and , respectively , @xmath565 . similarly to heegaard - floer link homology , these homologies come endowed with an _ alexander ( relative ) @xmath26-degree_. let @xmath0 be a homologically trivial knot in a contact three - manifold @xmath1 and suppose that @xmath38 is adapted to @xmath0 in the sense of subsection [ subsection : ech for manifolds with boundary ]
. recall in particular that there exist two concentric neighborhoods @xmath566 of @xmath0 whose boundaries are m - b tori @xmath567 and @xmath568 foliated by orbits of @xmath54 in the homology class of meridians for @xmath0 .
these two families of orbits are modified into the two couples of orbits @xmath569 and , respectively , @xmath570 .
let moreover @xmath571 .
consider the chain complex @xmath572 where , recall , the chain group is freely generated on @xmath371 by the orbit sets @xmath67 in @xmath573 and @xmath346 is the @xmath35-differential ( obtained by restricting the differential on @xmath364 ) given by equation [ equation : the complete ech differential with all the orbits ] .
a seifert surface @xmath226 for @xmath0 induces an _ alexander degree _
@xmath553 on the generators of @xmath574 exactly like in the case of @xmath575 .
let @xmath576 be the submodule of @xmath574 generated by the @xmath577 with @xmath578 . if @xmath579 we have the exhaustive filtration @xmath580 of @xmath574 .
proposition [ proposition : holomorphic curves preserve the filtration in general ] again implies that @xmath346 preserves the filtration .
let @xmath581 be the part of @xmath346 that strictly preserves the filtration degree @xmath461 , that is , the differential induced by @xmath582 on the quotient @xmath583 set @xmath584 we define the _ full embedded contact knot homology of @xmath497 _ by @xmath585 note that , as in the hat case , the only holomorphic curves counted by @xmath346 that do not strictly respect the filtration degree are the curves that contain the plain from @xmath307 to @xmath116 ( see observation [ observation : the only lowing - degree curve is the disk from h_+ in the hat case ] ) .
recalling the expression of @xmath346 given in equation [ equation : the complete ech differential with all the orbits ] , it follows that @xmath447 is given by @xmath586 where @xmath441 and any term is meant to be @xmath272 if it contains an orbit with total multiplicity that is negative or not in @xmath442 if the orbit is hyperbolic .
again the homology comes with an _ alexander degree _ , which is well defined once the an homology class for @xmath226 is fixed .
in fact we have the natural splitting : @xmath587 where @xmath588 recalling that @xmath589 is homeomorphic to @xmath475 , it is interesting to state the following : [ lemma : eck for knots full iso to ech y minus n(k ) ] if @xmath37 is a neighborhood of @xmath0 as above then @xmath590 by arguments similar to those in the proof of lemma [ lemma : how kill h+ and get the quotient e gamma sim gamma ] it is easy to prove that : @xmath591 where the last comes from the fact that @xmath592 for any @xmath441 .
note that so far we only assumed that @xmath38 is compatible with @xmath0 , while we did not suppose the condition * @xmath38 is compatible with a seifert surface @xmath226 for @xmath0 . as remarked in observation [ observation : alpha compatible or not with s to define eck ] , without @xmath593 we can not apply theorem [
theorem : ech of y iso to the relative versions ] , and so we do not know if the spectral sequence whose @xmath272-page is the @xmath6-chain complex limits to @xmath30 .
on the other hand we have the following [ lemma : no need of alpha adapted to s ] theorem [ theorem : ech of y iso to the relative versions ] holds even without assuming condition @xmath593 .
reading carefully the proof of theorem [ theorem : ech of y iso to the relative versions ] given in @xcite one can see that @xmath593 is not really necessary .
it is explicitly used only in section 9.7 to prove that the map @xmath594 is nilpotent , but this fact can be proved also without assuming @xmath593 . indeed @xmath595 strictly decreases the alexander degree .
then , if @xmath596 for every @xmath597 , for @xmath598 arbitrarily large @xmath599 would contain ( as factor ) an orbit set with arbitrarily large negative alexander degree , and so also with arbitrarily large action , which is not possible by lemma [ lemma : holomorphic curves low the actions of the orbits ] .
a rough explanation of last lemma is the following . by direct limit
arguments the orbits in the no man s land @xmath600 can be avoided also if @xmath593 is not assumed , so that we can still write @xmath601 ( up to the restriction on the action of the orbits made in ( * ? ? ?
* section 9 ) ) . the computations for @xmath602 in ( * ? ? ?
* section 8) do not use @xmath593 , and in fact here the hypothesis is not even assumed .
similarly , @xmath603 is still well defined as in ( * ? ? ?
* subsection 7.1 ) and does not depend on the choice of @xmath226 .
moreover the blocking lemma still implies that holomorphic curves with positive limit in @xmath256 can not cross @xmath389 , so that @xmath451 is again a subcomplex of @xmath364 .
this suggests that what happens in @xmath256 should not influence the direct limits computations in @xmath296 . in analogy with conjecture [ conjecture : eck iso to hfk hat ] we state the following : [ conjecture : eck iso to hfk full ] for any knot @xmath0 in @xmath8 : @xmath604 where @xmath38 is any contact form on @xmath8 adapted to @xmath0 . in this subsection
we extend the definitions of @xmath6 and @xmath7 to the case of homologically trivial links with more than one component . for us
a ( _ strongly _ ) _ homologically trivial @xmath27-link _ in @xmath8 is a disjoint union of @xmath27 knots , each of which is homologically trivial in @xmath8 .
suppose that @xmath605 is a homologically trivial @xmath27-link in @xmath8 .
we say that a contact form @xmath38 on @xmath8 is _ adapted to @xmath22 _ if it is adapted to @xmath238 for each @xmath241 . for any link @xmath22 and contact structure @xmath41 on @xmath8
there exists a contact form compatible with @xmath41 which is adapted to @xmath22 .
the proof of part 1 ) of lemma [ lemma : there exists alpha copatible with k and s ] is local near the knot @xmath0 and can then be applied recursively to each @xmath238 .
fix @xmath606 homologically trivial and @xmath38 an adapted contact form . since @xmath38 is adapted to each @xmath238 , there exist pairwise disjoint tubular neighborhoods @xmath607 of @xmath238 where @xmath38 behaves exactly like in the neighborhoods @xmath566 in subsection [ subsection : ech for manifolds with boundary ] .
in particular , for each @xmath241 , the tori @xmath608 and @xmath609 are m - b and foliated by families of orbits of @xmath54 in the homology class of a meridian of @xmath238 .
we will consider these two families as perturbed into two pairs @xmath610 and @xmath611 in the usual way .
let @xmath612 and set @xmath613 define moreover @xmath614 and let @xmath615 and @xmath616 be similarly defined .
consider now @xmath617 endowed with the restriction @xmath346 of the @xmath35 differential of @xmath1 and let @xmath618 be the associated homology .
@xmath618 is well defined and the curves counted by @xmath346 inside each @xmath243 are given by expressions analogue to those in [ equation : ech boundary near the binding ] . the blocking and trapping lemmas
can be applied locally near each component of @xmath389 and the proofs of lemmas 7.1.1 and 7.1.2 in @xcite work immediately in this context too .
this imply that the homology of @xmath619 is well defined . again the blocking and trapping lemmas together with the local homological arguments in lemmas 9.5.1 and 9.5.3 in @xcite ,
imply that the only holomorphic curves counted by @xmath346 inside each @xmath243 are as required ( see figure [ figure : dynamic near binding in w ] ) , and so that @xmath618 is well defined .
an explicit formula for @xmath346 can be obtained by generalizing equation [ equation : the complete ech differential with all the orbits ] in the obvious way .
for each @xmath620 , fix now a ( homology class for a ) seifert surface @xmath621 for @xmath238 .
these surfaces are not necessarily pairwise disjoint and it is even possible that @xmath622 for some @xmath623 .
consider then the _ alexander @xmath26-degree _ on @xmath617 given by the function @xmath624 define the partial ordering on @xmath26 given by @xmath625 proposition [ proposition : holomorphic curves preserve the filtration in general ] applied to each @xmath238 implies that if @xmath67 and @xmath58 are two orbit sets in @xmath626 , then for any @xmath471 @xmath627 this implies that @xmath346 does not increase the alexander degree , which induces than a @xmath26-filtration on @xmath628 .
reasoning as in the previous subsection , we are interested in the part of @xmath346 that strictly respects the filtration degree .
this can be defined again in terms of quotients as follows .
let @xmath629 and let @xmath630 be the submodule of @xmath631 freely generated by orbit sets @xmath632 such that @xmath633 define @xmath634 and let @xmath635 be similarly defined .
define the _
full @xmath6-differential in degree @xmath461 _ to be the map @xmath636 induced by @xmath637 on the quotient @xmath638 define then the _ full @xmath6-differential _ by @xmath639 [ observation : the only lowing - degree curves are the disks from h_+ for links ] observing the form of @xmath346 , it is easy again to see that the only holomorphic curves that are counted by @xmath346 and not by @xmath447 are the ones containing a holomorphic plane from some @xmath640 to @xmath116 .
the _ full embedded contact knot homology of @xmath563 _ is @xmath641 the fact that @xmath564 is well defined is a direct consequence of the good definition of @xmath642 and the fact that @xmath346 respects the alexander filtration .
note that also for links we have a natural splitting @xmath643 where @xmath644 the proof of the following lemma is the same of that of the analogous lemma [ lemma : eck for knots full iso to ech y minus n(k ) ] for knots applied to each component of @xmath22 .
[ lemma : eck iso to ech of int n ] if @xmath645 is a neighborhood of @xmath22 as above then @xmath646 consider now the submodule @xmath647 of @xmath631 endowed with the restriction of @xmath346 .
again its homology @xmath648 is well defined .
proceeding exactly like above , the choice of a seifert surface @xmath621 for each component @xmath238 of @xmath22 gives ( up to small perturbations of @xmath226 ) an alexander degree on the orbit sets defined by equation [ equation : multiple alexander filtration degree ] .
this induces a @xmath26-filtration on the chain complex @xmath649 . for any @xmath629 ,
define @xmath650 and @xmath651 exactly as above .
the _ hat version of embedded contact knot homology of @xmath563 _ is @xmath652 observation [ observation : the only lowing - degree curves are the disks from h_+ for links ] and a splitting like the one in equation [ equation : split of eck of l full ] hold also for @xmath565 .
moreover it is easy to see that if @xmath22 has only one connected component we get the same theories of subsections [ subsection : widehat ech for knots ] and [ subsection : the full eck ] .
we state the following [ conjecture : eck iso to hfk for links full and hat ] if @xmath22 is a link in @xmath8 : @xmath653 where @xmath38 is any contact form on @xmath8 adapted to @xmath22 .
note that the analogous conjectures stated before , as well as theorem [ theorem : hf and ech in intro ] , suggest that we should use the plus version of @xmath654 and not the minus one .
the problem is that in @xcite the authors define heegaard - floer homology for links only in the hat and minus versions . on the other hand
this switch is not really significant .
indeed one could define _ heegaard - floer cohomology _
groups by taking the duals , with coefficients in @xmath371 , of the chain groups @xmath655 , @xmath656 and @xmath657 in the usual way and get cohomology groups ( for the three - manifold @xmath8 ) @xmath658 since we are working in @xmath371 we have that each of this cohomology group is isomorphic to its respective homology group . on the other hand one can prove also that ( see proposition 2.5 in @xcite ) : @xmath659 analogous formulae hold also for knots .
the conjecture above is then consistent with those stated in the previous subsections . as in the definition of @xmath2 and @xmath660 also here we used the hypothesis that @xmath38 is adapted to @xmath22 , while , in view of lemma [ lemma : no need of alpha adapted to s ] , we dropped condition @xmath593 of last subsection .
one could wonder if it is possible to further relax the assumptions and get still a good definition of the @xmath6 homology groups .
the conjectures above suggest indeed that @xmath564 ( as well as the other homologies ) would be independent from @xmath38 and so , in particular , that we could be able to define it simply as the @xmath35 homology of the complement of ( any neighborhood of ) @xmath22 , provided that @xmath22 is a disjoint union of reeb orbits of @xmath38 .
indeed , even if we could not have an easy description of the curves counted by @xmath346 that cross @xmath22 , proposition [ proposition : holomorphic curves preserve the filtration in general ] still holds in this more general case .
on the other hand , technical aspects about contact flows and holomorphic curves suggest that the components of @xmath22 should be at least elliptic orbits .
this property will be necessary even in computing euler characteristics in next section , where we will need a circularity property of @xmath54 near @xmath22 that can not be assumed in an evident way if a component of @xmath22 is hyperbolic . in order to simplify the notation , in the rest of the paper
we will indicate the @xmath35 chain groups for the knot embedded contact homology groups of links and knots by : @xmath661 where @xmath256 and @xmath38 are as above .
in particular , if not stated otherwise , we will always assume that the contact form @xmath38 is adapted to @xmath22 .
these groups will implicitly come endowed with the differential @xmath447 .
we end this section by saying some word about a further generalization of @xmath6 to _ weakly homologically trivial _ links .
we say that @xmath662 is a weakly homologically trivial ( or simply _ weakly trivial _ ) @xmath27-component link if there exist surfaces with boundary @xmath663 , with @xmath664 and such that @xmath665
if @xmath623 and @xmath666 . also here we do not require that @xmath621 or even @xmath667 is disjoint from @xmath668 for @xmath669 .
clearly @xmath22 is a strongly trivial link if and only if it is weakly trivial with @xmath670 . in this case
we can not in general define a homology with a filtered @xmath27-degree .
if @xmath22 is a weakly trivial link with @xmath671 and @xmath38 is an adapted contact form , then there exists @xmath672 such that @xmath271 has more then one connected component .
suppose for instance that @xmath673 .
the arguments of proposition [ proposition : holomorphic curves preserve the filtration in general ] say then that if @xmath510 is a holomorphic curve from @xmath67 to @xmath58 , then @xmath674 so in this case we can still apply the arguments above and get well defined @xmath35 invariants for @xmath22 .
however this time they will come only with a filtered ( relative ) @xmath675-degree on the generators
@xmath67 of an @xmath35 complex of @xmath8 , which is given by the @xmath62-tuple @xmath676 .
[ example : eck for open books ] let @xmath230 be an open book decomposition of @xmath8 with , possibly , disconnected boundary . using a ( connected ) page of @xmath230 to compute the alexander degree and , with the notations of subsection [ subsubsection : ech from open books ] , we get @xmath677 for any @xmath678 .
in this section we compute the graded euler characteristics of the embedded contact homology groups for knots and links in homology three spheres @xmath8 with respect to suitable contact forms . the computations will be done in terms of the lefschetz zeta function of the flow of the reeb vector field . before proceeding we briefly recall what the graded euler characteristic is . given a collection of chain complexes @xmath679 where @xmath680 denotes a relative homological degree ,
its _ graded euler characteristic _ is @xmath681\ ] ] where @xmath682 is the standard euler characteristic of @xmath683 and the @xmath684 s are formal variables . by definition
, @xmath685 is a laurent polynomial and the properties of the standard euler characteristic imply @xmath686 in this case the homology @xmath687 is a _ categorification _ of the polynomial @xmath685 .
when we want to highlight the variables of these polynomials we will indicate them as subscripts of the symbol @xmath21 .
for example if @xmath22 is an @xmath27-link and we want to express its euler characteristic by a polynomial in the @xmath27 variables @xmath688 , we will write @xmath689 .
last theorem imply that _ the homology @xmath6 categorifies the alexander polynomial of knots and links in @xmath23_. this is the third known categorification of this kind , after the ones in heegaard - floer homology and in seiberg - witten - floer homology ( see @xcite and @xcite ) .
the last corollary implies that conjecture [ conjecture : eck iso to hfk for links full and hat ] ( which generalizes conjectures [ conjecture : eck iso to hfk hat ] and [ conjecture : eck iso to hfk full ] ) holds for links in @xmath23 at least at the level of euler characteristic .
given any link @xmath693 in @xmath23 we can associate to it its _ multivariable alexander polynomial _
@xmath694}{\pm t_1^{a_1}\cdots t_n^{a_n}}.\ ] ] with @xmath695 .
the quotient means that the alexander polynomial is well defined only up to multiplication by monomials of the form @xmath696 .
there are many possible definitions of the alexander polynomial @xmath48 . in this section
we give a formulation of @xmath48 in terms of the dynamics of suitable vector fields in @xmath700 .
the details about the proof of the statements can be found in the references .
the fact that the alexander polynomial is related to dynamical properties of its complement in @xmath23 origins with the study of fibrations of @xmath23 . for example in @xcite acampo studied the _ twisted lefschetz zeta function _ of the monodromy of an open book decomposition @xmath228 of @xmath23 associated to a milnor fibration of a complex algebraic singularity .
more in general , if @xmath255 is any open book decomposition of @xmath23 , one can easily prove ( see for example @xcite ) that @xmath701 where @xmath702 and @xmath703 are the identity map and , respectively , the application induced by @xmath248 , on @xmath704 .
the basic idea in this context is to express the right - hand side of equation above in terms of traces of iterations of @xmath703 ; then to apply the lefschetz fixed point theorem to get expressions in terms of periodic points , ( i.e. periodic orbits ) for the flow of some vector field in @xmath705 whose first return on a page is @xmath248 .
suppose now that @xmath22 is not a fibered link , so that its complement is not globally fibered over @xmath183 and let @xmath75 be a vector field in @xmath700 .
if one wants to apply arguments like above , it is necessary to decompose @xmath700 in `` fibered - like '' pieces with respect to @xmath75 , in which it is possible to define at least a local first return map of the flow @xmath706 of @xmath75 .
obviously some condition on @xmath75 is required . for example , in his beautiful paper @xcite , franks consider _ smale vector fields _ , that is , vector fields whose chain recurrent set is one - dimensional and hyperbolic ( cf .
@xcite ) . here
we are more interested in the approach used by fried in @xcite .
consider a three - dimensional manifold @xmath109 .
any _ abelian cover _
@xmath707 with deck transformations group isomorphic to a fixed abelian group @xmath708 is uniquely determined by the choice of a class @xmath709 . here
@xmath710 is determined by the following property : for any @xmath117 \in h_1(x)$ ] , if @xmath711 \rightarrow \widetilde{x}$ ] is any lifting of the loop @xmath712 \rightarrow x$ ] , then @xmath713)$ ] is determined by @xmath713)(\widetilde{\gamma}(0 ) ) = \widetilde{\gamma}(1)$ ] .
since the correspondence between abelian covers and cohomology classes is bijective , with abuse of notation sometimes we will refer to an abelian cover directly by identifying it with the corresponding @xmath710 .
[ example : abelian cover determined by a link ] let @xmath716 be an @xmath27-components link in a three manifold @xmath8 such that @xmath238 is homologically trivial for any @xmath241 and fix a seifert surface @xmath621 for @xmath238 .
let moreover @xmath717 be a positive meridian for @xmath238 . if @xmath718 is the inclusion , the isomorphism @xmath719 } \oplus \ldots \oplus { \mathbb{z}}_{[\mu_n]}\\ \ [ \gamma ] & \longmapsto & \left ( i_*([\gamma ] ) , \langle \gamma , s_1\rangle , \ldots , \langle \gamma , s_n \rangle \right ) \end{array}\ ] ] gives rise naturally to the abelian cover @xmath720 of @xmath721 defined by @xmath722 ) = \left(\langle \gamma , s_1\rangle , \ldots , \langle \gamma , s_n \rangle\right).\ ] ] setting @xmath723 \in h_1(y \setminus l,{\mathbb{z}})$ ] , we can regard @xmath724)$ ] as a monomial in the variables @xmath725 : @xmath722 ) = t_1^{\langle \gamma , s_1\rangle}\cdots t_n^{\langle \gamma , s_n\rangle}.\ ] ] in the rest of the paper we will often use this notation
. if @xmath75 is a vector field on @xmath109 satisfying some compatibility condition with @xmath710 ( and with @xmath727 if this is non - empty ) , the author relates the reidemeister - franz torsion of @xmath728 with the _ twisted lefschetz zeta function _ of the flow @xmath706 .
let @xmath75 be a vector field on @xmath109 and @xmath67 a closed isolated orbit of @xmath706 .
pick any point @xmath729 and let @xmath730 be a small disk transverse to @xmath67 such that @xmath731 . with this data it is possible to define the lefschetz sign of @xmath67 exactly like we did in section [ subsection : contact geometry ] for orbits of reeb vector fields associated to a contact structure @xmath41 , but using now @xmath732 instead of @xmath733 .
indeed it is possible to prove that the lefschetz sign of @xmath67 does not depend on the choice of @xmath143 and @xmath730 and it is an invariant @xmath734 of @xmath706 near @xmath67 .
we remark that in @xcite the author defines @xmath740 in a slightly different way and then he proves ( theorem 2 ) that , under some assumptions that we will state in the next subsection , the two definitions coincide .
suppose that @xmath741 is an abelian cover of @xmath109 and chose a generator @xmath742 of @xmath26 .
then , with a similar notation to that of example [ example : abelian cover determined by a link ] , we will often identify @xmath740 with an element of @xmath743 $ ] . in @xcite
fried relates the reidemeister torsion of an abelian cover @xmath710 of a ( non necessarily closed ) three - manifold @xmath109 with the twisted lefschetz zeta function of certain flows . in particular in section 5
he considers a kind of torsion that he calls _
alexander quotient _ and denotes by @xmath744 : the reason for the `` quotient '' comes from the fact that fried uses a definition of the reidemeister torsion only up to the choice of a sign ( this is the `` refined reidemeister torsion '' of @xcite ) , while @xmath744 is defined up to an element in the abelian group of deck transformations of @xmath710 ( see also @xcite ) . in fact one can check that @xmath744 is exactly the reidemister - franz torsion @xmath82 considered in @xcite . in particular , when @xmath109 is the complement of an @xmath27-component link @xmath22 in @xmath23 and @xmath710 is the universal abelian cover of @xmath109 , then @xmath745 where we removed @xmath746 from the notation ( see ( * ? ? ?
* section 8) and @xcite ) . given an abelian cover @xmath707 with deck transformations group @xmath708 , let @xmath737 be the corresponding cohomology class .
a circular vector field @xmath75 on @xmath109 is _ compatible _ with @xmath710 if there exists a homomorphism @xmath754 such that @xmath755 , where @xmath95 and @xmath756 are as above .
[ theorem : alex equals the twisted zeta function ] let @xmath109 be a three manifold and @xmath758 an abelian cover .
let @xmath75 be a non - singular , circular and non degenerate vector field on @xmath109 compatible with @xmath710 .
suppose moreover that , if @xmath759 , then @xmath75 is transverse to @xmath727 and pointing out of @xmath109
. then @xmath760 where the symbol @xmath20 denotes the equivalence up to multiplication for an element @xmath761 , @xmath762 .
if @xmath22 is any @xmath27-component link in @xmath23 , let @xmath645 be a tubular neighborhood of @xmath22 and pose @xmath763 .
let @xmath75 be a non - singular circular vector field on @xmath256 , transverse to @xmath389 and pointing out of @xmath256 . then @xmath764 in the next subsections we prove theorem [ theorem : euler characteristic of eck ] , that will be obtained as a consequence of the following more general result . recall that an @xmath27-link @xmath662 determines the abelian cover @xmath765 of @xmath766 given in example [ example : abelian cover determined by a link ] .
when @xmath8 is a homology three - sphere , we have @xmath767 in order to simplify the notations , we remove @xmath726 from the notations of the alexander quotient and of the twisted lefschetz zeta function : @xmath768 the proofs of theorems [ theorem : euler characteristic of eck ] and [ theorem : euler characteristics of eck and alex ] will be carried on in two main steps : in subsection [ subsection : fibered links ] we will prove the theorems in the case of fibered links , while the general case will be treated in subsection [ subsection : the general case ] .
in this subsection we prove theorems [ theorem : euler characteristic of eck ] and [ theorem : euler characteristics of eck and alex ] for fibered links .
let @xmath230 be an open book decomposition of a homology three - sphere @xmath8 and let @xmath38 be an adapted contact form on @xmath8 .
in particular , with our definition , @xmath38 is also adapted to @xmath22 . in order to prove the theorems above we want to express the euler characteristic @xmath772 in terms of the twisted lefschetz zeta function of the reeb flow @xmath773 of @xmath257 and
then apply theorem [ theorem : alex equals the twisted zeta function ] .
the first thing that one should do is then to check if @xmath773 and @xmath726 satisfy the hypothesis of that theorem .
unfortunately this is not the case .
the needed properties are in fact the following : in our situation only properties 1 and 2 are satisfied . indeed , by the definition of open book decomposition
, there is a natural fibration @xmath777 such that the surfaces @xmath778 are the pages of the open book .
the fact that @xmath38 is adapted to @xmath230 implies that @xmath75 is always positively transverse to the pages .
this evidently implies that @xmath779 so that @xmath75 is circular .
however properties 3 and 4 above are not satisfied . indeed , after the m - b perturbation of @xmath304 , @xmath75 is tangent to @xmath774 on @xmath780 and @xmath616
moreover , as observed in subsection [ subsection : morse - bott theory ] , the m - b perturbations near the two tori @xmath276 and @xmath304 may create degenerate orbits . what we will do is then to perturb @xmath75 to get a new vector field @xmath781 .
this vector field will be defined in @xmath782 , where @xmath783 is an open tubular neighborhood of @xmath22 defined by @xmath784 , where , using the coordinates of subsection [ subsection : open books ] , @xmath785 . 1 .
@xmath781 coincides with @xmath75 outside a neighborhood of @xmath645 ; 2 .
@xmath781 satisfies properties 1 - 4 above with @xmath786 replaced by @xmath787 ; 3 .
the only periodic orbits of @xmath781 in @xmath788 are the four sets of non - degenerate orbits @xmath789 .
observe that property ( i ) implies that the twisted lefschetz zeta functions of the restrictions of the flows @xmath773 and @xmath790 to @xmath589 coincide , while property ( ii ) allows to apply theorem [ theorem : alex equals the twisted zeta function ] to @xmath790 .
a perturbation of @xmath75 into an @xmath781 satisfying the conditions ( i)-(iii ) can be obtained in more than one way .
an example is pictured in figure [ figure : modification of the dynamics non contact ] ( cf . also figure [ figure : dynamic of the m - b perturbation near k ] ) .
we briefly explain how it is obtained .
since the modification of @xmath75 is non trivial only inside disjoint neighborhoods of each @xmath238 , we will describe it only for a fixed component @xmath0 of @xmath22 .
the characterization of the perturbation will be presented in terms of perturbation of the lines in a page @xmath226 of @xmath230 that are invariant under the first return map @xmath248 of @xmath706 : we will refer to these curves as to _ @xmath248-invariant lines on @xmath226_. note that these curves are naturally oriented by the flow . outside a neighborhood of @xmath791 one
can see this perturbation in terms of a perturbation of @xmath248 into another monodromy @xmath792 , and @xmath781 is the vector field @xmath793 in @xmath794}{(x,1 ) \sim ( \phi'(x),0)}$ ] , where @xmath752 is the coordinate of @xmath795 $ ] .
observe first that the only periodic orbit in the ( singular ) @xmath248-invariant line @xmath796 containing @xmath185 ( in correspondence to the singularity ) is exactly @xmath185 .
similarly , the only periodic orbit in the @xmath248-invariant singular flow line @xmath797 containing @xmath307 is precisely @xmath307 .
denote @xmath798 the mapping torus of @xmath799 , @xmath800 .
we modify @xmath75 separately inside the regions of @xmath801 as follows . in the region containing @xmath184 ( and with boundary @xmath802 ) ,
the set of @xmath248-invariant lines ( the elliptic lines in the picture at left ) is perturbed in a set of @xmath792-invariant spiral - kind lines ( at right ) , each of which is negatively asymptotic to @xmath796 and positively asymptotic to @xmath184 .
it is easy to see that after the perturbation the only periodic orbit in the interior of this region is @xmath184 .
moreover , we can arrange the perturbation in a way that the differential @xmath803 of the first return map on @xmath226 of @xmath790 along @xmath184 , coincides , up to a positive factor smaller then @xmath77 , with @xmath804 , so that the lefschetz sign @xmath805 of @xmath184 is still @xmath806 .
the perturbation in the region between @xmath802 and @xmath808 is done by slightly pushing the monodromy in the positive @xmath298-direction in a way that the set of @xmath248-invariant lines is perturbed into a set of @xmath792-invariant lines , each of which is negatively asymptotic to @xmath796 and positively asymptotic to @xmath797 ( and so in particular there can not exist periodic orbits in this region ) .
a similar perturbation is done also inside the region between @xmath808 and @xmath809 , but in this case each @xmath792-invariant line is negatively asymptotic to @xmath797 and intersects @xmath809 pointing out of the three - manifold .
note that the two basis of eigenvectors of @xmath811 and @xmath812 are contained in the tangent spaces of the curves @xmath796 and , respectively , @xmath797 .
since on these curves @xmath813 , the lefschetz signs of the two orbits are not changed by the perturbation .
[ corollary : zeta functions of r and r are equals ] if @xmath781 is obtained from @xmath75 as above , then @xmath815 ) ) .
\end{aligned}\ ] ] where @xmath117 $ ] is the homology class of @xmath67 in @xmath816 .
now we want to compute more explicitly the twisted lefschetz zeta function @xmath817 .
let us begin with the local lefschetz zeta function of the simple orbits ( see definition [ definition : local zeta function of an orbit ] ) .
remember that the lefschetz number of @xmath67 is @xmath819 if @xmath67 is elliptic or negative hyperbolic and @xmath820 if @xmath67 is positive hyperbolic .
we prove here only the case of @xmath67 positive hyperbolic , leaving to the reader the other similar computations .
let @xmath717 be a positive meridian of @xmath238 for @xmath620 and set @xmath824 \in h_1(y \setminus k)$ ] ; fix moreover a seifert surface @xmath621 for each @xmath238 .
recall that , for a given @xmath431 , @xmath108 denotes the set of simple reeb orbits contained in @xmath109 .
this is an easy computation .
it suffices to substitute the monomial representation of @xmath831 = [ \gamma]$ ] given in example [ example : abelian cover determined by a link ] in the expression of the lefschetz zeta function of lemma [ lemma : zeta functions of the orbits ] .
to finish the proof it remains essentially to prove that @xmath832)).\ ] ] this is easy to verify recursively on the set of simple orbits .
suppose @xmath833 is an orbit set and let @xmath67 be an orbit such that @xmath834 for any @xmath598 .
then the set of all multiorbits that we can build using @xmath58 and @xmath67 can be expressed via the product formulae : @xmath835 as remarked in subsection [ subsection : ech for three - manifolds ] , the index parity formula [ equation : index parity formula of the ech index ] implies that the lefschetz sign endows the @xmath35-chain complex with an absolute degree and it coincides with the parity of the @xmath35-index .
then the contribution to the graded euler characteristic of @xmath836 , for any @xmath837 ( @xmath838 if @xmath67 is elliptic and @xmath839 if @xmath67 is hyperbolic ) is : @xmath840 to prove the result in the hat version we reason again at the level of chain complexes . recall that , if @xmath847 , by the definition of the @xmath6-chain complexes : @xmath848 where the second line comes from the product formula [ equation : recurrence definition of the orbit sets ] and the fact that @xmath849 is elliptic for any @xmath241 . taking the graded euler characteristics as above we have : @xmath850 ) = \\ & = & \chi(\widehat{ecc}(l , y,\alpha ) ) \cdot \prod_{i= 1}^n \frac{1}{1-t_i } , \end{aligned}\ ] ] where the last equality comes from the fact that @xmath851 = [ \mu_i ] = t_i \in h_1(y \setminus l)$ ] . if @xmath852 , last equation and equation [ equation : euler characteristics of eck full ] evidently imply equation [ equation : euler characteristics of eck hat ] .
if @xmath230 is an open book decomposition of @xmath8 , one can think of @xmath564 and @xmath565 as invariants of the pair @xmath228 and the adapted @xmath38 .
it is interesting to note that the euler characteristic of @xmath853 with respect to the surface @xmath226 ( see example [ example : eck for open books ] ) coincides with the sum of the lefschetz signs of the reeb orbits of period @xmath77 in the interior of @xmath226 , i.e. the lefschetz number @xmath854 of @xmath248 .
in fact , given @xmath8 ( not necessarily an homology three - sphere ) we can say even more about this fact by relating @xmath853 to the _ symplectic floer homology @xmath855 _ of @xmath228 , whose euler characteristic is precisely @xmath854 . here
we are considering the version of @xmath855 for surfaces with boundary that is slightly rotated by @xmath248 in the positive direction , with respect to the orientation induced by @xmath226 on @xmath271 ( see for example @xcite and @xcite ) .
[ proposition : eck_1 iso to sh for open books ] let @xmath230 be an open book decomposition of a three - manifold @xmath8 and let @xmath38 be an adapted contact form .
then @xmath856 where the degree of @xmath564 is computed using a page of the open book .
the proof of last proposition passes through another floer homology theory closely related to @xmath35 , which is the _ periodic floer homology _ ,
denoted by @xmath857 , and defined by hutchings ( see @xcite ) . given a symplectic surface @xmath858 ( here with possibly empty boundary ) and a symplectomorphism @xmath859 , consider the mapping torus @xmath860}{(x,2 ) \sim ( \phi(x),0)}.\ ] ] then @xmath861 is defined in an analogous way than @xmath35 for an open book but replacing the reeb vector field with a stable hamiltonian vector field @xmath75 parallel to @xmath793 , where @xmath752 is the coordinate of @xmath862 $ ] : we refer the reader to @xcite or @xcite for the details .
the chain group @xmath863 is the free @xmath343 module generated by orbit sets of @xmath75 and the boundary map counts index 1 holomorphic curves in the symplectization ; then , under some condition on @xmath248 , the associated homology @xmath861 is well defined .
homology groups @xmath864 associated to the chain groups @xmath865 generated by degree-@xmath241 multiorbits are also well defined .
[ theorem : pfh_i iso to ech_i fo open books ] if @xmath38 is a contact form adapted to @xmath228 then there exists a stable hamiltonian structure such that for any @xmath866 , @xmath867 ( here we are using a simplified notation which is different from that in @xcite ) . @xmath868 is generated by orbits of period @xmath77 , which are in bijective correspondence with the set @xmath869 of the fixed points of @xmath248 via the map @xmath870 which moreover evidently respects the lefschetz signs .
then this correspondence induces an isomorphism between @xmath868 and a chain complex of @xmath855 . indeed the following holds ( see for example @xcite ) : this is an easy consequence of the definitions and the results above . by lemma [ lemma : eck iso to ech of int n ] we have @xmath872 observing that the proof of theorem [ theorem : pfh_i iso to ech_i fo open books ] given in @xcite works also if @xmath271 is disconnected we get @xmath873 the result then follows applying proposition [ proposition : pfh_1 iso to sh ] .
we get an interesting consequence of this fact when also the alexander degree of heegaard - floer knot homology of a fibered knot is computed with respect to ( the homology class of ) a page of the associated open book . indeed , using the symmetrized degree adopted by ozsvth and szab , we know that @xmath874 is isomorphic to a copy of @xmath371 generated by the class of the corresponding contact element .
moreover , whenever @xmath875 , we have also that _ @xmath876 categorifies @xmath854_. the first approach that one could attempt to apply theorem [ theorem : alex equals the twisted zeta function ] to a general link @xmath662 is to look for a contact form on @xmath8 that is compatible with @xmath22 and whose reeb vector field is circular outside a neighborhood of @xmath22 .
unfortunately we will not be able to find such a contact form .
the basic idea to solve the problem consists in two steps : 1 . find a contact form @xmath38 on @xmath8 which is compatible with @xmath22 and for which there exists a finite decomposition @xmath878 for which @xmath257 is circular in each @xmath879 ; 2
. apply theorem [ theorem : alex equals the twisted zeta function ] separately in each @xmath879 to get the result : this can be done using the ( more general results ) in sections 6 of @xcite .
torres formula , first proved in @xcite , is a classical result about alexander polynomial , which essentially explains how , starting from the alexander polynomial of a given link @xmath22 , to compute the alexander polynomials of any sub - link of @xmath22 .
[ prop : any l is part o a fibered l ] let @xmath880 be an @xmath27-components link and let @xmath41 be any fixed contact structure on @xmath8 .
then there exists an @xmath62-components link @xmath881 with @xmath882 and such that : as recalled in subsection [ subsection : open books ] , given a contact structure @xmath41 on @xmath8 , in @xcite giroux explicitly constructs an open book decomposition of @xmath8 that supports a contact form @xmath38 such that @xmath310 . in the proof of theorem [
theorem : giroux correspondance ] we saw that such an open book decomposition is built starting from a cellular decomposition @xmath311 of @xmath8 that is compatible with @xmath41 .
moreover we recalled that , up to taking a refinement , any cellular decomposition of @xmath8 can be made compatible with @xmath41 by an isotopy . using the simplicial approximation theorem
, it is possible to choose a triangulation @xmath311 of @xmath8 in a way that , up to isotopy , @xmath22 is contained in the @xmath77-skeleton @xmath314 of @xmath311 . up to take a refinement , we can suppose moreover that @xmath311 is adapted to @xmath41 .
let @xmath226 be the @xmath272-page of the associated open book built via theorem [ theorem : construction of an open book adapted to a contact structure ] : properties 1 and 2 of @xmath226 reminded during the proof of that theorem , imply that @xmath884 and that , if @xmath885 is a suitable neighborhood of @xmath316 , then it is possible to push @xmath886 inside @xmath226 to make it contained in @xmath271 .
note that in each strip composing @xmath887 we have only one possible choice for the direction in which to push @xmath886 to @xmath271 in a way that the orientation of @xmath22 coincides with that of @xmath271 .
we would like to extend this isotopy also to @xmath888 to make the whole @xmath22 contained in @xmath271 .
suppose that @xmath889 is a connected component ( homeomorphic to a ball ) of @xmath885 .
in particular we suppose that @xmath890 is connected .
then @xmath891 consists of two points @xmath892 and @xmath893 .
the extension is done differently in the following two cases ( see figure [ figure : making l contained in the binding ] ) : in @xmath885 : easy case at the left and general case at right .
the dotted lines are @xmath77-simplexes in @xmath314 , while the bold segments from @xmath892 to @xmath893 represent the push - offs of @xmath22 in @xmath885.,title="fig : " ] in @xmath885 : easy case at the left and general case at right .
the dotted lines are @xmath77-simplexes in @xmath314 , while the bold segments from @xmath892 to @xmath893 represent the push - offs of @xmath22 in @xmath885.,title="fig : " ] * _ easy case _ :
this is when @xmath892 and @xmath893 belong to the same connected component of @xmath894 .
the isotopy is then extended to @xmath889 by pushing @xmath895 to @xmath894 inside @xmath896 ( figure at left ) ; * _ general case _ : if @xmath892 and @xmath893 belong to ( the boundary of ) different connected components @xmath796 and @xmath797 of @xmath894 we proceed as follows .
+ let @xmath897 be a point in the interior of @xmath898 , @xmath800 .
let @xmath67 be a simple arc in @xmath896 from @xmath327 to @xmath328 ( there exists only one choice for @xmath67 up to isotopy ) .
let @xmath329 be obtained by positive giroux stabilization of @xmath226 along @xmath67 ( see figure at the right ) .
+ now we can connect @xmath892 with @xmath797 by an arc in @xmath899 crossing once the belt sphere of the @xmath77-handle of the stabilization ; let @xmath900 be the end point of this arc . since a giroux stabilization is compatible with the orientation of @xmath271 , @xmath900 and @xmath893 are in the same connected component of @xmath901 , so that we can connect them inside @xmath894 and we are done .
pushing @xmath22 to @xmath271 ( and changing @xmath22 and @xmath226 as before where necessary ) gives a link @xmath902 that is contained in @xmath271 . to see that @xmath902 is isotopic to @xmath22 we have to prove that , for any @xmath889 as before , the two kinds of push - offs we use do no change the isotopy class of @xmath22 .
clearly the isotopy class of @xmath22 is preserved in the easy case .
for the general case , it suffices to show that substituting the arc @xmath903 from @xmath892 to @xmath893 with an arc crossing once the belt sphere of the handle does not change the isotopy class of @xmath22 .
this is equivalent to proving that , if @xmath67 is the path of the giroux stabilization and @xmath904 , where @xmath333 is the core curve of the handle , then @xmath336 bounds a disk in @xmath721 .
this can be proved for example by using the particular kind of heegaard diagrams used in @xcite .
observe that , if @xmath905 is the co - core of the handle , then @xmath336 is isotopic in @xmath226 to @xmath906 , where @xmath792 is the monodromy on @xmath329 given by the giroux stabilization .
we finish by observing that , up to a small perturbation near @xmath271 , @xmath906 is isotopic to an attaching curve of a heegaard diagram of @xmath8 .
we recall now the torres formula that we will use in the second step of our proof of theorem [ theorem : euler characteristics of eck and alex ] .
since we need to consider the alexander quotient as a polynomial , we will use the same convention adopted for the graded euler characteristic and we will express the variables as subscripts of the symbol @xmath46 .
[ theorem : torres formula ] let @xmath716 be an @xmath27-link in a homology three - sphere @xmath8 , @xmath907 a knot in @xmath721 and @xmath908 .
let @xmath621 be a seifert surface for @xmath238 , @xmath909 .
then @xmath910 where @xmath911 indicates the polynomial @xmath912 evaluated in @xmath913 .
we refer the reader to @xcite for the original proof .
see also @xcite for a proof making use of techniques of dynamics .
we also mention that in @xcite a proof of this theorem is provided making use only of elementary techniques about seifert surfaces ; moreover a generalization of the formula to links in any three - manifold is given in @xcite .
the factor @xmath914 expresses then the fact that @xmath907 is the only orbit counted in @xmath45 and not in @xmath915 .
+ the condition @xmath913 comes from the fact that , if @xmath916 is a meridian for @xmath907 , so that @xmath917 $ ] , then @xmath918 ) ) = 1 $ ] .
[ observation : condition t_n = 1 ] one can see the condition @xmath913 also from a purely topological point of view .
image to take the manifold @xmath919 and then to glue back @xmath907 .
the effect on @xmath920 is that the generator @xmath921 $ ] is killed and now the homology class of a loop @xmath922 is determined only by the numbers @xmath923 , @xmath924 ( i.e. by @xmath925 ) .
let @xmath606 be a given link in @xmath8 .
proposition [ prop : any l is part o a fibered l ] implies that there exists an open book decomposition @xmath926 of @xmath8 with binding @xmath927 for some @xmath882 .
let @xmath38 be a contact form on @xmath8 adapted to @xmath926 .
let @xmath928 be its reeb vector field . as remarked in subsection [ subsection : fibered links ] , and using the same notations , @xmath75 is circular in @xmath929 where , recall , @xmath787 is an union of tubular neighborhoods @xmath930 , @xmath931 of @xmath22 . since @xmath38 is also adapted to @xmath340 , then each @xmath932 is , by definition , foliated by concentric tori , which in turn are linearly foliated by reeb orbits that intersect positively a meridian disk for @xmath238 in @xmath252 .
now , we can choose @xmath38 in a way that for each @xmath933 the tori contained in @xmath934 are foliated by orbits of @xmath75 with fixed irrational slope .
this condition can be achieved by applying the darboux - weinstein theorem in @xmath252 to make @xmath935 like in example 6.2.3 of @xcite .
this implies that , for each @xmath933 , the only closed orbit of @xmath75 in @xmath934 is @xmath238 .
define @xmath936 , where @xmath937 we have : @xmath938))\\ & = & \zeta_{\rho_l}(\phi_r|_{y \setminus u(l ' ) } ) ) \cdot \prod_{i = n+1}^m \zeta_{k_i}(\rho_l([k_i]))\\ & = & \zeta_{\rho_{l'}}(\phi_r|_{y \setminus u(l')}))|_{t_1,\ldots , t_n,1,\ldots,1 } \cdot \prod_{i = n+1}^m \zeta_{k_i}(\rho_l([k_i]))\\ & \doteq & \mathrm{alex}_{t_1,\ldots , t_n,1,\ldots,1}(y \setminus l ' ) \cdot \prod_{i = n+1}^m \zeta_{k_i}(\rho_l([k_i]))\\ & = & \mathrm{alex}_{t_1,\ldots , t_n,1,\ldots,1}(y \setminus l ' ) \cdot \prod_{i = n+1}^m \left(1 - \prod_{j=1}^n t_j^{\langle k_{i } , s_j \rangle } \right)^{-1}\\ & = & \mathrm{alex}_{t_1,\ldots , t_n}(y \setminus l),\end{aligned}\ ] ] where : * follow reasoning like in the proof of equation [ equation : product formula of ecc in terms of zeta functions ] ; * hold since @xmath238 , for @xmath933 , is the only reeb orbit of @xmath38 in @xmath934 ; * comes from the idea in observation [ observation : condition t_n = 1 ] : @xmath726 and @xmath939 coincide on the generators @xmath725 of @xmath940 for @xmath620 and @xmath824 = 1 \in h_1(y \setminus l)$ ] for @xmath933 ; * @xmath939 and @xmath941 satisfy hypothesis of theorem [ theorem : alex equals the twisted zeta function ] , up to slightly perturb @xmath75 near @xmath942 , @xmath620 , to make it non degenerate transverse to the boundary like in the proof in subsection [ subsection : fibered links ] ; * is due to the fact that the @xmath238 s are elliptic ; * is obtained by applying repeatedly the torres formula on the components @xmath943 . as mentioned at the beginning of this subsection , we could also apply the more general results in @xcite using the fact that the reeb vector field @xmath75 used is circular in each @xmath944 , since here @xmath75 is positively transverse to any meridian disk of @xmath238 in @xmath932 . | given a transverse knot @xmath0 in a three dimensional contact manifold @xmath1 , in @xcite colin , ghiggini , honda and hutchings define a hat version of embedded contact homology for @xmath0 , that we call @xmath2 , and conjecture that it is isomorphic to the knot floer homology @xmath3 .
we define here a full version @xmath4 and generalise the definitions to the case of links .
we prove then that , if @xmath5 , @xmath6 and @xmath7 categorify the ( multivariable ) alexander polynomial of knots and links , obtaining expressions analogue to that for knot and link floer homologies in the plus and , respectively , hat versions . |
one of the enduring mysteries of cosmic gamma - ray bursts ( grbs ) is that of the so - called `` dark '' bursts - those with no detected optical or near - infrared afterglows . in the fireball model of grbs
( for a review , see * ? ? ?
* ) , an explosive event produces a relativistically expanding blast wave .
internal shocks within the blast wave are the source of the high energy prompt emission while afterglows arise from external shocks when the blast wave run into an interstellar medium .
grbs may be divided into `` long - duration '' and `` short - duration '' categories ; no unambigous detection of an optical afterglow has yet been made for any short - duration event ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and only in one case has an x - ray afterglow been detected @xcite .
on the other hand , virtually all of the long - duration bursts display fading soft x - ray afterglows ( e.g. * ? ? ?
* ; * ? ? ?
* ) , and they have been shown to be associated with supernova explosions @xcite . for more than half of the long - duration grbs
no optical and/or radio afterglow has been detected ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) ; however , rapid and systematic follow - up of bursts has decreased the relative fraction of dark bursts in recent years @xcite .
recently , @xcite suggested using the value of the optical - to - x - ray spectral index as a diagnostic for early identification of dark bursts , and they found five dark bursts ( including grb 001025a ) in their sample of 52 bursts .
there are several possible explanations for dark bursts @xcite
. they could be at very high redshifts @xcite , or they could be heavily obscured by dust in their host galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
they could also occur in regions where the surrounding medium is tenuous , far from the centers of their host galaxies , where the strong shocks required for particle acceleration could not form @xcite .
observations of bursts detected by the _ bepposax _ , the _ high energy transient explorer _ , and the _ rossi x - ray timing explorer _ spacecrafts have indeed shown that bursts can be dark for a variety of reasons .
some bursts seem in fact to be dark due to extinction , e.g. grb 000210 @xcite , grb 970828 @xcite , and grb 001109 @xcite .
other bursts have faint optical afterglows either because of moderately high redshift , e.g. , grb 020124 @xcite , or because they are intrinsically faint without extinction , e.g. grb 980613 @xcite and grb 021211 @xcite .
analysis of bursts detected by _
swift _ will undoubtly shed more light on the nature of dark bursts .
already now it has become evident that afterglows of bursts detected by
_ swift _ on average are fainter than afterglows from bursts detected by previous instruments ( 1.7 magnitudes in the optical / near - infrared ( berger et al . , 2005 ) and a factor three in x - rays jakobsson , 2005 ) .
the rapid fading of some _ swift _ bursts demonstrates that at least these become intrinsically faint after the first few hundred seconds @xcite .
here we examine the case of the dark burst grb 001025a .
we report on our temporal and spectral analysis of the prompt emission , on our very deep optical upper limits on the afterglow , and on the results of our temporal and spectral analysis of the x - ray afterglow .
finally , we discuss the nature of grb 001025a within the framework of the fireball model , and the implications for dark bursts in general .
grb 001025a was first reported as an event detected by the _ rossi x - ray transient explorer ( rxte ) all sky monitor ( asm ) at 11405 s ut october 25 , 2000 @xcite .
its duration in the asm 1.512 kev energy band was @xmath215 s , and its peak flux was @xmath2 4 crab ( @xmath3 ) in the 512 kev band .
the event was detected by a single camera and localized to a @xmath4 error box .
it was also observed by _
ulysses _ , the x - ray / gamma - ray spectrometer ( xgrs ) aboard the _ near earth asteroid rendezvous mission ( near ) , and _ wind ( konus experiment ) in the interplanetary network ( ipn ) , and triangulated to a preliminary @xmath5 wide annulus which crossed the asm error box to form a 25 arcminute@xmath6 error box @xcite . _ _ _ we have used the final _ ulysses _ , near , and _
wind_-konus data to obtain a refined ipn error box for this burst
. the _ ulysses - konus annulus is centered at r.a.(2000)@xmath7 , decl.@xmath8 , and has a radius of @xmath9 ( @xmath10 ) .
the _ ulysses _ - near annulus is centered at r.a.@xmath11 , decl.(2000)@xmath12 , and has a radius of @xmath13 @xmath14 .
the intersection of these annuli defines a 96 arcminute@xmath6 error box .
however , the intersection of the _ ulysses_-konus annulus with the asm error box defines a 3.3 arcminute@xmath6 error box , whose coordinates are given in table [ error ] .
the combined asm - ipn error box is shown in figure [ xmmim ] . _ the asm , konus , and xgrs lightcurves are shown in figure [ all_time ] . with a t@xmath15 duration of 2.9 s (
i.e. , the time to accumulate between 5% and 95% of the photons ) in the 50200 kev energy range , this burst falls into the `` long - duration '' category .
figure [ all_time ] shows the lightcurve in several energy ranges , and indicates the intervals used for time - resolved spectral analysis .
this figure also shows the hardness ratio as a function of time .
a hard - to - soft evolution , which is often observed in grbs ( e.g. * ? ? ?
* ) , is evident .
the energy spectra are presented in figure [ konus_spec ] . the band @xcite model , which consists of two smoothly joined power laws , was used to perform spectral fits .
the band function is : @xmath16 this yields a peak energy in the range @xmath17 kev during the burst ( see table [ konus_spec_tab ] ) .
the 152000 kev peak flux of this burst was @xmath18 over 0.048 s , and the 152000 kev fluence was @xmath19 . neither the time history , nor the spectrum , nor the intensity of this event was exceptional in any way .
@xcite has proposed a simple redshift estimator based on the photon flux , energy spectrum , and duration of a grb . using this method ,
we obtain @xmath20 , with a statistical uncertainty of about 50% . in the weeks immediately following grb 001025a , optical , x - ray , and radio observations were carried out in attempts to identify an afterglow .
an optical afterglow was not identified , either in observations starting 1.16 days after the burst ( * ? ? ?
* @xmath21 ) , or in observations starting 1.66 days after the burst ( * ? ? ?
* @xmath22 ) . an _ xmm - newton
_ observation @xmath23 days after the burst revealed one apparently fading source ( s1 , see figure [ xmmim ] ) in the initial asm / ipn error box @xcite . in late - time follow - up
_ xmm - newton _
observations this source had disappeared , strongly indicating that ( s1 ) is the x - ray afterglow of grb 001025a . radio observations with the _ very large array _ ( vla ) were carried out at two epochs , but they did not reveal any radio afterglow .
the first vla observation took place november 1 , 2000 , approximately 6 days after the grb ( d. frail 2003 , private communication ) .
unfortunately , the _ xmm - newton _
attitude was subsequently revised @xcite , and the refined position of s1 was @xmath24 from the initial one . due to the off - set between the vla pointing and the revised position of s1 , the vla sensitivity at the position of s1 was relatively low .
vla observations took place again on november 21 and 24 , 2000 ( 2730 days after the burst ) , but given the timescales and sensitivies ( @xmath25 ) most grb radio afterglows would not have been detected .
thus it is not clear whether this burst was radio - quiet . in the following three sections , we present our analysis and results of optical observations and x - ray observations of the grb 001025a afterglow .
the field of grb 001025a was observed at the _ very large telescope _ ( vlt ) equipped with the fors1 camera at three epochs , under photometric and good seeing conditions ( see table [ vltlog ] for details ) .
the first epoch was a few hours after the release of the ipn error box ( hurley et al .
2000 ) , the second epoch was 24 hours later , and the final epoch was 5 months later . by comparing the combined @xmath1-band images of the three epochs we find no transient sources in the ipn error box and in particular near the _ xmm - newton _ error circle of s1 to a 3@xmath26 detection limit of @xmath0 .
figure [ host ] shows a 12@xmath2712 arcsec@xmath6 section of the last epoch vlt image with the _ xmm - newton _ 2@xmath26 error circle overplotted .
the 2@xmath26 error circle is consistent with the eastern part of what seems to be either a galaxy with two main components or a random projection of two unrelated objects .
following @xcite , we estimate that the probability for a chance alignment is less than 1% , so we consider this object the likely host galaxy of grb 001025a .
the total magnitude of this complex is @xmath280.04 and @xmath290.12 ( corrected for a modest foreground extinction of @xmath30 * ? ? ?
grb afterglows display a wide variety of fading behaviors ( see figure [ dark ] ) . of the pre-_swift
_ detected optical afterglows , all but grb 030115 would have been detected in the vlt observations .
the failure to detect the optical afterglow of grb 001025a is thus not because our observations are less sensitive than studies of other detected grb afterglows .
_ xmm - newton _ carried out observations of the region around the error box at three different epochs , spanning 1.92.3 days , 761.0761.2 days , and 910.2910.4 days after the grb ( see table [ xmmlog ] ) .
the _ xmm - newton _ target of opportunity observation @xmath23 days after the burst revealed at least four x - ray sources in the initial asm / ipn error box @xcite , one of which ( s1 ) appeared to be fading .
as shown in figure [ xmmim ] , only s1 is in the final asm / ipn error box .
this , along with marginal evidence for a fading behavior ( see section 4 below , and * ? ? ?
* ) , suggested that it was the x - ray afterglow of grb 001025a . in order to determine with certainty
whether s1 was indeed the fading afterglow ( as opposed to a random , variable x - ray source such as an active galactic nucleus ) , we carried out a follow - up _ xmm - newton _ observation .
this observation started on november 25 , 2002 , but had to be terminated because of high background radiation ; it was completed on april 23 , 2003 .
the first observation has already been reported on in @xcite , but for the analysis presented here , we have made two improvements : we have omitted data taken during periods with high background , and we have analyzed data from the second and third epochs as well .
the effective epic pn exposure times of the three _ xmm - newton _ observations , after screening out high background intervals , are 15 ks , 9 ks , and 18 ks respectively . the data were processed and analysed with sas version 5.4.1 and xspec version 11.3 . periods with flares were identified based on the 1012 kev lightcurve of the full epic detector and events registered during these periods ( where the pn count rate was above 1.5 counts / s , and where the mos count rate was above 0.25 counts / s ) were filtered out . for the pn detector ,
single and double events were included in the analysis and the two exposures taken in the first epoch were co - added . for the mos detectors , single , double , triple , and quadruple events were included in the analysis and the two mos data sets for each epoch were co - added . pn events in the energy interval 0.39 kev and mos events in the energy interval 0.47 kev were included in the analysis .
spectra and lightcurves of sources were extracted using a circular aperture of radius 18 .
background events were extracted from an annulus centered on the source in question with an inner radius of 18 and an outer radius of 50 .
the background regions are entirely on the same chip as the relevant sources , and they do not include other contaminating sources .
below we present results from our analysis of the _ xmm - newton _ images , lightcurves , and spectra of the grb 001025a x - ray afterglow . in the first epoch observation two sources
were detected , s1 r.a.(2000)@xmath31 , decl.(2000)@xmath32 , @xmath33 ) , and s2 , in and in the vicinity of the final asm / ipn error box , respectively ( see figure [ xmmim ] ) . over the first _ xmm - newton _ observation 264 net counts ( 0.312 kev ) were detected for source s1 by the pn detector , corresponding to a flux of @xmath34 erg @xmath35 s@xmath36 . for source s2 , 78 net counts ( 0.312 kev ) were detected , corresponding to a flux of @xmath37 erg @xmath35 s@xmath36 . by the time of the second epoch observation , s1 had faded beyond detection ( @xmath38(0.312 kev)@xmath39 erg @xmath35 s@xmath36 ) while the flux of s2 was consistent with that of the first epoch . in the third epoch observation s1
was still undetectable ( @xmath38(0.312 kev)@xmath40 erg @xmath35 s@xmath36 ) and the flux of s2 was consistent with the flux at the first two epochs .
hence , source s2 is persistent over several years , displaying flux variations less than about 25% from epoch to epoch .
s2 is coincident with an @xmath41 optical source which based on a spectrum obtained with the _ nordic optical telescope _ seems to be a qso at redshift @xmath42 ( m. i. andersen , private communication ) .
the pn lightcurve for s1 is well fitted by a power law with decay index @xmath43 ( @xmath44 for 8 d.o.f .
, 0.3 - 12 kev ) .
this is somewhat lower than , but consistent with , the steep decay index @xmath45 derived by @xcite .
the data quality of the corresponding co - added mos1 + 2 lightcurve does not allow us to further constrain the fading of s1 .
the decay of s1 is typical of x - ray afterglows having a decay index of @xmath46 @xcite . by extrapolating the best - fit pn lightcurve
we predict the total number of counts in the second and third epoch _ xmm - newton _
observations from s1 to be 0.04 counts and 0.02 counts , respectively ( i.e. , well below detectability ) .
hence , the second and third epoch observations establish that s1 is a non - persistent source ; hence we identify s1 as the x - ray afterglow of grb 001025a . in order to compare the x - ray afterglow to the prompt x - ray emission as observed by the rxte asm we extracted the pn 1.512 kev lightcurve .
a power law fit to the 1.512 kev light curve is not well constrained ( decay index @xmath47 ) .
hence , we extrapolated the best fit power law to the 0.312 kev light curve ( renormalized to the 1.512 kev flux level ) to the prompt emission phase .
this extrapolation reproduces the rxte asm flux within the uncertainties ( see figure [ xmm_lc ] ) .
however , fitting a power law to the fading of the prompt emission in the 1.512 kev band yields a steep power law ( decay index @xmath48 ) underestimating the pn data points .
such a steep decay has been reported in several early x - ray afterglows observed with the _ swift _ x - ray telescope @xcite .
it is not possible to fit a power law decay jointly to the fading prompt emission and the x - ray afterglow .
the x - ray afterglow observed by _ xmm - newton _ is therefore not a simple continuation of the prompt emission .
spectra of s1 were extracted from the first epoch observations and binned with a minimum of 20 counts per bin .
several spectral models were fitted to the pn and mos1 + 2 spectra simultaneously : ( i ) power law models , representing synchrotron emission from a population of relativistic electrons in the fireball @xcite , ( ii ) thermal plasma emission giving rise to x - ray emission lines as seen in several other grb afterglows @xcite ( the mekal model as is typically used to represent thermal emission * ? ? ?
* ; * ? ? ?
* ) , and ( iii ) combinations of ( i ) and ( ii ) . in all models galactic absorption was included , and additional absorption at the burst redshift was included in some of the models .
a summary of the spectral fits is given in tables [ pl],[mekal],[plmekal ] with errors quoted as 90% confidence intervals .
an absorbed synchrotron model is an acceptable fit to the spectrum ( see table [ pl ] ) , but the best - fit photon index is rather steep ( @xmath49 , consistent with the findings of * ? ? ?
* ) . the galactic column density in the direction of grb 001025a is @xmath50 @xmath35 and can in worst case be uncertain by up to 50% @xcite .
power law models indicate absorption in excess of galactic absorption , but the column density of extra - galactic absorbing material is not strongly constrained . in this model
the burst redshift is unconstrained .
an absorbed thermal plasma model ( mekal ) is as good a fit as the absorbed synchrotron emission model ( see table [ mekal ] ) and implies a plasma temperature around 3 kev and a burst redshift around @xmath51 .
@xcite favor a redshift in the range 0.51.2 from their mekal model fits , and we note that the thermal plasma could be outflowing resulting in a host galaxy redshift lower by @xmath52 ( e.g. * ? ? ?
hence , thermal emission models do not provide strong constraints on the burst redshift .
the absorbing column density at the burst redshift is at most twice the galactic column density .
the plasma abundances are not well constrained , but they are consistent with the solar values ( @xmath53 , @xmath54 ) . naturally , a model with a combination of a synchrotron emission component and a thermal emission component ( see figure [ xmmspec ] ) is also a good fit ( see table [ plmekal ] ) .
the relative normalization of the synchrotron emission component and the thermal emission component is not well constrained , but the synchrotron emission component contributes @xmath55 of the total flux and above @xmath23 kev the synchrotron emission component dominates . the best - fit photon index is lower than in the pure synchrotron emission model fits ( @xmath56 ) , but close to typical photon indices for x - ray afterglows @xcite .
the favored burst redshift is @xmath57 , and there is an indication of circumburst absorbtion ( @xmath58 @xmath35 ) .
the abundances of the thermal plasma are not well constrained , but they are consistent with the solar values ( @xmath59 ) . in conclusion , an absorbed synchrotron model , an absorbed thermal plasma model , and a combination of
the two all represent a good fit to the data .
consequently , we consider all these models in the following .
we derived the broadband spectral energy distribution by extrapolating the best - fit x - ray spectral models to the optical band .
we used the x - ray decay index @xmath60 to recast the x - ray spectrum to the epoch of the first @xmath1-band observations ( 1.20 days after the burst ) . for the synchrotron emission model ( spectral index @xmath61 ) an observed @xmath1-band extinction of at least 7 magnitudes
is required to bring the model in agreement with the @xmath1-band upper limit ( see figure [ sed ] ) . on the other hand , extrapolating the synchrotron emission component from the best - fit combined synchrotron emission / thermal emission model ( spectral index @xmath62 ) to
the @xmath1-band shows that the @xmath1-band upper limit is consistent with the model , if the cooling break is close to the x - ray band .
the spectral index of the ( x - ray ) afterglow is a key parameter for constraining the properties of the burst and its environment .
the pure synchrotron emission fit to the grb 001025a x - ray afterglow has a very steep spectral index for a grb x - ray afterglow . on the other hand
, the synchrotron component of the combined thermal plasma / synchrotron model has a spectral index in line with what is found for other grbs @xcite .
detection of thermal afterglow emission has been reported in several other grbs ( see * ? ? ? * for an account of grbs with thermal emission ) increasing the soft x - ray flux in the @xmath63 kev range .
one of the most prominent claims of thermal emission is in the grb 011211 x - ray afterglow where thermal emission is detected during the first 5 ks of the _ xmm - newton _
observation @xcite .
we investigated the effect on the spectral index of fitting an absorbed synchrotron model to a spectrum with thermal emission by fitting an absorbed synchrotron emission model to the first 5 ks of the grb 011211 _ xmm - newton _ spectrum .
we obtain a spectral index of @xmath64 ; this is very similar to the spectral index ( @xmath65 ) we find fitting the same model to grb 001025a .
the steep spectral index of a pure synchrotron emission fit to grb 001025a compared to the shallower index from a synchrotron emission plus thermal emission fit thus suggests that thermal emission may be steepening the x - ray afterglow spectrum of grb 001025a . due to the different best fit spectral indices of the synchrotron emission in the pure synchroton model and in the combined thermal plasma / synchrotron model , these two models have quite different implications for the deduced properties of the grb 001025a fireball and its environment .
the prompt soft x - ray emission has a steep decay as seen in the tails of other grbs @xcite .
observations with the _ swift _ xrt has revealed a population of bursts with an early steep decay and a shallower late time decay @xcite similar to the behavior of grb 001025a . several models predicts an early steep x - ray decay , including a hot cocoon surrounding a relativistic jet @xcite , emission from outside the relativistic beaming cone , @xmath66 , @xcite , and a single spherical shell emitting instantaneously @xcite .
the steepness of the prompt emission decay excludes that the tail of the prompt emission is the onset of the afterglow @xcite .
however , the afterglow could appear at any time @xmath67
s after the burst at which epoch the grb 001025a afterglow is expected to be below the sensitivity of rxte asm .
the x - ray afterglow has a steep spectral index , if it is due to synchrotron emission only .
the standard fireball model predicts in this case a temporal decay index of @xmath68 to @xmath69 for a spherical blast wave and an index of @xmath70 to @xmath71 for a jet @xcite . this is steeper than observed ( @xmath72 ) , but only the jet model can be ruled out .
if the x - ray afterglow is due to a combination of thermal plasma and synchrotron emission , the fireball model predicts a shallower afterglow decay index of the synchrotron component ( index @xmath73 to @xmath74 for a spherical blast wave , index @xmath75 to @xmath76 for a jet , * ? ? ? * ) more in line with the x - ray lightcurve . in this case models where the x - ray afterglow originates from fast cooling
are ruled out . in the combined synchroton / thermal plasma spectral model of the x - ray afterglow ,
the suggested burst redshift is @xmath57 .
this yields an isotropic prompt energy release of @xmath77 erg ( for a @xmath78 km / s / mpc , @xmath79 , @xmath80 cosmology ) which is quite typical ( e.g. * ? ? ?
the isotropic 0.312 kev afterglow luminosity is @xmath81 erg / s time - averaged over the first _ xmm - newton _ observation .
we now compare the x - ray luminosity of grb 001025a to the luminosities of the grb sample of @xcite .
we use our derived x - ray afterglow decay index , flux and spectral index to calculate the isotropic luminosity ( @xmath82 , eq .
( 1 ) in * ? ? ?
* using their cosmology ) of grb 001025a 10 hours after the burst . in the combined synchrotron / thermal plasma model with a burst redshift @xmath83
, we find @xmath84 erg / s .
this is close to the typical x - ray afterglow luminosity at this epoch for the @xcite burst sample .
however , in that sample , redshifts are not known for most of the bursts , in which case the luminosity has been derived assuming a redshift of @xmath85 .
if grb 001025a is at a redshift of @xmath85 its isotropic luminosity is @xmath86 erg / s which is in the low tail of the @xcite x - ray luminosity distribution .
if the burst redshift is @xmath87 , as predicted in the thermal plasma only model , the luminosity of grb 001025a @xmath88 erg / s is lower than all grbs in the @xcite sample .
the @xmath89-band detection of the likely host galaxy implies a redshift @xmath90 .
the magnitude and color of the galaxy are consistent with a redshift around two , but they do not constitute strong constraints on the redshift .
the pseudo - redshift estimate @xcite for this burst @xmath91 , and although not strongly constraining , this is in line with the lower redshift derived from the thermal emission models of the x - ray afterglow .
the appearance of the candidate host galaxy is similar to other grb hosts , i.e. , an irregular morphology and a color that for a redshift of @xmath92 is indicative of active star formation . from the absorbing column density in the soft x - ray band ,
the extinction towards the burst can be estimated by assuming a gas - to - dust ratio .
the best - fit column densities at the burst redshift are @xmath93 @xmath35 .
taking a galactic gas - to - dust ratio @xcite , this translates into a restframe extinction of @xmath94 for @xmath95 . using an smc gas - to - dust ratio results in @xmath96 , i.e. not enough to account for the extinction required in the pure synchrotron model for the x - ray afterglow .
a galactic gas - to - dust ratio is at variance with the findings for grbs with detected optical afterglows generally having smc - like gas - to - dust ratios @xcite .
either the pure synchrotron model is not appropriate for the grb 001025a x - ray afterglow or grb 001025a is much more obscured than bursts with detected optical afterglows ( typically having @xmath97 * ? ? ?
* ) . the dark bursts in the grb sample of @xcite ( updated as of july 8 , 2005 , * ?
* ) has a mean 1.610 kev flux of @xmath98 erg @xmath35 s@xmath36 11 hours after the burst .
extrapolating the best - fit x - ray lightcurve of grb 001025a to this epoch , the estimated 1.610 kev flux is @xmath99 erg @xmath35 s@xmath36 .
the x - ray afterglow of grb 001025a is fainter than any of the dark bursts in the @xcite sample and much fainter than the mean of grbs with detected optical afterglows in this sample ( @xmath100 erg @xmath35 s@xmath36 ) .
the low @xmath1-band flux of grb 001025a is not caused solely by a general faintness of the afterglow at all wavelengths .
the @xmath1-band - to-3 kev spectral index is @xmath101 so grb 001025a is clearly dark according to the definition of @xcite . for grbs with detected optical afterglows ,
the @xmath1-band - to-3 kev spectral index is @xmath102 and for _ swift _ detected bursts it is @xmath103 @xcite so the afterglow of grb 001025a is optically faint relative to the x - ray band .
the low optical - to - x - ray spectral index could be due to heavy obscuration in the burst environment .
alternatively , the cooling frequency could be close to the x - ray band , giving rise to a shallow spectral index requiring no optical obscuration .
the x - ray spectral index has been measured only for a few dark grbs . by re - normalizing the x - ray spectrum and the optical observations to a common epoch ( e.g. 11 h after the burst ) and extrapolating the x - ray spectral index to the optical band , assuming a position of the cooling break , the @xmath1-band flux can be predicted . for three of the five grbs with measured spectral index in the dark grb sample of @xcite ( grb 990704 , grb 990806 , grb 000214 )
the predicted @xmath1-band flux is lower than the observational upper limit , when invoking a cooling break at 1 kev .
these three grbs have steep x - ray spectral indices , possibly reflecting the presence of a thermal component on top of the synchrotron emission .
the true synchrotron spectral index may thus be lower than the index obtained from a pure synchrotron emission fit , giving rise to even lower predictions for the optical flux . for the remaining two grbs in the dark grb sample of @xcite ( grb 000210 and grb 001109 ) obscuration
is required in order to bring the observational upper limit in line with the model prediction ( or the derived x - ray spectral index is too step due to the presence of thermal emission ) .
two of the bursts detected by _ swift _ and with published x - ray spectral index ( grb 050401 and grb 050408 , * ? ? ?
* ) are dark according to the definition of @xcite . for both bursts obscuration
is required in order to be consistent with the upper limit on the @xmath1-band flux , even when invoking a cooling break at 1 kev .
grb 001025a was a long - duration grb with a prompt emission spectrum well fitted by the band model .
the tail of the prompt soft x - ray emission decays too steeply to be the beginning of the afterglow .
three epoch _ xmm - newton _ observations show that the grb 001025a afterglow is x - ray faint .
furthermore , the optical flux is relatively low compared to the x - ray band .
either ( i ) the optical afterglow suffers extinction of at least seven magnitudes , or ( ii ) a significant fraction of the x - ray afterglow flux originates from a thermal plasma , and the cooling frequency is close to the x - ray band at 1.2 days after the burst . in the latter case
, we predict that the burst redshift is around two , and we find that the x - ray luminosity and the spectral index are fairly typical of grb afterglows .
alternatively , the grb 001025a x - ray afterglow spectrum is characterized by a very steep synchrotron index , and grb 001025a is situated in an environment with an unusually large gas - to - dust ratio , consistent with the galactic ratio .
the present study of grb 001025a , and the properties of the few other dark grbs with measured spectral index , raises the possibility that some bursts appear optically dark because their afterglow is faint , and their cooling frequency is close to the x - ray band . the viability of this scenario can be tested from early observations of x - ray afterglows from dark grbs yielding the temporal and spectral behaviour of the x - ray afterglows , which in turn constrain the geometry of the blast wave and the underlying physical processes .
this would provide a crucial test for physical differences between dark bursts and bursts with an optical afterglow .
we are grateful to dale frail for making the results of his vla observations known to us .
we thank the anonymous referee for useful and very detailed comments that improved the presentation of our results .
kh is grateful for _ ulysses support under jpl contract 958056 , for support from the long term space astrophysics program under nag5 - 3500 , and for support as a participating scientist in the near mission under nag5 - 9503 .
kp and jpuf acknowledge support from the carlsberg foundation .
kp acknowledges support from instrument center for danish astrophysics .
pj acknowledges support from a special grant from the icelandic research council .
this work was supported by the danish natural science research council ( snf ) . from the russian side
this work was suported by the russian space agency contract and rbrf grant 03 - 02 - 17517 .
das is supported by an nsf astronomy and astrophysics postdoctoral fellowship under award ast-0105221 .
the authors acknowledge benefits from collaboration within the eu fp5 research training network `` gamma - ray bursts : an enigma and a tool '' .
based on observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and nasa , and based on observations made with eso telescopes at the la silla or paranal observatories under programme 66.a.-0386(a ) .
_ 00.256 & 0.26 @xmath112 & + 0.81 @xmath113 0.57 & -1.84 @xmath113 0.12 & 85 @xmath113 33 + 0.2560.768 & 0.37 @xmath114 & -0.39 @xmath113 0.11 & -2.75 @xmath113 0.2 & 140 @xmath113 19 + 0.7688.960 & 0.03 @xmath112 & -0.99 @xmath113 0.3 & -2.40 @xmath113 0.2 & 103 @xmath113 46 + 00.768 & 0.29 @xmath114 & -0.30 @xmath113 0.12 & -2.55 @xmath113 0.16 & 148 @xmath113 20 + 08.960 & 0.036 @xmath115 & -0.91 @xmath113 0.1 & -2.40 @xmath113 0.16 & 191 @xmath113 37 + [ konus_spec_tab ] 2000 oct 26.34 & 1.21 & @xmath1 & 12@xmath27200 & 0.500.75 + 2000 oct 27.32 & 2.19 & @xmath1 & 8@xmath27200 & 0.500.75 + 2001 mar 23.02 & 146.89 & @xmath89 & 3@xmath27200 & 0.450.55 + 2001 mar 23.04 & 146.91 & @xmath1 & 10@xmath27200 & 0.450.55 + [ vltlog ] fixed index , fixed galactic abs . & 2.0 & 0.61 & 0 & 66.8/33 + free galactic abs . &
@xmath116 & @xmath117 & 0 & 32.6/32 + fixed galactic abs , free abs . at z & @xmath118 & @xmath119 & @xmath120 & 31.3/31 + fixed index , free gal .
& 2.0 & @xmath129 & @xmath130 & @xmath131 & 17.9/27 + fixed gal .
& @xmath132 & @xmath133 & 0.61 & @xmath122 & 34.1/29 + free gal .
& @xmath134 & @xmath135 & @xmath136 & @xmath137 & 18.4/28 + fixed gal .
abs , free abs . at z & @xmath138 & @xmath139 & @xmath140 & @xmath141 & 19.6/28 + fixed index , free abs . at z & 2.0 & @xmath142 & @xmath143 & @xmath144 & 19.8/29 | we identify the fading x - ray afterglow of grb 001025a from _ xmm - newton _ observations obtained 1.92.3 days , 2 years , and 2.5 years after the burst .
the non - detection of an optical counterpart to an upper limit of @xmath0 , 1.20 days after the burst , makes grb 001025a a `` dark '' burst . based on the x - ray afterglow spectral properties of grb 001025a
, we argue that some bursts appear optically dark because their afterglow is faint and their cooling frequency is close to the x - ray band .
this interpretation is applicable to several of the few other dark bursts where the x - ray spectral index has been measured .
the x - ray afterglow flux of grb 001025a is an order of magnitude lower than for typical long - duration gamma - ray bursts .
the spectrum of the x - ray afterglow can be fitted with an absorbed synchrotron emission model , an absorbed thermal plasma model , or a combination thereof . for the latter , an extrapolation to optical wavelengths
can be reconciled with the @xmath1-band upper limit on the afterglow , without invoking any optical circumburst absorption , provided the cooling frequency is close to the x - ray band . alternatively ,
if the x - ray afterglow is due to synchrotron emission only , seven magnitudes of extinction in the observed @xmath1-band is required to meet the @xmath1-band upper limit , making grb 001025a much more obscured than bursts with detected optical afterglows . based on the column density of x - ray absorbing circumburst matter
, an smc gas - to - dust ratio is insufficient to produce this amount of extinction .
the x - ray tail of the prompt emission enters a steep temporal decay excluding that the tail of the prompt emission is the onset of the afterglow . to within the astrometric uncertainty ,
this afterglow was coincident with an extended object , seen in a deep vlt @xmath1-band image , which we identify as the likely host galaxy of grb 001025a . |
thanks to astronomical observations the modern cosmology has elaborated the concept of the standard cosmological model called the cold dark matter model with the cosmological constant ( @xmath6cdm model ) . from a methodological point of view cosmology has achieved the similar status as the particle physics with its standard model of particles .
the @xmath6cdm model acts as an effective theory describing the universe from redshift @xmath11 ( today ) to @xmath12 ( the epoch of primordial nuclesynthesis ) . in the very structure of the @xmath6cdm model
there are essentially two components : matter ( dark matter and baryonic matter ) and dark energy .
it is assumed that the universe is spatially homogeneous and isotropic , and that its evolution is governed by einstein field equations with the energy momentum tensor for the ideal fluid satisfying the barotropic equation of state @xmath13 , where @xmath14 is the energy density of fluid . from the observational point of view
it is convenient to use dimensionless density parameters @xmath15 , defined as the fractions of critical density @xmath16 which corresponds to a flat model .
these parameters are observables which can be determined from astronomical observations @xcite . in the @xmath6cdm model
it is assumed that all fluids are non - interacting .
this implies that the densities of baryonic matter and dark matter are scaling with respect to redshift as @xmath17 and dark energy has constant density .
the natural interpretation of the cosmological constant is to treat it as the energy of the quantum vacuum @xcite .
the cosmological model with the cosmological constant term and pressureless matter fits well to the observation data of measurements of snia luminosity distance as a function of redshift @xcite and observations microwave relic radiation ( wmap , planck ) .
measurements of large - scale structures also remain consistent with the @xmath6cdm model .
although the @xmath6cdm model describes well the present universe the nature of its basic constituents ( dark energy and dark energy ) remains unknown .
so dark energy and dark matter are like some useful fictions in the terminology of nancy cartwright @xcite . comparing the value of the cosmological constant required to explain the effect of the accelerated expansion of the universe observations of distant supernovae type snia with the value of the cosmological constant interpreted as the energy of the quantum vacuum
, we get the most incredible gap in history of physics @xmath18 . in this context
, the question is born why the value of the cosmological constant is the small , why not simply a zero ? this is the known problem of the cosmological constant . in the model under the consideration ,
the comparison of @xmath19 still gives 83 order of magnitude from the measurement value because @xmath20 .
another problem related closely with the problem of the cosmological constant is the coincidence problem @xcite .
this problem is caused by the lack of explanation of why in today s era of the density of dark matter and dark energy are comparable although it is assumed that they have different periods of recombination . in this paper
we construct a cosmological model in which it is assumed that the process of interaction between sectors of dark matter and dark energy is continuous .
relativistic diffusion describes the transfer of energy to the sector of dark matter . as a result
, we go beyond the standard model assuming from the outset that dark matter and dark energy interact .
this effect is described by the running cosmological constant and the modification of the standard scaling law of the dark matter density .
if we assume that general relativity is an effective theory which can be extrapolated to the planck epoch , then the interpretation of the cosmological constant parameter appeared in the @xmath6cdm model as a vacuum energy seems to be natural . by equating this density to the energy density of the zero point energy that is left in a volume after removing all particles , then we obtain that its value is about 120 orders of magnitude higher than the corresponding value required for explanation of acceleration of the universe in current epoch . in the universe with such a high value of cosmological constant ( dark energy )
we have a rapid inflation and galaxies would have no time to form .
the lack of explanation of this difference is called the cosmological constant problem .
its solution can be possible if we can find some physical mechanism lowering dramatically this value during the cosmic evolution .
of course this process should be defined in a covariant way following general relativity principles .
our hypothesis is that diffusion cosmology can offer the possibility of obtaining a low value of the cosmological constant today because the effects of diffusion effectively produce the running cosmological constant .
we study how a value of effective running cosmological constant parameter changes during the cosmic evolution and for late time is going to be a small constant value . from the astronomical observations of distant supernovae snia and measurements of cmb by planck , measurement of bao and other astronomical observations we obtain that the present value of the energy densities of both dark energy and dark matter are of the same order of magnitude @xcite .
if we assume that standard cosmological model ( @xmath6cdm model ) is an adequate way to describe the cosmic evolution , then the value @xmath21 will depend on the cosmological time or redshift and the question arises : why are two quantities with different time recombination comparable at the present epoch ?
it is called the cosmic coincidence problem
. we are looking for some physical relativistic mechanism which gives rise to this coincidence observed for the current universe . in the opposite case
very special initial conditions are required for its realization ( fine tunning problem ) . in the framework of diffusion cosmology our investigation of this problem
shows that while the values of dark matter and dark energy densities are comparable today they were significantly different in the past history of the universe . because the diffusion effects act effectively as fluids which interact each other during the cosmic evolution . in the consequence dark energy is running and a canonical rule of scaling dark matter @xmath22 proportional to @xmath23 is adjusted .
the main aim of our paper is to demonstrate how the coincidence problem can be naturally solved in the framework of diffusion cosmology .
the interacting dark energy models have been considered by many authors in the context of this problem .
one of reasons to study these models is to solve the cosmic coincidence problem @xcite . to this aim
different ad - hoc proposed models of an interacting term were postulated prori . in these models the covariance of general relativity is usually violated and therefore they have limited application to cosmology . in the present work we consider a unique relativistic diffusion model where an interaction mechanism is motivated physically . in the study of evolutional scenarios of the model under consideration we apply the dynamical systems methods @xcite .
our model belongs to a general class of jungle type of cosmologies represented by coupled cosmological models in a lotka - volterra framework @xcite .
the crucial role in the organization of the phase space plays the critical point located inside the physical region .
the possible bifurcation of this point are studied in details for extracting variability of de and dm density as the function of the cosmological time .
it is interesting that at this critical point @xmath24 ( scaling type solution ) .
haba et al . postulated a particular model of an energy - momentum exchange between dm and de sectors , while a baryonic sector was preserved @xcite . in this paper
we reconsider this model in the light of aforementioned cosmological problems .
we assume einstein equations in the form @xmath25 where @xmath26 is the metric , @xmath27 is the ricci tensor . in this paper
we use the natural units @xmath28 .
because of a cosmological application we assume that the universe has topology @xmath29 , where @xmath30 is homogeneous and isotropic space .
then the spacetime metric depends only on one function of the cosmic time @xmath31the scale factor @xmath32 .
additionally for simplicity we also assume flatness ( @xmath33 ) of sections @xmath34 .
we decompose the energy - momentum tensor on two parts @xmath35 we assume the conservation of the total energy momentum , which gives @xmath36 where @xmath37 is the diffusion constant and @xmath38 is the current , which represents a flow of stream of particles .
we also assume that energy density of the dark matter consisting of particles of mass @xmath39 is transfered by a diffusion mechanism in an environment described by a perfect fluid .
there is only unique diffusion which is relativistic invariant and preserves the particle - mass @xmath39 @xcite . the corresponding energy - momentum satisfies the conservation law ( [ con ] ) . the friedman equation in the frw metric with baryonic matter , dark matter and dark energy reads @xmath40 where @xmath41 and @xmath42 are determined by relations @xmath43 the current @xmath44 in equation ( [ con ] ) is conserved @xcite @xmath45 the above conservation condition for the frw metric reduces to @xmath46 with a positive constant @xmath47 which can be computed from the phase space distribution @xmath48 of diffusing particles @xcite .
the condition ( [ con ] ) after calculation of divergence reduces to the continuity conditions for energy density of both matter and dark energy @xmath49 where we assume the equation of state for dark energy as @xmath50 and for matter as @xmath51 ; a dot denotes differentiation with respect to the cosmological time @xmath31 .
equation ( [ darkmatter ] ) can be rewritten as @xmath52 where @xmath53 is the total energy of matter in the comoving volume @xmath54 . from relation ( [ energy ] ) , we can obtain that @xmath55 .
in our paper @xcite we considered one unique model of an energy transfer from dark energy ( de ) to dark matter ( dm ) with the diffusive interaction in the dark sector where de and dm can be treated as ideal fluids .
particles are scattering in an environment of other particles .
if we assume that the subsequent scattering events are independent , the particle motion is described by a markov process . in order , the assumption that the energy of the particle remains finite leads to the conclusion that the markov process must be a diffusion .
therefore diffusion is in some sense unique because there is only one diffusion which is relativistic invariant and preserves the mass @xcite . in consequence
the interaction between dm and de fluids is defined in a unique way .
in the investigation , the dark energy and dark matter interaction plays role in continuity equation .
this equation are special case of jungle cosmological models @xcite .
we assume that @xmath56 , where @xmath57 is density of baryonic matter and @xmath41 is density of dark matter .
the equation of state for dark energy is expressed by @xmath50 in our model , where @xmath58 is pressure of dark energy and the equation of state for matter is given by @xmath51 , where @xmath59 is pressure of matter .
the friedmann equation is expressed in the following form @xmath60 where @xmath61 , @xmath62 , @xmath63 . from the friedmann formula
we get a condition @xmath64 where @xmath65 and @xmath66 are dimensionless density parameters .
we can obtain equations ( [ darkmatter])-([friedmann ] ) in the form of the dynamical system @xmath67 , @xmath68 and @xmath69 , where @xmath70 , @xmath71 , @xmath72 and @xmath73 is a differentiation with respect to the reparametrized time @xmath74 .
for these variables , the dynamical system is in the following form @xmath75 from equation ( [ condition ] ) , we have the following relation @xmath76 then dynamical system ( [ dyn1])-([dyn3 ] ) is reduced to the two - dimension dynamical system . for analysis of the critical points in the infinity , we use the poincar sphere .
let us introduce new variables in which one can study dynamical behavior at infinity @xmath77 for these variables we get the dynamical system @xmath78,\label{poincare1 } \\
\delta'&=\delta\left[(1-\delta^2)(\frac{3}{2}x-\delta)-x^2 ( 3x+\delta-3\sqrt{1-x^2-\delta^2})\right],\label{poincare2}\end{aligned}\ ] ] where @xmath79 .
critical points , for equation ( [ poincare1])-([poincare2 ] ) , are presented in table [ table:1 ] . the phase portrait for the dynamical system ( [ poincare1])-([poincare2 ] )
is presented in figure [ fig : fig11 ] . in the phase portrait
there is an interesting class of trajectories labeled as `` i '' , starting from the critical point 6 and approaching the de sitter state .
because the diffusion has the physical sense only for interval @xmath80 the corresponding cosmological solution should be cut off for any @xmath81 from the de - sitter solution .
hence , we obtain that @xmath82 is positive number , i.e. solution which represent critical point ( 6 ) is non singular .
all trajectories starting from the de - sitter state can be treated as a models of extended idea of emergent cosmology @xcite @xcite .
while in the standard emergent cosmology universe is starting from the static einstein model , trajectories of type i are simple realization of extended idea of emergent universe in which universe is starting rather from the stationary state . the results of our previous paper @xcite show that the density parameter for total ( dark and visible ) @xmath83 and dimensionless parameter @xmath3 are constrained to @xmath84 , @xmath85 at the @xmath86 confidence level . this domain is represented in the phase space by a shaded rectangle . only these trajectories which intersect this domain are in good agreement with the observation at a 2@xmath10 confidence level .
therefore observation favored the cosmological models starting from the einstein - de sitter solution and going toward the de sitter attractor ( trajectory ii on the phase portrait ) .
note that on the phase portrait there are trajectories labeled as `` i '' starting from the de sitter state and approaching the de sitter state at late times .
they are going toward a saddle point representing a non - singular solution .
however all these trajectories do not intersect the rectangle and therefore they are not favored by observation .
the saddle point in the phase space is representing the milne universe ( see table [ table:1 ] ) .
therefore , the interacting term is proportional to @xmath87 and consequently is of the form @xmath88 .
the cosmological model with such a parametrization of dark energy was studied by szydlowski and stachowski @xcite .
figures [ fig : fig2a ] and [ fig : fig2b ] present the evolution of dark matter @xmath41 as a function of the cosmological time @xmath31 for trajectories of type ii .
the evolution of the cosmological time for matter is determined by the following formula @xmath89 the addictive form of the scaling relation for dark matter ( [ matter ] ) suggests that dark matter consists of two components : the first term scaling like @xmath90 and the seconf term scaling like @xmath91 .
the latter describes an amount of dark energy density which is transfered to dark energy sector by the diffusion process . in the unit of @xmath92
the canonically scaling dark matter is @xmath93 while transfered dark energy is about @xmath94 .
the amount of transfered dark energy today is of the order @xmath95 where @xmath96 is the age of the universe .
in consequence of equation ( [ matter ] ) , @xmath97 can be rewritten as @xmath98 therefore at the present epoch we have @xmath99 where @xmath100 and @xmath101 is the present age of the universe . note that while at late time @xmath102 for small time @xmath103 . if we give @xmath104 in equation ( [ matter ] )
then @xmath22 is scaling in the canonical way . from equation ( [ matter ] )
one can simply obtain that the density of dark matter is @xmath105 note that the interval of the values of @xmath41 is @xmath106 or @xmath107 , which depends on a type of trajectories .
the evolution of the scale factor @xmath32 with respect to the cosmological time , for trajectories of type ii , is demonstrated in figure [ fig : fig3 ] .
the function @xmath97 , for trajectories of type ii , is presented in figure [ fig : fig4 ] and the consideration for the maximum of this function is in the form @xmath108 where @xmath109 is corresponding the value of the cosmological time at the maximum .
the hubble function @xmath110 , for trajectories of type ii , is presented in figure [ fig : fig5 ] .
note that the hubble function in the late time is constant .
the evolution of @xmath8 , for trajectories of type ii , is shown in figure [ fig : fig10 ] and for late time @xmath8 is going toward constant value .
the evolution of @xmath111 , for trajectories of type ii , is demonstrated in figures [ fig : fig8 ] and [ fig : fig9 ] .
for the trajectories of type i , the functions @xmath110 , @xmath32 , @xmath41 , @xmath8 , @xmath111 and @xmath97 are presented in figures [ fig : fig6 ] , [ fig : fig7 ] , [ fig : fig12 ] , [ fig : fig13 ] , [ fig : fig14 ] , [ fig : fig15 ] .
these figures show that there are two distinct behavior of trajectories of type i and ii . while the trajectory of type ii represents matter dominating model with a singularity the trajectories of type i represents the model without an initial singularity .
lllllll & & & & & & + & & & & & & + 1 & & & & & & + & & & & & & + 2 & & & & & & + & & & & & & + 3 & & & & & & + & & & & & & + 4 & & & & & & + & & & & & & + 5 & & & & & & + & & & & & & + 6 & & & & & & + & & & & & & + 7 & & & & & & + & & & & & & + + , @xmath112 , where @xmath113 , @xmath114 and @xmath115 on the poincar sphere coordinates are @xmath116 , @xmath117 .
critical point ( 1 ) represents the de sitter universe a global attractor for all physical trajectories .
critical point ( 2 ) represents the scaling universe .
critical point ( 3 ) represents the einstein - de sitter universe .
critical points ( 4 ) and ( 5 ) represent the static universe .
critical point ( 6 ) represents the de sitter universe .
the gray region represents the domain of the present value of @xmath118 and @xmath119 , which is distinguished by astronomical data .
let us note trajectories lie in the domain with @xmath120 represent the contracting model but there is no symmetry with respect to the @xmath119-axis .
at critical point ( 6 ) , energy density of baryonic matter is negligible as well as density of dark matter and only effects of the relativistic diffusion are important . ] ) .
dark matter @xmath41 is expressed in [ 100@xmath121km/(s mpc)]@xmath122 .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] for the present epoch ) .
dark matter @xmath41 is expressed in [ 100@xmath121km/(s mpc)]@xmath122 .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 .
the value of the age of the universe for the best fit with errors are presented by the dashed lines . ]
function of cosmological time @xmath31 for trajectories of type ii ( for the best fitted values of model parameter together with confidence level at @xmath86 ) . for the present epoch
@xmath96 @xmath123 .
a universe is starting from the initial singularity toward a de sitter universe .
this type of behavior is favored by the observational data .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] of cosmological time @xmath31 for trajectories of type ii ( for the best fitted values of model parameter together with confidence level at @xmath86 ) .
note that as trajectory in the phase space achieved the state of the pericentrum located in the saddle point , this state is corresponding on the diagram the maximum .
note that the existence of a maximum value of @xmath3 parameter @xmath124 . for the late time @xmath97 function is decreasing function of @xmath31 and @xmath125=0 .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] ) . for late times
@xmath110 goes to constant values ( des@xmath126 ) .
the hubble function @xmath110 is expressed in [ 100@xmath121km/(s mpc ) ] .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] for trajectories of type ii ( for the best fitted values of model parameter together with confidence level at @xmath86 ) .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] for trajectories of type ii for the present epoch .
note that at the present epoch @xmath127 ( therefore coincidence problem is solved ) .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 .
the value of the age of the universe for the best fit with errors are presented by the dashed lines . ]
is expressed in [ 100@xmath121km/(s mpc)]@xmath122 .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] for typical trajectory of type i. the @xmath110 function is expressed in [ 100@xmath121km/(s mpc ) ] .
note that @xmath128 is finite therefore it is not a singularity .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 .
] for typical trajectory of type i. note that @xmath129 is finite therefore it is not a singularity .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] for typical trajectory of type i. note that @xmath130 is equal zero therefore it is not a singularity .
matter @xmath7 is expressed in [ 100@xmath121km/(s mpc)]@xmath122 .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] for typical trajectory of type i. note that @xmath131 is finite therefore it is not a singularity .
dark energy @xmath8 is expressed in [ 100@xmath121km/(s mpc)]@xmath122 .
we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] for typical trajectory of type i. we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ] for typical trajectory of type i. we choose ( s mpc)/(100@xmath121 km ) as a unit of the cosmological time @xmath31 . ]
dynamical system methods are especially suitable in investigation of dynamics of both fluids , dark energy and dark matter . presented here dynamical system approach to the study dm - de interaction in diffusion cosmology
can be simply generalized to the case when both dark energy and dark matter satisfy a general form of the equation of state @xmath132 where @xmath133 and @xmath134 are constant coefficients equation of state for dark energy and matter respectively . then the continuity equations for baryonic and dark matter and dark energy are presented by @xmath135 the corresponding dynamical system assumes the form of a 3-dimensional autonomous dynamical system @xmath136,\\ \frac{dy}{d\ln a}&=3y[(1+w)(y-1)+(1+\tilde{w})x]-xz,\\ \frac{dz}{d\ln a}&=z\left[3\tilde{w}-z+\frac{3}{2}[(1+\tilde{w})x+(1+w)y]\right],\end{aligned}\ ] ] where we choose state variables @xmath137 , and @xmath71 and @xmath138 like in previous considered case
. because @xmath139 the above dynamical system reduces to @xmath140,\label{dyn4}\\ \frac{dz}{d\ln a}&=z\left[3\tilde{w}-z+\frac{3}{2}[(1+\tilde{w})x+(1+w)(1-x)]\right].\label{dyn5}\end{aligned}\ ] ] critical points of dynamical system ( [ dyn4])-([dyn5 ] ) are completed in table [ table:2 ] . especially there is an interesting critical point inside the admissible region @xmath141 representing scaling solution : @xmath142 .
it is a saddle fixed point in the phase space @xmath143 .
this critical point is important in the context of solution of cosmic coincidence problem as well as scaling solution in the context of quintessence idea .
& critical point & type of the universe + 1 & @xmath144 , @xmath145 & de sitter universe + & & without diffusion + 2 & @xmath146 , @xmath145 & einstein - de sitter + 3 & @xmath147 , @xmath148 & scaling universe @xmath149 + 4 & @xmath144 , @xmath150 & de sitter universe + & & with diffusion + + above system possesses critical points on the planes of coordinate system or inside the phase space @xmath151 .
of course system under consideration is restricted to the submanifold @xmath152 , because the constrain condition @xmath153 .
the behavior of trajectories of dynamical system ( [ dyn4])-([dyn5 ] ) depends on the values of parameters @xmath133 , @xmath134 . by choosing different values of these parameters
one can study how phase space structure changes under change of values of parameters .
the equivalence of the phase portraits is establish following homeomorphism preserving direction of time along the trajectories . if there exist value of parameter for which phase is not topologically equivalent , then such value is bifurcation value .
the stability of critical points depends on the linearization matrix . at the critical point ( 3 ) , the linearization matrix has the following form @xmath154 where @xmath155 and @xmath156 are the right sides of equations ( [ dyn4 ] ) and ( [ dyn5 ] ) and @xmath157 and @xmath158 are coordinates of critical point ( 3 ) ( see table [ table:2 ] ) .
the determinant of matrix ( [ matrix ] ) can be expressed by the formula @xmath159 and the trace of matrix ( [ matrix ] ) is described by @xmath160 therefore the critical point ( 3 ) is stable when @xmath161 .
the characteristic equation for matrix @xmath162 at critical point ( 3 ) is in the following form @xmath163 from the characteristic equation ( [ chara ] ) , we can obtain the eigenvalues for critical point ( 3 ) ( see table [ table:2 ] ) . in figure
[ fig : fig16 ] we demonstrate the stability of critical point ( 3 ) , depending on @xmath133 and @xmath134 . and @xmath134 .
in the gray domains there is the focus type of critical point and the boundaries of this domains is given by the lines @xmath164 and @xmath165 . in the blue regions
there are the saddle type of critical point and is limited by lines @xmath166 and @xmath167
. in the white top and bottom regions there are the stable and unstable nodes , respectively . ] the linearized equation ( [ dyn4])-([dyn5 ] ) at critical point ( 3 ) is given by the following formulas @xmath168 @xmath169 where @xmath147 and @xmath148 .
the solutions of the above equations are presented by formulas @xmath170 @xmath171 where @xmath172 .
it is interesting to check how structure of phase space changes under changing coefficient equation of state for dark matter from 0 ( cold dark matter ) to @xmath173 ( hot dark matter ) .
results of dynamical investigation show that structure of the phase space is preserved under changes of the model parameter .
let us considered some details .
in this section we consider the case with relativistic dark matter ( @xmath173 ) and @xmath174 .
then the equation of state for dark matter is in form @xmath175 , where @xmath176 is the pressure of dark matter .
we get the following equations @xmath177
@xmath178 we can analyze the critical points in the infinity . in this case
we use the poincar sphere .
let @xmath179 , @xmath180 . for variables @xmath118 and @xmath119 , we get dynamical system @xmath181,\label{poincare3}\ ] ] @xmath182,\label{poincare4}\ ] ] where @xmath79 .
critical points , for the above equation , are presented in table [ table:3 ] . the phase portrait for the dynamical system ( [ poincare3])-([poincare4 ] )
are demonstrated in figure [ fig : fig1 ] .
& critical point & type of critical point & type of universe + 1 & @xmath183 , @xmath184 & saddle & de sitter universe without diffusion effect + 2 & @xmath185 , @xmath186 & saddle & scaling universe + 3 & @xmath187 , @xmath184 & unstable node & einstein - de sitter universe + 4 & @xmath188 , @xmath184 & stable node & static universe + 5 & @xmath187 , @xmath189 & saddle & static universe + 6 & @xmath183 , @xmath190 & unstable node & de sitter universe with diffusion effect + 7 & @xmath183 , @xmath191 & stable node & de sitter universe with diffusion effect + 8 & @xmath183 , @xmath192 & stable node & de sitter universe with diffusion effect + + ) -([dyn5 ] ) .
note that trajectories for the @xmath120 represent solutions with the negative value of @xmath193 . from the cosmological point of view trajectories representing expanding models with @xmath194
are physical .
critical points ( 1 ) represents the de sitter universe without diffusion effect .
critical point ( 2 ) represents the scaling universe .
critical point ( 3 ) represents the einstein - de sitter universe .
critical points ( 4 ) and ( 5 ) represent the static universe .
critical point ( 6 ) and ( 8) represent the de sitter universe with diffusion effect . ]
it is interesting to see how different solutions with and without an initial singularity are distributed in the phase space . to answer this question it would be useful to consider a phase space structure of the models under consideration . for this aim
we reduce the dynamics to the form of an autonomous 2d dynamical system . in such a system
the state variables are dimensionless parameters : the density parameter for radiation dark matter and the parameter @xmath3 characterizing the rate of energy transfer to the dark matter sector .
the main advantage of visualization global dynamics on the phase portrait is the possibility to see all solution of the system admitted for all initial conditions . on the phase portrait
there is the geometric representation of evolutional paths of both solution types .
critical points are representing asymptotic states of the system , i.e. stationary states . in order trajectories joining different critical points are representing the evolution of the system . similarly to dynamical investigations
presented in our previous paper @xcite we added to the plane circle at infinity via the construction of the poincar sphere .
hence we obtain a compact phase space and consequently a global phase portrait . in fig .
[ fig : fig1 ] we have identified linear solutions without the initial singularity as the representing by saddle critical point ( 2 ) . in the phase portrait there are critical points at a finite domain as well as located on the boundary at infinity .
note that the phase portrait has no symmetry with respect to the x - axis .
critical point ( 6 ) is representing an expanding stationary de sitter type solution determined by diffusion effects .
we denote as typical trajectories starting from this critical point and going toward the de sitter empty universe . this type of trajectories we called trajectories of type i. in the phase portrait there are also present trajectories of type ii .
these trajectories are starting from the einstein - de sitter universe with the initial singularity and coming toward the de sitter universe labeled as critical point ( 8) .
looking at the phase portrait one can observe that only critical points of type unstable and stable node ( global attractors and global repellers ) and saddle appear in the phase space .
therefore the model obtained is structurally stable , i.e. any small change of its r.h.s does not disturb the global phase portrait .
physically this means that corresponding model is realistic .
mathematically this fact has a nice interpretation in the context of the peixoto theorem @xcite that they are generic in this sense because they form open and dense subsets in the space of dynamical systems on the plane .
the standard cosmological model ( @xmath6cdm model ) is widely accepted but it has still some problems , namely the cosmological constant problem and the coincidence problem . in the standard cosmological model ( @xmath6cdm model ) it is assumed that all fluids are non - interacting .
this implies that the densities of baryonic matter and dark matter are scaling with respect to redshift as @xmath195 and dark energy has constant density . in this paper
we construct a cosmological model in which it is assumed that the process of interaction between sectors of dark matter and dark energy is continuous .
relativistic diffusion describes the transfer of energy to the sector of dark matter .
this effect is described by the running cosmological constant and the modification of the standard scaling law of the dark matter density to the form @xmath196 .
the dynamics of this model is studied for possible explanations of cosmological puzzles : the cosmological conatant problem and the coincidence problem . in the context of the coincidence problem
, our model can explain the present ratio of @xmath7 to @xmath8 , which is equal @xmath9 at a 2@xmath10 confidence level . in our model , the canonical scaling law of dark matter @xmath0
is modified by an additive @xmath1 to the form @xmath2 .
the analysis of the time dependence of density of dark energy and dark matter , we conclude that the value of effective energy of vacuum runs from an infinite value to a constant value , and the delta amendment to the scaling law goes from zero to zero and being different from zero in a long intermediate period
. this characteristic type of behavior is controlled by the diffusion effect .
the paper presents a detailed study of the behavior of a state of the system represented by the state variables @xmath197 . in this context , it was natural to consider the diffusion mechanism which controls the change of the ratio of both energy densities and the very dynamics of this process remains in analogy to the description of population changes of competing species @xcite .
a crucial role plays the saddle critical point @xmath198 , which is a scaling type of the solution ( @xmath199 ) .
the position of this point can not be disturbed by a small perturbation ( structurally stable point as well as whole system ) .
24ifxundefined [ 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/s10714 - 014 - 1753 - 8 [ * * , ( ) ] , link:\doibase 10.1086/185100 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1103/revmodphys.61.1 [ * * , ( ) ] @noop * * , ( ) , @noop _ _ ( , , ) link:\doibase 10.1140/epjc / s10052 - 014 - 3160 - 4 [ * * , ( ) ] , link:\doibase 10.1088/1475 - 7516/2005/03/002 [ * * , ( ) ] , link:\doibase 10.1103/physrevd.77.123504 [ * * , ( ) ] , link:\doibase 10.1142/s021827180600942x [ * * , ( ) ] , link:\doibase 10.1016/j.physletb.2005.08.134 [ * * , ( ) ] , in @noop _ _ , ( ) pp
. @noop ( ) , @noop * * , ( ) link:\doibase 10.1142/s0217732310033992 [ * * , ( ) ] , link:\doibase 10.1088/1475 - 7516/2011/11/016 [ * * , ( ) ] , link:\doibase 10.1016/j.geomphys.2012.06.004 [ * * , ( ) ] , link:\doibase 10.1002/cpa.20140 [ * * , ( ) ] link:\doibase 10.1103/physreve.79.021128 [ * * , ( ) ] , link:\doibase 10.1007/s10714 - 015 - 1852 - 1 [ * * , ( ) ] , link:\doibase 10.1088/1475 - 7516/2014/07/034 [ * * , ( ) ] , link:\doibase 10.1103/physrevd.91.123538 [ * * , ( ) ] , link:\doibase 10.1088/1475 - 7516/2015/10/066 [ * * , ( ) ] , link:\doibase 10.1007/978 - 1 - 4613 - 0003 - 8 [ _ _ ] , ed .
, , vol .
( , , ) | we study dynamics of cosmological models with diffusion effects modeling dark matter and dark energy interactions .
we show the simple model with diffusion between the cosmological constant sector and dark matter , where the canonical scaling law of dark matter @xmath0 is modified by an additive @xmath1 to the form @xmath2 .
we reduced this model to the autonomous dynamical system and investigate it using dynamical system methods .
this system possesses a two - dimensional invariant submanifold on which the dm - de interaction can be analyzed on the phase plane .
the state variables are density parameter for matter ( dark and visible ) and parameter @xmath3 characterizing the rate of growth of energy transfer between the dark sectors .
a corresponding dynamical system belongs to a general class of jungle type of cosmologies represented by coupled cosmological models in a lotka - volterra framework .
we demonstrate that the de sitter solution is a global attractor for all trajectories in the phase space and there are two repellers : the einstein - de sitter universe and the de sitter universe state dominating by the diffusion effects .
we distinguish in the phase space trajectories , which become in good agreement with the data .
they should intersect a rectangle with sides of @xmath4 $ ] , @xmath5 $ ] at the 95% cl .
our model could solve some of the puzzles of the @xmath6cdm model , such as the coincidence and fine - tuning problems . in the context of the coincidence problem
, our model can explain the present ratio of @xmath7 to @xmath8 , which is equal @xmath9 at a 2@xmath10 confidence level . |
the interaction of matter with magnetic fields has long been studied for the important consequences it has for particle dynamics . whether it is from the astrophysical implication of a magnetic field on the structure of a neutron star or to the zeeman shift of energy levels within an atom
, the presence of an external magnetic field can open the door to interesting effects not previously seen .
the effect of an homogeneous background field on the charged fermion propagator in qed was first studied by schwinger @xcite .
proper - time method provides the machinery for an exact solution to the altered propagator .
recently , a weak field expansion of this method has been applied to demonstrate that there exists enhancements to neutrino - photon interactions that involve charged fermion loops @xcite .
the goal of this article is to present what effects a magnetic field has on the decays of the @xmath2 which are otherwise forbidden .
the landau - yang theorem uses bose symmetry and rotational invariance arguments to prove that a vector particle can not decay into two massless vector particles @xcite .
that is , the decays @xmath3 and @xmath0 are not allowed .
however we can employ the field expansion of schwinger s method to calculate if and to what degree the magnetic field stimulates the decay .
we expand just to first order in the magnetic field @xmath4 .
this turns out to be a very good approximation considering that the relevant expansion parameter is @xmath5 , where @xmath6 is the mass of the fermion and @xmath7 is its charge . for the lightest of the possible fermions , the electron
, the critical field turns out to be of the order @xmath8 .
this field is a great deal larger than those magnetic fields produced by very dense white dwarfs , and near the neutron star magnetic field intensity .
the calculation is complete to first order , but because of this expansion parameter it is shown that the dominant contributions are from the lightest quarks ( @xmath9 ) for @xmath0 and from the electron in the @xmath10 case .
first we briefly introduce the reader to the methodology used in determining the changes in the fermion propagator .
next , the result of that technique is presented .
this result is then applied to the decay mode of the @xmath2 involving two gluons .
the results of this calculation are presented and contrasted with results from the similar decay into two photons .
schwinger s method uses an integration over a proper - time variable to solve for the green s function of the inhomogenous differential equation for a dirac field in the presence of the electromagnetic gauge field @xmath11 @xmath12 @xmath13 is considered the matrix element of an operator @xmath14 , such that @xmath15 . therefore eq . ( [ eqn : dirac ] ) is written as an operator equation with its indices suppressed @xmath16 or @xmath17}. \end{aligned}\ ] ] @xmath18 is the proper - time variable , and @xmath19 is a generalized momentum operator defined by @xmath20 this generalized momentum operator is characterized by the following commutation properties [ eqn : commutators ] @xmath21&=&ig_{\mu\nu}\\\label{eqn : com1 } [ \pi_{\mu},\pi_{\nu}]&=&-ieqf_{\mu\nu},\label{eqn : com2 } \end{aligned}\ ] ] where @xmath22 is the electromagnetic field - strength tensor . the quantity @xmath23 $ ] is regarded as an operator that evolves the system in accordance with the `` hamiltonian '' @xmath24 .
that is , the matrix element of the operator @xmath25 $ ] is the time development function that takes the state @xmath26 to the state @xmath27
@xmath28 this view of the transformation as an evolution of the state governed by a hamiltonian parallels the idea that the fermion moves from one space - time location to another in a way that is determined by this hamiltonian . accordingly , there exist `` equations of motion '' satisfied by the generalized position and momentum operators [ eqn : eqnofmot ] @xmath29=2\pi_{\mu}\\\label{eqn : eqnofmota } \frac{{\mathrm{d}}\pi_{\mu}}{{\mathrm{d}}s } & = & -i[\pi_{\mu},{\mathcal{h}}]=2qef_{\mu\nu}\pi^{\nu}.\label{eqn : eqnofmotb } \end{aligned}\ ] ] the relationship in eq .
( [ eqn : eqnofmotb ] ) only holds when @xmath22 is a constant .
this is precisely the case in which we are concerned . for calculation purposes the magnetic field
is considered to be constant and directed along the _
z_-axis ( @xmath30 , and @xmath31 for @xmath32 ) .
in addition there are three differential equations that describe the transformation function [ eqn : trandiffeq ] @xmath33 and a boundary condition @xmath34 using eqs .
( [ eqn : eqnofmot ] ) , ( [ eqn : trandiffeq ] ) , and ( [ eqn : boundcond ] ) one can solve for the explicit form of the green s function . for the purposes of this article , only the result will be given .
choosing a straight line to connect the space - time points @xmath35 and @xmath36 , the green s function is @xmath37 with @xmath38},\ ] ] and @xmath39}\nonumber\\ & & \times\biggl[e^{ieqbs\sigma_{3 } } \frac{\bigl(m + \not\!p_{\parallel}\bigr)}{\cos(eqbs ) } -\frac{\not\!p_{\perp}}{\cos^{2}(eqbs)}\biggr ] .
\end{aligned}\ ] ] the quantity @xmath40 indicates @xmath41 , while @xmath42 . similarly , @xmath43 and @xmath44 . by expanding eq .
( [ eqn : greenfunctmoment ] ) to first order in the magnetic field the integration simplifies , and our result is @xmath45 by making use of the fact that the magnetic field is homogeneous , the result in eq .
( [ eqn : gfm1o ] ) was cast into covariant notation .
at this point the general results from the previous section are used to calculate the decay rate of the neutral @xmath2 boson into to two gluons .
to lowest order in the perturbation theory there are two feynman diagrams which contribute to this process ; see fig . [
fig : zggs ] .
let us first turn our attention to the leftmost contribution of fig .
[ fig : zggs ] . with momentum @xmath47 into two gluons with momenta @xmath48 and @xmath49.,title="fig:",width=151 ] with momentum @xmath47 into two gluons with momenta @xmath48 and @xmath49.,title="fig:",width=151 ]
the _ t - matrix _ for the first of our diagrams is written as @xmath50 \nonumber\\ & & \times\epsilon_{\mu}^{*a}(k_{1})\epsilon_{\nu}^{*b}(k_{2 } ) \epsilon_{\sigma}(k)\exp{[-i(kz - k_{1}x - k_{2}y)]}. \end{aligned}\ ] ] the sum over @xmath51 indicates the sum over all possible fermions that contribute to this loop diagram .
the sums over @xmath52 and @xmath53 indicate the sums over the possible states of the gluons .
however , one of these sums is eliminated by virtue of the trace over the two su(3 ) generators @xmath54 . from this point the summation over color will be represented by the repeated color index @xmath52 .
the couplings @xmath55 and @xmath56 are the vector and axial vector couplings , respectively ; both are dependent on the particular fermion that couples to the @xmath2 boson .
@xmath57 is the strong coupling constant evaluated at @xmath58 . because there is the combination of the three green s functions in eq .
( [ eqn : ta ] ) , there exists a product of three phases of the same form as in eq .
( [ eqn : phase ] ) .
as indicated previously , the integration from two space - time points is along a straight line .
a convenient choice of gauge for this integration is one in which @xmath59 generally , the zeroth and third components of @xmath11 can be any constant but the above is chosen for its simplicity . with the choice of gauge found in eq .
( [ eqn : gauge ] ) the total phase factor of the diagram is written @xmath60}\nonumber\\ & \approx&1+\!\frac{ieq}{2}(z - y)^{\mu } f_{\mu\nu}(z - x)^{\nu}. \end{aligned}\ ] ] the fact that the magnetic field is constant and along the _
z_-axis has been exploited .
when the first order corrections to the green s function found in eqs .
( [ eqn : gfm1o ] ) and ( [ eqn : totphase ] ) are substituted into eq .
( [ eqn : ta ] ) , the result is thought of as the sum of five separate contributions @xmath61 there is the zeroth order contribution to the diagram @xmath62 which is completely independent of the magnetic field .
there is a first order correction to each of the legs of the fermion loop @xmath63 , @xmath64 and @xmath65 . and lastly
, there is a first order correction to the overall phase of the diagram @xmath66 . because the invariant amplitude of a given diagram @xmath67 is defined through the relationship @xmath68 , it is straightforward to show that the zeroth order contribution from the first diagram is @xmath69 .
\end{aligned}\ ] ] for the first order correction to one of the legs of the loop , the contribution to the amplitude is @xmath70 .
\end{aligned}\ ] ] the corrections to the remaining legs of the fermion loop are similar to eq .
( [ eqn : mapr1 ] ) .
however , for the contribution due to the phase it is necessary we recognize that @xmath71 , \end{aligned}\ ] ] where @xmath72 is the momentum carried by the fermion as it moves from point @xmath73 to point @xmath74 , and @xmath75 is its momentum from point @xmath35 to @xmath73 .
the relationship in eq .
( [ eqn : tphpro ] ) is rewritten in terms of derivatives with respect to the fermion momenta @xmath76 .
\end{aligned}\ ] ] by making use of integration by parts and the identity @xmath77 the first order contribution of the correction to the phase to the amplitude is @xmath78 .
\end{aligned}\ ] ] we follow the same procedure for the second diagram of fig .
[ fig : zggs ] as we did for the first .
the results obtained are similar to those found in eqs .
( [ eqn : ma0 ] ) , ( [ eqn : mapr1 ] ) , and ( [ eqn : maph ] ) with only the labels @xmath79 interchanged .
the sum of the zeroth order contributions from the two diagrams vanishes .
this is consistent with the fact that the process @xmath0 is forbidden by the landau - yang theorem in the absence of a magnetic field @xcite .
the use of the charge conjugation operator simplifies the contributions that are first order in the magnetic field .
we find that the axial vector parts of the two diagrams cancel while the vector parts are identical and add .
tracing over the dirac indices and integrating over the loop momentum results in an amplitude that is proportional to forty - four separate tensor structures formed by possible contractions of the polarization vectors , the field - strength tensor , and the gluon momenta .
since the amplitude is symmetric under interchange of the labels @xmath79 , we find that the tensor structures can be grouped into twenty - two pairs ; each pair having its own unitless coefficient which depends on the possible fermion masses , charges , and couplings in the charge loop . limiting the gluons to be only transversely
polarized allows us to drop all terms proportional to @xmath80 .
this cuts the number of tensor structures in half to eleven .
thus , the amplitude is written as @xmath81(1/m_{z } ) \nonumber \\ & & + n_{3}\bigl[(\epsilon_{1}^{*a}\cdot\epsilon)(\epsilon_{2}^{*a}f k_{2})+(\epsilon_{2}^{*a}\cdot\epsilon)(\epsilon_{1}^{*a}f k_{1})\bigr](1/m_{z } ) \nonumber \\ & & + n_{4}(\epsilon_{1}^{*a}\cdot\epsilon_{2}^{*a } ) \bigl(\epsilon f ( k_{1}+k_{2})\bigr)(1/m_{z } ) \nonumber \\ & & + n_{5}(\epsilon_{1}^{*a}\cdot\epsilon_{2}^{*a } ) \bigl(\epsilon\cdot(k_{1}-k_{2})\bigr)(k_{1 } f k_{2 } ) ( 1/m_{z})^{3 } \nonumber \\ & & + n_{6}(\epsilon_{1}^{*a}\cdot k_{2})(\epsilon_{2}^{*a}\cdot k_{1 } ) \bigl(\epsilon\cdot(k_{1}-k_{2})\bigr)(k_{1 } f k_{2 } ) ( 1/m_{z})^{5 } \nonumber \\ & & + n_{7}\bigl[(\epsilon_{1}^{*a}\cdot\epsilon ) ( \epsilon_{2}^{*a}\cdot k_{1})-(\epsilon_{2}^{*a}\cdot\epsilon ) ( \epsilon_{1}^{*a}\cdot k_{2})\bigr](k_{1 } f k_{2 } ) ( 1/m_{z})^{3 } \nonumber \\ & & + n_{8}\bigl[(\epsilon_{2}^{*a}\cdot k_{1 } ) ( \epsilon\cdot k_{2})(\epsilon_{1}^{*a}f k_{2})+(\epsilon_{1}^{*a}\cdot k_{2 } ) ( \epsilon\cdot k_{1})(\epsilon_{2}^{*a}f k_{1})\bigr](1/m_{z})^{3 } \nonumber \\ & & + n_{9}\bigl[(\epsilon_{2}^{*a}\cdot k_{1 } ) ( \epsilon\cdot k_{1})(\epsilon_{1}^{*a}f k_{2})+(\epsilon_{1}^{*a}\cdot k_{2 } ) ( \epsilon\cdot k_{2})(\epsilon_{2}^{*a}f k_{1})\bigr](1/m_{z})^{3 } \nonumber \\ & & + n_{10}\bigl[(\epsilon_{2}^{*a}\cdot k_{1 } ) ( \epsilon\cdot k_{2})(\epsilon_{1}^{*a}f k_{1})+(\epsilon_{1}^{*a}\cdot k_{2 } ) ( \epsilon\cdot k_{1})(\epsilon_{2}^{*a}f k_{2})\bigr](1/m_{z})^{3 } \nonumber \\ & & + n_{11}(\epsilon_{1}^{*a}\cdot k_{2 } ) ( \epsilon_{2}^{*a}\cdot k_{1})\bigl(\epsilon f ( k_{1}+ k_{2})\bigr ) ( 1/m_{z})^{3}\bigr\ } , \end{aligned}\ ] ] where @xmath58 is the mass of the @xmath2 , @xmath82 is the weinberg angle , and the @xmath83 are the unitless coefficients .
the reader should see the appendix for a detailed explanation of how the coefficients are defined . at this point
we confirm that our amplitude is gauge invariant by replacing a polarization vector by its associated momentum @xmath84 and verifying that the amplitude vanishes .
for these purposes we drop the color index @xmath52 .
the gauge invariance implies six independent relations among the coefficients that reduce the number of independent coefficients from eleven to five .
these six relations and five coefficients are given explicitly in eqs.([eqn : girelat ] ) and ( [ eqn : coeff ] ) of the appendix .
what is of real interest , however , is the square of this amplitude .
when the square is taken and evaluated in the center of mass frame of the @xmath2 , we find that the square of the amplitude depends on only three of our five independent tensor coefficients @xmath85 where @xmath86 is the polar angle as measured from the direction in which the magnetic field points .
it is useful to extract from eq .
( [ eqn : amp2 ] ) the effective expansion parameter for the magnetic field .
that is , it was discussed earlier that the perturbation theory for the magnetic field is an expansion in @xmath5 . because this is fermion dependent , it is useful to pick the lightest fermion and make @xmath87 the expansion parameter .
recall that @xmath88 . making this substitution , eq .
( [ eqn : amp2 ] ) becomes @xmath89 at this point it is instructive to note that the leading contributions of @xmath90 go as @xmath91 , while for both @xmath92 and @xmath93 the leading contributions go as @xmath94 . ) , ( [ eqn : n2 ] ) , ( [ eqn : n3 ] ) , ( [ eqn : at ] ) , and ( [ eqn : bt ] ) of the appendix for large values of @xmath94 . ]
qualitatively , this means that only the lightest of the fermions make significant contributions to the the decay rate .
it also means that the @xmath2 is many times more likely to decay such that the gluons travel perpendicular to the magnetic field rather than parallel .
computation reveals that this factor is @xmath95 the large range in values is a symptom of the very large uncertainty in the current quark masses .
if the constituent quark masses are used , the ratio drops to a factor of @xmath96 .
the decay rate for this process is @xmath97 because of the large current quark mass ranges quoted by the particle data group @xcite , there is a corresponding large range in possible values for the decay rate .
we choose to use the mean values of these masses for computation .
seperate calculations are carried out using the constituent quark masses .
given below are the results of the decay rate for both the current and constituent mass calculations , , @xmath98 , @xmath99 , @xmath100 , @xmath101 , and @xmath102 .
the constituent quark mass values used are : @xmath103 , @xmath104 , @xmath105 , @xmath106 , and @xmath102 @xcite .
] respectively , @xmath107 the values of the branching ratio of the above decay rate with the total width of the @xmath2 in the absence of a magnetic field are @xmath108 for comparison purposes it is convenient to contrast the results of the two gluon case with the two photon case .
the results of the latter decay were determined in a manner completely analogous to that which has been outlined in this article for the decay @xmath0 .
the values of the decay rate for the process @xmath109 are @xmath110 notice that the degree to which the two values differ is much smaller than in the two gluon case .
this is because the major contribution to the process is due to the fermion mass being that of the electron , which is known very well .
the corresponding values for the branching ratio are @xmath111
this article has demonstrated that the existence of a magnetic field can stimulate decays that were otherwise forbidden .
schwinger s powerful proper - time method was used in conjunction with a first order expansion in the field to determine what corrections exist for the dirac green s function .
these new contributions were analyzed in the specific decays @xmath0 and @xmath1 . though the resulting decay rates are small , they are non - zero .
in addition it was also found that there exists a very large asymmetry as to whether the decay products will exit perpendicular or parallel to the magnetic field . as noted in the body of the article , the amplitude for the decay @xmath0 to first order in an external magnetic field ( fig .
[ fig : zggs ] ) is the sum the contributions from the zeroth order amplitude ( eq .
( [ eqn : ma0 ] ) ) , first order corrections to each leg of the charged ferion loop ( eq .
( [ eqn : mapr1 ] ) and the like ) , the first order correction to the phase of the amplitude ( eq . ( [ eqn : maph ] ) ) , and similar corrections for the crossed process which one obtains by interchange of the labels ( 1,2 ) .
[ eqn : coeff ] @xmath113
\\\label{eqn : n1 } n_{2}=\sum_{i}\frac{q_{i}g_{v}^{i}}{\bigl(t_{i}-4\bigr)^{2}}&\times & \bigl[\bigl(-8\,t_{i}^{3}+32\,t_{i}^{2}-256\,t_{i}+1024\bigr ) \nonumber\\ & & -\bigl(80\,t_{i}^{2}-448\,t_{i}+512\bigr)a(t_{i})\nonumber\\ & & + \bigl(16\,t_{i}^{2}-64\,t_{i}-256 + 1024\,t_{i}^{-1}\bigr)b(t_{i } ) \bigr]\\\label{eqn : n2 } n_{3}=\sum_{i}\frac{q_{i}g_{v}^{i}}{\bigl(t_{i}-4\bigr)^{2}}&\times & \bigl[\bigl(-(8/3)\,t_{i}^{3}+(160/3)\,t_{i}^{2}-(512/3)\,t_{i}\bigr ) \nonumber\\ & & + \bigl(48\,t_{i}^{2}-320\,t_{i}+512\bigr)a(t_{i})\nonumber\\ & & -\bigl(16\,t_{i}^{2}-128\,t_{i}+256\bigr)b(t_{i})\bigr ] \\\label{eqn : n3 } n_{4}=\sum_{i}\frac{q_{i}g_{v}^{i}}{\bigl(t_{i}-4\bigr)^{2}}&\times & \bigl[\bigl(8\,t_{i}^{3}-64\,t_{i}^{2}+128\,t_{i}\bigr)\nonumber\\ & & -\bigl(48\,t_{i}^{2}-384\,t_{i}+768\bigr)a(t_{i})\nonumber\\ & & + \bigl(16\,t_{i}^{2}-128\,t_{i}+256\bigr)b(t_{i})\bigr ] \\\label{eqn : n4 } n_{5}=\sum_{i}\frac{q_{i}g_{v}^{i}}{\bigl(t_{i}-4\bigr)^{2}}&\times & \bigl[\bigl(16\,t_{i}^{3}+64\,t_{i}^{2}-1024\,t_{i}+2048\bigr ) \nonumber\\ & & -\bigl(224\,t_{i}^{2}-1920\,t_{i}+4096\bigr)a(t_{i})\nonumber\\ & & + \bigl(32\,t_{i}^{2}-128\,t_{i}-512 + 2048\,t_{i}^{-1}\bigr)b(t_{i})\bigr],\label{eqn : n5 } \end{aligned}\ ] ] where @xmath114 is defined by the square of the ratio of the @xmath2 mass to the quark mass @xmath115 and the functions @xmath116 and @xmath117 are defined by @xcite @xmath118&= & 2\biggl[\sqrt{\frac{t-4}{t}}\sinh^{-1}\sqrt{\frac{-t}{4}}-1\biggr ] , & t\leq 0 \\[5 mm ] & = & 2\biggl[\sqrt{\frac{4-t}{t}}\sin^{-1}\sqrt{\frac{t}{4}}-1\biggr ] , & 4\geq t\geq 0 \\[5 mm ] & = & 2\biggl[\sqrt{\frac{t-4}{t } } \biggl(\cosh^{-1}\sqrt{\frac{t}{4}}-\frac{i\pi}{2}\biggr)-1\biggr ] , & t\geq 4 , \end{array}\ ] ] and @xmath119}{x(1-x)}&= & 4\biggl[\sinh^{-1}\sqrt{\frac{-t}{4}}\biggr]^{2 } , & t\leq 0 \\[5 mm ] & = & -4\biggl[\sin^{-1}\sqrt{\frac{t}{4}}\biggr]^{2 } , & 4\geq t\geq 0 \\[5 mm ] & = & 4\biggl[\cosh^{-1}\sqrt{\frac{t}{4}}\biggr]^{2}-\pi^{2 } -4i\pi\cosh^{-1}\sqrt{\frac{t}{4 } } , & t\geq 4.\\ \end{array}\ ] ] notice that in eqs .
( [ eqn : coeff ] ) there is a sum over all possible quarks in the charge loop . for the decay
@xmath3 this could also include the electron .
@xmath120 and @xmath121 are the charge and the associated vector coupling of the fermion , repectively .
the factors of @xmath58 present are due to the evaluation of the dot product between the two gluon momenta @xmath122 where @xmath47 is the four - momentum of the @xmath2 . | an investigation of the processes @xmath0 and @xmath1 in a background magnetic field is presented . for homogeneous fields corrections to the charged fermion propagator
can be calculated in leading orders of the magnetic field .
this work examines the first order contributions of the corrected propagator to decays that are otherwise zero .
results of the decay rates for varying field strengths are included . |
nlz is supported by a _ spitzer _
the observations reported here were obtained with the nasa / esa hubble space telescope and at the mmt observatory , a facility operated jointly by the smithsonian institution and the university of arizona .
public access time is available at the mmt observatory through an agreement with the national science foundation .
funding for the creation and distribution of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , nasa , the nsf , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
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 university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the korean scientist group , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
abazajian , k. , et al .
2005 , aj , 129 , 1755 antonucci , r.r.j . & miller , j.s .
1985 , apj , 297 , 621 hao , l. , et al .
2005 , aj , 129 , 1783 lacy , m. , et al .
2004 , apjs , 154 , 166 ptak , a. , zakamska , n.l . ,
strauss , m.a . ,
krolik , j.h . , heckman , t.m . ,
schneider , d.p . , & brinkmann , j. 2006 , apj , 637 , 147 szokoly , g.p .
, et al . 2004 , apjs , 155 , 271 urry , c.m . & padovani , p. 1995 , pasp , 107 , 803 york , d.g . , et al .
2000 , aj , 120 , 1579 zakamska , n.l . ,
2003 , aj , 126 , 2125 zakamska , n.l . ,
strauss , m.a . ,
heckman , t.m .
, ivezi , . , & krolik , j.h . 2004 , aj , 128 , 1002 zakamska , n.l . , et al . 2005 ,
aj , 129 , 1212 | type ii quasars are luminous active galactic nuclei whose centers are obscured by large amounts of gas and dust . in this contribution
we present 3-band hst images of nine type ii quasars with redshifts @xmath0 selected from the sloan digital sky survey based on their emission line properties .
the intrinsic luminosities of these quasars are thought to be in the range @xmath1 , but optical obscuration implies that host galaxies can be studied unencumbered by bright nuclei .
each object has been imaged in three filters ( ` red ' , ` green ' and ` blue ' ) placed between the strong emission lines .
the spectacular , high quality images reveal a wealth of details about the structure of the host galaxies and their environments .
most galaxies in the sample are ellipticals , but strong deviations from de vaucouleurs profiles are found , especially in the blue band .
we argue that most of these deviations are due to the light from the nucleus scattered off interstellar material in the host galaxy .
this scattered component can make a significant contribution to the broad - band flux and complicates the analysis of the colors of the stellar populations in the host galaxy .
this extended component can be difficult to notice in unobscured luminous quasars and may bias the results of host galaxy studies .
galaxies : active , galaxies : structure , quasars : general , scattering 98.54.aj , 98.54.cm , 98.58.ay , 98.62.lv one of the key long - term questions in quasar studies is to identify just which properties of the galactic host ( overall morphology ?
presence of bars ?
close neighbors ?
star - formation history ?
gas content ? ) are associated with activity in its nucleus .
however , observations of quasar hosts are made difficult by the bright source in the center of the host galaxy the quasar itself ; in some cases the light from the quasar completely overwhelms the host galaxy , preventing its detection altogether . therefore , a significant advantage can be gained by studying the hosts of quasars in which the optical and uv light from the nuclear source are obscured by large amounts of gas and dust near the nucleus so called type ii quasars @xcite .
such objects have only recently been discovered in large numbers in x - ray , mid - ir and optical surveys ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the optical , the strong blue continuum and broad emission lines typical of unobscured ( type i ) quasars are not observed in type ii quasars , so that narrow emission lines that originate above and below the plane of obscuration are the only signature of the presence of a hidden active galactic nucleus ( agn ) .
we therefore searched the spectroscopic database of the sloan digital sky survey ( sdss ; * ? ? ?
* ; * ? ? ?
* ) for objects with narrow emission lines with high - ionization line ratios and selected a few hundred type ii quasar candidates in the redshift range @xmath2 @xcite .
we have been conducting sensitive follow - up observations of the objects in this sample to determine their multi - wavelength properties .
we found that their ir and x - ray properties are consistent with their interpretation as powerful ( bolometric luminosity @xmath3 erg sec@xmath4 ) obscured ( hydrogen column densities @xmath5 @xmath6 ) agn @xcite . even though the direct emission from the nucleus is completely blocked by the obscuring material ,
some light emitted along other directions is scattered into our line of sight by electrons or dust particles in the host galaxy @xcite .
the scattering efficiency ( the ratio of the apparent luminosity due to scattering to the intrinsic luminosity ) is very poorly known , but some observations suggest that it can be as large as a few per cent @xcite . for an intrinsically luminous quasar
such scattering efficiency implies that the scattered light can make a significant contribution to the observed flux .
this contribution is very difficult to separate from the stellar light , since the scattered light can form complicated shapes in the plane of the sky .
imaging polarimetry helps resolve this ambiguity , but in the absence of such observations , spectropolarimetry or broad - band polarimetry allows one to determine the major direction of scattering and therefore indicates where in the galaxy most of the scattered light is coming from . in this contribution
we describe results of the imaging of nine type ii quasars using the advanced camera for surveys ( acs ) aboard the hubble space telescope ( hst ) , in combination with polarimetric data from ground - based observations .
we selected nine type ii agn with @xmath7([oiii]5007)@xmath8 from the samples by @xcite and @xcite in the redshift range @xmath0 .
based on the correlation between the narrow line luminosities and the broad - band luminosities that exists for unobscured quasars @xcite , we estimate that the intrinsic luminosities of the objects in our program are about @xmath1 .
to avoid contamination from the strongest emission lines ( [ oii]3727 , [ oiii]4959,5007 , h@xmath9 , h@xmath10 ) , we carefully chose redshifts and acs filters so that none of the filters used in our program contain any of these lines
. each object was imaged in three filters : ` blue ' ( blueward of the [ oii]3727 line ) , ` green ' ( between [ oii]3727 and h@xmath9 ) and ` red ' ( redward of the [ oiii]5007 emission line ) .
the three images were combined to produce color - composite images shown in figures [ fig_1 ] and [ fig_2 ] . within each figure ,
the filters , redshifts and image combining parameters were similar , so that the difference in colors of objects in the images reflect differences in their optical spectral energy distributions .
is marked with a solid black line , and grey lines mark the @xmath11 confidence limits if the uncertainty in the angle is significant . ] , except a different set of filters was used for these three objects .
polarimetric data are not available for sdss j1301@xmath120058 and sdss j1243@xmath120232 . ]
polarimetric or spectropolarimetric data were obtained for seven of the objects using the ccd spectropolarimeter ( spol ) at the 6.5 m multiple mirror telescope ( mmt ) and bok telescope .
the observed degrees of polarization vary from 1% to 10% ( in the green band ) , and the polarization position angles are indicated with solid lines ( marked ` @xmath13 ' ) in figures [ fig_1 ] and [ fig_2 ] . by fitting ellipses to the isophotes for each object in each band , we constructed one - dimensional brightness profiles and found that in most cases they are consistent with the de vaucouleurs law , except sdss j1106 + 0357 ( figure [ fig_2 ] ) which has a prominent exponential component in the outer parts .
once the best - fitting two - dimensional de vaucouleurs profile is subtracted off the image , significant deviations are revealed .
they are especially prominent in the blue band , and the brightest of them can be seen in figures [ fig_1]-[fig_2 ] as blue patches , either biconical ( as in sdss j1323 - 0159 ) or irregularly shaped ( as in sdss j1039 + 6430 ) .
we find that in most cases the line connecting the center of the galaxy and the brightest blue patch is orthogonal to the measured polarization position angle , and we therefore interpret these regions as emission from the obscured nucleus scattered into our line of sight by the material in the host galaxy .
notable exceptions include sdss j0920 + 4531 , in which the blue region near the center is misaligned with the scattering orientation as measured using spectropolarimetry .
given the overall morphology of the object which shows several components interacting with the main galaxy , it is likely that the bright blue patch near the center is just a starforming region rather than scattered light .
in this case , the scattering region can not be easily identified on the basis of morphology .
the example of sdss j0920 + 4531 shows that in the absence of polarimetric data , star formation in the host can not be easily distinguished from the scattered light contribution .
this difficulty is further underscored by the case of sdss j0123 + 0044 , where the ambiguity is not resolved even by the polarimetric data . in this object ,
several patches of blue surround the nucleus . while the brightest of them is orthogonal to the polarization position angle ( and therefore is likely to be scattered rather than stellar light ) , the fainter spots to the east of the nucleus lie on the extension of a faint thin structure which extends out to 30 kpc to the north of the galaxy ( this structure is too faint to be seen on figure [ fig_1 ] ) .
given this morphological connection with what looks like a tidal tail , the eastern spots are probably patches of star formation in a companion in the process of disruption .
scattered light can make a significant contribution to the broad - band fluxes of quasar host galaxies .
one striking example is sdss j1039 + 6430 , in which most of the blue emission is due to the scattered light , and the measured polarization reaches 17% in this band . scattering from different parts of the host galaxy can wash out the polarization signal , so that scattered light can be important even if the observed polarization is not very high . if one were to take the colors of the extended emission at face value to estimate the age of the stellar population in the host galaxy , the measured ages would be strongly biased toward small values .
as we demonstrated above , polarimetric data can help distinguish between scattered light and star formation , but does not eliminate the ambiguity entirely .
this difficulty may be even more severe in host studies of unobscured quasars , in which the presence of a very bright blue nuclear source can make a detailed morphological analysis difficult and can dilute the polarization signal .
our other findings include the presence of kpc - scale dust lanes across the centers of some host galaxies ( e.g. , sdss j1323@xmath120159 and sdss j0123 + 0044 ) , extended tidal structures in three of the objects in our sample , and faint companions within 20 projected kpc in each of the remaining six objects .
the statistical significance of these results will be addressed in our future work . |
recent proper motion studies of infrared - luminous stars in the galactic center ( gc ) ( @xcite ) have convincingly demonstrated the existence of a compact @xmath20 dark mass in the dynamical center of the gc , which is located within @xmath21 of the radio source @xmath0 ( @xcite ) ( @xmath22pc at 8kpc ) .
lower bounds on the compactness of this mass concentration , together with dynamical considerations , argue against the possibility of a massive cluster , and point towards a super - massive black hole as the likely alternative ( @xcite ; @xcite ) .
this conclusion has important implications for basic issues such as the prevalence of massive black holes in the nuclei of normal galaxies , and the nature of the accretion mechanism that makes @xmath0 so much fainter than typical active galactic nuclei ( @xcite ; @xcite ) .
wardle & yusef - zadeh ( 1992 ) considered several gravitational lensing effects that could be induced by a massive black hole in the gc .
in particular , wardle & yusef - zadeh pointed out that gravitational lensing would occasionally lead to the amplification and splitting of the stellar images of stars which happen to move behind the black hole , and that such events would typically last from months to years , depending on the distance of such stars from the black hole .
wardle & yusef - zadeh also estimated that an optical depth of order unity for such events requires an observed central surface density @xmath23 stars arcsec@xmath24 , and that this in turn would require an angular resolution of @xmath25 to individually separate the lensed images from the crowded background of faint stars .
however , the photometric sensitivities and spatial resolutions required for such observations are far beyond current observational capabilities .
presently , the deepest near - infrared @xmath1-band ( 2.2 @xmath2 m ) images of the central arcsec reach down to @xmath26 ( @xcite ) .
as we will show , at this limiting magnitude the expected central surface density is @xmath27 stars arcsec@xmath24 ( @xcite ) .
the highest spatial resolutions obtained so far are the diffraction limited resolutions of @xmath28 ( @xcite ; @xcite ) and @xmath29 ( @xcite ) achieved in the proper motion surveys . in this paper
we investigate a different possibility , namely _ microlensing _ amplification of faint stars by the central black hole .
such events occur when the amplified but _ unresolved _ images of faint stars rise above the detection threshold and then fade again as such stars move behind the black hole close to the line - of - sight to @xmath0 .
we present a quantitative study of such microlensing events , and we compute in detail the microlensing rates as functions of the event durations , for a wide range of assumed detection thresholds . in addition , we also consider the possible amplification of sources which lie above the detection thresholds , and we re - examine the limit in which the two lensed images can actually be resolved .
our study is motivated by several recent developments .
deep infrared star counts in the inner gc ( @xcite ; @xcite ) and infrared and optical star counts in baade s window ( @xcite ; @xcite ) now make it possible to reliably model the infrared stellar luminosity function in the vicinity of the galactic center .
the on - going proper motion monitoring campaigns of the inner few arcsec of the gc record both the positions and fluxes of individual stars in the field at a sampling rate of 12 observing runs per year . as a by - product
, these measurements can be used to search for microlensing events , albeit at a low sampling rate .
such events would appear as time varying sources very close to @xmath0 .
it is therefore intriguing that one or two variable ir sources may have already been detected close to , or coincident with , the position of @xmath0 ( @xcite ; @xcite ) .
these sources may have brightened to @xmath30 before fading from view , and appear to have varied on a timescale of @xmath16 yr .
as we will argue , this behavior is consistent with the expected behavior of bright long - duration microlensing events .
however , we will also argue that the probability that a single such event could have been detected during the course of the recent proper - motion monitoring campaigns is small ( @xmath17 )
. the detection probabilities will increase considerably as the observational sensitivities improve ( e.g. with the advent of adaptive optics ) , and if @xmath0 is monitored more frequently than has been done so far .
the structure of our paper is as follows . in
[ s : rate ] we set up the formalism required for our computations .
some of the more technical aspects are discussed in the appendix . in [ s : gc ] we discuss our treatment of the stellar densities , velocity fields , and @xmath1-band luminosity functions , which enter into the computation of the lensing rates .
we present our results in [ s : results ] , where we provide computations of the lensing rates as functions of the event durations for a wide range of assumed detection thresholds . in [ s : discuss ] we present a discussion and summary .
the effective size of a gravitational lens at the lens plane is set by the einstein radius , @xmath31 , @xmath32 where @xmath33 is the gravitational constant , @xmath34 is the speed of light , @xmath35 is the lens mass ( here the black hole mass ) , @xmath36 and @xmath37 are the observer lens and lens source distances , respectively ( see e.g. review by bartelmann & narayan 1998 ) . we assume , as will be justified below , that @xmath38 , and define @xmath39 and @xmath40pc .
the effective size of the lens at the source plane is @xmath41 .
the angular size of the einstein radius , @xmath42 , is @xmath43 where @xmath44 is the sun s galactocentric distance ( @xcite ) . in our study
we assume that any ir - extinction associated with an accretion disk or an ` atmosphere ' near the event horizon is negligible .
we begin by showing that in our problem the stars can be treated as point sources and that the linear approximation ( small light - bending angle approximation ) holds .
first , when @xmath45 , a star can be approximated as a point source as long as its radius , @xmath46 , is much smaller than @xmath47 , where @xmath48 is the required amplification to observe the lensed source above the detection threshold ( see [ s : mlrate ] ) . as we show below (
[ s : plklf ] ) , the faintest stars that contribute significantly to the lensing are low - mass stars with @xmath49 cm , which require amplification factors of @xmath50 .
brighter stars require progressively smaller amplification factors , down to @xmath51 , which for detection thresholds @xmath52 correspond to stars with @xmath53 cm . assuming a mean stellar mass of @xmath54 , a central mass density of @xmath55pc@xmath56 in the gc ( @xcite ) implies a mean stellar separation of @xmath57pc . even at such a small distance from the black hole , @xmath58 cm , so that the condition @xmath59 is generally satisfied for all relevant stars in the system .
we thus conclude that the point source approximation holds over the entire range of interest .
second , the linear approximation holds as long as the einstein radius is much larger than the schwarzschild radius of the black hole , @xmath60 , @xmath61 which even for @xmath62 is as high as @xmath63 .
a point lens produces two images , one within and one outside the einstein radius .
the angular separation between the two images is @xmath64 and their mean angular offset from the optical axis is @xmath65 , where @xmath65 is the angular separation between the unlensed source and the lens . as will be shown in
[ s : gc ] , @xmath66 for amplification above present - day detection thresholds in the gc , so that @xmath67 . three angular scales in the problem
determine the way the lensing will appear to the observer : the einstein angle @xmath68 , the fwhm angular resolution of the observations , @xmath69 , and the mean projected angular separation between the observed stars , @xmath70 , where @xmath11 is the detection threshold .
the lensed images have to be detected against the background of a very dense stellar system .
this background will be low as long as as the central surface density of observed stars , @xmath71 , is small enough so that @xmath72 .
thus , for a given detection threshold , an angular resolution of at least @xmath73 is required to detect the lensed star .
lower resolutions correspond to the regime of ` pixel lensing ' ( @xcite ) , which we do not consider here . when @xmath74 , the two images will not be resolved is required to resolve the two microlensed images .
the exact value of the minimal separation probably lies in the range @xmath69 to @xmath75 , and depends on the details of the procedure for faint source separation in the crowded inner @xmath76 of the gc .
] , and the lensed source will appear as a _ microlensing _ event . for a given angular resolution , there is a maximal lens source distance @xmath77 for microlensing , @xmath78 more distant stars will have @xmath79 and their two images will be separately resolved .
the unresolved images of a faint star at @xmath80 close to the line of sight will appear as a ` new source ' at the position of @xmath0 , to within the observational resolution .
the total source amplification , @xmath48 , is related to @xmath81 by ( @xcite ) @xmath82 and for small @xmath81 , @xmath83 .
the maximal amplification along the star s trajectory is reached when its projected position is closest to the lens .
a star of stellar type @xmath84 and an absolute @xmath1 magnitude @xmath85 , which is observed on a line of sight with an extinction coefficient @xmath86 and distance modulus @xmath87 , will be detected above the flux threshold @xmath11 if it is amplified by at least @xmath88 .
this corresponds to a maximal impact parameter of @xmath89 . as the minimal required amplification approaches 1
, @xmath90 diverges .
this , and the fact that in practice the threshold is not sharply defined , leads us to introduce a cutoff at @xmath91 , which implies a minimal amplification of @xmath92 .
the basic quantity that is required for predicting the lensing properties is the differential lensing rate as function of stellar type , event duration , amplification , and source distance from the black hole .
we derive an explicit expression for the differential lensing rate in the appendix . here
it is instructive to discuss the relations between these properties by considering the simpler case of the total lensing rate , regardless of the event duration .
the total integrated rate for microlensing amplification of background stars above the detection threshold is @xmath93 where @xmath94 is the inner radius of the central stellar cluster , @xmath95 the maximal radius for producing unresolved lensed images , @xmath96 is the transverse 2d stellar velocity averaged over the velocity distribution function , and @xmath97 is the total number density of stars . here and below , the notation @xmath98 designates the average of the bracketed quantity over the stellar types with @xmath99 , weighted by their fraction in the stellar population , @xmath100 , where it is assumed that @xmath100 does not depend on @xmath101 .
the properties of the stellar population enter the integrated rate only through the mean impact parameter @xmath102 , which for @xmath103 is simply @xmath104 where @xmath105 is the observed ( dust extinguished ) stellar @xmath1-band flux and @xmath106 is the detection threshold flux corresponding to @xmath11 .
we characterize the microlensing time - scale for stars of type @xmath84 as the average time they spend _ above the detection threshold _
, @xmath107 where a @xmath108 factor comes from averaging over all impact parameters with @xmath109 .
the lensing rate , amplification and event duration are inter - related . for @xmath110 , the _
median _ value of the distribution of maximal amplifications is simply @xmath111 .
note , however , that the _ mean _ maximal amplification , @xmath112 , diverges logarithmically .
generally , a fraction @xmath113 of the events will have a maximal amplification of @xmath114 or more ( i.e. a peak magnitude above the threshold of @xmath115 or less ) .
the rate of such events is smaller , @xmath116 the time - scale of events amplified by a factor of @xmath117 above the threshold is somewhat longer than the average time - scale ( eq . [ e : t0 ] ) , since they must cross closer to the line of sight , @xmath118 which approaches @xmath119 for large amplifications .
conversely , for a given @xmath120 and @xmath89 , even a small increase in @xmath121 is associated with a considerable increase in the maximal amplification .
when @xmath79 at @xmath122 , the two images can be resolved . as we show in [ s : results ] , @xmath77 is already quite small for present - day angular resolutions , and will become smaller still as the resolution improves .
this implies that there may be a non - negligible contribution to the lensing rate from regions beyond @xmath77 .
we therefore have to consider also the case of resolved images . for @xmath123
, the two images will appear at an offset of @xmath124 from @xmath0 , on opposite sides of it .
the amplifications of the individual images are related by @xmath125 and @xmath126 , which for the high amplifications that are requires in the gc can be approximated as @xmath127 .
the formalism used for calculating the lensing of unresolved images can therefore be applied in this case simply by raising the effective detection threshold by a factor of two ( @xmath128 ) .
this makes resolved images harder to detect , but on the other hand , if both images are observed , the identification of the event as lensing is much more certain .
the formalism for describing unresolved and resolved lensing of stars from below the detection threshold can be also extended to cases where the microlensed source is an observed bright , @xmath129 star . at the current detection threshold ,
the observed central surface density is @xmath130 arcsec@xmath24 ( @xcite ; @xcite ) .
for such a small number of stars , whose orbits can be tracked individually , a statistical treatment is not very useful .
however , our calculations indicate that a ten - fold improvement in the detection threshold will yield an observed central surface density of at least @xmath131 arcsec@xmath24 ( eq . [ e : dp ] ) .
for such a large sample , a statistical treatment is more meaningful .
we therefore present below , for completeness , results for microlensing amplification of stars fainter than the detection threshold ( ` faint - star lensing ' ) as well as for stars brighter than the detection threshold ( ` bright - star lensing ' ) . for bright - star microlensing
we assume that the effective minimum magnification factor for detection is @xmath92 ( @xmath91 ) .
three basic quantities enter into the computation of the microlensing rate : the stellar number density distribution , the stellar velocity field , and the @xmath1-band luminosity function ( klf ) .
the stellar population in the central @xmath132 pc appears to consist of a mixture of old bulge stars , and ` central cluster ' stars which may have been produced in various star - formation episodes during the lifetime of the galaxy ( @xcite ; @xcite ) .
evidence for recent star formation in the gc has been provided by krabbe et al .
( 1995 ) who found a concentration of luminous early - type stars within a few arcsec of @xmath0 , implying a recent ( @xmath133 yr ) starburst in which @xmath134 stars were formed .
additional young stellar systems , the ` arches ' and ` quintuplet ' clusters also exist close the gc and contain large numbers of massive stars ( @xcite )
. a further indication of on - going star - formation in the gc is the fact that the klf in the central cluster extends to more luminous stars than the klf of the galactic bulge as observed via baade s window ( @xcite ; @xcite ; @xcite ; @xcite ) .
serabyn & morris ( 1996 ) suggested that continuous star formation in the central cluster is maintained by molecular clouds in the gc , and that the @xmath135 radial light profile of the central cluster reflects the distribution of the star - forming molecular clouds . in our analysis
we make the simplifying assumption that the discrete young clusters in the gc can be modeled by a volume averaged and smoothly distributed stellar population . in particular , we do not consider here lensing events that might be associated with the innermost stellar cusp , which has been identified by eckart & genzel ( 1997 ) and ghez et al .
( 1998 ) in the immediate vicinity of @xmath0 .
an over - density of stars above the smoothed distribution very near the black hole may contribute to very short duration microlensing events .
however , the total stellar mass and luminosity function of the stars that are associated with this ` @xmath0 cusp ' are poorly constrained at the present time , so that computations of the lensing rates require more extensive modeling and analysis .
we will present such computations elsewhere . because the lensing time - scales increase with distance , long duration events , which are relevant for low sampling rates of the current proper motion studies , tend to be associated with more distant stars .
we therefore consider both stars in the central star - forming cluster , as well as more distant old - population bulge stars in our analysis .
we now proceed to discuss the stellar densities , velocities and klfs of these two components .
genzel et al .
( 1996 , 1997 ) derived density and velocity models for the central cluster based on fits to the observed star counts , stellar radial velocities and proper motions in the inner few pc .
their mass density distribution is parameterized by a softened isothermal distribution @xmath136 where @xmath137 is the core radius and @xmath138 is the central density , with best fit values of @xmath139pc and @xmath140 .
the 1d velocity dispersion of the late type stars in the core , which are dynamically relaxed , is modeled by ( @xcite ) @xmath141 where @xmath142 and @xmath143 is the projected distance corresponding to @xmath144 . in the inner gc ,
the mass density is strongly constrained by the dynamics , whereas the @xmath1 luminosity density @xmath145 ( @xmath146 ) is not well defined due to the patchy nature of the extinction .
the situation is reversed on scales larger than 1kpc , where the observed rotation curve and velocity dispersion are harder to interpret , but @xmath145 can be de - projected from the observed surface brightness .
kent ( 1992 ) proposed that both the rotation curve and the surface brightness along the major axis of the galactic bulge in the mid - plane of the galaxy at @xmath147kpc can be described as a superposition of bulge and disk components with a mass - to - light ratio , where the solar monochromatic luminosity at @xmath148 m is @xmath149ergs@xmath7@xmath2m@xmath7 . note that genzel et al .
( 1996 ) define this quantity as @xmath150 , with @xmath151 m .
the two definitions are related by @xmath152 . ]
@xmath153 and a luminosity density @xmath154 where here @xmath11 is a modified bessel function ( not to be confused with the detection threshold ) , @xmath155pc and @xmath156pc .
kent also suggested a @xmath157 extrapolation of @xmath145 towards the center based on observations of the intensity variation in the inner 10 pc .
we replace this extrapolation with an updated and non - divergent model by adopting the genzel et al .
( 1996 ) mass density model ( eq .
[ e : nstar ] ) for the core and by assuming @xmath153 for the bulge and disk , so that @xmath158 and @xmath159 .
we further assume that the bulge is axisymmetric , and model the mass density over the entire distance range by @xmath160 figure [ f : rho ] shows our mass density model , and in particular the emergence of the bulge component at @xmath132 pc .
the central cluster completely dominates the total mass density in the inner 10 pc , but then falls to @xmath161% of the total at 50pc and to @xmath162% at @xmath163pc . the velocity field on this scale includes both the bulk rotation and the random motion .
we approximate kent s models for the galactic rotation and the velocity dispersion in the inner 300pc along the major axis of the bulge by log - linear fits .
we assume that the rotation velocity is perpendicular to the line of sight , and that its contribution to the transverse velocity is @xmath164 and that the 1d velocity dispersion in the bulge is @xmath165 we model the 1d velocity dispersion over the entire distance range by @xmath166 note that both the proper motion of @xmath0 , which is @xmath167 ( @xcite ) , and the sun s galactocentric rotation have a negligible contribution to the relative proper motion of the source and lens because of the high stellar velocities near the dynamic center and because @xmath45 .
in our analysis we consider a model for the klf which is based on the observed portions of the klfs in the central cluster and in the bulge , and on a theoretical klf computed in a population synthesis model for the central cluster .
the klf of the bulge has been observed through baade s window , which at @xmath168 , @xmath169 , is @xmath170kpc from the gc at the tangential point .
the bulge klf approximately follows a single power - law @xmath171 , with @xmath172 ( @xcite ) , from @xmath173 to @xmath174 , mag for @xmath175 in erg@xmath176@xmath177 . ] assuming a distance modulus @xmath178 and @xmath179 .
the faintest observed stars are at the detection threshold , and the power - law klf likely extends , or steepens slightly , to lower luminosity stars near the main - sequence turn - off point .
recent _ hst _ observations of the optical luminosity function in baade s window ( @xcite ) probe it down to very low stellar masses ( @xmath180 ) , well below the main - sequence turn - off point . the @xmath181-band luminosity function ( vlf ) presented by holtzman et al .
( 1998 ) shows that a sharp break occurs at the turn - off point @xmath182@xmath183 , corresponding to @xmath184 stars .
we extrapolated the power - law klf of tiede et al .
( 1995 ) down to @xmath185 ( assuming @xmath181-@xmath186 for @xmath187 stars ) and compared it to the holtzman et al .
( 1998 ) vlf at @xmath188 .
we find that the predicted @xmath1 star - counts somewhat underestimate the @xmath181 star - counts , but agree with them to within a factor of 2 .
we thus conclude that the power - law klf can be extended down to at least @xmath189 .
the observed vlf can be used to determine the behavior of the klf at even lower luminosities by using the @xmath181-to-@xmath1 conversions for low mass stars presented by henry & mccarthy ( 1993 ) .
doing this shows that the klf likely flattens at a break point close to @xmath190 , with @xmath191 in the range @xmath192 , and turns over at @xmath193 .
as we have discussed above , the stellar population in the central cluster appears to be consistent with continuous star - formation throughout the galaxy lifetime ( see e.g. serabyn & morris 1996 ) .
however , despite the many differences between the populations in the bulge and the central cluster , the observed klf in the central @xmath194 of the gc follows a power - law similar to that of the bulge klf ( @xcite ) , but extends to more luminous stars ( @xcite ) .
the exact upper luminosity cutoff is not well - defined due to statistical fluctuations in the counts and the contribution of asymptotic giant branch ( agb ) stars , which do not follow the power - law .
the central cluster s klf is not known beyond the current detection threshold of @xmath195 , which is the region of interest for the microlensing calculations .
however , the similarity of the klf power - law indices in the central cluster and bulge klfs suggests that the two klfs are similar , apart for having different upper luminosity cutoffs .
the much lower extinction in baade s window , which is only @xmath196 as compared to @xmath197 in the gc ( @xcite ) , makes it possible to extend the klf in the gc to lower luminosities . indeed , blum et al .
( 1996 ) find that the bulge klf , after correcting for the @xmath86 difference , can be smoothly joined to the gc lf in the central @xmath198 .
their resulting composite lf has a best fit power - law index @xmath199 , and extends from @xmath200 down to @xmath201 .
our theoretical model for the central cluster , which we discuss below , indicates that this power - law character is further maintained down to @xmath202 .
we therefore adopt the same @xmath199 power - law index for both the central cluster and bulge klfs . for the central cluster klf
we set the upper luminosity cutoff at @xmath203 , and for the bulge klf at @xmath204 .
we set an effective low luminosity cutoff for both at @xmath205 , which corresponds to stars with masses @xmath206 .
as we show in [ s : fit ] , the lensing rates are insensitive to the exact choice of the low luminosity cutoff .
we note that the observed bulge klf is better fitted by a somewhat flatter power - law than @xmath199 , which implies less faint - star lensing candidates . on the other hand
, the comparison with the optical luminosity function suggests that the klf is a conservative estimate of the total number of stars .
in addition , the flatter power - law fails to account for the strong excess of horizontal branch ( hb ) stars above the power - law ( @xcite ) .
these are important potential microlensing sources , as they lie in a magnitude range that is just below the threshold if the bulge population is observed through the gc extinction of @xmath207 .
we therefore consider these small discrepancies in slope and number to be within the limitations of the power - law approximation and the observational uncertainties . as a check on the empirically based pure power - law klf ,
we have also computed a series of theoretical klfs using our ` population synthesis ' code ( see e.g. @xcite ) . in our models
we used the geneva stellar evolutionary tracks ( @xcite ) , and concentrated on stellar models with twice - solar metallicities , as is indicated by the enhanced abundances of the gas in the gc ( @xcite ) .
however , recent measurements of stellar spectra of cool luminous stars in the gc point to solar metallicities ( @xcite ) .
we therefore considered also such stellar models , and verified that our results do not depend strongly on the assumed metallicity .
we computed the stellar @xmath1 luminosities using the empirical bolometric corrections and @xmath181-@xmath1 colors for dwarfs , giants and supergiants compiled by schmidt - kaler ( 1982 ) , and tokunaga ( 1998 ) .
the geneva tracks for intermediate mass stars ( @xmath208@xmath209 ) do not extend beyond the end of the early agb phase .
we extended these tracks to include also the more luminous thermal - pulsing agb phases following prescriptions described by charlot & bruzual ( 1991 ) and bedijn ( 1988 ) . in our models
we assume explicitly that stellar remnants and stars less massive than @xmath210 are negligible sources of @xmath1-band luminosity .
this low - mass luminosity cutoff roughly corresponds to the low luminosity cutoff of the empirically based power - law klf .
we constructed theoretical models for a range of cluster parameters assuming continuous star - formation lasting for 10gyr .
we considered models with pure power - law imfs , and the miller - scalo imf ( @xcite ; @xcite ) for a range of lower- and upper - mass cutoffs .
we selected the model which yields a klf which best matches the observed klf for the central cluster measured by davidge et al .
( 1997 ) .
we find that a model with a miller - scalo imf ranging from @xmath211 to @xmath212 provides the best overall fit to the data . in figure
[ f : klf ] we compare our model klf for the central cluster with the @xmath199 power - law fit of blum et al .
( 1996 ) to their composite klf .
the overall agreement between the theoretical klf and the power - law klf is remarkable , for both stellar models with solar and twice - solar metallicities .
the model successfully reproduces both the power - law character and slope of the observed klf of the central cluster .
furthermore the model shows that the power - law klf likely extends down to at least @xmath213 . in this model ,
the most probable lensed sources are @xmath214 stars with @xmath215 , @xmath216 cm just off the main sequence , which require a magnification of order @xmath217 to be detected above the threshold .
our population synthesis model predicts a mass - to - light ratio @xmath218 , in excellent agreement with @xmath219 measured in the central cluster by genzel et al .
( 1996 ) . as shown by eq .
[ e : rate ] , the lensing rate depends on , @xmath97 , the total number density of stars which are effective sources of k - band luminosity .
we refer to such stars as ` k - emitting ' stars , and define @xmath220 where @xmath221 is the total dynamical mass density , @xmath222 is the mean stellar mass of the k - emitters , and @xmath223 is the fraction of the total dynamical mass contained in @xmath1-emitting stars , i.e. excluding low - mass stars and remnants ( i.e. objects which lie below the effective low - luminosity cutoff of the klf ) and gas clouds .
the observed star counts , @xmath224 , within an area s , can be used to obtain a best fit value of @xmath225 given the constraint @xmath226 where the klf @xmath227 is normalized to 1 .
the mass - to - light ratio over the integration volume used in eq .
[ e : norm ] can then be deduced from the fit by @xmath228 where @xmath229 is the average k - band luminosity of the k - emitting stars .
the power - law klf we adopt in our computations ( @xmath230 ) is characterized by @xmath231 , @xmath232 and @xmath233 , where @xmath234 and @xmath235 are the effective lower and upper luminosity cutoffs to the klf , and @xmath236 is the @xmath1 luminosity corresponding to the detection threshold . by approximating @xmath237 ( where @xmath48 is the required amplification ) the mean stellar luminosity and mean impact parameter for such klfs are given by the simple approximate relations @xmath238 it also follows that @xmath97 , @xmath239 and the lensing rate @xmath240 scale as @xmath241 eq . [ e : plrho ] shows that the total lensing rate is _ insensitive to the low - luminosity cutoff _ of the klf .
this simply reflects the fact that an increase in the number of @xmath1-emitting stars as @xmath234 decreases is offset by a correspondingly smaller mean lensing impact parameter for the stellar system , and vice - versa .
this important property allows us to robustly compute the lensing rate even if the effective lower luminosity cutoff of the power - law klf is not well determined then the rate does depend on @xmath234 with @xmath242 .
however , as we have discussed , the observations of holtzman et al .
( 1998 ) strongly suggest that the klf flattens , rather than steepens , below our assumed value for @xmath234 . ] .
we note that since the differential contribution of stars with luminosity @xmath243 to the mean impact parameter @xmath102 scales like @xmath244 , the integrated contribution of stars from a @xmath245-wide bin is , for @xmath246 , a very weakly decreasing function of the bin s @xmath1 magnitude .
we therefore expect that the lensed stars will exhibit a wide range of intrinsic @xmath1 magnitudes , with a weak trend towards lensing by stars close to the detection threshold .
it is uncertain at which radius the stellar population makes the transition from a population that is characteristic of a star - forming cluster to a more bulge - like population .
however , as we argued above , the observed properties of the klfs in the gc and the bulge suggest that they are very similar for @xmath247 . since the normalization of the klf does not depend strongly on the upper luminosity cutoff ( eq . [ e : plrho ] ) , and since the very high luminosity tail of the klf is not relevant for the lensing calculations ( the @xmath248 stars have a negligible contribution to the lensing rate of observed stars ) , we infer the value of @xmath225 using the star - counts observed in the core , and we adopt this normalization for the entire inner 300 pc . in carrying out this procedure
was carried out by approximating the rectangular field with a circular field of projected radius @xmath249 having the same area .
the integration was done over a cylinder of radius @xmath249 , centered on the black hole and extending along the line of sight @xmath250pc in each direction , a distance large enough to ensure that the surface density reaches its asymptotic value .
] we used the observed number counts in the @xmath251 bin ( stars per mag per arcsec@xmath24 ) , averaged over the @xmath252 field observed by davidge et al .
the star counts in this luminosity range are likely complete , and this range is also well separated from the regions that are affected by agb and hb stars , which cause deviations from the power - law behavior ( @xcite ) . using eq .
[ e : norm ] , we then find that @xmath253 .
this value for @xmath225 can be reproduced , for example , by a choice of @xmath254 and @xmath255 , which is comparable with the values @xmath256 and @xmath257 of our population synthesis model for the central cluster . for the central cluster power - law klf , @xmath258 , so that eq .
[ e : mtol ] yields @xmath259 , again in excellent agreement with @xmath219 inferred by genzel et al .
( 1996 ) for the central cluster .
we have carried out detailed computations of the lensing event rates using the formalism described in [ s : rate ] and in the appendix , and using the stellar number and velocity distributions and the power - law @xmath1-band luminosity function , as discussed in [ s : gc ] .
the integrations were carried out from a minimum distance @xmath260pc , equal to the mean central stellar separation , to a distance @xmath261pc , where the integrated lensing rates approach their asymptotic values .
the normalized central cluster klf allows us to estimate the stellar surface density as a function of the detection threshold , and from it derive the mean angular separation of the stars , @xmath262 , the required angular resolution of the observations @xmath69 , and the maximal distance for observing ( unresolved ) microlensed stars @xmath77 ( as given by eqs .
[ e : phi ] and [ e : dmu ] ) .
we find that for the central cluster klf @xmath263 for @xmath262 in arcsec and @xmath264 in arcsec@xmath24 .
thus , for a detection threshold of @xmath265 we expect a central surface density of 21 stars per arcsec@xmath24 for complete counts and 15 stars arcsec@xmath24 , respectively . however , the star counts are incomplete close to the detection thresholds , and the measured @xmath264 actually indicate a significant density enhancement ( the ` @xmath0 cusp ' ) relative to the immediate surroundings . ] .
we present our results in a way which makes it possible to flexibly estimate the actual detection probabilities for a broad range of observing strategies .
we consider the generic monitoring campaign to consist of a series of @xmath266 very short observing runs which are carried out during a total time period @xmath267 ( typically @xmath268 several years ) , with a mean interval @xmath269 between each observing run ( typically @xmath270 a month to a year ) .
we now wish to distinguish between ` long ' and ` short ' events , where we define the event duration as the time the source spends above the detection threshold .
long events are those with durations @xmath271 , and would appear as time varying sources that brighten and fade during the course of several observing runs .
the microlensing origin of long events could be verified , in principle , from the shape of the light curve and its achromatic behavior .
short events are those with durations @xmath272 , and would be observed as a single ` flash ' provided they occur within a time @xmath273 of any one of the @xmath266 observing runs . in the limit of small event rates the detection probability may be written as @xmath274 where @xmath275 and @xmath276 are the rates of short and long events , and @xmath277 is the rate averaged lensing duration for short events . in the ( ideal ) limit of continuous monitoring , @xmath276 approaches the total event rate and @xmath278 .
the results of our computations are displayed in figs .
[ f : cumr ] and [ f : dk ] . in fig .
[ f : cumr ] we plot the cumulative rates , @xmath279 , for all lensing events with durations @xmath273 longer than the timescale @xmath280 , as functions of @xmath280 .
we present results for events which produce unresolved and resolved images for stars which are either intrinsically below or above the detection thresholds .
we present computations for detection thresholds ranging from 16 to 19 magnitudes .
the total lensing rates , can be read off the plot from the asymptotic values of the curves as @xmath281 .
the curves are flat for timescales less than @xmath282 1 week , which is shorter than most of the event durations .
as @xmath280 increases @xmath276 decreases as a smaller fraction of the lensing events satisfy @xmath283 . as an example , fig .
[ f : cumr ] shows that for @xmath5 , the total lensing rate is equal to @xmath6 yr@xmath7 , and that for events with durations greater than 1 yr the rate is equal to @xmath9 yr@xmath7 . in fig .
[ f : cumr ] we also plot the rate - averaged lensing timescale , @xmath277 , for events with @xmath284 , which is needed to estimate the detection probability of short events .
the values of @xmath277 are almost independent of @xmath11 , since the shape of the cumulative rate function is insensitive to @xmath11 .
we note also that for small timescales @xmath285 , as would be expected for a rate which is independent of the timescale .
the rate of short events is simply @xmath286 .
thus , for @xmath5 the rate of events lasting less than 1 yr is @xmath287 yr@xmath7 , and the average duration of such events is about 3 months . in fig .
[ f : dk ] we plot @xmath288 , the median value of the maximal amplifications for long events ( @xmath289 ) , as a function of @xmath280
. long events which amplify sub - threshold stars are associated with stars at greater distances from the black hole , stars at the low velocity tail of the velocity distributions , or stars that cross closer to the line of sight to @xmath0 .
because of the latter effect , @xmath290 is greater than @xmath128 , which is the median value for all the events .
this effect is less marked for resolved lensing , which is characterized by longer timescales , and is very weak for lensing of stars which lie above the detection threshold , where @xmath290 approaches the limit @xmath291 .
figure [ f : dqdr ] shows the contributions to the lensing rate from different regions in the gc , for a specific example where @xmath292 and @xmath293yr .
several important features are apparent in the results displayed in fig .
[ f : dqdr ] and [ f : cumr ] .
first , it is evident that the cumulative rate @xmath294 approaches its asymptotic value at @xmath295pc , and that short lensing events due to stars near the black hole dominate the total rate .
for example , fig .
[ f : cumr ] shows that the median timescale of unresolved amplifications of sub - threshold stars for is @xmath208 months , and fig .
[ f : dqdr ] shows that the median distance from the black hole is @xmath296pc .
the run of the differential rate for long events , @xmath297 , with distance illustrates that the inner regions hardly contribute any long events .
second , unresolved lensing of sub - threshold stars have the shortest timescales .
this is because such events are due mainly to stars close to the black hole where the velocities are high and the lensing cross sections are small .
unresolved lensing of stars which lie above the threshold , for which the cross sections are larger , have somewhat longer timescales , and resolved events , which are due to stars at larger distances from the black hole , are longer still .
third , the rates for resolved lensing are about an order of magnitude smaller than those for unresolved lensing events .
we now apply our results to estimate the probability that the variable @xmath1-band source ( or sources ) reported by genzel et al .
( 1997 ) ( source s12 ) and ghez et al .
( 1998 ) ( source s3 ) was a microlensing event .
the monitoring of proper motions in the gc has been going on for about @xmath298yr , at a sampling interval of @xmath293yr , with a detection threshold of @xmath299 and a fwhm resolution of @xmath300 ( @xcite ) . for this detection threshold , @xmath301 ( which corresponds to a required spatial resolution @xmath302 ) , and @xmath303pc .
we can now use figs .
[ f : cumr ] and [ f : dk ] and eq .
[ e : p ] to estimate the detection probability and typical amplification of a lensing event in this experiment .
the rate of unresolved and long events which amplify sub - threshold stars is @xmath304yr@xmath7 , with a resulting detection probability @xmath305 .
the median amplifications of such events is @xmath306 above the detection threshold .
the probability for detecting a short amplification of sub - threshold stars is @xmath307 .
the probabilities for detecting unresolved events involving above threshold stars is of the same order of magnitude .
the probabilities for detecting long resolved events are an order of magnitude smaller , and the probabilities for detecting short resolved events are negligible in this experiment .
thus , we find that the behavior of the variable source ( or sources ) at @xmath0 , i.e. a brightening of a previously undetected source to @xmath308@xmath15 above the threshold with an event duration of @xmath16 yr , is the typical behavior that would be expected for a microlensing event .
however , we also find that the a - posteriori probability for detecting such an event is only @xmath17 .
the probability for detecting a lensing event can be increased considerably by carrying out more sensitive observations at higher sampling rates .
for example , 10 years of monthly observations with a detection threshold of @xmath309 will require a resolution of @xmath310 , which is already available ( @xcite ) , and will increase the total detection probability of long lensing events to @xmath311 .
these estimates are uncertain by a factor of few due to uncertainties in the stellar density distribution , the @xmath1 luminosity function and the extinction .
in the past several years , very high resolution , very deep ir observations of the gc have regularly monitored stellar motions within few arcsec of the radio source @xmath0 .
the primary objective of these efforts is to dynamically weigh the central dark mass and set lower bounds on its compactness . as a by product
, these observations can detect and record the light curves of faint time - varying sources in the inner gc over timescales of years .
these proper motion studies have recently revealed the possible presence of one or two variable @xmath1-band sources very close to , or coincident with , the position of @xmath0 ( @xcite ; @xcite ) .
the first issue to resolve , when considering results from such technologically challenging observations , is whether these sources are real , or merely artifacts of the complex procedures for obtaining deep diffraction limited images in the crowded gc .
assuming that such a source is indeed real , and that it is not simply a variable star that happens to lie very near to the line of sight to @xmath0 , there are two interesting possibilities to consider , both directly linked to the existence of a super - massive compact object in the gc .
a time variable ir source , coincident with @xmath0 , may be the ir counterpart of the radio source , with the ir flare resulting from fluctuations in the accretion process , which has up to now eluded detection in any band other than the radio ( see e.g. @xcite ) .
another possibility is that the new source is a faint star in the dense stellar cluster in the gc , which is microlensed by the black hole .
such amplification events would appear as time varying sources very close to the position of @xmath0 .
we note that a detection of microlensing could provide an independent probe of the compactness of the central dark mass .
the innermost observed stars in the gc set an upper limit on the size of the dark mass , @xmath312 , where @xmath60 is the schwarzschild radius ( @xcite ) .
microlensing has the potential , given a well sampled light curve , to improve on the dynamical limit , since for typical values of the lens source distance , @xmath313 ( see eq .
[ e : rsch ] ) .
if the central mass is not a black hole , but rather an extended object , then the microlensing light curve will deviate from that of a point - mass lens when @xmath314 . because the innermost observed stars already constrain @xmath315 , only lensed stars in the inner few pc have a small enough @xmath42 to probe possible structure in the dark mass distribution ( eq . [ e : oe ] ) .
therefore , the question whether there is a significant chance to detect such events is related to the yet unresolved question of whether there is a strongly peaked stellar cusp in the innermost gc . we will discuss these issues elsewhere . in this paper we investigated the possibility of microlensing amplification of faint stars in the dense stellar cluster in the gc by the super - massive black hole , which is thought to coincide with the radio source @xmath0 .
we calculated in detail the rates , durations and amplifications of such events , and considered separately the cases of unresolved and resolved images and of intrinsically faint and bright sources .
we presented our results in a general way that can be used to estimate the detection probabilities of microlensing for a wide range of observing strategies . the background stellar surface density increases with the detection threshold @xmath11 .
this determines the observational angular resolution required to detect a microlensing event , and fixes the maximal distance behind the black hole for which the two lensed images of a star will appear unresolved .
this maximal distance occurs before the integrated lensing rate reaches its asymptotic value , even for present - day detection thresholds .
we therefore considered two manifestations of the lensing : unresolved microlensing of stars near the black hole , and resolved lensing of stars farther away .
we find that short lensing events , due to stars close to the black hole , dominate the total lensing rate .
this reflects the fact that the high stellar density and velocities near the black hole more than compensate for the smaller lensing cross - sections there .
for this reason , and because unresolved images are on average twice as bright as the resolved images , unresolved microlensing dominates the lensing rate .
we have also considered the lensing amplification of bright observed stars .
the contribution of this type of microlensing to the total rate becomes progressively more important as the detection threshold decreases , and at low sampling rates , which are primarily sensitive to longer events .
we find that low sampling rates significantly bias the detection towards high amplitude events .
our predicted lensing rates are small , but not so small as to be negligible .
in particular , longer , deeper proper motion monitoring done at higher rates , e.g. 10 years of monthly monitoring with @xmath309 , may have a significant chance of detecting such an event . finally , could either of the variable sources reported by genzel et al . ( 1997 ) ( source s12 ) and by ghez et al . ( 1998 ) ( source s3 ) be the amplified microlensed image of a faint star ?
the lack of evidence for related variability in the radio and x - rays , as would be expected in some accretion scenarios if the @xmath1-flare were due to fluctuations in the accretion process , argues against the possibility that the new source is the ir counterpart of @xmath0 .
another possibility is that the new sources are variable stars , which are below the detection limit in their low state .
long period variable stars ( lpvs ) , which are probably luminous mira variables , are observed in the gc ( @xcite ; @xcite ) .
typical amplitude variations of @xmath316@xmath317 are observed over @xmath318@xmath319 yr , although in some cases the variations are as large as @xmath320 .
haller & rieke ( 1989 ) find 12 lpvs in a 4.5 arcmin@xmath321 survey of the gc ( not including the central @xmath322 ) down to @xmath323 ( @xmath324 ) . of these ,
only one exhibited high amplitude variations ( @xmath245 in 4 months ) . at @xmath325 ,
the new variable @xmath1 source is much fainter than the lpvs observed in this survey .
if it is an intrinsically bright lpv , it must be lie on a highly extinguished line of sight . using the observed surface density of @xmath326 , @xmath327 lpvs , it is possible to make a rough estimate of the probability for finding such a star within @xmath328 of @xmath0 . even after taking into account a factor @xmath329 difference in the surface mass density between the dynamical center and the survey area at @xmath330 from the center ,
the probability is only @xmath331 .
this of course does not rule out the possibility of an intrinsically faint but highly variable star . in any case , if the new source is a variable star , future observations should detect continued variability from this source .
we can also argue statistically against the possibility that these flares are due to microlensing of a star by another star in the gc , which happens to be close to the line - of - sight to @xmath0 . because @xmath332 , the typical lensing timescale by a star will be of the order of @xmath318 hr , @xmath333 times shorter than that of lensing by the black hole ( eqs .
[ e : re ] and [ e : rate ] ) .
furthermore , it is straightforward to show that the total rate of lensing of a star by a star within an angle @xmath21 of the line - of - sight to @xmath0 is @xmath334 yr@xmath7 for @xmath265 , which is 100 time smaller than the rate of lensing by the black hole .
our analysis has shown that the behavior of the variable @xmath1-band source ( or sources ) at @xmath0 , in particular a brightening of a previously undetected source to @xmath335@xmath15 above the threshold , on a timescale of @xmath16 yr , is the typical behavior that would be expected for a microlensing event .
however , we also estimate that the probability that such an event could have been observed during the course of the proper motion studies that have been carried out thus far is only @xmath336 . while this probability is small , it is not so small as to rule out this possibility entirely .
the probability of detecting a microlensing event at @xmath0 will increase considerably in the future as the observational sensitivities and the monitoring sampling rates improve .
we thank d. backer , r. blandford , a. eckart , r. genzel , a. ghez , i. king , e. maoz , b. paczyski and e. serabyn for helpful discussions and comments .
this work was supported in part by a grant from the german - israeli foundation ( i-551 - 186.07/97 ) .
a. s. thanks the radio astronomy laboratory at u.c . berkeley , and the center for star formation studies consortium for their hospitality and support .
figure [ f : cumr ] shows that the total lensing rate is dominated by short events ( shorter than a few month ) .
such events are shorter than the current mean time between observations , and can therefore be observed only in the rare cases where they occur simultaneously with an observing run ( [ s : rate ] ) . of more relevance
are long events , which span two or more observing runs .
longer events occur when the lensed star is either slow , or when it is far away from the black hole , so that its trajectory through the einstein ring is long ( eq . [ e : re ] ) , or when it passes close to the optical axis ( eq . [ e : t0c ] ) . in this appendix
we calculate the rate of events that are longer than some specified time , and estimate their median maximal amplification .
microlensing events with @xmath289 are those with transverse velocity @xmath120 and impact parameter @xmath337 such that @xmath338 where @xmath84 is the stellar type and @xmath89 is the maximal impact parameter for amplification above the threshold .
the local rate per stellar type is @xmath339 where @xmath340 and @xmath341 is the 2d distribution function of velocities .
the 2d velocity @xmath342 is composed of an ordered component @xmath343 , which we assume to be perpendicular to the line of sight , and a random isotropic component , which we assume to be gaussian with a 1d dispersion @xmath344 .
the 2d distribution function of the projected velocity in polar coordinates is @xmath345 where @xmath346 , @xmath347 and @xmath348 is the angle between @xmath342 and @xmath343 . upon substitution into eq .
[ e : tmin ] and averaging over all stellar types , one obtains @xmath349 where @xmath350 , and @xmath351 where @xmath352 is a modified bessel function of order @xmath266 . in the simple case where @xmath353 , @xmath354 in the limit of @xmath355 ( @xmath356 ) , @xmath357\,,\ ] ] which is independent of @xmath89 ,
so that @xmath358 .
it is straightforward to verify that in this limit , the mean transverse velocity approaches its correct asymptotic values , since @xmath359 a fraction of the long duration events are due to stars on trajectories with smaller than average impact parameters .
therefore , the median impact parameter of long events is smaller than @xmath360 and the median amplification above the threshold is greater than 2 .
the local average impact parameter is @xmath361 where the weight function @xmath362 can be written as @xmath363 it is straightforward to verify that @xmath364 the total rate averaged impact parameter is @xmath365 and the median amplification above the threshold can be estimated by @xmath366 .
backer d. c. 1996 , in iau symp .
169 , unsolved problems of the milky way , ed .
blitz l. & teuben p. , ( kluwer : dordrecht ) , 193 bartelman m. & narayan r , 1997 , in formation of structure in the universe , ed .
dekel a. & ostriker j. p. , ( cambridge university press : cambridge ) , in press ( astro - ph/9606001 ) bedijn p. j. , 1988 , , 205 , 105 blum r. d. , sellgren k. , & depoy d. l. , 1996 , apj , 470 , 864 carney b. , fulbright j. p. , terndrup d. m. , suntzeff , n. , & walker , a. 1995 , aj , 110 , 1674 charlot s. , & bruzual g. a. , 1991 , , 367 , 121 crotts a. s. p. , 1992 ,
apj , 399 , l43 davidge t. j. , simons d. a. , rigaut f. , doyon r. , & crampton d. , 1997 , aj , 114 , 2586 eckart a. , genzel r. , 1996 nature , 383 , 415 eckart a. , genzel r. , hofmann r. , sams b. j. & tacconi - garman l. e. 1993 , apj , 407 , l77 eckart a. , genzel r. , hofmann r. , sams b. j. & tacconi - garman l. e. 1995 , apj , 445 , l23 genzel r. , hollenbach d. , & townes c. h. , 1994 , rep .
phys . , 57 , 417 genzel r. , eckart a. , ott t. & eisenhauer f. 1997 , mnras , 291 , 219 genzel r. , thatte , n. , krabbe a. , kroker h. & taconi - garman l. e. 1996 apj , 472 , 153 ghez a. m. , klein b. l. , morris m. , & becklin e. e. 1998 , apj , in press haller j. w. , & rieke m. j. , 1989 , in iau symp .
136 , the center of the galaxy , ed .
morris m. , ( kluwer : dordrecht ) , 487 henry t. j. & mccarthy jr .
d. w. , 1993 , aj , 106 , 773 holtzman j. a. , et al . , 1998 ,
aj , 115 , 1946 kent s. m. , 1992 , apj , 387 , 181 krabbe a. , et al . , 1995 ,
apj , 447 , l95 lebofsky m. j. & rieke g.h . , 1987 , the galactic center , ed .
backer d. c. , ( aip : new york ) , 79 maoz e. , 1998 , apj , 494 , l181 melia f. , 1994 , apj , 426 , 577 miller g. e. , & scalo j. m. , 1979 , apjs , 41 , 513 morris m. , & serabyn e. , 1996 , ara&a , 34 , 645 narayan r. , yi i. , & mahdevan r. 1995 , nature , 374 , 623 paczyski b. 1986 , apj , 304 , 1 ramirez , s. v. sellgren k. carr j. balachandran s. blum r. d. terndrup d. , 1998 , the central parsecs galactic center workshop , eds . h. falcke a. cotera w. huschl f. melia & m. rieke , to appear in the asp conf .
. rieke g. h. , rieke m. j. & paul a. e. , 1989 , apj , 336 , 752 scalo j. m. , 1986 , fundam .
phys . , 11 , 1 schaerer d. , charbonnel c. , meynet g. , maeder a. , & schaller g. , 1993 , s , 102 , 339 schmidt - kaler t. h. , 1982 , in landolt - bornstein new series , astronomy and astrophysics , vol
. 2b , ( springer : new york ) , 451 serabyn e. , & morris m. 1996 , nature , 382 , 602 serabyn e. , shupe d. , & figer d. f. 1998 , nature , 394 , 448 sternberg a. , 1998 , apj , 506 , 721 tamura m. , werner m. w. , becklin e. e. , phinney e. s. , 1996 , apj , 467 , 645 tiede g. p. , frogel j. a. , & terndrup d. m. , 1995 , aj , 110 , 2788 tokunaga a. t. , 1998 , in astrophysical quantities , 4th ed . , ed .
a. cox , ( aip press ) , in preparation wardle m. & yusef - zadeh f. 1992 , apj , 387 , l65
| we investigate microlensing amplification of faint stars in the dense stellar cluster in the galactic center by the super - massive black hole , which is thought to coincide with the radio source @xmath0 .
such amplification events would appear very close to the position of @xmath0 , and could be observed , in principle , during the monitoring of stellar proper motions in the galactic center .
we use observations of the near - infrared @xmath1-band ( 2.2@xmath2 m ) luminosity function in the galactic center and in baade s window , as well as stellar population synthesis computations , to construct empirical and theoretical @xmath1 luminosity function models for the inner 300pc of the galaxy .
these , together with the observed dynamical properties of the central cluster and inner bulge , are used to compute the rates of microlensing events which amplify stars with different intrinsic luminosities above specified detection thresholds .
we present computations of the lensing rates as functions of the event durations , which range from several weeks to a few years , for detection thresholds ranging from @xmath3 to @xmath4 .
we find that events with durations shorter than a few months dominate the lensing rate because of the very high stellar densities and velocities very near the black hole , where the effective lens size is small .
for the current detection limit of @xmath5 , the total microlensing rate is @xmath6 yr@xmath7 . the rate of events with durations @xmath8 yr is @xmath9 yr@xmath7 .
the median value of the peak amplification for short events is @xmath10 above the detection threshold , and is only weakly dependent on @xmath11 .
long events are rarer , and are associated with more distant stars , stars at the low velocity tail of the velocity distribution , or stars that cross closer to the line - of - sight to @xmath0 .
therefore , the median peak amplifications of long events are larger and attain values @xmath12 above the threshold .
recent proper motion studies of stars in the galactic center have revealed the possible presence of one or two variable @xmath1-band sources very close to , or coincident with , the position of @xmath0 ( eckart & genzel 1996 ; genzel et al . 1997 ; ghez et al .
1998 ) .
these sources may have attained peak brightnesses of @xmath13 , about @xmath14@xmath15 above the observational detection limits , and appear to have varied on a timescale of @xmath16 yr .
this behavior is consistent with long - duration microlensing amplification of faint stars by the central black hole .
however , we estimate that the probability that a single such an event could have been detected during the course of the recent proper motion monitoring campaigns is @xmath17
. a ten - fold improvement in the detection limit and @xmath18yr of monthly monitoring could increase the total detection probability to @xmath19 . |
understanding the relationship between agn , which are powered by accretion onto supermassive black holes , and the stellar mass galactic black hole binary systems ( gbhs ) is currently one of the major research areas in high energy astrophysics .
if we can understand the relationship , then we can predict how agn should behave on cosmological time - scales by observing the brighter , and much faster varying , gbhs .
the comparison is complicated by the fact that ( at the risk of some oversimplification ) , gbhs can exist in a number of states , defined originally in terms of their medium energy ( 210 kev ) x - ray properties , particularly : ( i ) the hard spectrum , low - flux ( ` hard ' ) state , ( ii ) the soft spectrum , high flux ( ` soft ' ) state and ( iii ) the ` very high state ' ( vhs - sometimes also described as the high - intermediate state ) where the flux is very high and the spectrum very soft . in general the accretion rate increases as we go from the hard to soft to vhs states .
see @xcite for a comprehensive description of gbh states and behaviour .
the various states have quite distinct x - ray timing properties which may , in fact , provide a more fundamental state discrimator than the spectral properties .
timing is usually quantified in terms of the power spectral densities ( psds ) of the x - ray lightcurves . where significant variability is observed , ` soft ' state psds are characterised by a power law of slope of -1 at low frequencies ( i.e. the power , @xmath0 with @xmath1 ) , breaking to a slope @xmath2 above a characteristic frequency @xmath3 ( or time - scale @xmath4 ) .
hard state and vhs state gbh psds , which are similar to each other , are more complex .
they are best described by the combination of two or more lorentzian - shaped components ( e.g. * ? ? ?
* ) , although at low signal to noise they can be approximated by a power law with a second , lower frequency break , approximately 1.5 to 2 decades lower in frequency than the high frequency break .
so far , almost all well observed agn , which are x - ray bright and have moderately high accretion rates , have soft state psds , as we would expect from comparison with gbhs of similar accretion rate ( measured relative to the eddington accretion rate , @xmath5 ) ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however the comparison for the hard and vhs states is largely unknown . until very recently ,
ngc3783 had been suggested to have two breaks in its psd @xcite and so be a hard state agn , although its accretion rate is similar to that of the well observed soft state agn .
however recent work , based on improved data @xcite , has shown that , in fact , ngc3783 is another soft state agn .
if soft states are linked to the accretion disc extending in to very small radii , then perhaps their much cooler accretion discs compared to gbhs may allow the optically thick disc to survive without evaporation to smaller radii .
thus maybe all agn are soft state objects and the comparison with gbhs does not extend to other states ? such a conclusion would have a fundamental impact on our understanding of the accretion process onto black holes and so it is very important to confirm , from timing properties , a definite non - soft state agn .
the only remaining case where there is at least some evidence for a non - soft state psd is . from observations with , evidence for a low frequency break in the psd was presented by @xcite .
evidence for a second break , at high frequencies , was presented by @xcite from analysis of observations .
@xcite interpreted the combined and psd as a hard state psd but did note that the high frequency break does not scale linearly with mass to the hard state of the archtypal gbh cyg x-1 ; the frequency difference is too small .
however as is one of the highest accretion rate agn known ( @xmath6 , @xcite ) , a vhs state remains a strong possibility .
unfortunately proper modelling of the combined long and short time - scale psd is hampered by the gaps which occur in the light curves , and hence in the psd , at the orbital period ( @xmath7s ) .
we therefore made a 100 ks continuous observation with to fill that gap in the psd .
spectral analysis of this observation has been presented by @xcite and @xcite and cross - spectral analysis ( i.e. lags between energy bands as a function of fourier frequency or time - scale ) has been presented by @xcite .
here we present a full power spectral analysis of the combined , and data and find firm evidence for two breaks in the combined psd .
we also show that the psd can be parameterised as the sum of two lorentzian components .
we note particularly that the highest frequency lorentzian , for which good spectral information is available from both the and observations , is stronger at higher energies , exactly as is found for lorentzian components in gbhs .
in gbhs in the hard and vhs states , changes in the lag between bands occur at approximately the time - scales at which the psd changes from one lorentzian component to another , supporting the model that each lorentzian comes from a separate physical emission region . in this paper
we therefore re - examine the lag spectrum and compare it with the psd , to test whether a two - component emission region may apply in and also as a further diagnostic of the state of the agn .
the paper is arranged as follows : in sec .
[ data ] , we describe the observational data sets which we used and discuss the individual and psds in secs [ s_ascapsd ] and [ s_xmmpsd ] .
we then discuss the joint and power spectrum and its energy dependence in sec .
[ s_combinedpsd ] and compare the powerspectrum with the lag spectrum in sec [ s_lags ] .
we summarise our conclusions in sec .
[ s_conclusions ] .
[ [ section ] ] was observed by from 1999 january 1 to 2003 march 4 . for most of this
period observations were carried out approximately once every 4 days but from 2000 may 31 to 2000 july 01 , observations were carried out 8 times per day .
the observations from 1999 january 1 to 2000 september 19 are described in detail in @xcite .
we have extracted all of the data from the archive and produced 210 kev lightcurves in the same manner as we have described for other agn monitoring programmes ( e.g. see * ? ? ? * for details ) . for use in power
spectral analysis we produce a long time - scale light curve , covering the full observing period with 4-day sampling , and an intermediate time - scale lightcurve , covering the one month period of 3-hour sampling .
[ [ section-1 ] ] was monitored with between 2000 june 1 and july 5 , producing a month - long light curve .
this observation was only interrupted by the regular earth occultations of the satellite on the orbital time - scale of 5632s . binning the data in orbit - long bins
produces an evenly sampled light curve , containing 551 points , that probes time - scales from a few times @xmath8 to @xmath9 s. additionally , each of the 551 individual orbits cover time - scales of @xmath10 s , whose power spectra can be combined to produce a high quality high frequency psd .
we downloaded sis data from the tartarus database in 0.62 and 210 kev energy bands and produced background subtracted light curves , with average count rates of 2.7 and 0.6 c / s respectively .
the average photon energies of these two bands are 1.18 and 3.2 kev respectively .
the lightcurves , psd and other variability properties of this data set have been studied by e.g. @xcite .
[ [ section-2 ] ] was observed by from 2005 january 5 , 19:47 , to 2005 january 6 , 23:16 .
the european photon imaging cameras ( epic ) pn , mos1 and mos2 cameras were all operated in small window mode using a medium filter , as described by @xcite .
the total exposures for the pn and mos cameras were 98.8 ks and 99.1 ks respectively . for the pn camera , we extracted source photons from a @xmath11 region and the background was selected from a source - free region of equal area on the same chip .
only single and double events , with quality flag=0 were retained .
for the mos1 and mos2 cameras , source photons were extracted from a circular region of @xmath12 in radius and single , double , triple and quadruple events were used .
the pn data was free of pile - up , but both mos cameras suffered significant pile - up in the cores of the target psfs .
therefore we discarded data from the central @xmath13 diameter of each mos exposure . the average pn background - subtracted count rate in the 0.210 kev band was @xmath14 c / s .
the background level was low and almost constant throughout the observation .
we produced light curves in the 0.22 and 210 kev energy bands in 48 s bins , and combined the light curves from the different detectors to produce weighted average light curves .
the 0.210 kev lightcurve is shown in fig [ fig : xmmlc ] ( see also * ? ? ?
as the variability properties normally depend on energy band , we also constructed light curves from the pn data alone in the 0.52 and 28.8 kev bands , to match the average energies of the light curves .
we used these lightcurves in all quantitative analyses .
the superposed psds derived from the 210 kev and 0.62 kev data are shown in fig . [ fig : ascapsds ] .
data below @xmath15hz result from the light curves binned on the orbital time - scale .
data above @xmath16hz are the sum of the psds derived from the many short sub - orbital light curves .
note the frequency gap that these data do not cover .
the poisson noise level has been subtracted . on time - scales
longer than the orbital time - scale , the low and high energy psds have a similar normalisation .
however on shorter time - scales the 210 kev psd lies above the 0.62 kev psd , except perhaps at the very highest frequencies ( @xmath17hz ) where the two psds may be converging .
this frequency dependent excess of the high energy psd over the low energy psd both below and above the high frequency break ( @xmath18hz ) , which is different from the behaviour of other well observed agn , ( e.g. ngc4051 , * ? ? ?
* ) , will be modelled properly in section [ s_combinedpsd ] .
the light curves were almost completely continuous except for 5 small gaps of duration 20 - 30 secs each . following the procedure established in @xcite these gaps were filled by linear interpolation , together with the addition of noise characteristic of the surrounding data points , in order to eliminate spurious high frequency power in the power spectrum .
power spectral densities were derived from these light curves using a standard direct fourier transform . in fig .
[ fig : xmm210psds ] we show the 0.22 and 210 kev psds and we see close agreement with the psds in the same energy bands . as with the psds ,
the low and high energy psds coincide very closely at frequencies below @xmath15hz , but at higher frequencies the 210 kev psd exceeds the lower energy psd .
note that here , for the low energy band , we show the 0.22 kev psd rather than the 0.52 kev psd and note that the bump centred around @xmath19hz is even less pronounced in the 0.22 kev band , strengthening our observation that the excess has a strongly energy - dependent shape .
although not shown , the 0.52 kev psd is almost identical to the 0.62 kev psd .
a simple power law fit to both low and high energy psds is a very poor fit , leaving a large residual bump centred around @xmath19hz .
a bending powerlaw , however , which provides a good fit to the high / soft state of cyg x-1 and to the psds of ngc4051 @xcite and 6 @xcite , is a good fit ( table [ tab : brpl ] ) . tying the break frequencies to be the same in both energy band psds , we note ( as can also be seen directly from fig .
[ fig : xmm210psds ] ) , that although the slopes above the break are similar in both energy bands , the slope below the break is noticeably flatter at higher energies .
the psds , although less well defined because of the gap on the orbital time - scales , are quite consistent with the fit parameters given in table [ tab : brpl ] .
this behaviour is different from that of ngc4051 and mcg-6 - 30 - 15 ( e.g. see table 6 of * ? ? ?
* ) where , although the slopes below the break are not very well determined , there is no evidence for a variation with energy . above the break , however , the slope is flatter at higher energies in ngc4051 and mcg-6 - 30 - 15 .
the psd slope variations seen here in can be more naturally interpreted in terms of a power law ( which is necessary to explain the power at low frequencies ) together with a lorentzian component which is stronger at higher energies .
such a model , in which the power law slope and normalisation , and lorentzian central frequency are tied , provides an equally good fit to the data . in the case of the latter model
, we note that the lorentzian component is approximately three times stronger in the 2 - 8.8 kev than 0.52 kev band ( table [ tab : lor ] ) .
an increasing lorentzian amplitude at higher energies is also typically observed in qpos and some broad lorentzian components in gbhs ( e.g. * ? ? ?
* ) . as the exact values of the fit parameters depend on proper fitting of the lower frequency data from and we leave further discussion to section [ s_combinedpsd ] .
there is a residual to the power law plus lorentzian fit to the psd at @xmath20hz
. this residual can be fitted ( using the unbinned data ) by the addition of a second , narrow ( width @xmath21hz ) , lorentzian with normalisation approximately half that of the broad lorentzian .
the fit is shown in fig .
[ fig : xmm2lor ] . the fit is improved but , adjusting the f - test probabilities to take account of the fact that the line could have appeared in any of the available frequency bins , we find that the improvement is significant only at the 90% confidence level . [ cols="^,>,>,^,^,^ " , ] [ tab : lor ]
the data , which covers time - scales from @xmath22 years to 3 hours , has already been analyzed by @xcite and , including further data , by @xcite who both fit a breaking power law model to the psd .
@xcite derive a slope at the lowest frequencies @xmath23 of @xmath24 , a break frequency , @xmath25hz and an intermediate slope above the break of @xmath26 .
the earlier results of @xcite are consistent with these values .
we begin by refitting these data using our standard simulation - based modelling method ( e.g. see * ? ? ?
we fit a simple bending power law model ( e.g. see * ? ? ?
* ; * ? ? ?
we bin the long time - scale light curve into 4-d bins , and the short time - scale light curve into 3-h bins .
the resultant best - fitting values are similar to those of @xcite and the errors in the fit parameters are similarly large . as our value of @xmath23
is quite consistent with zero and as , in both their hard and very high states , gbhs have @xmath27 , we fix @xmath27 in subsequent power law fits .
we then fit the combined , and data sets with a doubly bending power law model , including a second , higher frequency break at @xmath28 and slope @xmath29 at the highest frequencies .
we include the 210 kev sis data in 512s bins and the pn data in 48s bins in the 28.8 kev band to match the average energy of the band .
the pn bin value was chosen so that the resultant light curve has @xmath30 points so that we can use the fast fourier transform , rather than the slower direct fourier transform , for that part of the simulation , without loss of accuracy .
we obtain a very good fit ( acceptance probability p=75% - see table [ tab:2lor ] and fig . [ bendingpsd ] ) with @xmath31 , @xmath32hz and @xmath33 hz .
these break frequencies are in good agreement with the values obtained by @xcite from analysis of data . in order to make a combined psd fit of the soft x - ray band , we first compared the low frequency variability power of the sis data in the 0.62 and 210 kev bands .
the resultant power spectra showed identical amplitudes , so both energy bands can be fit together at low frequencies without needing to rescale their psds .
we assume that the close similarity of the 0.62 and 210 kev lightcurves continues to even lower frequencies , i.e. those sampled by the long term data .
we therefore fitted the 210 kev long term data together with the sis and 0.62 kev data to cover the same frequency range as above .
the fit is again very good ( p=85% - table [ tab:2lor ] ) giving parameters @xmath34 , @xmath35 hz and @xmath36 hz . the steeper intermediate slope reflects the lower high - frequency power in the soft band compared to the hard band .
the break frequencies in both bands are consistent within 1 @xmath37 uncertainties , as can be seen from the confidence contours plotted in fig .
[ dbknpo_contour ] motivated by the strong evidence for a lorentzian component in the data alone , we have fitted the combined , and data with a model consisting of two lorentzian components , @xmath38 where the lorentzian peak frequency @xmath39 is related to the centroid frequency @xmath40 by @xmath41 and the subscripts l and h refer to the low and high frequency components , respectively .
we fitted the combined psd in soft and hard energy bands independently , allowing all parameters to vary .
the low frequency component converged to very low values of the quality factor @xmath42 , i.e. to very wide lorentzians where the lorentizan shape is almost insensitive to the value of @xmath43 .
we therefore fixed this parameter at @xmath44 ( with 90% and 99% upper confidence limits of @xmath45 and 0.5 respectively ) for both energy bands to enable a more detailed parameter grid search for the other parameters , and hence to obtain better constraints on these parameters .
the best - fitting values are listed in table [ tab:2lor ] .
the lorentzian peak frequencies are similar to , but not exactly the same as , the corresponding break frequencies , with the lorentzian frequencies lying closer to the centre of the band - limited power . figures [ hardpsd ] and [ softpsd ] show the best - fitting two - lorentzian models to the hard and soft band respectively . in this interpretation ,
the increase in variability power at high frequencies in the hard band is entirely due to the greater strength of the high frequency lorentzian in this band .
the peak frequencies in both energy bands are consistent within the 90% confidence limits ( see contour plot in fig.[2lor_contour ] ) .
note that figs .
[ hardpsd ] and [ softpsd ] show the psd unfolded through the distortions produced by sampling effects .
these distortions also depend on the underlying psd shape so the resultant unfolded psd depends on the model being fitted .
if the two - lorentzian model is the correct underlying psd , the strong power component around @xmath46hz will lead to spurious apparent power at lower frequencies in a psd which has not been unfolded from the distortions , and where the low frequency part of the psd has been derived from observations which do not sample the @xmath46hz region .
this effect is known as aliasing ( e.g. see * ? ? ? * for more details ) and here particularly affects the long timescale data ( blue crosses in figs [ bendingpsd ] , [ hardpsd ] and [ softpsd ] ) .
when we unfold the psd we remove this spurious low frequency power .
thus the lowest frequency part of the psds in figs .
[ hardpsd ] and [ softpsd ] lies below the same part in the bending power law model ( fig.[bendingpsd ] ) , where the model did not contain so much power at higher frequencies .
therefore the low break frequency in fig .
[ bendingpsd ] is lower than in figs .
[ hardpsd ] and [ softpsd ] .
since the normalisations of low and high - frequency lorentzians are highly correlated with each other and with the lorentzian frequency , we only treat the ratio of the low and high - frequency lorentzian normalisations , @xmath47 , as a free parameter and do not fit the high and low - frequency normalisations separately . although the 1-dimensional errors quoted in table 3 seem to indicate that the normalisation ratios for the hard and soft band overlap , we show in fig . [ relnormcon ] that the 90% confidence contours of @xmath48 versus @xmath47 do not overlap .
therefore , for the same @xmath49 , the high - frequency lorentzian is always relatively stronger , compared to the low - frequency lorentzian , in the hard band than in the soft band .
in fact , the low - frequency lorentzian normalisation is similar in both bands and relatively well - constrained . for the best fitting model parameters , the values of fractional rms in the low frequency lorentzian are @xmath50% and @xmath51% in hard and soft bands respectively , and in the high frequency lorentzian the values are @xmath52% and @xmath53% in hard and soft bands .
thus the low frequency lorentzian rms values are consistent with being the same , while the high frequency lorenztian is weaker in the soft band , which explains the different hard and soft band psd shapes at high frequencies .
ccccccc * doubly - bending power law * + energy band & @xmath54[hz]&@xmath55[hz]&@xmath56&@xmath57&@xmath58 & fit probability + 0.62.0&3.6@xmath59 & 1.8@xmath60 & 0.0 & 1.5@xmath61&3.4@xmath62&0.85 + 2.08.8 & 7.5@xmath63 & 2.4@xmath64 & 0.0&1.2@xmath65&4.2@xmath66&0.75 + + * two - lorentzian model * + & & + energy band & @xmath67 [ hz]&@xmath68 & @xmath69[hz]&@xmath70 & @xmath71&fit probability + 0.62.0&@xmath72&@xmath73&@xmath74&@xmath75&@xmath76&0.37 + 2.08.8 & @xmath77 & 0.01&@xmath78 & @xmath79&@xmath80&0.40 + * the errors are large and are not very well defined .
+ all errors are 68% confidence ( @xmath81 ) .
[ tab:2lor ]
in galactic black hole systems , lags are observed between light curves in different energy bands , with the higher energy variations generally lagging the lower energy variations ( referred to here as a ` positive ' lag ) .
if the light curves in the different bands are separated into different fourier components , it is generally found that the time lags between bands vary as a function of the fourier frequency , i.e. time - scale , of the component , with larger lags being associated with longer time - scales .
( see @xcite for a full discussion of the relevant analysis techniques ) .
such lags have now been measured in some agn ( e.g * ? ? ?
* ; * ? ? ?
* ) , including in @xcite .
@xcite show that , for , the positive time lag between the 0.52 and 28.8 kev bands is approximately 800 s at low fourier frequencies and @xmath82s at high frequencies , with a sharp drop in time lag values at a frequency of @xmath83 hz .
this stepped appearance of the lag spectrum resembles that seen in cyg x-1 in the low / hard and intermediate states , where the steps in the lag spectrum correspond to the transition frequencies between peaks in the corresponding psds @xcite .
the correspondence between lag spectrum steps and psd peaks suggests that each constant - lag value is associated with a given variability component , seen as a lorentzian component in the psd .
therefore , the shape of the lag spectrum in suggests that the two - component interpretation of the psd is correct . to test the above hypothesis we directly fitted the lag spectrum with a model corresponding to the lags expected from two lorentzian variability components , each with a single distinct time lag , which is constant for all fourier frequencies .
the observed lag at any particular frequency is then a function of the overlap between the two lorentzians , and must be calculated in the complex fourier domain , by evaluating the cross - spectrum as a function of fourier frequency , @xmath84 .
the cross - spectrum is given by @xmath85 , where @xmath86 and @xmath87 are the fourier transforms of the hard and soft band light curves and the asterisk denotes the complex conjugate .
@xmath88 where the subscripts @xmath89 and @xmath90 denote the low and high - frequency lorentzians respectively , and similarly for the hard band , @xmath91 . for the low and high frequency lorentzian components ,
the soft band light curve is correlated with the hard band light curve except with a phase lag @xmath92 , @xmath93 respectively .
we determine the phase lag from the fixed low and high - frequency lorentzian _
time _ lags @xmath94 and @xmath95 , by using @xmath96 . to avoid clutter , we now drop the frequency - dependence of the various parameters and write the real part of the cross - spectrum as follows : @xmath97[re(h_l)+re(h_h)]\nonumber\\ & + & [ im(s_l)+im(s_h)][im(h_l)+im(h_h)]\\ & = & [ |s_l| \cos(\phi_0+\phi_l)+|s_h| \cos(\phi_1+\phi_h)]\nonumber\\ & \times&[|h_l| \cos(\phi_0)+|h_h| \cos(\phi_1)]\nonumber\\ & + & [ |s_l| \sin(\phi_0+\phi_l)+|s_h| \sin(\phi_1+\phi_h)]\nonumber\\ & \times&[|h_l| \sin(\phi_0)+|h_h| \sin(\phi_1 ) ] , \end{aligned}\ ] ] where @xmath98 and @xmath99 denote real and imaginary components . @xmath100 and @xmath101 are simply random phase values for the low and high frequency lorentzians respectively , measured for each realisation of the light curves , and will be different for different realisations .
they are not related to the phase lags of interest between bands , i.e. @xmath102 and @xmath103 for the low and high frequency lorentzians . for a noise process ,
such as the broad lorentzians observed here , @xmath100 and @xmath101 , in independent fourier frequencies , are independently and identically distributed over a uniform distribution between 0 and @xmath104 .
the coefficients of the cross terms @xmath105 and @xmath106 will average to 0 when taking the average value of the cross spectrum for many realizations of the light curves , or many frequencies within a frequency bin , and the average value of @xmath107 is then @xmath108 similarly we can show that @xmath109 is given by @xmath110 @xmath111 , @xmath112 , @xmath113 and @xmath114 correspond to the square root of the power of those lorentzians in each energy band , at the frequency bin to be evaluated . the observed phase lag is simply the argument of the cross - spectrum , so the observed frequency - dependent time lag , @xmath115 can be evaluated from equations [ eqn : rec ] and [ eqn : imc ] using : @xmath116 which can be fitted directly to the observed lag spectrum using the best - fitting lorentzian psd model obtained in section 5.2 and letting the low and high - frequency time lags be free parameters .
this simple model reproduces well the shape of the observed lag spectrum , shown in fig .
[ fitlags ] , with a reduced @xmath117 for 17 d.o.f ( fit probability=0.15 ) .
the best - fitting values are @xmath118 s for the low - frequency lag and @xmath119 s for the high - frequency lag .
the low frequency components fitted to the psd data in both energy bands are identical , but the high frequency components had different @xmath43 and peak frequency values , although they agree at the 1@xmath37 level . forcing the high frequency components to be equal in both energy bands , using @xmath120 and @xmath121 hz ( both within 1@xmath37 of the best - fitting values ) , produces a better fit to the lag spectrum , with reduced @xmath122 for 17 d.o.f .
( fit probability = 0.32 ) , and lags consistent with the previous fit .
the consistency between the lag spectrum and psd significantly strengthens the interpretation of the psd of as the summation of lorentzian components , and hence as a very high state system .
the variability in the soft and hard bands is relatively coherent .
measuring the degree of linear correlation between energy bands as a function of fourier frequency shows that the light curves remain coherent ( coherence , @xmath123 where @xmath124 for uncorrelated light curves and @xmath125 for perfectly correlated light curves ) up to frequencies of at least @xmath19 hz ( see figure 3 in @xcite ) .
the coherence is consistent with the two - component interpretation , even if the variability patterns of the low and high frequency lorentzians are uncorrelated .
for example , if one energy band is fully modulated by one of the lorentzians , while at the same time - scales the other energy band responds equally to both lorentzians , the coherence would drop to a value of @xmath126 .
this extreme case never occurs in the present data , so the coherence can remain over 50 % even if the lorentzians are uncorrelated .
the relative contribution of the lorentzians to each energy bands is most different at frequencies @xmath127 hz , where both lorentzians have similar power in the hard band ( figs.[hardpsd ] and [ softpsd ] ) , but the high - frequency lorentzian is 10 times stronger than the low frequency component in the soft band . in this case
, the coherence is expected to drop to @xmath128 , as measured by @xcite ( their figure 3 ) .
it is possible that the low - frequency variability component that we fit to the psd as a very broad lorentzian actually contains more components .
the available data , however , are not good enough to disentagle extra components in the low frequency psd . unfortunately the lag spectrum can not help in this case as the long term monitoring was performed only by , which has a hard energy response .
we are therefore unable to compute a lag spectrum between hard and soft bands as we have been able to on intermediate and short time - scales .
we have fitted the psd of over the @xmath129 hz frequency range , combining long term archival monitoring data , a month - long observation by covering intermediate frequencies and a new 100 ks observation by which covers the shortest time - scales .
we confirm the existence of two breaks in the psd , which make this the only agn so far observed to have band - limited variability . the psd can be fitted either with a double - bending power law , where the variability power drops at low and high frequencies , or with a combination of two distinct variability components , peaking close to the low and high frequency breaks , respectively . using two lorentzian - shaped components
, we obtain a good fit in both the soft 0.52 and hard 210 kev bands , with self - consistent lorentzian parameters between the energy bands .
we note that , when fitting the upper break in the psd , the frequency of the fitted lorentzian is a factor of 4 lower than the fitted bend frequency as , in any band - limited psd , the lorentzian fits will tend to move towards the centre of the band . this difference should be considered when combining values of psd break timescales derived from different fitting methods . in the lorentzian interpretation of the psd , the increase in variability power at high frequencies in the hard band is explained by a strengthening of the high frequency lorentzian , as is seen in gbhs ( e.g. * ? ? ?
* ) , while the low frequency lorentzian remains approximately at the same level . the two - component interpretation is strongly supported by the lag spectrum calculated between these energy bands using and data ( arvalo et al .
the lag spectrum has a step - like shape , where the time lags are approximately constant at low frequencies and drop sharply to a different constant value at around @xmath15 hz .
the drop in the lag spectrum occurs exactly at the frequency expected if each lorentzian component has a single time lag value associated with it .
this characteristic lag spectral shape therefore provides strong evidence that the different lorentzian components in the psd arise from distinct physical regions , e.g. different radii in the accretion disc .
the lag spectrum of ark 564 has a very similar shape to that of cyg x-1 in its intermediate state ( see figure 8 of * ? ? ?
* ) , although the frequencies are , of course , very different .
however our recent 3d fit @xcite shows that psd break frequencies scale not only inversely with black hole mass , @xmath130 , but also approximately with @xmath5 . if we apply the same scaling to the lag frequencies , then ark 564 and cyg x-1 in its intermediate state coincide almost exactly .
ark 564 was included in our 3d fit , using the value of @xmath4 given by @xcite , and lay close to the best - fitting plane .
the better measured value of @xmath4 which we derive here , on the assumption of a bending power law , is identical to the value of @xmath4 given by @xcite thus ark 564 remains consistent with our previous scaling relationship .
all previous studies of the scaling between agn and gbhs have concentrated on the high frequency part of the psds as the ( scaled ) low frequency part of most agn psds is not well measured .
on the basis of just one agn where we are able to fit a low frequency lorentzian component we are therefore extremely cautious in any comparison , and we also note that intermediate state gbh psds show many and varied shapes where the separation between high and low frequency lorentzians is not constant ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however the separation of approximately 1.5 decades in frequency between the two lorentzians seen here is not wildly different from typical separations seen in intermediate state gbhs .
thus some sort of mass ( and probably also accretion rate ) related scaling probably applies to the lower , as well as to the upper , psd frequencies .
the psd and lag spectrum indicate a vhs / intermediate state , rather than a soft state , with a hard state being ruled out because of the very high accretion rate .
is this state classification consistent with other state measures , e.g. radio luminosity ?
vhs / intermediate state gbhs can be quite luminous , although transient , radio sources ( e.g. grs1915 + 105 , * ? ? ?
* ) whilst soft state gbhs are undetected in deep radio observations .
ark 564 ( black hole mass @xmath131 , * ? ? ?
* ) is considerably more radio luminous than agn with well defined soft - state psds and similar black hole masses , e.g. ngc4051 ( @xmath132 , * ? ? ?
* ) and mcg-6 - 30 - 15 ( @xmath133 , * ? ? ?
* ) . both ngc4051 ( et al . in preparation
; see also * ? ? ?
* ) and @xcite contain flat spectrum vlbi cores for which @xmath134 ( assuming redshift - derived distances ) .
these observations are certainly consistent with a vhs / intermediate state interpretation for ark 564 although the present lack of radio detection of soft state gbhs prevents a more quantitative comparison .
if the same disc - jet coupling applies in ngc4051 and ark 564 and the radio luminosity of the jet scales as @xmath135 ( absolute , not eddington , units ; * ? ? ?
* ) , then we would expect @xmath136 ( deriving @xmath137 from * ? ? ?
thus it is not yet clear whether the higher radio luminosity of ark 564 relative to ngc4051 is purely due to a higher accretion rate or to some difference in jet structure .
if we take the lorentzian peak frequency ( @xmath138 hz , i.e. 1666s , in the 28.8 kev band ) as being the best measurement of the underlying physical process responsible for the sharp drop in power , we can consider what that timescale might correspond to .
the dynamical time - scale around the ark 564 black hole ( @xmath139 ) is considerably shorter than the lorentzian peak time - scale , unless we assume that the time - scale arises far ( few tens of @xmath140 ) from the black hole .
there is evidence from the broad iron x - ray flourescence lines @xcite that the accretion disc often reaches down to @xmath141 , close to the innermost stable orbit of a maximally spinning black hole black hole .
if the lorentzian peak time - scale arises from the inner edge of such a spinning black hole disc , it may well correspond to the viscous time - scale ( assuming a viscosity parameter , @xmath142 ) . in that case
the ratio of disc scale height to radius , @xmath143 , = 1.2 , which is reasonable given the very high accretion rate .
a possible interpretation of both the psd and lag spectrum is that the variability originates mainly in two distinct regions .
note that by ` variability ' we mean the source of the variations , not the source of the x - rays .
the variability may be related to accretion rate fluctuations in the accretion disc .
it is reasonable to assume that variability time - scales are larger at larger radii ( e.g. * ? ? ?
once produced , variations would then propagate inwards and , on reaching the x - ray emitting region , would modulate the emission .
it is also reasonable to assume that the emitted energy spectrum becomes harder towards smaller radii so , for example , the bulk of the 0.52 kev emission will be emitted from a larger radius than the bulk of the 210 kev emission .
if the region responsible for most of the variability at low frequencies lies outside all of the x - ray emitting region , then the time lag between the soft and hard bands may represent the time taken for the fluctuations to propagate from the ` soft ' to ` hard ' emission region ( e.g. see @xcite , @xcite ) .
if all of the frequencies in the low frequency variability component are produced in the same region , for example a narrow annulus of the disc , then all of those frequencies will have the same soft - hard lag .
if the source of the higher frequency variations lies within the x - ray emission region , e.g. between the soft and hard emission regions , then the high frequency variations will affect only the harder emission region and the resultant x - ray variations will be stronger at higher energies , as we observe . in this scenario
the predominantly softer , outer emission region , which is subject only to the slower variations , may form part , or possibly all , of the so - called ` soft excess ' , i.e. the excess in the energy spectrum below @xmath144kev above an extrapolation of a power law to low energies .
this softer region , although lying outside the harder emitting region is , however , probably still quite close ( ie within @xmath145 rather than @xmath146 ) to the black hole .
this scenario is consistent with the observations of @xcite and @xcite of ark 564 .
@xcite show that , on @xmath147 day time - scales , the soft excess emission from ark 564 varies at least as much as the power law emission .
@xcite show , by fitting a combination of bremsstrahlung ( not necessarily a physical bremsstrahlung component but simply to fit the ` soft excess ' ) and power law components to 500s sections of the data discussed here , that the ` bremsstrahlung ' component at low energies ( 0.31 kev ) is less variable than the power law component in the same band , and that the power law component at higher energies ( 310 kev ) is more variable still .
however a full understanding of how the soft excess is related to the rest of the x - ray emission , e.g. whether it might actually be caused by absorption @xcite requires detailed frequency - resolved spectroscopy ( e.g. see * ? ? ?
* ; * ? ? ?
* ) for which the present data are insufficient .
we finally comment on the negative phase lag , @xmath148s , at the highest observable frequencies .
as the high frequency x - ray variations at lower energies will also be produced at small radii , e.g. as the low energy tail of the harder emission region , there should be very little lag between the soft and hard emission , as we see .
however the lag is negative , i.e. the hard band leads , in the higher frequency component .
this observation suggests that the high - frequency variability in the soft band might arise from reflection / reprocessing within @xmath149 of the illuminating hard x - ray source , implying the existence of cold gas close to the central x - ray source .
this work is based on observations with , an esa science mission with instruments and contributions directly funded by esa member states and the usa ( nasa ) .
imch and pa acknowledge support from stfc under rolling grant pp / d001013/1 and pu acknowledges support from an stfc advanced fellowship .
we acknowledge redshift data from the nasa / ipac extragalactic database . | we present a power spectral analysis of a 100 ksec observation of the narrow line seyfert 1 galaxy ark 564 . when combined with earlier and observations
, these data produce a power spectrum covering seven decades of frequency which is well described by a power law with two very clear breaks .
this shape is unlike the power spectra of almost all other agn observed so far , which have only one detected break , and resemble galactic binary systems in a soft state .
the power spectrum can also be well described by the sum of two lorentzian - shaped components , the one at higher frequencies having a hard spectrum , similar to those seen in galactic binary systems .
previously we have demonstrated that the lag of the hard band variations relative to the soft band in ark 564 is dependent on variability time - scale , as seen in galactic binary sources .
here we show that the time - scale dependence of the lags can be described well using the same two - lorentzian model which describes the power spectrum , assuming that each lorentzian component has a distinct time lag .
thus all x - ray timing evidence points strongly to two discrete , localised , regions as the origin of most of the variability .
similar behaviour is seen in galactic x - ray binary systems in most states other than the soft state , i.e. in the low - hard and intermediate / very high states .
given the very high accretion rate of ark 564 the closest analogy is with the very high ( intermediate ) state rather than the low - hard state .
we therefore strengthen the comparison between agn and galactic binary sources beyond previous studies by extending it to the previously poorly studied very high accretion rate regime .
[ firstpage ] galaxies : active , powerspectra : lorentzian components , accretion : discs , spectral timing : phase lags |
it is well - known that the free schrdinger propagator @xmath2 in @xmath3 is not bounded on the lebesgue spaces @xmath4 , except for @xmath5 .
this has motivated the study in other function spaces arising in harmonic analysis . among them
recently much attention has been given to the so - called modulation spaces .
they can be defined similarly to the besov spaces , but dyadic annuli in the frequency domain are replaced by isometric boxes .
actually , for our purposes it will be more useful an equivalent definition , in terms of the short - time fourier tranform , or bargmann transform , of temperate distributions ( see @xcite or section 2 below , where the weighted variant is considered ) .
namely , for @xmath6 , consider the translation and modulation operators @xmath7 as well as the time - frequency shifts @xmath8 the short - time fourier tranform ( stft ) of a temperate distribution @xmath9 with respect to a schwartz function @xmath10 is defined as @xmath11 ( with the pairing @xmath12 conjugate linear in the second factor ) .
then , for @xmath13 and @xmath14 , we define the modulation spaces @xmath15 and @xmath16 ( changing window yields equivalent norms ) .
in particular we have @xmath17 , whereas @xmath18-based sobolev spaces can be regarded as weighted modulation spaces as well . in short
, the modulation space norm measures the position - momentum ( or time - frequency ) concentration in phase space of a function .
now , it was proved in @xcite that the propagator @xmath2 is in fact bounded on @xmath19 , @xmath13 ; see also @xcite . as shown in @xcite , in the case of the fractional laplacian @xmath20 , with @xmath21 ,
a loss of derivatives instead occurs .
local well - posedness of the corresponding nonlinear equations , with nonlinearity of power - type , or even entire real - analytic , were also considered in @xcite .
remarkably , modulation spaces revealed to provide a good framework for the global wellposedness as well , as shown recently in @xcite for several dispersive equations .
the main results in this connection are now available in the recent book @xcite or in the survey @xcite .
a strictly related issue is the sparsity of the gabor matrix representation , in phase space , for the corresponding propagator ; this property can be in fact considered as a microlocal form of boundedness and implies the boundedness on modulation spaces in the usual sense .
however it contains much more refined information which is essential , e.g. , for the problem of propagation of singularities ; we refer to @xcite and the references therein for more detail . a case of special interest is given by the schrdinger equation with potential , i.e. @xmath22 , with @xmath23 real - valued .
the literature about wellposedness in @xmath18-based sobolev spaces is enormous and we refer e.g. to @xcite and the references therein . concerning the wellposedness in modulation spaces , it was proved in @xcite that if @xmath23 is smooth and has quadratic growth , i.e. @xmath24\times\rd)$ ] , for @xmath25 , the corresponding propagator is bounded on @xmath26 , @xmath27 ( this result also follows from @xcite ) .
instead , if @xmath23 is `` first order '' , i.e. @xmath24\times\rd)$ ] , for @xmath28 , one has boundedness on every @xmath19 , @xmath13 . the main motivation of the present paper is an extension of these results in the case of non - smooth potential , say with minimal regularity .
indeed , our result will apply to more general equations .
looking for optimal results , one is led to consider potentials with derivatives in the so - called sjstrand class , which is nothing but the modulation space @xmath29 .
in fact , in the theory of pseudodifferential operator that function space was first introduced in @xcite as a natural symbol class with minimal regularity which still gives rise to bounded operators on @xmath30 , and also on @xmath19 @xcite ( similar results hold for certain fourier integral operators as well @xcite ) . several deep results in micolocal analysis , such as the fefferman - phong inequality , keep valid for symbols ( with a certain number of derivatives ) in @xmath31 @xcite .
hence that class looks the natural choice when dealing with boundedness of linear operators in modulation spaces .
let us now state our results in a simplified form for a model equation ; we refer to section 2 below for general statements , which involve weighted modulation spaces as well .
denote by @xmath32 and @xmath33 the closure of the schwartz space in the corresponding modulation spaces ( hence @xmath34 if @xmath35 ) .
let @xmath36 be fixed .
consider the initial value problem @xmath37 with @xmath38 , @xmath39 $ ] , @xmath40 ( @xmath41 ) .
suppose that @xmath42 for @xmath43 , @xmath44 for @xmath45 , @xmath46 , with @xmath47 and @xmath48 real - valued .
assume moreover a mild continuity dependence with respect to @xmath49 ( narrow convergence ; see section 2 below ) and let @xmath50 .
then for every @xmath51 , there exists a unique solution @xmath52,\mathcal{m}^p(\rd))$ ] to .
moreover the corresponding propagator is bounded on @xmath53 .
in the case of first order potentials we have a better conclusion .
suppose in addition @xmath54 and let @xmath55 .
then for every @xmath56 there exists a unique solution @xmath52,\mathcal{m}^{p , q}(\rd))$ ] to .
moreover the corresponding propagator is bounded on @xmath57 .
it is shown in remark [ remark5.2 ] that similar results do not hold if one replaces @xmath58 by the larger space @xmath59 . for comparison ,
observe that @xmath60 , but functions in @xmath1 generally do not possess any derivative .
hence , the above results represent a significant improvement of those in @xcite .
notice that in the special case of @xmath61th order potentials @xmath62 more refined results were obtained in @xcite , where the propagator was shown to be a generalized metaplectic operator ; see also @xcite for the corresponding problem of propagation of singularities .
actually , as anticipated , we will consider much more general equations , namely of the form @xmath63 , where @xmath64 is a second order pseudodifferential operator , whose symbol has some derivatives in @xmath31 . in this connection our results can be regarded as the evolution counterpart of the boundedness results for pseudodifferential operators with symbols in @xmath31 proved in @xcite .
moreover , to our knowledge , they contain as special cases all the known continuity results of schrdinger propagators on modulation spaces .
the proof of wellposedness relies on the construction of a parametrix for the forward cauchy problem , in the spirit of @xcite .
namely , one decomposes the initial datum in coherent states @xmath65 , i.e. gabor atoms , where @xmath66 is a fixed window ; then a parametrix is constructed as a generalized localization operator in phase space ( a type of operators introduced in their basic form in @xcite ) that moves the gabor atoms in phase space according to the hamiltonian flow , together with a phase shift .
notice that at this low level of regularity the more classical approach via wkb expansions and fourier integral operators turns out inapplicable . finally , we will consider the corresponding nonlinear equations , with a nonlinearity @xmath67 which is entire real - analytic in @xmath68 ( e.g. a polynomial in @xmath69 ) and we will prove that all the above results extend in that setting if the initial datum is in @xmath70 or even in @xmath71 ( in the case of first order potentials ) , at least for small time .
briefly , the paper is organized as follows . in section 2
we recall the main definition and properties of modulation spaces and we state the results in full generality .
section 3 is devoted to some preliminary estimates , whereas in section 4 we introduce a class of generalized localization operators which will appear subsequently .
section 5 is devoted to the proof of the main results ( linear case ) .
finally section 6 deals with the extension to nonlinear equations , at least for small time .
the fourier transform is normalized as @xmath72 and the weyl quantization of a symbol @xmath73 is correspondingly defined as @xmath74 we recall the following easy properties , which can be checked directly : @xmath75=(d_{\xi_j } a)^w,\qquad [ a^w , d_j]=-(d_{x_j}a)^w\ ] ] @xmath76 @xmath77 we have already defined in the introduction the time - frequency shifts and the stft of a temperate distribution , as well as the unweighted modulation spaces @xmath19 .
here we extend the definition in the presence of a weight .
we consider a positive submultiplicative even continuous function @xmath78 in @xmath79 ( @xmath80 , @xmath81 ) ; moreover we suppose in the sequel that @xmath78 satisfies @xmath82 as well as @xmath83 for some @xmath84 .
we denote by @xmath85 the space of @xmath78-moderate weight @xmath86 , i.e. positive continuous function @xmath87 in @xmath79 such that @xmath88 , satisfying in addition the following condition : for every constant @xmath89 there exists @xmath90 such that , for @xmath91 , @xmath92 if @xmath93 is any invertible transformation , lipschitz together with its inverse . as prototype
one can consider the standard weights @xmath94 for which we have @xmath95 if and only if @xmath96 .
now , for @xmath55 , @xmath97 and @xmath14 , we define the spaces @xmath98 it can be proved that changing windows in @xmath99 ( on even in @xmath100 ) yields equivalent norms .
briefly we write @xmath101 . also , we set @xmath102 for the closure of the schwartz space in @xmath103 ( so that @xmath104 when @xmath35 ) , and @xmath105 .
we now recall the definition of the narrow convergence . in this connection
there are several related definitions in the literature ; the present one has first appeared in @xcite and is different e.g. from that in @xcite .
let @xmath106 be a subset of some euclidean space .
we say that a map @xmath107 is continuous for the narrow convergence if it is continuous in @xmath108 ( weakly ) and if there exists a function @xmath109 such that @xmath110 for every @xmath111 and almost every @xmath112 . here
@xmath113 is a fixed window and the definition is independent of the choice of @xmath66 , as a consequence of lemma [ changewind ] below .
it is also clear from the very definition that the set of symbols @xmath114 is then bounded in @xmath115 .
let @xmath36 be fixed .
consider the cauchy problem @xmath116 with @xmath117 $ ] , @xmath118 , and @xmath119 specified below .
this is our main result .
[ teo1 ] let @xmath120 , @xmath117 $ ] , @xmath121 , where @xmath122 , @xmath123 are real - valued and suppose that , for @xmath124 and @xmath125 , the map @xmath126 let @xmath50 , @xmath127 . for every @xmath128 , there exists a unique solution @xmath52,\mathcal{m}^p_m(\rd))$ ] to .
moreover the corresponding propagator is bounded on @xmath129 . in the case of symbols that can be written as a sum of symbols depending only on @xmath130 or @xmath131 less regularity may be assumed .
[ teo2 ] let @xmath132 , @xmath117 $ ] , @xmath121 , where @xmath133 , @xmath134 , @xmath135 , are real - valued and suppose that , for @xmath124 and @xmath136 , the map @xmath137 let @xmath50 , @xmath127 . for every @xmath128 , there exists a unique solution @xmath52,\mathcal{m}^p_m(\rd))$ ] to .
moreover the corresponding propagator is bounded on @xmath129 .
[ teo3 ] under the same assumption as in theorem [ teo2 ] , suppose moreover that @xmath54 .
let @xmath55 , @xmath97 .
for every @xmath138 , there exists a unique solution @xmath52,\mathcal{m}^{p , q}_m(\rd))$ ] to .
moreover the corresponding propagator is bounded on @xmath139 .
theorem [ teo2 ] applies of course to problems of the form @xmath140 under the hypotheses given there on the potentials @xmath134 , @xmath124 . if @xmath54 then theorem [ teo3 ] applies as well .
we can also consider the case of the fractional laplacian , i.e. @xmath141 with @xmath38 .
[ teo4 ] consider potentials @xmath142 , @xmath143 with @xmath48 , @xmath47 real - valued , and suppose that , for @xmath124 , @xmath136 , @xmath144 , the map @xmath145 let @xmath50 , @xmath146 .
for every @xmath128 , there exists a unique solution @xmath52,\mathcal{m}^p_m(\rd))$ ] to , the corresponding propagator being bounded on @xmath129 .
suppose moreover @xmath54 .
let @xmath55 , @xmath147 .
for every @xmath138 , there exists a unique solution @xmath52,\mathcal{m}^{p , q}_m(\rd))$ ] to , the corresponding propagator being bounded on @xmath139 .
in the sequel we will use the following covariance property of the stft , which can be verified by direct inspection : @xmath148 if @xmath149 such that @xmath150 and @xmath9 , then the inequality @xmath151 holds pointwise for all @xmath152 .
the following lemma is proved in ( * ? ? ?
* lemma 3.1 ) .
[ lemmag ] let @xmath10 , @xmath153 . for some window
@xmath154 it turns out , setting @xmath155 , @xmath156 [ proregolarita ] let @xmath157\ni t\mapsto a_t$ ] be a continuous map in @xmath158 for the narrow convergence .
then @xmath159 is continuous as a function of @xmath160 .
clearly it suffices to consider the un - weighted case ( @xmath161 ) , since @xmath162 .
let @xmath163 , @xmath164 .
then latexmath:[\[\label{30-u } consider now the first term in the right - hand side of .
if @xmath10 , @xmath166 , and @xmath167 , it turns out @xmath168 this last expression tends to zero as @xmath169 by the dominated convergence theorem , because @xmath170 in @xmath108 implies that @xmath171 pointwise , whereas the assumption about narrow convergence yields @xmath172,\ u , v\in\rdd\ ] ] for some function @xmath173 .
similarly , for the second term in the right - hand side of we have @xmath174 for some @xmath173 , and one concludes as above . the following result is essentially known .
we shall give the proof for the benefit of the reader .
[ proiniziale ] let @xmath97 , @xmath13 .
the operators @xmath175 , @xmath176 , are bounded @xmath177 , @xmath178 .
similarly , the operators @xmath179 , @xmath180 , are bounded @xmath181 , @xmath182 it suffices to prove the first part of the statement .
consider the case of @xmath183 ( similar arguments apply to @xmath184 ) .
let @xmath10 , @xmath166 .
we have @xmath185 on the other hand , with @xmath186 , @xmath187 for every @xmath188 , where we used and the fact that the stft of schwartz functions is schwartz .
hence we get @xmath189 and therefore @xmath190 which gives the desired boundedness of @xmath191 ( @xmath192 if @xmath193 is large enough , by ) .
similarly , from we have @xmath194 which gives the boundedness of @xmath195 the following result shows the usefulness of the notion of narrow convergence . [ pro3.2 ]
let @xmath107 be a continuous map from the narrow convergence .
let @xmath97 .
then the map of operators @xmath196 is strongly continuous on @xmath139 , @xmath13 .
it is well known , see e.g. @xcite , that the set @xmath197 is bounded in the space of linear continuous operators on @xmath198 .
hence it is sufficient to prove the strong continuity on @xmath199 , since @xmath200 with inclusion continuous and dense .
now , we can suppose that @xmath201 , say for @xmath202 , for the narrow convergence .
let @xmath203 , @xmath166 .
for @xmath204 , we get @xmath205 hence @xmath206 now , we have @xmath207 pointwise if @xmath202 . on the other hand it follows from lemma [ lemmag ] that for a new window @xmath154 , with @xmath155 , @xmath208 for some @xmath109 , where for the last inequality we used the hypothesis of narrow continuity of @xmath209 .
since we also have @xmath210 by assumption , we get @xmath211 from and the dominated convergence theorem .
the next results will be used often in the subsequent sections .
[ prod ] let @xmath157\ni t\mapsto a_t\in m^{\infty,1}_{1\otimes v}(\rdd)$ ] be continuous for the narrow convergence , and @xmath212 be a continuous function of @xmath213 $ ] . then the map @xmath214\times \rdd\ni ( t,\zeta)\mapsto b_{t,\zeta}(w):=\int_0 ^ 1 \gamma(\tau ) a_t(\zeta+\tau w)\ , d\tau,\quad w\in\rdd\ ] ] is still continuous in @xmath158 for the narrow convergence .
the same holds true for the map @xmath214\times \rdd\ni ( t,\zeta)\mapsto \tilde{b}_{t,\zeta}(w):=a_t(\zeta+w).\ ] ] we prove only the first part of the statement , because the last part follows similarly from an easier argument .
the continuity of @xmath215 in @xmath108 is clear .
let us estimate the stft of @xmath216 .
let @xmath66 be a gaussian function , with @xmath166 .
we have @xmath217(\zeta+\tau z,\tau^{-1}w)|\ , d\tau,\end{aligned}\ ] ] where @xmath218 .
using the change - of - window formula in lemma [ changewind ] we get @xmath219|\ast |v_{g } g_\tau\otimes v_g g_\tau|\big)(\zeta+\tau z,\tau^{-1}w)|\ , d\tau.\ ] ] we now take the supremum with respect to @xmath220 and we get , by young inequality , @xmath221(z,\cdot)|\ast |v_g g_\tau|\big)(\tau^{-1}w)\ , d\tau,\ ] ] where we used the estimate ( see e.g. ( * ? ? ?
* formula ( 19 ) ) ) @xmath222 now , we have @xmath223(z , w)|\leq h(w)$ ] for some @xmath109 , by assumption .
hence we obtain @xmath224 it suffices to prove that this last expression is in @xmath225 .
this follows at once from the fubini - tonelli theorem , using the following two estimates : it turns out @xmath226 which follows because @xmath78 is submultiplicative and satisfies , and we also have @xmath227 in fact , an explicit computation ( see e.g. ( * ? ? ?
* lemma 3.1 ) ) shows that @xmath228 with @xmath229 gaussian , whereas @xmath78 has polynomial growth .
[ pro3.4 ] let @xmath230 satisfy the assumption in theorem [ teo1 ] , and @xmath231 , @xmath13 .
then the operator @xmath232 is bounded @xmath233 , @xmath182 and the map @xmath234 is strongly continuous on the these spaces .
since the operators @xmath235 , @xmath180 , are bounded @xmath182 ( proposition [ proiniziale ] ) , by a @xmath236-th order taylor expansion of @xmath237 at @xmath238 , @xmath135 , we are reduced to prove the result for the remainder @xmath239 using repeatedly , and we can write @xmath240 where @xmath241 and @xmath242 are continuous maps in @xmath158 for the narrow convergence by proposition [ prod ] and the assumption . hence the corresponding operators are strongly bounded on @xmath102 by proposition [ pro3.2 ] .
this concludes the proof .
we consider the hamiltonian flow @xmath243 , as a function of @xmath39 $ ] , @xmath118 , given by the solution of @xmath244 under the assumptions of theorems [ teo1]-[teo4 ] we see that the functions @xmath245 , @xmath43 , are continuous by proposition [ proregolarita ] .
moreover they are bounded , because @xmath246 .
hence the solution of the above initial value problem exists globally in time for every initial datum , and the flow is a map of class @xmath247 in all variables @xmath248 . the map @xmath249 is moreover symplectic .
further consider the real - valued phase @xmath250 defined by @xmath251 we now introduce the class of operators used in the next section for the construction of the parametrix ( cf .
@xcite ) .
[ lemma2 ] let @xmath252 .
let @xmath157\times\rdd\ni ( t , x,\xi)\mapsto g(t , x,\xi,\cdot)$ ] be a continuous map in @xmath253 ( weakly ) , such that @xmath254 for every @xmath248 , for some function @xmath109 ( in particular we have @xmath255 ) . for every @xmath256 $ ]
, the operator @xmath257 is strongly continuous on @xmath258 , @xmath27 , @xmath231 .
we first apply the change of variable @xmath259 , which has jacobian @xmath260 , and we call again @xmath261 for the new variables ; we obtains @xmath262 we can write @xmath263 where @xmath264 is of course bounded , and the operators @xmath265 , @xmath266 are defined as follows .
the operator @xmath267 is strongly continuous on @xmath268 , @xmath269 , and on the closure of the schwartz space in @xmath270 ; this is straightforward to check .
the operator @xmath271 is bounded on @xmath268 , because by the assumption on @xmath272 and we can dominate its integral kernel by a convolution kernel in @xmath273 : @xmath274 to prove the strong continuity of @xmath265 we can assume @xmath275 , and the conclusion then follows by the dominated convergence theorem , as in the proof of proposition [ pro3.2 ] .
we first construct a parametrix for the cauchy problem , in the form of generalized localization operator as in the previous section .
[ teopara ] assume the hypotheses of theorem [ teo1 ] .
for @xmath276 there are operators @xmath277 , @xmath278 satisfying the following properties : * @xmath277 and @xmath278 are strongly continuous on @xmath258 , @xmath27 ; * @xmath279 for every @xmath280 ; * @xmath281 for every @xmath282 , @xmath276 .
we set @xmath283 where @xmath66 is any schwartz function , with @xmath166 .
property ( 2 ) then follows from the inversion formula ( * ? ? ?
* corollary 3.2.3 ) @xmath284 and the change of variable @xmath285 , which has jacobian @xmath260 .
the strong continuity of @xmath277 in ( 1 ) is a consequence of proposition [ lemma2 ] .
let us compute @xmath286 .
we have by an explicit computation , using and , @xmath287 \\ & = e^{i\psi(t , x,\xi)}\pi(x^t,\xi^t)[-a_2(t , x^t,\xi^t)g-\nabla_x a_2(t , x^t,\xi^t)\cdot y g\nonumber\\ & \qquad\qquad\qquad\qquad\qquad\qquad\qquad-\nabla_{\xi } a_2(t , x^t,\xi^t)\cdot ( -i\nabla_y ) g - a_1(t , x^t,\xi^t)g]\nonumber\end{aligned}\ ] ] whereas ( cf .
the symplectic invariance of the weyl calculus in ( * ? ? ?
* theorem 18.5.9 ) ) @xmath288 hence we obtain @xmath289,\ ] ] with @xmath290 using repeatedly and , the assumption and proposition [ prod ] we can write @xmath291 where @xmath292 are a family of symbols in @xmath158 , continuous with respect to the narrow convergence as a function of @xmath248 . hence we get @xmath293 with @xmath294 v_g f(x^s,\xi^s)\ , dx\,d\xi.\ ] ] to see that this operator enjoys the properties in ( 1 ) , by proposition [ lemma2 ] we have to verify that @xmath295 is continuous as a function of @xmath248 in @xmath253 , which is clear , and that @xmath296,\ , ( x,\xi),w\in\rdd\ ] ] for some function @xmath109 independent of @xmath248 . to this end
observe that , using , we are left to prove that if @xmath297 is a family of symbols in @xmath158 , continuous for the narrow convergence with respect to @xmath248 , it turns out , for @xmath298 , @xmath299 for some function @xmath109 .
lemma [ changewind ] reduces matters to the case @xmath300 , where one can then conclude by applying lemma [ lemmag ] .
let us prove that there exists a strongly continuous operator @xmath301 on @xmath129 such that @xmath302 , and @xmath303 first we observe that for @xmath304 as in theorem [ teopara ] , for every @xmath128 there exists a unique solution @xmath305,\mathcal{m}^p_m(\rd))$ ] to the volterra equation @xmath306 this is a consequence of the contraction mapping theorem applied in the space @xmath307,\cm^p_m(\rd))$ ] endowed with the norm @xmath308}e^{-\lambda t}\|u(t)\|_{\cm^p_m}$ ] , with @xmath309 moreover one obtains @xmath310 , @xmath117 $ ] .
if @xmath78 is such a solution , we then set @xmath311 where @xmath277 is the parametrix in theorem [ teopara ] .
it is then straightforward to check that holds true , as well as the claimed boundedness of the propagator .
we have to prove that if @xmath52,\mathcal{m}^p_m(\rd))$ ] is a solution to the equation in and @xmath312 , then @xmath313 .
let @xmath10 .
by proposition [ pro3.4 ] we have in particular @xmath314 , \mathcal{m}^p _ { m / v_2}(\rd))\subset c^1([0,t ] , \cs'(\rd))$ ] , so that @xmath315 is continuously differentiable as a function of @xmath248 . by the fundamental theorem of calculus we have the pointwise equality @xmath316 on the other hand it turns out @xmath317 where we used that the map @xmath318 is continuously differentiable in @xmath99 .
using and we obtain and @xmath319 ; cf .
proposition [ pro3.4 ] . ]
@xmath320 where @xmath321 ( same notation as in the proof of theorem [ teopara ] ) .
summing up we get @xmath322 using we can estimate the last term as @xmath323 since holds true for some @xmath109 , by young s inequality we obtain @xmath324 we can then estimate in @xmath325,\ ] ] and therefore @xmath326 for every @xmath39 $ ] , by gronwall s inequality .
a carefully inspection of the proof of theorem [ teo1 ] and the needed preliminary results shows that the only point were we used the condition on @xmath327 for @xmath328 , and @xmath329 for @xmath43 , is to pass to the algebraic expression in to the corresponding quantization in the form where , using , additional derivatives fall on the symbol and similarly in the proof of theorem [ teopara ] ( from to ) .
however , when the symbol @xmath64 has the special form in theorem [ teo2 ] the factorization at the level of symbol for @xmath330 , @xmath142 gives a corresponding exact factorization at the level of operators , and therefore the conditions @xmath331 , @xmath43 , and @xmath332 , @xmath45 , suffice in that case .
it suffices to prove that , under the present assumptions , the results in proposition [ lemma2 ] and therefore those in theorem [ teopara ] still hold for the modulation spaces @xmath102 , @xmath13 . in fact , one is left to prove that the map @xmath333 is bounded of @xmath334 .
now , we have @xmath335 , so that the symplectic map @xmath336 ( we omit the dependence on @xmath337 for brevity ) has the special form @xmath338 , @xmath339 , with @xmath340 .
hence the change of variable @xmath341 ( for fixed @xmath342 ) gives @xmath343 however @xmath344 since @xmath336 is a lipschitz map , and holds by assumption .
this concludes the proof .
we need the following result from ( * ? ? ?
* proposition 2.5 ) .
[ sopra ] let @xmath345 be positively homogeneous of degree @xmath346 , i.e. @xmath347 for @xmath348 , @xmath349 , and @xmath350 .
set @xmath351 .
then , for @xmath10 there exists a constant @xmath352 such that @xmath353 one can choose a cut - off function @xmath354 , @xmath355 in a neighborhood of the origin , and write the equation as @xmath356 with @xmath120 , where @xmath357 , @xmath358 , @xmath359 . by proposition [ sopra ]
we have @xmath360 for every @xmath144 .
hence , the conclusion follows at once from theorems [ teo2 ] , [ teo3 ] ( with @xmath361 ) .
[ remark5.2]let us observe that the initial value problem for @xmath362 is not wellposed in @xmath363 .
in fact , if the initial datum is @xmath364 , the solution @xmath365 verifies ( by the invariance of @xmath59 under fourier transform ) @xmath366 now , using the window @xmath367 , an explicit computation ( cf .
* theorem 6 ) , modified according to our normalization ) shows that @xmath368 which , if @xmath369 , tends to @xmath61 as @xmath370 for every fixed @xmath371 . on the other hand
, @xmath372 is independent of @xmath130 .
hence we see that @xmath373 for every @xmath371 . as a consequence
, @xmath374 does not tend to zero as @xmath375 .
in this section we briefly discuss the extension of the above results in presence of a non - linearity of the type @xmath67 , where the function @xmath376 is entire real - analytic , with @xmath377 ( @xmath378 has a taylor expansion in @xmath379 , valid in the whole complex plane ) . in particular we can take a polynomial in @xmath379 .
each of the theorems [ teo1][teo4 ] admits a nonlinear variant .
namely , under those assumptions , we have local wellposedness in @xmath380 or even in @xmath381 ( according to the corresponding linear result ) , for @xmath27 , @xmath97 , @xmath382 , of the initial value problem for the nonlinear equation @xmath383 , where @xmath384 is the linear operator , with @xmath67 as specified above .
the argument is standard ; cf .
* theorem 6.1 ) .
namely , we apply the contraction mapping theorem in the space @xmath385,\cm^{1}_m(\rd))$ ] , with @xmath386 small enough ( or @xmath385,\cm^{p,1}_m(\rd))$ ] , when appropriate ) . by the duhamel principle we can rewrite the semilinear equation in integral form @xmath387 where @xmath301 , @xmath388 , is the linear propagator corresponding to initial data at time @xmath389 .
the classical iteration scheme works in @xmath390 if the following properties are verified : * @xmath301 is strongly continuous on @xmath391 for @xmath276 ( which also implies a uniform bound for the operator norm with respect to @xmath337 , by the uniform boundedness principle ) ; * we have @xmath392 , for @xmath393 in every fixed ball .
the estimate in * b ) * was proved in ( * ? ? ?
* formula ( 28 ) ) in the case @xmath394 , @xmath395 , but the same proof extends to any weight @xmath127 satisfying @xmath382 , since one just uses the fact that @xmath380 is a banach algebra for pointwise multiplication ( the same holds for @xmath396 ) . as far as * a ) * is concerned , it follows from the above linear results that @xmath301 is bounded on @xmath397 uniformly with respect to @xmath337 and it is strongly continuous in @xmath397 as a function of @xmath49 , for fixed @xmath389 .
the same holds for @xmath337 exchanged , because the equation is time - reversible . to prove its strong continuity jointly in @xmath398 we observe that if @xmath399 , @xmath400 , @xmath401 hence , the map @xmath402 is in fact continuous in @xmath403 uniformly with respect to @xmath49 .
this yields the strong continuity of @xmath301 on @xmath380 , as a function of @xmath337 and concludes the proof .
h. g. feichtinger .
modulation spaces on locally compact abelian groups , in `` _ wavelets and their applications _ '' , m. krishna , r. radhaand , s. thangavelu eds . , chennai , india , allied publishers , new delhi , 2003 , pp .
updated version of a technical report , university of vienna , 1983 .
n. lerner , y. morimoto , on the fefferman - phong inequality and a wiener - type algebra of pseudodifferential operators , _ publications of the research institute for mathematical sciences _
( kyoto university ) , * 43 * ( 2007 ) , 329371 .
r. rochberg and k. tachizawa , pseudodifferential operators , gabor frames , and local trigonometric bases , in _
gabor analysis and algorithms , _ appl .
numer . harmon .
, birkhuser boston , boston , ma , 171192 , 1998 .
m. ruzhansky , m. sugimoto and b. wang , modulation spaces and nonlinear evolution equations , in _ evolution equations of hyperbolic and schrdinger type _ , 267 - 283 , progress in mathematics , vol .
301 , birkhuser , 2012 .
d. tataru , phase space transforms and microlocal analysis , in _ phase space analysis of partial differential equations _ , vol .
ii , 505524 , pubbl .
ennio giorgi , scuola norm .
sup . , pisa , 2004 . | we consider a class of linear schrdinger equations in @xmath0 with rough hamiltonian , namely with certain derivatives in the sjstrand class @xmath1 .
we prove that the corresponding propagator is bounded on modulation spaces .
the present results improve several contributions recently appeared in the literature and can be regarded as the evolution counterpart of the fundamental result of sjstrand about the boundedness of pseudodifferential operators with symbols in that class .
finally we consider nonlinear perturbations of real - analytic type and we prove local wellposedness of the corresponding initial value problem in certain modulation spaces .
elena cordero fabio nicola luigi rodino ( communicated by the associate editor name ) |
lepton flavour violation ( lfv)is an important signature into physics beyond the standard model ( sm ) . in the sm with massive neutrinos , neutrinos have the yukawa interaction .
lepton flavour is violated like quark flavour . however , lfv is strongly suppressed by the neutrino masses .
supersymmetry ( susy ) with massive neutrinos makes the situation drastically changed .
it can predict sizable lfv effects , because the alternative source of lfv is generated from the slepton mixing through the renormalization group effect on the right - handed neutrino yukawa interaction .
the lfv processes @xmath1 of charged lepton are being measured .
the meg experiment @xcite gives the upper bound on the @xmath2 process as @xmath3 the forthcoming experiment reaches to @xmath4 .
the @xmath5 decay processes are also measured at b - factories @xcite . on the other hand , it has been already discovered that lepton flavour is violated in the neutrino sector .
the oscillation parameters will be measured in detail . in this letter
, we study the phase of the neutrino yukawa matrix and investigate its effect on the magnitude of the lfv processes .
in the framework based on the see - saw mechanism , we can parameterize the neutrino yukawa matrix @xmath6 in terms of physical quantities as @xcite @xmath7 where @xmath8 is the vacuum expectation value of the higgs boson , @xmath9 is the observed maki - nakagawa - sakata ( mns ) matrix including two majorana phases @xmath10 ; i.e. , @xmath11 here we neglect the @xmath12 element of @xmath9 .
@xmath13 is the neutrino mass matrix and @xmath14 is the right - handed majorana neutrino mass matrix in each diagonal base .
an arbitrary complex orthogonal matrix @xmath15 is expressed as @xmath16 where @xmath17 and @xmath18 the elements @xmath19 give a drastic effect on structure of the yukawa matrix as hyperbolic functions .
we call @xmath20 as _ r - phase _ to distinguish it from the majorana phase
. in the following , we assume that the neutrino masses are _ approximately _ degenerate and the right - handed neutrino masses are degenerate @xmath21 where , @xmath22 and @xmath23 are the solar and atmospheric neutrino mass squared differences , @xmath24 is the size of right - handed neutrinos , and @xmath25 is undetermined light neutrino mass parameter . we then obtain the neutrino yukawa matrix @xmath6 in its diagonal base as @xmath26 where @xmath27 this expression indicates the characteristic relation @xmath28 among the neutrino yukawa couplings , which is similar to those of the quarks and the charged leptons @xmath29 . for @xmath30 and @xmath31 , @xmath32 is close to @xmath33 , so that the neutrino yukawa couplings become hierarchical .
the parameter @xmath32 is written in terms of the combination of _ r - phase _
( see eqs.([r1 ] ) and ( [ r2 ] ) ) , and it determines the size of the neutrino yukawa hierarchy .
we consider the susy models , especially the minimal supersymmetric standard model ( mssm ) with right - handed neutrinos . in the slepton sector
, the mssm lagrangian has an alternative source of lfv through the following soft susy breaking terms , @xmath34 where @xmath35 are the slepton tri - linear couplings and @xmath36 are the soft mass parameters for the sleptons .
we assume that these couplings are universal at the grand unified theory ( gut ) scale ( @xmath37 ) , i.e. , @xmath38 in this framework , the lfv processes @xmath1 can appear due to the slepton mixing @xcite , which comes from the renormalization group effect on @xmath6 between the scales of @xmath37 and @xmath24 @xmath39 the branching ratios for these processes are expressed by @xmath40 where @xmath41 is the typical susy scale , @xmath42 is the ratio of the vacuum expectation values of higgs bosons , @xmath43 is the fine structure constant , and @xmath44 is fermi constant .
let us consider the @xmath32 dependence of lfv with @xmath45 , the effect of _ r - phase _ contributes to only @xmath2 process .
a comparison of each magnitude of lfv processes is already meaningless , and we plotted the branching ratio for @xmath46 process ( see figure [ br - r ] . ) , event for @xmath47 ( upper solid line ) and @xmath48 ( lower solid line ) .
the experimental upper limit from meg is also shown ( dashed line).,width=283 ] where we take @xmath49 , and @xmath50 .
the lfv branching ratio is approximately proportional to @xmath51 , so that a large value of @xmath32 can change the branching ratio by several orders of magnitude . on the other hand
, we show the dependence on the majorana phase @xmath52 ( see figure [ br - a ] . ) .
when @xmath45 , the branching ratio does not depend on @xmath53 .
event for @xmath54 ( upper solid line ) and @xmath55 ( lower solid line ) .
the upper limit is also shown ( dashed line).,width=283 ] figure [ br - a ] shows that the branching ratio is a periodic function on @xmath52 , and that the effect of majorana phase is smaller than that of @xmath32 .
we have analyzed the structure of the neutrino yukawa matrix and have discussed the magnitude of lfv processes from the phase effect on the neutrino yukawa matrix in the mssm with the right - handed neutrinos .
the neutrino yukawa matrix has two types of phases , majorana phases and _ r - phases_. in the case that neutrino masses are degenerate and the right - handed neutrino masses are degenerate , the eigenvalues of the neutrino yukawa matrix become hierarchical spectrum and the _ r - phases _ determine the size of yukawa hierarchy . in the susy models ,
sizable lfv can arise due to the slepton mixing from the renormalization group effect on the neutrino yukawa matrix between @xmath37 and @xmath24 .
the majorana phases can change the lfv branching ratios by a factor , and these magnitudes become periodic as the function of majorana phases . on the other hand ,
_ r - phases _ enhance the magnitude of the lfv branching ratios by several orders .
j. hisano , t. moroi , k. tobe and m. yamaguchi , phys .
d * 53 * , 2442 ( 1996 ) j. hisano , t. moroi , k. tobe , m. yamaguchi and t. yanagida , phys .
b * 357 * , 579 ( 1995 ) s. kanemura , k. matsuda , t. ota , t. shindou , e. takasugi and k. tsumura , arxiv : hep - ph/0409283 . | we estimate the magnitude of lepton flavour violation ( lfv ) from the phase of the neutrino yukawa matrix . in the minimal supersymmetric standard model with right - handed neutrinos ,
the lfv processes @xmath0 can appear through the slepton mixing , which comes from the renormalization group effect on the right - handed neutrino yukawa interaction between the grand unified theory scale and the heavy right - handed neutrino mass scale .
two types of phases exist in the neutrino yukawa matrix .
one is the majorana phase , which can change the magnitude of the lfv branching ratios by a few factor .
the other phases relate for the size of the yukawa hierarchy and its phase effect can change the lfv branching ratios by several orders of magnitude . |
information is about making distinctions or differences . in james gleick s book , _ the information : a history , a theory , a flood _ , he noted the focus on differences in the seventeenth century polymath , john wilkins , who was a founder of the royal society . in 1641 ,
the year before newton was born , wilkins published one of the earliest books on cryptography , _ mercury or the secret and swift messenger _ , which not only pointed out the fundamental role of differences but noted that any ( finite ) set of different things could be encoded by words in a binary code .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for in the general we must note , that whatever is capable of a competent difference , perceptible to any sense , may be a sufficient means whereby to express the cogitations .
it is more convenient , indeed , that these differences should be of as great variety as the letters of the alphabet ; but it is sufficient if they be but twofold , because two alone may , with somewhat more labour and time , be well enough contrived to express all the rest .
xvii , p. 69 ) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ wilkins explains that a five letter binary code would be sufficient to code the letters of the alphabet since @xmath0 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ thus any two letters or numbers , suppose a.b .
being transposed through five places , will yield thirty two differences , and so consequently will superabundantly serve for the four and twenty letters ... .
xvii , p. 69 ) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ as gleick noted : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ any difference meant a binary choice .
any binary choice began the expressing of cogitations . here , in this arcane and anonymous treatise of 1641 , the essential idea of information theory poked to the surface of human thought , saw its shadow , and disappeared again for [ three ] hundred years . @xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in this paper , we will start afresh by deriving an information - as - distinctions notion of logical entropy @xcite from the new logic of partitions @xcite that is mathematically dual to the usual boolean logic of subsets .
then the usual shannon entropy @xcite will be essentially derived from the concepts behind logical entropy as another way to measure information - as - distinctions .
this treatment of the various notions of shannon entropy ( e.g. , mutual , conditional , and joint entropy ) will also explain why their interrelations can be represented using a venn diagram picture @xcite .
the logic normally called `` propositional logic '' is a special case of the logic of subsets originally developed by george boole @xcite . in the boolean logic of subsets of a fixed non - empty universe set @xmath1 ,
the variables in formulas refer to subsets @xmath2 and the logical operations such as the join @xmath3 , meet @xmath4 , and implication @xmath5 are interpreted as the subset operations of union @xmath6 , intersection @xmath7 , and the conditional @xmath8 .
then `` propositional '' logic is the special case where @xmath9 is the one - element set whose subsets @xmath10 and @xmath11 are interpreted as the truth values @xmath12 and @xmath11 ( or false and true ) for propositions . in subset logic ,
a _ valid formula _ or _ tautology _ is a formula such as @xmath13 \rightarrow t$ ] where for any non - empty @xmath1 , no matter what subsets of @xmath1 are substituted for the variables , the whole formula evaluates to @xmath1 by the subset operations .
it is a theorem that if a formula is valid just for the special case of @xmath9 ( i.e. , as in a truth table tautology ) , then it is valid for any @xmath1 .
but in today s textbook treatments of so - called `` propositional '' logic , the truth - table version of a tautology is usually given as a definition , not as a theorem in subset logic . what is lost by restricting attention to the special case of propositional logic rather than the general case of subset logic ?
at least two things are lost , and both are relevant for our development . * firstly if it is developed as the logic of subsets , then it is natural , as boole did , to attach a quantitative measure to each subset @xmath14 of a finite universe @xmath1 , namely the normalized counting measure @xmath15 which can be interpreted as the _ logical probability _
@xmath16 ( where the elements of @xmath1 are assumed equiprobable ) of randomly drawing an element from @xmath14 . *
secondly , the notion of a subset ( unlike the notion of a proposition ) has a mathematical dual in the notion of a quotient set , as is evidenced by the dual interplay between subobjects ( subgroups , subrings , ... ) and quotient objects throughout abstract algebra .
this duality is the `` turn - around - the - arrows '' category - theoretic duality , e.g. , between monomorphisms and epimorphisms , applied to sets @xcite .
the notion of a quotient set of @xmath1 is equivalent to the notion of an equivalence relation on @xmath1 or a partition @xmath17 of @xmath1 .
when boole s logic is seen as the logic of subsets ( rather than propositions ) , then the notion arises of a dual logic of partitions which has now been developed @xcite .
a partition @xmath17 on a finite set @xmath1 is a set of non - empty disjoint subsets @xmath18 ( `` blocks '' of the partition ) of @xmath1 whose union is @xmath1 .
the idea of information - as - distinctions is made precise by defining a _ distinction _ or _ dit _ _ of a partition _
@xmath19 of @xmath1 as an ordered pair @xmath20 of elements @xmath21 that are in different blocks of the partition .
the notion of `` a distinction of a partition '' plays the analogous role in partition logic as the notion of `` an element of a subset '' in subset logic .
the set of distinctions of a partition @xmath22 is its _ dit set _ @xmath23 .
the subsets of @xmath1 are partially ordered by inclusion with the universe set @xmath1 as the top of the order and the empty set @xmath10 as the bottom of the order .
a partition @xmath19 _ refines _ a partition @xmath24 , written @xmath25 , if each block @xmath26 is contained in some block @xmath27 .
the partitions of @xmath1 are partially ordered by refinement which is equivalent to the inclusion ordering of dit sets .
the discrete partition @xmath28 , where the blocks are all the singletons , is the top of the order , and the indiscrete partition @xmath29 ( with just one block @xmath1 ) is the bottom .
only the self - pairs @xmath30 of the diagonal @xmath31 can never be a distinction .
all the possible distinctions @xmath32 are the dits of @xmath33 and no dits are distinctions of @xmath34 just as all the elements are in @xmath1 and none in @xmath10 . in this manner
, we can construct a table of analogies between subset logic and partition logic .
[ c]|c||c|c| & subset logic & partition logic + ` elements ' & elements @xmath35 of @xmath14 & dits @xmath20 of @xmath22 + order & inclusion @xmath36 & refinement : @xmath37 + top of order & @xmath1 all elements & @xmath38 , all dits + bottom of order & @xmath10 no elements & @xmath39 , no dits + variables in formulas & subsets @xmath14 of @xmath1 & partitions @xmath22 on @xmath1 + operations & subset ops . &
partition ops .
@xcite + formula @xmath40 holds & @xmath35 element of @xmath41 & @xmath42 dit of @xmath43 + valid formula & @xmath44 , @xmath45 & @xmath46 , @xmath47 + table of analogies between subset and partition logics a dit set @xmath23 of a partition on @xmath1 is a subset of @xmath48 of a particular kind , namely the complement of an equivalence relation .
an _ equivalence relation _ is reflexive , symmetric , and transitive .
hence the complement is a subset @xmath49 that is : 1 .
irreflexive ( or anti - reflexive ) , @xmath50 ; 2 .
symmetric , @xmath51 implies @xmath52 ; and 3 .
anti - transitive ( or co - transitive ) , if @xmath53 then for any @xmath54 , @xmath55 or @xmath56 , and such binary relations will be called _ partition relations _ ( also called _ apartness relations _ ) . given any subset @xmath57 ,
the _ reflexive - symmetric - transitive ( rst ) closure _ @xmath58 of the complement @xmath59 is the smallest equivalence relation containing @xmath59 , so its complement is the largest partition relation contained in @xmath14 , which is called the _ interior _
@xmath60 of @xmath14 .
this usage is consistent with calling the subsets that equal their rst - closures _
closed subsets _ of @xmath48 ( so closed subsets = equivalence relations ) so the complements are the _ open subsets _ (= partition relations ) .
however it should be noted that the rst - closure is not a topological closure since the closure of a union is not necessarily the union of the closures , so the `` open '' subsets do not form a topology on @xmath48 .
the interior operation @xmath61 provides a universal way to define operations on partitions from the corresponding subset operations : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ apply the subset operation to the dit sets and then , if necessary , take the interior to obtain the dit set of the partition operation . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ given partitions @xmath17 and @xmath62 on @xmath1 , their _ join _
@xmath63 is the partition whose dit set @xmath64 is the interior of @xmath65 ( since the union @xmath66 is the subset join operation ) .
but the union of partition relations ( open subsets ) is a partition relation ( open subset ) so that : @xmath67 .
this gives the same join @xmath63 as the usual definition which is the partition whose blocks are the non - empty intersections @xmath68 for @xmath26 and @xmath27 .
to define the _ meet _ @xmath69 of the two partitions , we apply the subset meet operation of intersection to the dit sets and then take the interior ( which is necessary in this case ) :
@xmath70 $ ] .
this gives the same result as the usual definition of the partition meet in the literature .
defined here with refinement as the ordering relation . ]
perhaps surprisingly , the other logical operations such as the implication do not seem to be defined for partitions in the literature .
since the subset operation of implication is @xmath8 , we define the _ partition implication _ @xmath71 as the partition whose dit set is : @xmath72 $ ] .
is the partition that is like @xmath22 except that whenever a block @xmath26 is contained in a block @xmath27 , then @xmath18 is ` discretized ' in the sense of being replaced by all the singletons @xmath73 for @xmath74 .
then it is immediate that the refinement @xmath25 holds iff @xmath75 , as we would expect from the corresponding relation , @xmath36 iff @xmath76 , in subset logic . ]
the refinement partial order @xmath25 is just inclusion of dit sets , i.e. , @xmath25 iff @xmath37 . if we denote the lattice of partitions ( using the refinement ordering ) as @xmath77 , then the mapping : @xmath78 dit set representation of partition lattice represents the lattice of partitions as the lattice @xmath79 of open subsets ( under inclusion ) of @xmath80 . for any finite set @xmath81 , a ( finite ) _ measure _
@xmath82 is a function @xmath83 such that : 1 .
@xmath84 , 2 . for any @xmath85 , @xmath86 , and 3 . for any disjoint subsets @xmath87 and @xmath88 , @xmath89 .
any finite set @xmath81 has the _ counting measure _ @xmath90 and _ normalized counting measure _
@xmath91 defined on the subsets of @xmath81 .
hence for finite @xmath1 , we have the counting measure @xmath92 and the normalized counting measure @xmath93 defined on @xmath94 .
boole used the normalized counting measure @xmath95 defined on the power - set boolean algebra @xmath96 to define the logical probability @xmath97 of an event @xmath2.@xcite in view of the analogy between elements in subset logic and dits in partition logic , the construction analogous to the logical probability is the normalized counting measure applied to dit sets . that is the definition of the : @xmath98 _ logical entropy of a partition _ @xmath22 .
thus the logical entropy function @xmath99 is the dit set representation composed with the normalized counting measure : @xmath100 .
logical entropy function one immediate consequence is the inclusion - exclusion principle : @xmath101 which provides the motivation for our definition below of @xmath102 as the `` logical mutual information '' of the partitions @xmath22 and @xmath103 . in a random ( i.e. , equiprobable ) drawing of an element from @xmath1 , the event @xmath104
occurs with the probability @xmath16 . if we take two independent ( i.e. , with replacement ) random drawings from @xmath1 , i.e. , pick a random ordered pair from @xmath48 , then @xmath105 is the probability that the pair is a distinction of @xmath22 , i.e. , that @xmath22 distinguishes .
these analogies are summarized in the following table which uses the language of probability theory ( e.g. , set of outcomes , events , the occurrence of an event ) : [ c]|c||c|c| & subset logic & partition logic + ` outcomes ' & elements @xmath35 of @xmath14 & ordered pairs @xmath106 + ` events ' & subsets @xmath14 of @xmath1 & partitions @xmath22 of @xmath1 + ` event occurs ' & @xmath107 & @xmath108 + norm .
counting measure & @xmath109 & @xmath110 + interpretation & prob .
event @xmath14 occurs & prob .
partition @xmath22 distinguishes + table of quantitative analogies between subset and partition logics .
thus logical entropy @xmath111 is the simple quantitative measure of the distinctions of a partition @xmath22 just as the logical probability @xmath112 is the quantitative measure of the elements in a subset @xmath14 . in short , information theory is to partition logic as probability theory is to ordinary subset logic . to generalize logical entropy from partitions to finite probability distributions ,
note that : @xmath113 .
using @xmath114 , we have : @xmath115 .
an ordered pair @xmath116 for some @xmath26 is an _ indistinction _ or _ indit _ of @xmath22 where @xmath117 . hence in a random drawing of a pair from @xmath48 ,
@xmath118 is the probability of drawing an indistinction , while @xmath119 is the probability of drawing a distinction .
entropies will be defined both for partitions on finite sets and for finite probability distributions ( i.e. , finite random variables ) . given a random variable @xmath35 with the probability distribution @xmath120 over the @xmath121 distinct values @xmath122 , a distinction of the discrete partition on @xmath1 is just a pair @xmath123 with @xmath124 and with the probability @xmath125 . applying the previous notion to the logical entropy of a partition to this case with @xmath126 ( where @xmath127 ) , we have the : @xmath128 _ logical entropy of a finite probability distribution _ @xmath129 .
of the random variable @xmath35 but since the values of @xmath35 are irrelevant ( other than being distinct for @xmath124 ) , we can take the logical entropy @xmath130 as a function solely of the probability distribution @xmath129 of the random variable . ] since @xmath131 , we again have the logical entropy @xmath132 as the probability @xmath133 of drawing a distinction in two independent samplings of the probability distribution @xmath129 .
this is also clear from defining the product measure on the subsets @xmath57 : @xmath134 _ product measure _ on @xmath48 then the logical entropy @xmath135 is just the product measure of the@xmath136dit set of the discrete partition on @xmath1 .
there is also the obvious generalization to consider any partition @xmath22 on @xmath1 and then define for each block @xmath26 , @xmath137 . then the logical entropy @xmath138 is the product measure of the dit set of @xmath22 ( so it is still interpreted as the probability of drawing a distinction of @xmath22 ) and that is equivalent to @xmath139 . for the uniform distribution @xmath140 , the logical entropy has its maximum value of @xmath141 .
regardless of the first draw ( even for a different probability distribution over the same @xmath121 outcomes ) , the probability that the second draw is different is @xmath141 .
the logical entropy has its minimum value of @xmath12 for @xmath142 so that : @xmath143 . an important special case is a set @xmath1 of @xmath144 equiprobable elements and a partition @xmath22 on @xmath1 with @xmath121 equal - sized blocks of @xmath145 elements each .
then the number of distinctions of elements is @xmath146 which normalizes to the logical entropy of @xmath147 and which is independent of @xmath148 .
thus it holds when @xmath149 and we take the elements to be the equal blocks themselves .
thus for an equal - blocked partition on a set of equiprobable elements , the normalized number of distinctions of elements is the same as the normalized number of distinctions of blocks , and that quantity is the : @xmath150 _ logical entropy of an equiprobable set of _ @xmath121 _ _
elements__. it might be noted that no averaging is involved in the interpretation of @xmath105 .
it is the number of distinctions @xmath151 normalized .
the definition of the logical entropy @xmath152 of a probability distribution @xmath120 is in the form of the average value of the random variable which has the value @xmath153 with the probability @xmath154 .
hence the formula can be arrived at by applying the law of large numbers in the form where the finite random variable @xmath81 takes the value @xmath155 with probability @xmath154 : @xmath156 . at each step @xmath157 in repeated independent sampling @xmath158 of the probability distribution @xmath120 , the probability that the @xmath159 result @xmath160 was _ not _
@xmath160 is @xmath161
so the _ average _ probability of the result being different than it was at each place in that sequence is : @xmath162 . in the long run
, the _ typical _ sequences will dominate where the @xmath163 outcome is sampled @xmath164 times so that we have the value @xmath165 occurring @xmath164 times : @xmath166 . the logical entropy @xmath167 is usually interpreted as the _ pair - drawing probability of getting distinct outcomes _ from the distribution @xmath168 .
now we have a different interpretation of logical entropy as _ the average probability of being different_. the logical entropy formula @xmath169 is the probability of getting distinct values @xmath170 in two independent samplings of the random variable @xmath35 .
the complementary measure @xmath171 is the probability that the two drawings yield the same value from @xmath1 .
thus @xmath172 is a measure of heterogeneity or diversity in keeping with our theme of information as distinctions , while the complementary measure @xmath173 is a measure of homogeneity or concentration .
historically , the formula can be found in either form depending on the particular context . the @xmath154 s might be relative shares such as the relative share of organisms of the @xmath163 species in some population of organisms , and then the interpretation of @xmath154 as a probability arises by considering the random choice of an organism from the population . according to i. j. good
, the formula has a certain naturalness : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if @xmath174 are the probabilities of @xmath175 mutually exclusive and exhaustive events , any statistician of this century who wanted a measure of homogeneity would have take about two seconds to suggest @xmath176 which i shall call @xmath177 .
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ as noted by bhargava and uppuluri @xcite , the formula @xmath178 was used by gini in 1912 ( @xcite reprinted in @xcite ) as a measure of mutability or diversity .
but another development of the formula ( in the complementary form ) in the early twentieth century was in cryptography .
the american cryptologist , william f. friedman , devoted a 1922 book ( @xcite ) to the `` index of coincidence '' ( i.e. , @xmath179 ) .
solomon kullback ( of the kullback - leibler divergence treated later ) worked as an assistant to friedman and wrote a book on cryptology which used the index .
@xcite during world war ii , alan m. turing worked for a time in the government code and cypher school at the bletchley park facility in england .
probably unaware of the earlier work , turing used @xmath180 in his cryptoanalysis work and called it the _ repeat rate _ since it is the probability of a repeat in a pair of independent draws from a population with those probabilities ( i.e. , the identification probability @xmath181 ) .
polish cryptoanalyists had independently used the repeat rate in their work on the enigma @xcite . after the war , edward h. simpson , a british statistician , proposed @xmath182 as a measure of species concentration ( the opposite of diversity ) where @xmath22 is the partition of animals or plants according to species and where each animal or plant is considered as equiprobable . and simpson gave the interpretation of this homogeneity measure as `` the probability that two individuals chosen at random and independently from the population will be found to belong to the same group.''@xcite hence @xmath183 is the probability that a random ordered pair will belong to different species , i.e. , will be distinguished by the species partition . in the biodiversity literature @xcite ,
the formula is known as `` simpson s index of diversity '' or sometimes , the _ gini - simpson index @xcite_. however , simpson along with i. j. good worked at bletchley park during wwii , and , according to good , `` e. h. simpson and i both obtained the notion [ the repeat rate ] from turing . ''
@xcite when simpson published the index in 1948 , he ( again , according to good ) did not acknowledge turing `` fearing that to acknowledge him would be regarded as a breach of security . ''
@xcite in 1945 , albert o. hirschman ( @xcite and @xcite ) suggested using @xmath184 as an index of trade concentration ( where @xmath154 is the relative share of trade in a certain commodity or with a certain partner ) .
a few years later , orris herfindahl @xcite independently suggested using @xmath176 as an index of industrial concentration ( where @xmath154 is the relative share of the @xmath163 firm in an industry ) . in the industrial economics literature , the index @xmath185 is variously called the hirschman - herfindahl index , the hh index , or just the h index of concentration . if all the relative shares were equal ( i.e. , @xmath186 ) , then the identification or repeat probability is just the probability of drawing any element , i.e. , @xmath187 , so @xmath188 is the number of equal elements .
this led to the `` numbers equivalent '' interpretation of the reciprocal of the h index @xcite .
in general , given an event with probability @xmath189 , the _ numbers - equivalent interpretation _ of the event is that it is ` as if ' an element was drawn out of a set @xmath190 of @xmath191 equiprobable elements ( it is ` as if ' since @xmath192 need not be an integer ) .
this interpretation will be used later in the dit - bit connection . in view of the frequent and independent discovery and rediscovery of the formula @xmath180 or its complement @xmath193 by gini , friedman ,
turing , hirschman , herfindahl , and no doubt others , i. j. good wisely advises that `` it is unjust to associate @xmath177 with any one person . ''
@xcite two elements from @xmath194 are either identical or distinct .
gini @xcite introduced @xmath195 as the distance between the @xmath163 and @xmath159 elements where @xmath196 for @xmath197 and @xmath198 . since @xmath199 , the logical entropy , i.e. , gini s index of mutability , @xmath200 , is the average logical distance between a pair of independently drawn elements .
but one might generalize by allowing other distances @xmath201 for @xmath197 ( but always @xmath198 ) so that @xmath202 would be the average distance between a pair of independently drawn elements from @xmath1 . in 1982 , c. r. ( calyampudi radhakrishna ) rao introduced precisely this concept as _ quadratic entropy _ @xcite .
in many domains , it is quite reasonable to move beyond the bare - bones _ logical distance _ of @xmath196 for @xmath197 ( i.e. , the complement @xmath203 of the kronecker delta ) so that rao s quadratic entropy is a useful and easily interpreted generalization of logical entropy . : @xmath204 . ]
the shannon entropy will first be motivated in the usual fashion and then developed from the basic logical notion of entropy .
shannon , like ralph hartley @xcite before him , starts with the question of how much `` information '' is required to single out a designated element from a set @xmath1 of equiprobable elements .
this is often formulated in terms of the search @xcite for a hidden element like the answer in a twenty questions game or the sent message in a communication . but being able to always find the designated element is equivalent to being able to distinguish all elements from one another .
that is , if the designated element was in a set of two or more elements that had not been distinguished from one another , then one would not be able to single out the designated element .
thus `` singling out '' or `` identifying '' an element in a set is just another way to conceptualize `` distinguishing '' all the elements of the set .
intuitively , one might measure `` information '' as the minimum number of yes - or - no questions in a game of twenty questions that it would take in general to _ distinguish _ all the possible `` answers '' ( or `` messages '' in the context of communications ) .
this is readily seen in the simple case where @xmath205 , i.e. , the size of the set of equiprobable elements is a power of @xmath206 . then following the lead of wilkins over three centuries earlier , the @xmath207 elements could be encoded using words of length @xmath208 in a binary code such as the digits @xmath209 of binary arithmetic ( or @xmath210 in the case of wilkins ) .
then an efficient or minimum set of yes - or - no questions needed to single out the hidden element is the set of @xmath208 questions : `` is the @xmath159 digit in the binary code for the hidden element a @xmath11 ? '' for @xmath211 .
each element is distinguished from any other element by their binary codes differing in at least one digit .
the information gained in finding the outcome of an equiprobable binary trial , like flipping a fair coin , is what shannon calls a _ bit _ ( derived from `` binary digit '' ) .
hence the information gained in distinguishing all the elements out of @xmath207 equiprobable elements is : @xmath212 bits where @xmath213 is the probability of any given element ( henceforth all logs to base @xmath206 ) .
this is usefully restated in terms of partitions .
given two partitions @xmath17 and @xmath24 of @xmath1 , their _ join _
@xmath63 is the partition of @xmath1 whose blocks are the non - empty intersections @xmath68 for @xmath26 and @xmath27 .
the determination of the @xmath159 digit in the binary code for the hidden element defines a binary partition @xmath214 of @xmath1 . then to say that the answers to the @xmath208 questions above distinguish all the elements means that the join , @xmath215 , is the discrete partition on the set @xmath1 with cardinality @xmath207 .
thus we could also take @xmath216 as the minimum number of binary partitions necessary to distinguish the elements ( i.e. , to single out any given element ) . in the more general case where @xmath217 is not a power of @xmath206
, we extrapolate to the definition of @xmath218 where @xmath219 as : @xmath220 _ shannon - hartley entropy for an equiprobable set _ @xmath1 of @xmath121 elements .
the definition is further extrapolated to the case where we are only given a probability @xmath189 so that we say that @xmath221 binary partitions are needed to distinguish a set of @xmath191 elements when @xmath191 is not an integer .
this interpretation of the special case of @xmath207 or more generally @xmath222 equiprobable elements is extended to an arbitrary finite probability distribution @xmath120 by an averaging process . for the @xmath163 outcome ( @xmath223 ) , its probability @xmath154 is `` as if '' it were drawn from a set of @xmath224 equiprobable elements ( ignoring that @xmath224 may not be an integer for this averaging argument ) so the shannon - hartley information content of distinguishing the equiprobable elements of such a set would be @xmath225 .
but that occurs with probability @xmath154 so the probabilistic average gives the usual definition of the : @xmath226 _ shannon entropy of a finite probability distribution _ @xmath129 . for the uniform distribution @xmath140 , the shannon entropy has it maximum value of @xmath227 while the minimum value is @xmath12 for the trivial distribution @xmath228 so that : @xmath229
. shannon makes this averaging argument rigorous by using the law of large numbers .
suppose that we have a three - letter alphabet @xmath230 where each letter was equiprobable , @xmath231 , in a multi - letter message .
then a one - letter or two - letter message can not be exactly coded with a binary @xmath232 code with equiprobable @xmath12 s and @xmath11 s . but any probability can be better and better approximated by longer and longer representations in the binary number system . hence we can consider longer and longer messages of @xmath148 letters along with better and better approximations with binary codes .
the long run behavior of messages @xmath158 where @xmath233 is modeled by the law of large numbers so that the letter @xmath234 will tend to occur @xmath235 times and similarly for @xmath236 and @xmath237 . such a message is called _
typical_. the probability of any one of those typical messages is : @xmath238 ^{n}$ ] or , in this case , @xmath239 ^{n}=\left ( \frac{1}{3}\right ) ^{n}$ ] . hence the number of such typical messages is @xmath240 .
if each message was assigned a unique binary code , then the number of @xmath232 s in the code would have to be @xmath81 where @xmath241 or @xmath242 .
hence the number of equiprobable binary questions or bits needed per letter of the messages is : @xmath243 .
this example shows the general pattern . in the general case ,
let @xmath120 be the probabilities over a @xmath121-letter alphabet @xmath244 .
in an @xmath148-letter message , the probability of a particular message @xmath158 is @xmath245 where @xmath246 could be any of the symbols in the alphabet so if @xmath247 then @xmath248 . in a _
typical _ message , the @xmath163 symbol will occur @xmath164 times ( law of large numbers ) so the probability of a typical message is ( note change of indices to the letters of the alphabet ) : @xmath249 ^{n}$ ] . since the probability of a typical message is @xmath250 for @xmath251 , the typical messages are equiprobable .
hence the number of typical messages is @xmath252 ^{n}$ ] and assigning a unique binary code to each typical message requires @xmath81 bits where @xmath253 ^{n}$ ] where : @xmath254 ^{n}\right\ } = n\log\left [ \pi_{k=1}^{n}p_{k}^{-p_{k}}\right ] $ ] @xmath255 @xmath256 .
hence the shannon entropy @xmath257 is interpreted as the limiting _
average number of bits necessary per letter in the message_. in terms of distinctions , this is the _ average number of binary partitions necessary per letter to distinguish the messages_. it is this averaging result that allows us to consider `` the number of binary partitions it takes to distinguish the elements of @xmath1 '' when @xmath258 is not a power of @xmath206 since `` number '' is interpreted as `` average number . ''
shannon entropy can also be defined for a partition @xmath17 on a set @xmath1 .
if the elements of @xmath1 are equiprobable , then the probability that a randomly drawn element is in a block @xmath26 is @xmath259 . in a set of @xmath260
equiprobable elements , it would take ( on average ) @xmath261 binary partitions to distinguish the elements . averaging over the blocks
, we have the : @xmath262 _ shannon entropy of a partition _ @xmath22 .
the functional form of shannon s formula is often further `` justified '' or `` motivated '' by asserting that it is the same as the notion of entropy in statistical mechanics , and hence the name `` entropy . ''
the name `` entropy '' is here to stay but the justification of the formula by reference to statistical mechanics is not quite correct .
the connection between entropy in statistical mechanics and shannon s entropy is only via a numerical approximation , the stirling approximation , where if the first two terms in the stirling approximation are used , then the shannon formula is obtained .
the first two terms in the stirling approximation for @xmath263 are : @xmath264 .
the first three terms in the stirling approximation are : @xmath265 .
if we consider a partition on a finite @xmath1 with @xmath144 , with @xmath121 blocks of size @xmath266 , then the number of ways of distributing the individuals in these @xmath121 boxes with those numbers @xmath267 in the @xmath163 box is : @xmath268 . the normalized natural log of @xmath269
, @xmath270 is one form of entropy in statistical mechanics .
indeed , the formula `` @xmath271 '' is engraved on boltzmann s tombstone . the entropy formula
can then be developed using the first two terms in the stirling approximation .
@xmath272 $ ] @xmath273 -\sum _ { i}n_{i}\left [ \ln\left ( n_{i}\right ) -1\right ] \right ] $ ] @xmath274 = \frac{1}{n}\left [ \sum n_{i}\ln\left ( n\right ) -\sum n_{i}\ln\left ( n_{i}\right ) \right ] $ ] @xmath275 where @xmath276 ( and where the formula with logs to the base @xmath277 only differs from the usual base @xmath206 formula by a scaling factor ) .
entropy @xmath278 is in fact an excellent numerical approximation to @xmath270 for large @xmath148 ( e.g. , in statistical mechanics ) .
but the common claim is that shannon s entropy has the _ same functional form _ as entropy in statistical mechanics , and that is simply false .
if we use a three - term stirling approximation , then we obtain an even better numerical approximation : , mackay @xcite also uses stirling s approximation to give a `` more accurate approximation '' ( using the next term in the stirling approximation ) to the entropy of statistical mechanics than the shannon entropy . ] @xmath279 but no one would suggest using that `` more accurate '' entropy formula in information theory .
shannon s formula should be justified and understood by the arguments given previously , and not by over - interpreting the approximate relationship with entropy in statistical mechanics .
the basic datum is `` the '' set @xmath280 of @xmath121 elements with the equal probabilities @xmath219 . in that basic case of an equiprobable set
, we can derive the dit - bit connection , and then by using a probabilistic average , we can develop the shannon entropy , expressed in terms of bits , from the logical entropy , expressed in terms of ( normalized ) dits , or vice - versa .
given @xmath280 with @xmath121 equiprobable elements , the number of dits ( of the discrete partition on @xmath280 ) is @xmath281 so the normalized dit count is : @xmath282 normalized dits .
that is the dit - count or logical measure of the information is a set of @xmath121 distinct elements . or the normalized count @xmath283 . ]
but we can also measure the information in the set by the number of binary partitions it takes ( on average ) to distinguish the elements , and that bit - count is : @xmath284 bits . by solving the dit - count and the bit - count for @xmath189 and equating
, we can derive each measure in terms of the other : @xmath285 and @xmath286 the dit - bit conversion formulas .
the common thing being measured is an equiprobable @xmath280 where @xmath287 .
the dit - count for @xmath280 is @xmath288 and the bit - count for @xmath280 is @xmath289 , and the bit - dit connection gives the relationship between the two counts .
using this dit - bit connection between the two different ways to measure the `` information '' in @xmath280 , each entropy can be developed from the other .
we start with the logical entropy of a probability distribution @xmath168 : @xmath290 .
it is expressed as the probabilistic average of the dit - counts or logical entropies of the sets @xmath291 with @xmath224 equiprobable elements . , there is no necessity that @xmath292 is an integer so the dit - counts for @xmath291 are extrapolations while the bit - counts or binary partition counts for @xmath293 are already extrapolations even when @xmath121 is an integer but not a power of @xmath206 . ]
but if we switch to the binary - partition bit - counts of the information content of those same sets @xmath291 of @xmath224 equiprobable elements , then the bit - counts are @xmath294 and the probabilistic average is the shannon entropy : @xmath295 . both entropies have the mathematical form : @xmath296 and differ by using either the dit - count or bit - count to measure the information in @xmath291 .
clearly the process is reversible , so one can use the dit - bit connection in reverse to develop the logical entropy @xmath130 from the shannon entropy @xmath297 .
thus the two notions of entropy are simply two different ways , using distinctions ( dit - counts ) or binary partitions ( bit - counts ) , to measure the information in a probability distribution .
moreover the dit - bit connection carries over to the compound notions of entropy so that the shannon notions of conditional entropy , mutual information , and joint entropy can be developed from the corresponding notions for logical entropy . since the logical notions are the values of a probability measure , the compound notions of logical entropy have the usual venn diagram relations such as the inclusion - exclusion principle .
there is a well - known analogy between the `` venn diagram '' relationships for the shannon entropies and the relationships satisfied by any measure on a set ( @xcite , @xcite ) . as l. l. campbell puts it , the analogy : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ suggests the possibility that @xmath298 and @xmath299 are measures of sets , that @xmath300 is the measure of their union , that @xmath301 is the measure of their intersection , and that @xmath302 is the measure of their difference .
the possibility that @xmath303 is the entropy of the `` intersection '' of two partitions is particularly interesting .
this `` intersection , '' if it existed , would presumably contain the information common to the partitions @xmath304 and @xmath305.@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ all of campbell s desiderata are precisely true when : * `` sets '' = dit sets , and * `` entropies '' = normalized counting measure of the ( dit ) sets , i.e. , the logical entropies . since the logical entropies are the values of a measure , by developing the corresponding shannon notions from the logical ones , we have an explanation of why the shannon notions also exhibit the same venn diagram relationships . the expository strategy is to first develop the shannon and logical compound notions of entropy separately and then to show the relationship using the dit - bit connection .
given two partitions @xmath17 and @xmath62 on a finite set @xmath1 , how might one measure the new information that is provided by @xmath22 that was not already in @xmath103 ?
campbell suggests associating sets with partitions so the conditional entropy would be the measure of the difference between the sets .
taking the information as distinctions , we take the difference between the@xmath136dit sets , i.e. , @xmath306 , and then take the normalized counting measure of that subset of @xmath307 : @xmath308 _ logical conditional entropy of _ @xmath22 _ _ given _ _
@xmath103 .
when the two partitions @xmath22 and @xmath103 are joined together in the join @xmath63 , whose blocks are the non - empty intersections @xmath68 , their information as distinctions is also joined together as sets , @xmath67 ( the `` union '' mentioned by campbell ) , which has the normalized counting measure of : @xmath309 $ ] _ logical entropy of a partition join _ @xmath63 .
this logical entropy is interpreted as the probability that a pair of random draws from @xmath1 will yield a @xmath22-distinction _ or _ a @xmath310-distinction ( where `` or '' includes both ) .
then the relationships between the logical entropy concepts can be read off the venn diagram inclusion - exclusion principle for the dit sets : @xmath311 so that @xmath312 .
figure1.eps figure 1 : venn diagram for subsets of @xmath48 the shaded area in the venn diagram has the dit - count measure : @xmath313 @xmath314 . for
the corresponding definitions for random variables and their probability distributions , consider a random variable @xmath315 taking values in the product @xmath316 of finite sets with the joint probability distribution @xmath317 , and thus with the marginal distributions : @xmath318 and @xmath319 .
for notational simplicity , the entropies can be considered as functions of the random variables or of their probability distributions , e.g. , @xmath320 . for the joint distribution
, we have the : @xmath321 $ ] _ logical entropy of the joint distribution _ which is the probability that two samplings of the joint distribution will yield a pair of _ distinct _ ordered pairs @xmath322 , @xmath323 , i.e. , with an @xmath81-distinction @xmath324 _ or _ a @xmath325-distinction @xmath326 . for the definition of the conditional entropy @xmath327
, we simply take the product measure of the set of pairs @xmath315 and @xmath328 that give an @xmath81-distinction but not a @xmath325-distinction .
thus given the first draw @xmath315 , we can again use a venn diagram to compute the probability that the second draw @xmath328 will have @xmath329 but @xmath330 . to illustrate this using venn diagram reasoning ,
consider the probability measure defined by @xmath317 on the subsets of @xmath316 .
given the first draw @xmath315 , the probability of getting an @xmath315-distinction on the second draw is @xmath331 and the probability of getting a @xmath332-distinction is @xmath333 . a draw that is a @xmath332-distinction is , a fortiori , an @xmath322-distinction so the area @xmath334 is contained in the area @xmath335 .
then the probability of getting an @xmath315-distinction that is not a @xmath332-distinction on the second draw is the difference : @xmath336 .
figure2.eps figure 2 : @xmath337 = probability of an @xmath338-distinction but not a @xmath332-distinction on @xmath316 .
since the first draw @xmath315 was with probability @xmath317 , we have the following as the product measure of the subset of @xmath339 ^{2}$ ] of pairs @xmath340 $ ] that are @xmath81-distinctions but not @xmath325-distinctions : @xmath341 $ ] _ logical conditional entropy of _ @xmath338 _ _ given _ _ @xmath332 .
then a little algebra quickly yields : @xmath341 $ ] @xmath342 -\left [ 1-\sum _ { y}p\left ( y\right ) ^{2}\right ] = h\left ( x , y\right ) -h\left ( y\right ) $ ] .
the summation over @xmath317 recasts the venn diagram to the set @xmath343 where the product probability measure ( for the two independent draws ) gives the logical entropies : figure3.eps figure 3 : @xmath344 .
it might be noted that the logical conditional entropy , like the other logical entropies , is not just an average ; the conditional entropy is the product probability measure of the subset : @xmath345 : x\neq x^{\prime},y = y^{\prime}\right\ } \subseteq\left ( x\times y\right ) \times\left ( x\times y\right ) $ ] .
the shannon conditional entropy for partitions @xmath22 and @xmath103 is based on subset reasoning which is then averaged over a partition . given a subset @xmath346 @xmath347 ,
a partition @xmath348 induces a partition of @xmath346 with the blocks @xmath349 .
then @xmath350 is the probability distribution associated with that partition so it has a shannon entropy which we denote : @xmath351 .
the shannon conditional entropy is then obtained by averaging over the blocks of @xmath103 : @xmath352 _ shannon conditional entropy of _ @xmath22 _ _ given _ _
@xmath103 .
since the join @xmath63 is the partition whose blocks are the non - empty intersections @xmath68 , @xmath353 .
developing the formula gives : @xmath354 = h\left ( \pi \vee\sigma\right ) -h\left ( \sigma\right ) $ ] .
thus the conditional entropy @xmath355 is interpreted as the shannon - information contained in the join @xmath63 that is not contained in @xmath103 .
figure4.eps figure 4 : @xmath356 `` venn diagram picture '' for shannon conditional entropy of partitions given the joint distribution @xmath317 on @xmath316 , the conditional probability distribution for a specific @xmath357 is @xmath358 which has the shannon entropy : @xmath359 . then the conditional entropy is the average of these entropies : @xmath360 _ shannon conditional entropy of _ @xmath338 _ _ given _ _ @xmath332 . expanding as before gives @xmath361 with a similar venn diagram picture ( see below ) .
now we can develop the shannon conditional entropy from the logical conditional entropy and thereby explain the venn diagram relationship .
the logical conditional entropy is : @xmath341 $ ] where @xmath335 is the normalized dit count for the discrete partition on a set @xmath362 with @xmath363 equiprobable elements .
hence that same equiprobable set requires the bit - count of @xmath364 binary partitions to distinguish its elements .
similarly @xmath334 is the normalized dit count for ( the discrete partition on ) a set @xmath365 with @xmath366 equiprobable elements , so it requires @xmath367 binary partitions to make those distinctions .
those binary partitions are included in the @xmath368 binary partitions ( since a @xmath332-distinction is automatically a @xmath315-distinction ) and we do nt want the @xmath332-distinctions so they are subtracted off to get : @xmath369 bits . taking the same probabilistic average ,
the average number of binary partitions needed to make the @xmath338-distinctions but not the @xmath332-distinctions is : @xmath370 = \sum_{x , y}p\left ( x , y\right ) \log\left ( \frac{p\left ( y\right ) } { p\left ( x , y\right ) } \right ) = h\left ( x|y\right ) .$ ] replacing the dit - counts by the bit - counts for the equiprobable sets , and taking the probabilistic average gives the same venn diagram picture for the shannon entropies .
figure5.eps figure 5 : @xmath371 .
if the atom of information is the distinction or dit , then the atomic information in a partition @xmath22 is its dit set , @xmath372 . following again campbell s dictum about the mutual information , the information common to two partitions @xmath22 and @xmath103 would naturally be the intersection of their@xmath136dit sets : @xmath373 _ mutual information set_. it is an interesting and not completely trivial fact that as long as neither @xmath22 nor @xmath103 are the indiscrete partition @xmath34 ( where @xmath374 ) , then @xmath22 and @xmath103 have a distinction in common .
[ non - empty dit sets intersect]given two partitions @xmath22 and @xmath103 on @xmath1 with non - empty dit sets , @xmath375 . , if every pair of elements @xmath21 is equated by one or the other of the relations , i.e. , @xmath376 , then either @xmath377 or @xmath378 . ]
since @xmath22 is not the indiscrete partition , consider two elements @xmath35 and @xmath379 distinguished by @xmath22 but identified by @xmath103 [ otherwise @xmath380 .
since @xmath103 is also not the indiscrete partition , there must be a third element @xmath381 not in the same block of @xmath103 as @xmath35 and @xmath379 .
but since @xmath35 and @xmath379 are in different blocks of @xmath22 , the third element @xmath381 must be distinguished from one or the other or both in @xmath22 .
hence @xmath382 or @xmath383 must be distinguished by both partitions and thus must be in their mutual information set @xmath384.@xmath385 the dit sets @xmath23 and their complementary indit sets (= equivalence relations ) @xmath386 are easily characterized as : @xmath387 the mutual information set can also be characterized in this manner .
[ structure of mutual information sets]given partitions @xmath22 and @xmath103 with blocks @xmath388 and @xmath389 , then @xmath390 .
the union ( which is a disjoint union ) will include the pairs @xmath20 where for some @xmath26 and @xmath27 , @xmath391 and @xmath392 .
since @xmath379 is in @xmath346 but not in the intersection @xmath68 , it must be in a different block of @xmath22 than @xmath18 so @xmath393 .
symmetrically , @xmath394 so @xmath395 . conversely if @xmath396 then take the @xmath18 containing @xmath35 and the @xmath346 containing @xmath379
. since @xmath20 is distinguished by both partitions , @xmath397 and @xmath398 so that @xmath399.@xmath385 the probability that a pair randomly chosen from @xmath48 would be distinguished by @xmath22 _ and _ @xmath103 would be given by the normalized counting measure of the mutual information set which is the : @xmath400 @xmath401 probability that @xmath22 and @xmath103 distinguishes _ mutual logical information of _ @xmath22 _ _ and _ _ @xmath103 . by the inclusion - exclusion principle : @xmath402 .
normalizing , the probability that a random pair is distinguished by both partitions is given by the inclusion - exclusion principle : @xmath403 inclusion - exclusion principle for logical entropies of partitions this can be extended after the fashion of the inclusion - exclusion principle to any number of partitions .
it was previously noted that the intersection of two dit sets is not necessarily the dit set of a partition , but the interior of the intersection is the dit set @xmath404 of the partition meet @xmath69 .
hence we also have the : @xmath405 submodular inequality for logical entropies .
consider again a joint distribution @xmath317 over @xmath316 for finite @xmath81 and @xmath325 .
intuitively , the mutual logical information @xmath406 in the joint distribution @xmath317 would be the probability that a sampled pair @xmath315 would be a distinction of @xmath407 _ and _ a distinction of @xmath408 .
that means for each probability @xmath317 , it must be multiplied by the probability of not drawing the same @xmath338 _ and _ not drawing the same @xmath332 ( e.g. , in a second independent drawing ) . in the venn diagram ,
the area or probability of the drawing that @xmath338 or that @xmath332 is @xmath409 ( correcting for adding the overlap twice ) so the probability of getting neither that @xmath338 nor that @xmath332 is the complement @xmath410 + \left [ 1-p\left ( y\right ) \right ] -\left [ 1-p\left ( x , y\right ) \right ] $ ] .
figure6.eps figure 6 : @xmath411 + \left [ 1-p\left ( y\right ) \right ] -\left [ 1-p\left ( x , y\right ) \right ] $ ] = shaded area in venn diagram for @xmath316 hence we have : @xmath412 + \left [ 1-p\left ( y\right ) \right ] -\left [ 1-p\left ( x , y\right ) \right ] \right ] $ ] _ logical mutual information in a joint probability distribution_. the probability of two independent draws differing in _ either _ the @xmath338 _ or _
the @xmath332 is just the logical entropy of the joint distribution : @xmath413 = 1-\sum_{x , y}p\left ( x , y\right ) ^{2}$ ] . using a little algebra to expand the logical mutual information : @xmath414 + \left [ 1-{\textstyle\sum\nolimits_{x , y } } p\left ( x , y\right ) p\left ( y\right ) \right ] -\left [ 1-{\textstyle\sum\nolimits_{x , y } } p\left ( x , y\right ) ^{2}\right ] \\ & = h\left ( x\right ) + h\left ( y\right ) -h\left ( x , y\right)\end{aligned}\ ] ] inclusion - exclusion principle for logical entropies of a joint distribution .
figure7.eps figure 7 : @xmath415 = shaded area in venn diagram for @xmath343 .
it might be noted that the logical mutual information , like the other logical entropies , is not just an average ; the mutual information is the product probability measure of the subset : @xmath345 : x\neq x^{\prime},y\neq y^{\prime}\right\ } \subseteq\left ( x\times y\right ) \times\left ( x\times y\right ) $ ] .
the usual heuristic motivation for shannon s mutual information is much like its development from the logical mutual information so we will take that approach at the outset .
the logical mutual information for partitions can be expressed in the form : @xmath416 $ ] so if we substitute the bit - counts for the dit - counts as before , we get : @xmath417 = \sum_{b , c}p_{b\cap c}\log\left ( \frac{p_{b\cap c}}{p_{b}p_{c}}\right ) $ ] _ shannon s mutual information for partitions_. keeping the log s separate gives the venn diagram picture : @xmath418 \\ & = h\left ( \pi\right ) + h\left ( \sigma\right ) -h\left ( \pi\vee \sigma\right)\end{aligned}\ ] ] inclusion - exclusion analogy for shannon entropies of partitions . to move from partitions to probability distributions , consider again the joint distribution @xmath317 on @xmath316 . then developing the shannon mutual information from the logical mutual information amounts to replacing the block probabilities @xmath419 in the join @xmath63 by the joint probabilities @xmath317 and the probabilities in the separate partitions by the marginals ( since @xmath420 and @xmath421 ) , to obtain : @xmath422 _ shannon mutual information in a joint probability distribution_. then the same proof carries over to give the : @xmath423 figure8.eps figure 8 : inclusion - exclusion `` picture '' for shannon entropies of probability distributions .
the logical mutual information formula : @xmath412 + \left [ 1-p\left ( y\right ) \right ] -\left [ 1-p\left ( x , y\right ) \right ] \right ] $ ] develops via the dit - count to bit - count conversion to : @xmath424 = \sum _ { x , y}p\left ( x , y\right ) \log\left ( \frac{p\left ( x , y\right ) } { p\left ( x\right ) p\left ( y\right ) } \right ) = i\left ( x , y\right ) $ ] .
thus the genuine venn diagram relationships for the product probability measure that gives the logical entropies carry over , via the dit - count to bit - count conversion , to give a similar venn diagram picture for the shannon entropies .
two partitions @xmath22 and @xmath103 are said to be ( stochastically ) _ independent _ if for all @xmath26 and @xmath27 , @xmath425 . if @xmath22 and @xmath103 are independent , then : @xmath426 , so that : @xmath427 shannon entropy for partitions additive under independence . in ordinary probability theory , two events @xmath428 for a sample space @xmath1 are said to be _ independent _ if @xmath429 .
we have used the motivation of thinking of a partition - as - dit - set @xmath23 as an event in a sample space @xmath48 with the probability of that event being @xmath105 , the logical entropy of the partition .
the following proposition shows that this motivation extends to the notion of independence .
[ independent partitions have independent dit sets]if @xmath22 and @xmath103 are ( stochastically ) independent partitions , then their dit sets @xmath23 and @xmath430 are independent as events in the sample space @xmath48 ( with equiprobable points ) . for independent partitions @xmath22 and @xmath103 , we need to show that the probability @xmath431 of the event @xmath432 is equal to the product of the probabilities @xmath105 and @xmath433 of the events @xmath23 and @xmath434 in the sample space @xmath48 . by the assumption of stochastic independence , we have @xmath435 so that @xmath436 . by the previous structure theorem for the mutual information set : @xmath437 , where the union is disjoint so that : @xmath438 so that : @xmath439.@xmath385 hence the logical entropies behave like probabilities under independence ; the probability that @xmath22 _ and _ @xmath103 distinguishes , i.e. , @xmath440 , is equal to the probability @xmath441 that @xmath22 distinguishes times the probability @xmath442 that @xmath103 distinguishes : @xmath443 logical entropy multiplicative under independence .
it is sometimes convenient to think in the complementary terms of an equivalence relation `` equating '' or identifying rather than a partition distinguishing .
since @xmath105 can be interpreted as the probability that a random pair of elements from @xmath1 are distinguished by @xmath22 , i.e. , as a distinction probability , its complement @xmath444 can be interpreted as an _ identification probability _ ,
i.e. , the probability that a random pair is equated by @xmath22 ( thinking of @xmath22 as an equivalence relation on @xmath1 ) . in general , @xmath445 \left [ 1-h\left ( \sigma\right ) \right ] = 1-h\left ( \pi\right ) -h\left ( \sigma\right ) + h\left ( \pi\right ) h\left ( \sigma\right ) = \left [ 1-h\left ( \pi\vee\sigma\right ) \right ] + \left [ h\left ( \pi\right ) h\left ( \sigma\right ) -m(\pi , \sigma\right ] $ ] which could also be rewritten as : @xmath446 -\left [ 1-h\left ( \pi\right ) \right ] \left [ 1-h\left ( \sigma\right ) \right ] = m(\pi , \sigma)-h\left ( \pi\right ) h\left ( \sigma\right ) $ ] . thus if @xmath22 and @xmath103 are independent , then the probability that the join partition @xmath63 identifies is the probability that @xmath22 identifies times the probability that @xmath103 identifies : @xmath445 \left [ 1-h\left ( \sigma\right ) \right ] = \left [ 1-h\left ( \pi\vee\sigma\right ) \right ] $ ] multiplicative
identification probabilities under independence .
a joint probability distribution @xmath317 on @xmath316 is _ independent _ if each value is the product of the marginals : @xmath447 . for an independent distribution , the shannon mutual information @xmath448
is immediately seen to be zero so we have : @xmath449 shannon entropies for independent @xmath317 .
for the logical mutual information , independence gives : @xmath450 \\ & = { \textstyle\sum\nolimits_{x , y } } p\left ( x\right ) p\left ( y\right ) \left [ 1-p\left ( x\right ) -p\left ( y\right ) + p\left ( x\right ) p\left ( y\right ) \right ] \\
& = { \textstyle\sum\nolimits_{x } } p\left ( x\right ) \left [ 1-p\left ( x\right ) \right ] { \textstyle\sum\nolimits_{y } } p\left ( y\right ) \left [ 1-p\left ( y\right ) \right ] \\ & = h\left ( x\right ) h\left ( y\right)\end{aligned}\ ] ] logical entropies for independent @xmath317 .
this independence condition @xmath451 plus the inclusion - exclusion principle @xmath452 also implies that : @xmath453 \left [ 1-h\left ( y\right ) \right ] & = 1-h\left ( x\right ) -h\left ( y\right ) + h\left ( x\right ) h\left ( y\right ) \\ & = 1-h\left ( x\right ) -h\left ( y\right ) + m\left ( x , y\right ) \\ & = 1-h\left ( x , y\right ) \text{.}\ ] ] hence under independence , the probability of drawing the same pair @xmath315 in two independent draws is equal to the probability of drawing the same @xmath338 times the probability of drawing the same @xmath332 .
given two probability distributions @xmath120 and @xmath454 on the same sample space @xmath455 , we can again consider the drawing of a pair of points but where the first drawing is according to @xmath129 and the second drawing according to @xmath456 .
the probability that the points are distinct would be a natural and more general notion of logical entropy that would be the : @xmath457 _ logical _ _ cross entropy of _
@xmath129 _ and _ @xmath456 which is symmetric
. the logical cross entropy is the same as the logical entropy when the distributions are the same , i.e. , if @xmath458 , then @xmath459 .
the notion of _ cross entropy _ in shannon entropy can be developed by applying dit - bit connection to the logical cross entropy @xmath460 to obtain : @xmath461 which is not symmetrical due to the asymmetric role of the logarithm , although if @xmath458 , then @xmath462 .
since the logical cross entropy is symmetrical , it could also be expressed as @xmath463 which develops to the shannon cross entropy @xmath464 so it might be more reasonable to use a _ symmetrized cross entropy _ : @xmath465 $ ] . the _ kullback - leibler divergence _ ( or _ relative entropy _ )
@xmath466 is defined as a measure of the distance or divergence between the two distributions where @xmath467 .
a basic result is the : @xmath468 with equality if and only if @xmath458 _ information inequality _
@xcite . given two partitions @xmath22 and @xmath103 , the inequality @xmath469
is obtained by applying the information inequality to the two distributions @xmath470 and @xmath471 on the sample space @xmath472 : @xmath473 with equality iff independence . in the same manner
, we have for the joint distribution @xmath474 : @xmath475 with equality iff independence . the _ symmetrized kullback - leibler divergence _
is : @xmath476 = h_{s}\left ( p||q\right ) -\left [ \frac{h\left ( p\right ) + h\left ( q\right ) } { 2}\right ] $ ] .
but starting afresh , one might ask : what is the natural measure of the difference or distance between two probability distributions @xmath120 and @xmath454 that would always be non - negative , and would be zero if and only if they are equal ?
the ( euclidean ) distance between the two points in @xmath477 would seem to be the logical answer
so we take that distance ( squared with a scale factor ) as the definition of the : @xmath478 @xmath479 _ logical divergence _ ( or _ logical _ _ relative entropy _ ) . ] which is symmetric and we trivially have : @xmath480 with equality iff @xmath458 logical information inequality .
we have component - wise : @xmath481 -\left [ \frac{1}{n}-p_{i}^{2}\right ] -\left [ \frac{1}{n}-q_{i}^{2}\right ] $ ] so that taking the sum for @xmath223 gives : @xmath482 -\frac{1}{2}\left [ \left ( 1-{\textstyle\sum\nolimits_{i } } p_{i}^{2}\right ) + \left ( 1-{\textstyle\sum\nolimits_{i } } q_{i}^{2}\right ) \right ] \\ & = h\left ( p\vert q\right ) -\frac{h\left ( p\right ) + h\left ( q\right ) } { 2}\text{.}\ ] ] logical divergence = _ jensen difference _
@xcite between probability distributions .
then the information inequality implies that the logical cross - entropy is greater than or equal to the average of the logical entropies : @xmath483 with equality iff @xmath458 .
the half - and - half probability distribution @xmath484 that mixes @xmath129 and @xmath456 has the logical entropy of @xmath485 $ ] so that : @xmath486 with equality iff @xmath458
. mixing different @xmath129 and @xmath456 increases logical entropy . the logical divergence can be expressed as : @xmath487 -\frac{1}{2}\left [ \left ( \sum_{i}p_{i}\left ( 1-p_{i}\right ) \right ) + \left ( \sum_{i}q_{i}\left ( 1-q_{i}\right ) \right ) \right ] $ ] that develops via the dit - bit connection to : @xmath488 $ ] @xmath489 = \frac{1}{2}\left [ d\left ( p||q\right ) + d\left ( q||p\right ) \right ] $ ] @xmath490 .
thus the logical divergence @xmath491 develops via the dit - bit connection to the symmetrized version of the kullback - leibler divergence .
the following table summarizes the concepts for the shannon and logical entropies .
we use the case of probability distributions rather than partitions , and we use the abbreviations @xmath492 , @xmath493 , and @xmath494 .
[ c]l|c|c| & @xmath495 & @xmath496 + & @xmath497 & @xmath498 + & @xmath499 & @xmath500 + & @xmath501 & @xmath502 + & @xmath503 & @xmath504 \left [ 1-h\left ( y\right ) \right ] $ ] + & @xmath505 & @xmath506 $ ] + & @xmath507 & @xmath508 + & @xmath509 & @xmath510 + & @xmath511 & @xmath512 + & @xmath513 & @xmath514 /2 $ ] + & @xmath515 & @xmath516 + the above table shows many of the same relationships holding between the various forms of the logical and shannon entropies due ultimately to the dit - bit connection . the dit - bit connection between the two notions of entropy is based on them being two different measures of the `` amount of information - as - distinctions , '' the dit - count being the normalized count of the distinctions and the bit - count being the number of binary partitions required ( on average ) to make the distinctions .
logical entropies arise naturally as the normalized counting measure for partition logic just as probabilities arise as the normalized counting measure for subset logic , where the two logics are dual to one another .
all the forms of logical entropy have simple interpretations as the probabilities of distinctions .
shannon entropy is a higher - level and more refined notion adapted to the theory of communications and coding where it can be interpreted as the average number of bits necessary per letter to code the messages , i.e. , the average number of binary partitions necessary per letter to distinguish the messages .
ricotta , carlo and laszlo szeidl 2006 . towards a unifying approach to diversity measures : bridging the gap between the shannon entropy and rao s quadratic index .
_ theoretical population biology_. 70 : 237 - 43 . | the logical basis for information theory is the newly developed logic of partitions that is dual to the usual boolean logic of subsets .
the key concept is a `` distinction '' of a partition , an ordered pair of elements in distinct blocks of the partition .
the logical concept of entropy based on partition logic is the normalized counting measure of the set of distinctions of a partition on a finite set just as the usual logical notion of probability based on the boolean logic of subsets is the normalized counting measure of the subsets ( events ) .
thus logical entropy is a measure on the set of ordered pairs , and all the compound notions of entropy ( join entropy , conditional entropy , and mutual information ) arise in the usual way from the measure ( e.g. , the inclusion - exclusion principle)just like the corresponding notions of probability .
the usual shannon entropy of a partition is developed by replacing the normalized count of distinctions ( dits ) by the average number of binary partitions ( bits ) necessary to make all the distinctions of the partition . |
an hypernucleus is a bound system of nonstrange and strange baryons and it is an important source of information about baryon - baryon strangeness - changing weak interactions . in the present contribution ,
we focus on the @xmath0-hypernuclei , which can be produced by an hadronic reaction - such as @xmath6 - or by an electromagnetic one - such as @xmath7 . the @xmath0 is generally formed in an excited state and then , via the emission of a series of @xmath8-rays , it reaches its ground @xmath9-state .
the @xmath0 decays itself mainly via the weak interaction by two decay mechanisms : the so - called mesonic decay @xmath10 and the non - mesonic one @xmath11 , where no meson is present in the final state @xmath12 .
the mesonic decay is the main decay process for the free @xmath0 , but when the @xmath0 is within the nuclear environment it is strongly inhibited by the pauli principle . in this case , the non - mesonic weak decay becomes the most important decay mechanism .
this decay can be stimulated either by protons or neutrons , where the corresponding decay widths are denoted as @xmath13 and @xmath14 , respectively . just after the @xmath0-decays ,
the resultant nucleons are still within the nucleus and in their way out they can interact strongly with any nucleon of the nucleus . finally , two or more nucleons are ejected from the nucleus and it is the nucleon spectra of these emitted nucleons the magnitude which can be measured . for review articles one can see @xcite , while some of the experimental works are @xcite . for the non - mesonic @xmath0-decay , there are two quantities which deserves special attention . the first one is the ratio @xmath15 where theory predicts a value smaller than 0.5 , while the so - called experimental result has a value closer to one .
the second quantity is the asymmetry of the protons emitted in the @xmath16 decay of polarized hypernuclei . in this case , data indicate a value close to zero , while most of the theoretical works predict a large negative number .
just recently , the incorporation of the @xmath17-meson suggests a solution to this problem ( see @xcite , and references therein ) . in the present work , however , we will not deal with the asymmetry . for @xmath18 ,
the connection between the theory and the experimental results is not straightforward because there is no direct measurement of this ratio .
both @xmath19 and @xmath20 are the decay widths of the so - designed primaries disintegration and as mentioned above , data give us results on the the nucleon spectra which emerge from the nucleus .
the connection between @xmath18 and the nucleon spectra is a theoretical problem in itself . before going on , it is interesting to resume some of the theoretical efforts in dealing with @xmath18 .
the main ingredients required in the evaluation of the transition rate @xmath12 , are the transition potential and the wave functions which represent the @xmath0 and the nucleons .
the first microscopic scheme for @xmath18 has been proposed by adams @xcite , who has used the nuclear matter framework , one pion exchange model ( ope ) , @xmath21 piece of the @xmath22 coupling and short range correlations ( src ) . while this model fairly reproduce the total @xmath16 decay width @xmath23 , it produce ratios smaller than 0.20
. some of the improvements over adams model are : @xmath24 the inclusion of heavier mesons than the pion in the @xmath25transition potential @xcite ; @xmath26 the inclusion of interaction terms that violates the isospin @xmath21 rule has been considered in @xcite .
alternatively , @xmath27 the transition potential can be described in terms of the quark degree of freedom @xcite , which automatically introduces the @xmath28 contribution . and @xmath29 the employment of finite nucleus wave functions instead of plane waves , a scheme usually called wave function method ( wfm ) @xcite . this list does not pretend to be complete .
however , we should mention that in all these works the discrepancy between theory and experiment remains .
we turn now to the interpretation of data . as mentioned , it is the spectra of nucleons emerging from the nucleus , the quantity which is measured .
the nucleons originated in the @xmath0-decay interact strongly with others nucleons before leaving the nucleus .
these interactions are called final state interactions @xmath30 .
again , we can distinguish two issues : @xmath24 the primary decays and @xmath26 a model for the @xmath2 . regarding the first point
, we should note that @xmath12 is not the only @xmath16-weak decay mechanism : the @xmath16-decay width can be also stimulated by two nucleons @xmath31 .
the corresponding decay width is called @xmath32 .
this two - body induced decay is originated from ground state correlation in the hypernuclei @xcite .
an estimation of @xmath32 is important in the analysis of data , because this process is a source of nucleons which can not be distinguished from those stemming from @xmath33 .
there are several models for the @xmath2 among which we focus on two of them : the intranuclear cascade code ( inc ) @xcite and the microscopic model developed in @xcite .
the inc is a semi - phenomenological approach .
the starting point in the inc is the microscopic evaluation of @xmath19 , @xmath20 and @xmath32 .
afterwards , the nucleons produced in the weak decay are followed in a semi - classical manner until they leave the nucleus . by means of this emulation of the physical conditions of the hypernuclear decay ,
a more accurate agreement between the theoretical results and the data is achieved .
the analysis of the experimental information using the inc , produce ratios @xmath34 . however , one limitation of the inc is that the quantum interference terms between the primary weak decay reaction @xmath35 and any other reaction which has the same initial and final state , such as @xmath4 , can not be included . on the other hand and to the best of our knowledge ,
the microscopic model described in @xcite , is the only microscopic model which puts in the same level of theoretical effort the weak decay mechanism and the @xmath2 . due to its character
, the microscopic model automatically includes the quantum interference terms .
some details of the microscopic model are given in the next section .
while the theoretical prediction starts with the primary decay and should end in the nucleon spectra , the analysis of the experiment begins with the spectra , goes back and a value for @xmath18 is determined .
this is done using several models , among which the inc is certainly one of the more elaborated ones . it should be noted , that the inc is employed in two ways : for the theoretical prediction of the nucleon spectra and for the interpretation of data . in any case , for the extraction of the @xmath18-experimental value one needs two theoretical inputs : @xmath32 and the @xmath2 .
now , if several models for these inputs produce the same ratio @xmath18 , it is reasonable to name it as the experimental value for @xmath18 .
therefore , an accurate evaluation of the quantum interference terms together with an alternative formalism to the inc , is important . in the present contribution , we further developed the microscopic model of @xcite presenting results for the nucleon spectra , with emphasis in the quantum interference terms .
the paper is organized as follows . in section
ii we present the microscopic model for the nucleon spectra .
this is done in general terms , including the two - body induced contribution . in section iii ,
explicit expression within the first order contribution to the rpa are shown . in section
iv , the numerical results are discussed and finally , in section v we give our conclusions .
in this section the microscopic model developed in @xcite is briefly summarized with the addition of @xmath32 .
in fact , @xmath32 is the sum of three terms : @xmath36 , @xmath37 and @xmath38 ( for details see @xcite ) .
we are interested in reporting expressions for the single and double nucleon spectra : @xmath39 and @xmath40 , which represent the number of nucleons of kind @xmath41 vs. it kinetic energy , @xmath42 and number of pairs of nucleons @xmath43 vs. @xmath44 or vs. the relative angle between @xmath45 and @xmath46 , @xmath47 , respectively . without @xmath2 ,
the expressions for @xmath39 and @xmath40 are , @xmath48 where we have used the normalization , @xmath49 , with @xmath50 .
the index @xmath51 , indicates that there is no @xmath2 .
the multiplicative factors in the right hand side of these equations are the number of particles ( pairs of particles ) of kind @xmath45 ( @xmath43 ) produces by the primary decay .
explicit expressions for @xmath52 and @xmath32 are reported in @xcite and @xcite , respectively . as mentioned
, the nucleon spectra is the number of particles ( or pairs of particles ) which angle or kinetic energy lays within a certain range .
therefore , @xmath52 and @xmath32 are cut into pieces which correspond to certain angles- or kinetic energies - ranges .
this is easily implemented by adding steps functions ( in the integrand of @xmath52 and @xmath32 ) , which limit the integration to such ranges .
the next step is the inclusion of the @xmath2 . to this end
, we introduce the quantity @xmath53 . this function result from evaluating any possible goldstone diagram for the @xmath0-weak decay , where the strong interaction is present after the weak decay takes place .
the index , @xmath54 , is the final state ( _ i.e. _ the emitted nucleons ) , taking the values : @xmath54=@xmath55 , @xmath56 , @xmath57 , @xmath58 , etc . at variance
, the indices @xmath59 refer to the two primary weak decays of each diagram and can have the values @xmath60or @xmath61 ; which stand for the transitions amplitudes @xmath62 , @xmath63 , @xmath64 , @xmath65 and @xmath66 , respectively .
it is important to be aware of the fact that the decay widths , @xmath19 , @xmath20 , @xmath67 , etc .
, are the square of a transition amplitude ( @xmath68 or @xmath31 ) . in this sense , @xmath53 does not represent only a decay width , but also some interference terms , which are pure quantum mechanical effects . by the addition of the @xmath2 in eqs .
( [ nn10]-[npp10 ] ) we obtained , @xmath69 where the factors @xmath70 are the numbers of nucleons of the type @xmath45 in the state @xmath54 . in the same way , @xmath71 are the numbers of pairs of nucleons of the type @xmath43 in the state @xmath54 .
let us give two examples . if @xmath72 , @xmath73 , @xmath74 , @xmath75 , @xmath76 , @xmath77 ; and if @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath82 , @xmath83 . in these expressions ,
the summation over @xmath59 and @xmath54 , runs over the values of these indices mentioned above . as a final comment for this section
we make some further considerations about the quantum interference terms .
all decay widths come from the square of a transition amplitude .
when no @xmath2 are included , the only transitions amplitudes are @xmath68 and @xmath31 . for simplicity ,
we focus on @xmath68 .
there is no interference term between @xmath62 and @xmath63 because the final state is different .
when the @xmath2 come into play , the strong interaction allows many others transitions amplitudes . in this case , as different reactions can end in the same final state and as the initial state is the same hypernuclei state for all the processes , the total transition amplitude is the sum of all these terms .
the squares of the individual terms , are the decay widths , while the crossed products are the interference terms . within our model ,
@xmath53 contain both decay widths and interference terms .
note that the strong interaction is present in @xmath32 as a ground state correlation , that is , the strong interaction acts before the weak transition potential . as an additional comment about the interference terms
, they can be grouped into two categories : @xmath84 the ones with @xmath85 and @xmath86 the ones with @xmath87 .
it is perhaps more convenient to explain these categories by means of an example .
let us propose three transition amplitudes : @xmath88 , @xmath89 and @xmath90 , where in the last two expressions the strong interaction is responsible for the second reaction .
there is an interference term between @xmath91 and @xmath92 ( @xmath85 ) , where the strong interaction appears in first order .
it is clear that the square of @xmath93 is positive , but the interference terms between both terms can be either positive or negative .
finally , the interference terms between @xmath91 and @xmath94 is an example of the @xmath87-category .
in this section we present expressions for the functions @xmath53 , using non - relativistic nuclear matter together with the local density approximation ( lda ) @xcite , which allows us to discuss any particular hypernucleus .
the eqs .
( [ nn1f]-[npp1f ] ) are general expressions for @xmath39 and @xmath40 . although @xmath53 is completely defined , one has to choose some set of goldstone diagrams which represent the @xmath0-decay to obtain the explicit expressions for this function .
this set has an infinite number of diagrams and _ a priori _ any sub - set of diagrams could be equally important . in this work and as a first step in the evaluation of the nucleon spectra , we have decided to study the rpa - like diagrams .
the direct part of the rpa , known as ring approximation , is perhaps the simplest manner to implement the strong interaction into this problem .
the pauli exchange terms are known to be important for these particular set of diagrams ( see @xcite ) and consequently the rpa looks as a natural first step . in
@xcite we have presented expressions for @xmath53 within the ring approximation .
however , no comparison with data has been done , because of the absence of the pauli exclusion principle .
let us call by @xmath95 a generic function which can be either @xmath52 , @xmath32 or @xmath53 . instead of giving the expression for @xmath95 ,
it is more convenient to work with the partial decay width @xmath96 , where , @xmath97 is the @xmath0 energy - momentum , @xmath98 and @xmath99 are the fermi momentum for neutrons and protons , respectively . to evaluate @xmath100 for
a particular nucleus one can use either an effective fermi momentum or the local density approximation . in this work
the lda is adopted , which make @xmath98 and @xmath99 position - dependent .
they are defined as @xmath101 , where @xmath102 and @xmath103 , with @xmath104 , @xmath45 and @xmath105 being , respectively , the density profile , number of neutrons and number of protons of the nuclear core of the hypernucleus . in the last case , it is equivalent to write the function @xmath96 in terms of the densities as @xmath106 .
the lda reads , @xmath107 where for the @xmath0 wave function @xmath108 , we take the @xmath109 wave function of a harmonic oscillator . the final result is obtained by averaging over the @xmath0 momentum distribution , @xmath110 , as follows , @xmath111 where @xmath112 is the fourier transform of @xmath108 and @xmath113 , being @xmath114 the binding energy for the @xmath0 . as remarked in the last section ,
the expressions for @xmath53 are obtained as if they were decay widths . in this spirit , it is convenient to overview the derivation of @xmath52 , because this simplifies the derivation of @xmath53 . in fig .
[ fig1rpa ] we show the direct and exchange @xmath52-contributions .
the distinction between @xmath19 and @xmath20 can be ascribed to the isospin of the hole line , where when @xmath115 is a neutron ( proton)-hole we are considering the @xmath19 ( @xmath20)-decay width . to get the analytical expressions for these diagrams we have employed the golstone rules ( for a more detailed version , see
@xcite ) , @xmath116 decay width , respectively . ]
where for simplicity , @xmath117 represents the spin ( @xmath118 ) , isospin ( @xmath119 ) and energy - momentum of the particle @xmath120 . the energy - momentum carried by the transition potential , @xmath121 ,
can not be specified until we separate the direct and the exchange terms .
the meaning of the subindexes @xmath122 and @xmath123 is shown in fig .
[ fig1rpa ] , where due to the momentum conservation , @xmath124 .
the direct and exchange matrix elements are , @xmath125 where the minus sign comes from the crossing of the fermionic lines . due to energy - momentum conservation
, we have @xmath126 and @xmath127 .
the first ( second ) term in the the right hand side of eq .
( [ me0ant ] ) originates the direct ( exchange ) contribution to the decay width .
the particle and hole propagators are , @xmath128 and @xmath129 where @xmath130 is the nucleon total free energy and @xmath131 is the nucleon binding energy .
the value of the fermi momentum depends on whether we have a proton or a neutron .
the @xmath132 transition potential @xmath121 , is , @xmath133 where we have included the nuclear residual interaction @xmath134 , which is employed soon .
the isospin dependence is given by , @xmath135 the values @xmath136 stand for the isoscalar and isovector parts of the interaction , respectively .
the spin and momentum dependence of the transition potential is , @xmath137 where the quantities @xmath138 , @xmath139 , @xmath140 , @xmath141 , @xmath142 and @xmath143 contain short range correlations ( src ) and are given in appendix b of @xcite .
they are built up from the full one - meson - exchange potential ( omep ) , which involves the complete pseudoscalar and vector meson octets ( @xmath144 ) .
it is self evident that the @xmath145 ( @xmath146)-terms are the parity violating ( parity conserving ) terms of the transition potential . the nuclear residual interaction is drawn as , @xmath147 where the functions @xmath148 , @xmath149 and @xmath150 are adjusted to reproduce any effective omep - nuclear residual interaction .
we have now all the elements required for the evaluation of @xmath53 within the rpa approximation .
the ring approximation has been discussed in @xcite , where the advantage of using that approximation is that the ring series can be summed up to infinite order in a very simple way .
the situation is different in the rpa , where the corresponding series can be summed up only in some particular cases : when the nuclear residual interaction is represented by a contact or by a separable interaction . for a general finite range interaction
there is no way out but to evaluate each exchange term individually .
the problem is quite involved because matrix elements must be antisymmetrized for both @xmath121 and @xmath134 .
the lowest order rpa - contribution is the one in which @xmath134 appears in first order , being the only one reported in the present contribution . in the lowest rpa - contribution , there are two matrix elements with @xmath121 and one with @xmath134 .
each matrix element has a direct and an exchange part .
this makes a total of eight different diagrams , which are shown in fig .
[ fig2rpa ] .
unfortunately , in nuclear matter each diagram must be evaluated individually . in the next section and by means of the numerical analysis , we discuss this approximation . using the standard goldstone rules , this lowest order rpa - contribution
is written as , @xmath151 the summation runs over all spins , while the isospin sum can not be specified until one sets the final state @xmath54 . the energy - momentum carried by each fermionic line
is shown in fig .
[ fig3rpa ] for the direct term . due to the energy - momentum conservation in each vertex , for all the contributions we have @xmath152 and @xmath153 ;
while the energy - momentum carried by @xmath154 and @xmath121 depends on the topology of each diagram .
these values , together with some more details on the exchange terms are specified in the appendix . and @xmath134 , respectively .
an up ( down ) going arrow represents a particle ( hole ) , while an arrow with a wide line represents the @xmath0 . ]
it is convenient to re - write eq .
( [ gamgold ] ) as , @xmath155 where the super - indexes @xmath156 and @xmath157 , refer to the direct ( @xmath158 ) and exchange ( @xmath159 ) matrix elements @xmath160 , @xmath161 and @xmath162 , respectively .
for instance , for the direct contribution , we have @xmath163 .
note that @xmath164 ( @xmath165 and @xmath157 ) , can take the value @xmath158 or @xmath159 , which makes the total of eight different contributions , being @xmath166 the first order term to the ring series . by inspection of the diagrams in fig .
[ fig2rpa ] , we notice that the strong interaction split these graphs into two pieces . in the lower piece particles are named as @xmath167 and @xmath168 , while in the upper portion we have @xmath167 and @xmath169 .
each diagram gives two contributions : the first one with @xmath167 and @xmath168 ( @xmath167 and @xmath169 ) on ( off ) the mass shell , and the inverse situation .
then , it is convenient to differentiate these two possibilities by the super - index @xmath170 and @xmath171 , respectively , as follows , @xmath172 the origin of these two terms comes from the energy integration in eq .
( [ gamgold ] ) , which produces an expression with two poles : one in the upper and the other in the lower part of the diagram . clearly , the distinction between @xmath170 and @xmath171 , is redundant for the direct contribution , due to the symmetry of the diagram .
-contribution from fig .
[ fig3rpa ] , in order to assign values to the momentum transfer .
the remainder contributions have the same pattern . ] in order to perform the summation on spin and isospin quantum numbers it is convenient to write , @xmath173 where @xmath174 in the isospin summation , @xmath170 ( @xmath171 ) means that the final state @xmath54 is in the lower ( upper ) piece of the diagram . to avoid confusion , we give an example : let @xmath175 , @xmath176 and @xmath72 .
the @xmath170 sum is zero because for the final state @xmath54 , there are no protons in the lower portion of the diagram .
the @xmath171 sum reduces to one term which isospin projections for each particles are @xmath177 and @xmath178 . performing the energy integration , the spin summation and after some algebra , we obtain , @xmath179 in these expressions , @xmath180 is the number of crossing of fermionic lines and @xmath181 indicates the principal value .
explicit expressions for the functions @xmath182 can be found in the appendix . here
@xmath183 ( @xmath184 ) is the energy - momentum carried by the transition potential in the lower ( upper ) part of the diagram , while @xmath119 corresponds to the nuclear interaction .
we present in this section explicit expressions for the direct contribution , whereas the exchange ones are shown in the appendix .
the @xmath185 reads , @xmath186 where for the direct contribution we have , @xmath187 . by means of eqs .
( [ decwpar3 ] ) and ( [ decwpar2 ] ) , the dependence on the fermi and @xmath0-momenta in the partial widths is eliminated . for the @xmath188-diagrams
, we have @xmath189 . by performing the summation over isospin for all possible
primary decays and final states , we have , @xmath190 where in the right hand side of these equations , we have put an additional super - index ( @xmath191 or @xmath192 ) which is the isospin projection of the @xmath167 particle . as we are using different fermi momenta for protons and neutrons , the knowledge of the isospin projection of each particle is required .
this information about @xmath167 together with the transition @xmath193 is enough to this end .
the particular @xmath194 contribution can be further simplified by the identification of a part of the integral in eqs.([rpa4 ] ) and ( [ rpa4p ] ) , with the lindhard function ( details can be found in @xcite ) .
finally , the expressions for @xmath195 with @xmath196 , can be found in the appendix .
for the first order contribution ( in the nuclear residual interaction , @xmath134 ) to the rpa , which from now on is named as @xmath197 , we can write , @xmath198 the final step is to show the expressions for the spectra . in the present contribution
the two - body induced decay is not evaluated .
then , it is convenient to re - write eqs .
( [ nn1f]-[npp1f ] ) , within the first order contribution to the rpa as the only @xmath2 , @xmath199 where the normalization is now , @xmath200 .
these are already explicit expressions for the @xmath39 and @xmath201-spectra , once some step functions limiting the range of the momentum ( or angle ) of the final particles are incorporated in eqs .
( [ gamdirpi0 ] ) , ( [ rpa4 ] ) and ( [ rpa4p ] ) . also an additional step function is required if an energy threshold is considered .
it should be noted that the @xmath197 , together with any other contribution in which the nuclear residual interaction appears in first order ( or in any odd order ) , is an interference term .
the diagrams where the nuclear residual interaction is present an even number of times , has both interference terms and decay widths , being the decay widths the ones where the cut is in the middle of the diagram . in the next section we present numerical results for these expressions .
in this section we present the numerical results for the @xmath202 , @xmath203 , @xmath204 and @xmath205-spectra using the first order contribution to the rpa .
the lda allows us to discuss the @xmath1 hypernucleus .
the transition potential is represented by the exchanges of the @xmath206 , @xmath207 , @xmath208 , @xmath209 , @xmath210 and @xmath211-mesons , which formulation has been taken from @xcite and the values of the different coupling constants and cutoff parameters appearing in the transition potential have been taken from @xcite .
for the nuclear residual interaction , we have used the bonn potential @xcite in the framework of the parametrization presented in @xcite , which contains the exchange of @xmath206 , @xmath209 , @xmath17 and @xmath210 mesons , while the @xmath207 and @xmath157-mesons are neglected . in implementing the lda ,
the hyperon is assumed to be in the @xmath109 orbit of a harmonic oscillator well with frequency @xmath212 mev .
as already stated , we have employed different values for the proton and neutron fermi momenta , @xmath213 and @xmath214 , respectively . to start with ,
let us discuss the free spectra , that is , the spectra without final state interaction . to this end in eqs .
( [ nn1fr]-[npp1fr ] ) all @xmath197-terms are eliminated . for the free spectra
, the inc and the microscopic scheme should give the same result .
therefore , we make a comparison of our values with those of the inc , in order to test our values for the spectra . from a technical point of view
, both methods evaluate the spectra differently . while in the inc ,
the emitted particles are classified according to their energies ( or relative angles ) , in the present microscopic model the integral which represents the decay width is partitioned in energy ( or relative angles ) regions .
these two procedures are equivalent . however , some differences show up , which are explained soon in this section . and @xmath203-free kinetic energy spectra for @xmath1 .
the continuous line is our result , while the dashed line represents the inc one .
the energy is given in mev and the spectra has been evaluated using 10 mev steps for the kinetic energy .
we have used @xmath215 and @xmath216 , given in units of @xmath217 ev , where @xmath218 is the decay width of the @xmath0 in free space . ] in fig .
[ espnplib ] we present our result for the @xmath202 and @xmath203-spectra and compare it with the corresponding inc - result from @xcite . related to the inc ,
we notice that our value for the @xmath202-maximum is shifted towards a higher energy , while the opposite occurs for @xmath203 .
in fact , this shift is more marked in @xmath203 .
to understand the origin of this behavior in fig .
[ espcomp1 ] we have split @xmath203 into their two isospin components . the first isospin component is the charge - exchange one , where there is a proton in the @xmath0-vertex ( that is , in the @xmath167-position in fig . [ fig1rpa ] ) .
the second is the charge - conserving contribution , which has a neutron in the just mentioned place .
the first contribution is multiplied by a factor four ( due to the isospin ) , while we have a factor one for the charge - conserving term .
this makes the first contribution to be the dominant one .
notice that the transition potential is not the same for both contributions , because only the isovector terms in the transition potential act for the charge - exchange contribution .
in addition , our results show that the particle attach to the @xmath0-vertex is slower than the other particle .
-free kinetic energy spectra .
the dashed line is the contribution where a proton is attached to the @xmath0-decay vertex , the dotted - dashed line represents the contribution where a neutron is attached to the same vertex , while the continuous line is the total contribution .
units are the same as in fig .
[ espnplib ] . ]
this fact has a simple physical interpretation . in the expression for @xmath20 ( or @xmath19 ) , we take the principal value for the mesons propagators in the transition potential .
as we are getting closer to the pion pole , the energy left to the @xmath167-particle becomes smaller , while the proximity to the pole , makes the contribution more important .
as mentioned above , the inc weighs this contribution with @xmath20 , making no distinction between the inner isospin components .
this fact is not mandatory in the inc , but the available calculations are performed in this way .
we have implemented an isospin average to get rid of the different isospin summation employed in the inc .
we do not reproduce these results , as they simply show the agreement between both methods .
we turn now to the inclusion of @xmath2 . in the present contribution , we have limited the @xmath2 to the first order contribution to the rpa . in table i , we present our results for @xmath219 , within three different approximations : the first order contribution ( in the nuclear residual interaction ) to the rpa ( @xmath197 ) , the same first order term to the ring series ( @xmath220 ) , and the full ring approximation taken from @xcite .
the values quoted in this table result from the integration of the spectra over the whole energy region .
the difference between @xmath197 and @xmath220 is the inclusion of the exchange terms . for @xmath221 ,
the exchange terms increase this quantity by @xmath222 , while this percentage is @xmath223 for @xmath224 .
although they are different quantities , it is interesting to comment that the exchange terms produce also an increase of @xmath225 and @xmath226 in @xmath19 and @xmath20 , respectively @xcite .
for the two remainders terms , we have an increase ( decrease ) of @xmath227 ( @xmath226 ) for @xmath228 ( @xmath229 ) .
the next point is to make an estimation of the accuracy of the first order rpa - contribution to account for the full rpa series . to this end , we compare the full ring series with @xmath220 .
we can see that only for @xmath224 , the @xmath220 is an adequate approximation , while there is a strong discrepancy for all the others terms .
the reason for this behavior relays upon the inner isospin structure of the ring ( or rpa ) approximation .
as already stated , the @xmath224 is dominated by graphs where a proton is attached to the @xmath0-weak decay vertex .
the particle - hole bubbles ( which propagates as the ring series ) , are built up from a neutron particle and a proton hole .
the charge - conservation does not allow any other particle - hole configuration .
the situation is different for @xmath221 , @xmath228 , @xmath229 and the @xmath224-contribution with a neutron in the @xmath0-vertex .
as an example , we consider the second order contribution to @xmath221 .
this contribution has three particle - hole bubbles .
the upper and lower bubbles can be only a neutron particle - neutron hole configuration .
but the intermediate bubble can be either a neutron particle - neutron hole or a proton particle - proton hole one .
higher order contributions can have more complex configurations ( for details , see @xcite ) .
the convergence of the ring series is fast only when the particle - hole bubbles are of the same kind . from this analysis
, it can be said that the @xmath197 should be seen as an approximation in itself .
the ring series can be summed up to infinite order in an easy way , but there is no ground to neglect the exchange terms .
.numerical results for @xmath219 , within several approximations .
note that @xmath230 .
the @xmath197 ( @xmath220)-line is the first order contribution to the @xmath231 ( @xmath232 ) , while the @xmath232-line represents the total ring approximation take from @xcite .
all values are given in units of @xmath233 . [ cols="^,^,^,^,^",options="header " , ] as a final point for this section and beyond the limitations in phase space in the rpa , the nuclear residual interaction employed in the rpa , should be discussed .
to perform this calculation we have chosen the bonn potential where we have neglected src in all results except for the ones in fig .
[ espnptot ] . from this figure
, we can see that the addition of src reduce significantly the @xmath197-contribution .
the interaction without src , has been selected to explore the rpa results under extreme conditions .
the final outcome shows us that the rpa does not alter the free spectra very much .
the addition of src makes this contribution even less important . from this
, we can conclude that the @xmath197 is adequate enough to explore the rpa contribution and from our whole analysis , it is clear that the rpa does not represent the relevant kind of fsi required to describe the @xmath0-weak decay spectra . from the inc
we know that the @xmath2 are important , which states our problem as the finding of the most significant diagrams to account for the @xmath2 .
we do not known _ a priori _ the answer to this question . to obtain some insight about this problem
, the rpa has been a good starting point due to it simplicity .
the search of the relevant set of diagrams has an additional difficulty , which is the selection of the nuclear residual interaction , together with the adjustment of some potential - parameters .
the bonn potential is one possible choice , but the truth is that we do not know the strong interaction in this energy - momentum region .
however , the inc provides us with a powerful tool to settle the strong interaction : in our microscopic scheme it is easy to turn off the interference terms not contained in the inc . under these conditions , and once the adequate set of diagrams are found , the inc - result can be used to fix the strong interaction .
note that the inc does not employ an effective strong interaction but the experimental scattering cross section between nucleons .
this would give us some confidence on the strong interaction so adjusted .
for this particular purpose , the use of the inc is better than the data themselves , because the theoretical model for the weak transition potential would be the same in both the inc and in our microscopic model .
afterwards , the interference terms should be added and a comparison with data should be done . in our opinion , it is this whole schedule the one which should be employed to test our knowledge of the baryon - baryon strangeness - changing weak interactions .
the present contribution is a step forward in this direction . in @xcite , we have presented our scheme in general terms .
that model has been developed in the present work , but for a particular kind of @xmath2 . in a forthcoming work ,
the @xmath234-decay channel which is not contained in the rpa and can be represented by second order self - energy contributions , will be included . at variance with the present calculation
, this kind of @xmath2 contains both decay widths and interference terms .
the present work address the problem of the theoretical interpretation of the spectra of nucleons emitted from the non - mesonic weak decay of a @xmath0-hypernucleus . to the best of our knowledge
, we have presented for the first time a scheme which deals with this issue using a microscopic formalism .
we have presented results for the single and double kinetic energy spectra and also for the opening angle distribution of @xmath55 and @xmath56 pairs .
the @xmath2 have been incorporated by means of the first order contribution ( in the nuclear residual interaction ) to the rpa .
our formalism naturally contains the quantum interference terms between any pair of decay amplitudes which end in the same final state .
the quantum interference terms appear once the @xmath2 are incorporated . from the inc we know that the @xmath2 are important . in a microscopic scheme
there is a huge amount of possible diagrams which constitute these @xmath2 .
beforehand , we do not known which set of diagrams is the most significant one .
the rpa has been considered as a starting point in this analysis because it simplicity gives us some insight on the kind of the contributions required .
however , the rpa itself does not represent an important contribution .
the comparison of our results with data suggests that the @xmath2 should connect @xmath235 with @xmath234 configurations .
this kind of study is much more complex and it is beyond the scope of the present contribution . we have payed a particular attention to the non - diagonal interference terms which are important within the rpa . in fact , they are very significant in the @xmath204-spectra , while they can be neglected in @xmath205 . the effect on the @xmath202-spectra is a moderate increase and it is also small for the @xmath203-spectra .
the origin of this behavior does not relays upon any particularity of the rpa , but on some isospin factors .
we agree that only when the relevant @xmath2-diagrams are found and incorporated , we would be able to know if the quantum interference terms are important or not .
but from our results , it is clear that they can not be ignored .
the existing and more successful formalism which takes care of the same problem is the inc .
the inc is a semi - phenomenological approach , where the trajectory of the particles emitted in the weak decay is tracked in their way out of the nucleus .
the @xmath2 are incorporated in the inc using free path and experimental cross sections between nucleons .
unfortunately , it is not possible to establish a biunique relation between the processes in the inc and in our microscopic model .
the limitation of our microscopic model is the technical difficulty in finding the relevant @xmath2-diagrams .
however , our scheme improves the inc in two main points .
the first one is the incorporation of the quantum interference terms mentioned above .
the second one refers to the description of light hypernuclei , where the employment of the inc is not appropriate .
we have presented our results using nuclear matter together with the lda , which is also a bad approximation for light hypernuclei .
however , our expressions are not limited to nuclear matter and they can be used in a finite nucleus calculation .
the incorporation of @xmath2 in this kind of work is certainly a quite involve task .
our aim is to describe the nuclear spectra of medium hypernuclei where the use of nuclear matter is possible .
however , once this goal is eventually accomplished , the knowledge of the relevant @xmath2-diagrams would encourage the evaluation of a finite nucleus calculation for light hypernuclei .
we would like to comment now on the @xmath236-ratio .
an experimental determination of this ratio suffers from two ambiguities : the possible contribution of two - body induced non mesonic weak decay ( which are originated from ground state correlations , @xmath237 ) and the effect of @xmath2 .
it is worth to mention here that these two processes should not be confused .
the two - body induced , is part of the primary decay as shown in eqs .
( [ nn10]-[npp10 ] ) . after
the primary decay takes place , the strong interaction between nucleons ( i.e. , the @xmath2 ) starts acting .
the @xmath2 affects nucleons from both one- and two - body induced decays .
a final @xmath234-state , for example , can be originated either from a @xmath237 ( without @xmath2 ) or from the action of @xmath2 over a one - body induced decay .
moreover , there is an interference term between these processes , as the initial and final state is the same .
this interference term can not be evaluated within the inc as it is a pure quantum mechanic effects . beyond technical difficulties ,
this kind of contributions are formally part of the @xmath219 in eqs .
( [ nn1f]-[npp1f ] ) .
this is just one example , the theoretical problem of the inclusion of both @xmath237 and @xmath2 , is certainly a very difficult problem .
any experimental determination of the @xmath236-ratio , assumes a model for this theoretical problem . referring to the inc ,
it is used to predict nucleon spectra and to extract the so - called experimental value for the @xmath236-ratio . for the extraction of the ratio ,
the experimental value for @xmath238 or @xmath239 ( together with the corresponding error bars ) is used as an input . to put it in simple terms , a typical theoretical value for the ratio is , @xmath240 , while data analyzed by means of the inc , gives a result @xmath241 ( we have focused on the @xmath242 region , for @xmath1 ) . based on this analysis ,
it is reasonable to assert that the @xmath236-puzzle has been solved .
however , one should be aware of the fact that the experimental information used as an input in the inc is the @xmath243 ( or @xmath244 ) ratio , where here @xmath39 ( or @xmath201 ) is the integral of the spectra over energy or angle .
clearly , this integration erase some physical information .
in fact , while the inc reproduce fairly well the @xmath202-kinetic energy spectra , some discrepancies remains for the @xmath203-kinetic energy spectra one .
having in mind this element , together with the above mentioned ambiguities , it is our opinion that the real theoretical problem is the reproduction of the spectra . from the present contribution , we have indications that the quantum interference terms should be included in the spectra evaluation .
the next step , is to found the relevant @xmath2-diagrams .
as discussed , the @xmath197-results has given us some hints on this issue .
i would like to thank f. krmpoti , g. garbarino and a. ramos , for fruitful discussions and for the critical reading of the manuscript .
this work has been partially supported by the conicet , under contract pip 6159 .
in this appendix we present the explicit form for @xmath195 with @xmath196 . in order to simplify the expressions , all functions which has the isospin @xmath245 ( @xmath246)-index , carries an energy - momentum @xmath183 ( @xmath184 ) , while the nuclear strong interaction has an isospin index @xmath247 and has a @xmath119-energy momentum transfer .
moreover , as each function is shown in a separate subsection , the super - index @xmath248 is written only when it is necessary .
[ [ the - gammadde_i - i - rightarrow - j - and - gammaedd_i - i - rightarrow - j - contributions ] ] the @xmath255 and @xmath256-contributions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [ [ the - gammadee_i - i - rightarrow - j - and - gammaeed_i - i - rightarrow - j - contributions ] ] the @xmath269 and @xmath270-contributions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ w. m. alberico and g. garbarino , phys . rep . * 369 * , 1 ( 2002 ) ; in _ hadron physics _ , ios press , amsterdam , 2005 , p. 125 . edited by t. bressani , a. filippi and u. wiedner .
proceedings of the international school of physics enrico fermi " , course clviii , varenna ( italy ) , june 22 july 2 , 2004 .
t. inoue , s. takeuchi and m. oka , nucl .
phys . * a 577 * , 281c ( 1994 ) ; _ ibid . _ * a 597 * , 563 ( 1996 ) .
t. inoue , m. oka , t. motoba and k. itonaga , nucl .
phys . * a 633 * , 312 ( 1998 ) .
t. inoue , m. oka , t. motoba and k. itonaga , nucl .
phys . * a 633 * , 312 ( 1998 ) . | single and double coincidence nucleon spectra in the @xmath0-hypernuclei weak decay are evaluated and discussed using a microscopic formalism .
nuclear matter is employed together with the local density approximation which allows us to analyze the @xmath1 hypernucleus non - mesonic weak decay .
final state interactions ( @xmath2 ) are included via the first order ( in the nuclear residual interaction ) terms to the rpa , where the strong residual interaction is modelled by a bonn potential . at this level of approximation , these @xmath2 are pure quantum interference terms between the primary decay @xmath3 and @xmath4 , where the strong interaction is responsible for the last piece in the second reaction .
also the pauli exchange contributions are explicitly evaluated .
we show that the inclusion of pauli exchange terms is important .
a comparison with data is made .
we conclude that the limitations in phase space in the rpa makes this approximation inadequate to reproduce the nucleon spectra .
this fact , does not allow us to draw a definite conclusion about the importance of the interference terms .
_ pacs number : _ 21.80.+a , 25.80.pw .
_ keywords : _
@xmath0-hypernuclei , non - mesonic decay of hypernuclei , @xmath5 ratio .
v#1 # 1|#1 # 1#1| |
the purpose of this paper is to study the behavior of the thin free boundary as it approaches to the fixed boundary in the so - called ( scalar ) _ signorini problem _ ( also know as the _ thin obstacle problem _ ) .
the signorini problem consists in minimizing the dirichlet energy functional @xmath5 on a closed convex set @xmath6 for a given function @xmath7 . here and everywhere in the paper
we use the following notations : @xmath8 for a subset @xmath9 .
we assume @xmath10 . using direct methods of calculus of variation
one can verify that a minimizer @xmath11 exists and satisfies the following variational inequality : @xmath12 the problem above goes back to the foundational paper @xcite on variational inequalities .
it is has been known for quite some time that the minimizers are in the class @xmath13 for some @xmath14 ( see @xcite and also @xcite ) and even @xmath15 in the dimension @xmath16 , see @xcite . besides , the minimizers satisfy @xmath17 the latter are known as the _ signorini _ or _ complementarity boundary conditions_. the problem features the following apriori unknown subsets of @xmath18 : @xmath19 the study of the geometric and analytic properties of the free boundary is one of the objectives of the signorini problem .
sometimes it is said that the free boundary @xmath20 is _ thin _ , to indicate that it is ( expected to be ) of dimension @xmath21 .
recent years have seen some interesting new developments in the problem , starting with the proof in @xcite that the minimizers @xmath22 are in the class @xmath15 , in any dimension @xmath23 , which is the optimal regularity .
this opened up the possibility of studying the free boundary @xmath24 , which has been done in @xcite , see also ( * ? ? ?
* chapter 9 ) .
an effective tool in the study of the free boundary is _
almgren s frequency formula _
@xmath25 it originated in the work of almgren on multi - valued harmonic functions @xcite and has an important property of being monotone in @xmath26 , even for solutions of the signorini problem .
one then classifies the free boundary points according to the value @xmath27 it is known that @xmath1 for @xmath28 in the signorini problem and more precisely @xmath29 or @xmath30 @xcite .
this results in a decomposition @xmath31 the set @xmath32 is known as the _ regular set_. it has been recently shown that @xmath32 is real analytic @xcite by using a partial hodograph - legendre transform from @xmath33 regularity proved in @xcite . see also @xcite for a different proof of @xmath34 regularity , based on a generalization of the boundary harnack principle .
the only other free boundary points studied in the literature are the ones in @xmath35 , @xmath36 which correspond to the points where the coincidence set @xmath37 has a zero @xmath38 density , see @xcite .
such points are known as _
singular points_. it was proved in @xcite that @xmath35 is contained in a countable union of @xmath39 manifolds .
an interesting question is finding all possible values for @xmath40 . in dimension
@xmath16 the answer to that question is known ( proof is a simple exercise ) : @xmath41 must be one of the following values : @xmath42 however , this is still an open problem in dimensions @xmath43 .
the objective in this paper is the study of the behavior of the free boundary @xmath24 in the signorini problem as it approaches a set where @xmath22 is forced to be zero .
more precisely , consider a closed subset @xmath44 of the set @xmath45 in , defined by @xmath46 ( 150,150 ) ( 0,0 ) near the contact points @xmath47 with the fixed boundary @xmath48 considered in the hyperplane @xmath49.,title="fig : " ] ( 79,130)@xmath50 ( 140,43 ) @xmath51 ( 34,43)@xmath52 ( 46,72)@xmath47 ( 70,94 ) @xmath53 ( 77,43)@xmath54 ( 20,75)@xmath55 ( 110,75)@xmath56 ( 49,69)(1,-1)10 ( 53,73)(2,-1)13 ( 78,95)(1,0)16 and minimize the dirichlet energy @xmath57 in over @xmath44 .
that is , compared to the signorini problem , we have an additional constraint that the functions must vanish on @xmath58 . if we think of the solution of the signorini problem as an elastic membrane that is forced to stay above zero in @xmath18 , the new constraint in @xmath44 can be thought of as `` clamping down '' or `` gluing '' the membrane on @xmath58 .
the boundary of the latter set in @xmath18 is @xmath59 which we call the _ fixed boundary_. note that the coincidence set @xmath37 will contain now @xmath58 and the truly free part of @xmath24 is @xmath60 .
the points in @xmath61 are categorized as _ contact points _ , and the ones in @xmath62 are _ non - contact points _ , see fig .
[ fig : contact ] .
we note that the minimizers in @xmath44 still solve the signorini problem in small halfballs @xmath63 with @xmath64 and therefore we will have that @xmath65 and that it satisfies @xmath66 there are many papers in the literature dealing with the contact of the free and fixed boundaries in various free boundary problems .
the case of the classical obstacle problem , for instance , was studied by @xcite .
we also refer to ( * ? ? ?
* chapter 8) and references therein for some of these results , including also extensions to other obstacle - type problems .
in contrast to the case of the classical obstacle problem , where the presence of the fixed boundary actually helps for instance , to avoid a geometric `` thickness '' condition on coincidence set needed for the regularity of the free boundary in the signorini problem the presence of the fixed boundary introduces a serious handicap . indeed , as we have mentioned earlier , the optimal regularity of the signorini problem is @xmath67 .
this regularity is exhibited by the following explicit solution : @xmath68 on the other hand , it is easy to see that @xmath69 is a minimizer of @xmath57 over @xmath44 ( simply because it is harmonic in @xmath70 ) , thus limiting the generally expected regularity of minimizers of @xmath57 to at most @xmath0 .
( see fig . [
fig : rez ] for the illustration of these solutions . )
this lower regularity of minimizers undercuts many techniques used for the signorini problem , calling for caution even when dealing with the first derivatives of the solution .
luckily , however , one of the most important tools in our analysis , almgren s frequency formula , still works : one of the steps in the proof is based on a rellich - type identity , which in our case becomes an inequality in the correct direction and allows the proof to go through .
( 320,140)(0,0 ) ( 0,20 ) is an explicit solution of the signorini problem and @xmath71 is a minimizer over @xmath44 with worst possible regularity.,title="fig : " ] ( 160,20 ) is an explicit solution of the signorini problem and @xmath71 is a minimizer over @xmath44 with worst possible regularity.,title="fig : " ] ( 160,0)@xmath72 ( 0,0)@xmath73 the first main result in this paper establishes the optimal regularity of the minimizers . [ thm : c^1/2 ] if @xmath22 is a minimizer of the functional @xmath57 in over @xmath44 in , then @xmath74 with @xmath75 the regularity above implies that for any @xmath28 we have @xmath76 the knowledge of the possible values of @xmath41 is important for the classification of free boundary points ( as we discussed at the end of subsection [ sub : the_problem ] ) .
concerning these values we have the following results .
[ thm : contact points ] if @xmath22 is a minimizer of @xmath57 over @xmath44 , then for a contact point @xmath77 we have @xmath78 at non - contact points we give a more complete picture .
[ thm : noncontact ] if @xmath22 is a minimizer of @xmath57 over @xmath44 , then for a non - contact point @xmath79 we have that @xmath80 can take only the following values : @xmath81
it will be convenient for our considerations to extend every function @xmath82 by even symmetry in @xmath50-variable to the entire ball @xmath83 : @xmath84 with such extension in mind , the energy @xmath57 in can be replaced with @xmath85 as the first result towards the optimal regularity , we show that the minimizers are @xmath86 regular for some @xmath14 .
[ prop : holder ] if @xmath22 is a minimizer of @xmath57 over @xmath44 , then @xmath87 , with a dimensional constant @xmath14 and @xmath88 we start by showing that the positive and negative parts of the minimizer @xmath22 are subharmonic .
note that at this stage we have not yet established the continuity of @xmath22 , so we will resort to the energy methods .
[ lem : subharmonicity ] @xmath89 are subharmonic functions in @xmath83 .
[ pf : subharmonicity ] proving the lemma is equivalent to showing that for any nonnegative test function @xmath90 we have @xmath91 let @xmath92 be a nondecreasing function such that @xmath93 then for a fixed @xmath94 and sufficiently small @xmath95 we have @xmath96 and thus @xmath97 are admissible functions from @xmath44 . since @xmath22 is a minimizer , we have @xmath98 , yielding @xmath99 since the second integral is nonnegative , sending @xmath100 to @xmath101 we obtain . once we know that @xmath102 are subharmonic in @xmath83 , we immediately obtain that @xmath22 is locally bounded .
[ lem : loc - bound ] if @xmath22 is a minimizer of @xmath57 over @xmath44 , then @xmath103 and more precisely @xmath104 we can now proceed to the proof of hlder continuity . using the local boundedness and the fact that @xmath102 vanish on @xmath58 , by the comparison principle we can write that @xmath105 where @xmath106 and @xmath107 solves @xmath108 with boundary values changing continuously from @xmath101 to @xmath109 in @xmath110 .
we next claim that the barrier function @xmath107 above is in @xmath111 .
indeed , we can use a bi - lipschitz transformation to map @xmath112 to @xmath113 preserving the distance from the origin .
then @xmath107 will transform into @xmath114 , which would be a solution of a uniformly elliptic equation in divergence form with measurable coefficients : @xmath115 by the de giorgi - nash theorem , we know @xmath116 , and since the transformation is bi - lipschitz we also get @xmath117 , which provides @xmath118 the latter , together with gives @xmath119 combined with the next lemma , this implies @xmath87 .
[ lem : holder - reg ] let @xmath22 be a minimizer of @xmath57 over @xmath44 . if for a @xmath120 and all @xmath121 the following property holds : @xmath122 then @xmath123 with @xmath124 depending only on @xmath125 , @xmath126 , @xmath127 .
[ pf : holder - reg ] denote @xmath128 .
take any @xmath129 . without loss of generality
we can assume @xmath130 and @xmath131 .
we will consider three cases : 1 .
using we get @xmath133 2 .
@xmath134 and the @xmath126-th coordinate of @xmath135 , @xmath136 . in this case
we observe that @xmath137 and thus @xmath22 is harmonic there , @xmath138 and the interior gradient estimates for harmonic functions imply @xmath139 3 .
@xmath134 and @xmath140 . in this case @xmath141 .
thus @xmath22 solves the signorini problem in @xmath142 and @xmath143 . using the interior lipschitz regularity for the solutions of the signorini problem , see ( * ? ? ?
* theorem 1 ) , we have @xmath144 and we complete the proof as in the previous case . as we observed in the introduction , we know that the function @xmath145 restricts the regularity of our solutions to @xmath0 . in order to rigorously obtain that @xmath0 is also the minimum expected ( and thus optimal ) regularity , we need the following monotonicity formula for the halfball , first introduced in @xcite . [
lem : half - monotonicity ] for any @xmath146 satisfying @xmath147 then the function @xmath148 is nondecreasing for @xmath149 .
[ pf : half - monotonicity ] the proof is a verbatim repetition of that of ( * ? ? ?
* lemma 4 ) , despite of the slight difference in the assumptions .
namely , instead of asking the convexity of the set @xmath150 , we note that it is only used to show that the complement set of the support of @xmath114 contains the lower dimensional halfball @xmath151 , which is automatically satisfied in the setting of our problem .
we are now ready to proof our first main result .
we apply the monotonicity formula in lemma [ lem : half - monotonicity ] to the minimizer @xmath22 of @xmath57 to obtain @xmath152 here , the last inequality is standard for non - negative subharmonic functions ( for a proof see for example @xcite ) . applying this for @xmath153 we obtain the corresponding inequality for @xmath22 . now using the fact that @xmath22 vanishes on @xmath154 we also have the poincar inequality for the halfball @xmath155 then by the scaling of lemma [ lem : loc - bound ] we have @xmath156 let us notice that the above estimate holds also for any ball @xmath157 with a center @xmath158 , and @xmath159 @xmath160 yielding @xmath161 using lemma [ lem : holder - reg ] we obtain @xmath162 . without loss of generality
we will further assume that @xmath163 .
as we mentioned in the introduction , almgren s frequency formula plays and important role in the signorini problem .
since we have an additional constraint for functions in @xmath44 , it is not automatic that it will still be monotone . fortunately , however , it is still the case .
[ thm : almgren ] if @xmath22 is a minimizer of @xmath57 over @xmath44 , then @xmath164 is monotone in @xmath26 for @xmath165 and @xmath166 such that @xmath167 .
moreover , @xmath168 for all @xmath169 iff @xmath22 is homogeneous of degree @xmath41 in @xmath170 , with respect to the center @xmath171 .
the following notations will be used in the proof : @xmath172 now if we consider the logarithm of @xmath173 and formally differentiate it , we obtain @xmath174 in order to prove the theorem , we need to show that the right hand side is nonnegative .
we accomplish this by proving differentiation formulas / inequalities in lemmas [ lem : first - id ] , [ lem : hprime ] , and [ lem : rellich ] , following similar proofs in @xcite or @xcite .
we start with the following alternate formula for @xmath175 .
[ lem : first - id ] for the minimizers @xmath22 of @xmath57 over @xmath44 , the following identity holds for @xmath176 with @xmath177 : @xmath178 [ pf : first ] to prove the lemma we note that for any test function @xmath179 which vanishes in a neighborhood of @xmath180 then we have @xmath181 for a small @xmath94 , choose @xmath182 , where @xmath183 and @xmath184 is such that @xmath185 we want to plug @xmath186 into and let @xmath187 .
we first claim that @xmath188 which is the same as @xmath189 indeed @xmath190 multiplying both sides of the above by @xmath191 and integrating over @xmath83 , we obtain @xmath192 using that @xmath193 and @xmath194 .
since the first integral on the right hand side goes to @xmath101 as @xmath187 , it remains only to estimate the second one .
we have @xmath195 again the first integral goes to @xmath101 , and to estimate the second one we use the @xmath0 regularity of @xmath22 to obtain @xmath196 besides , we also have that @xmath197 , which gives @xmath198 and establishes .
now , to complete the proof of the lemma , we let @xmath186 in and pass to the limit as @xmath187 .
using the fact that @xmath199 we obtain .
[ lem : hprime ] for the minimizer @xmath22 of @xmath57 over @xmath44 the following identity holds for @xmath176 with @xmath177 : @xmath200 the differentiation formula should be understood in the sense that @xmath201 is an absolutely continuous function of @xmath26 and that the differentiation formula holds for a.e . @xmath26 .
[ pf : hprime ] we have @xmath202 hence , we obtain @xmath203 which yields the desired identity .
while the above two identifies were the same as in the signorini problem , the third one becomes actually an inequality , which suffices for our purposes .
[ lem : rellich ] for the minimizer @xmath22 of @xmath57 over @xmath44 the following inequality holds for @xmath176 with @xmath204 : @xmath205 or , equivalently , @xmath206 we explicitly observe that the center @xmath171 of the ball @xmath207 must be in the upper thin halfball @xmath208 for the inequality to hold .
[ pf : relich ] the proof of this lemma uses the domain variation in radial direction similar to the one in @xcite . the main difference is that our constraints allow us to make perturbations that increase the distance from the origin , thus yielding an inequality ( with the correct sign ) instead of the equality in the non - constrained case .
we consider the function @xmath209 then for @xmath210 , we have @xmath211 note that the same will not be true for negative @xmath100 ( which is why we only have an inequality ) , that variation will bring over the zero values of @xmath22 from @xmath212 into @xmath213 , rendering the variation not an admissible function .
once we established the admissibility of @xmath214 , we can translate @xmath171 into the origin and continue the rest of the proof for balls centered at the origin . using the minimality of @xmath22
, we have @xmath215 letting @xmath187 this gives @xmath216 sending this time @xmath217 , we obtain @xmath218 which is equivalent to .
we can now prove the monotonicity of almgren s frequency .
[ pf : almgren ] the three lemmas proved above imply @xmath219 the last inequality follows from the cauchy - schwartz inequality , the equality case of which has to be satisfied if @xmath220 and provides that @xmath22 is homogeneous ( see @xcite or @xcite ) . from the scaling properties of @xmath221
we can also see that it is constant when the function @xmath22 is homogeneous , thus the theorem is proved .
an important tool for us will be the following rescaling of the minimizers at some points @xmath222 : @xmath223 the limits of the rescaled functions @xmath224 as @xmath225 will be called _
blowups _ of @xmath22 at the point @xmath171 .
the above definition normalizes the @xmath226 norm of the rescaled functions to be one : @xmath227 another useful property is the following identity @xmath228 we next want to let @xmath229 and study the convergence of the rescaled functions @xmath230 .
we start by showing that such convergence will be strong in @xmath231 .
[ lem : convergence of rescalings ] let @xmath232 be a minimizer of @xmath57 over @xmath233 with some @xmath234 .
let also @xmath235 , and @xmath236 weakly in @xmath237 and @xmath238 in @xmath239 .
then @xmath240 strongly in @xmath241 : @xmath242 moreover @xmath243 minimizes @xmath57 over @xmath244 with boundary values @xmath245 .
\1 ) we first prove that for any two solutions @xmath246 and @xmath247 , @xmath248 are subharmonic : @xmath249 for all nonnegative test functions @xmath250 .
we will show only the subharmonicity of @xmath251 , the other one being analogous .
now since the only complications can occur on @xmath252 , without loss of generality we may assume that @xmath253 is nonempty .
then from the signorini conditions on @xmath18 we have that @xmath254 for any point @xmath255 , let @xmath256 be such that @xmath257 . then from harmonicity of @xmath258 in @xmath259
, we have that for any test function @xmath260 , @xmath261 , @xmath262 this implies the subharmonicity of @xmath251 in a neighborhood of any point @xmath255 , implying the subharmonicity in @xmath83 .
\2 ) take the sequence @xmath263 and @xmath243 as in the statement of the lemma .
the previous step shows that @xmath264 are subharmonic .
letting @xmath265 we get @xmath266 is also subharmonic . now using the energy inequality we obtain @xmath267 which implies the strong convergence in @xmath268 .
\3 ) recall now that @xmath232 minimizes @xmath57 over @xmath233 . since
@xmath232 are bounded in @xmath237 and the trace mapping is compact , we can take a subsequence such that @xmath269 in @xmath270 as @xmath271 .
taking the minimizer @xmath272 of @xmath57 on @xmath244 and letting @xmath243 be the strong limit of @xmath232 obtained in previous step and using that @xmath273 is subharmonic , we obtain @xmath274 thus @xmath232 converges uniformly to @xmath272 on @xmath275 for any @xmath276 , meaning @xmath277 in @xmath83 .
[ pf : blowups are homog .
] indeed , using and theorem [ thm : almgren ] for a @xmath280 we obtain @xmath281 using and the above estimate we arrive at @xmath282 which shows that the sequence @xmath283 is bounded in @xmath237 .
thus , we can choose a weakly converging subsequence @xmath284 in @xmath237 .
from lemma [ pf : lem - convergence of rescalings ] we also have the strong convergence @xmath285 in @xmath241 , which means in particular that @xmath286 provided @xmath287 .
now suppose @xmath288 .
then by the maximum principle we would have that the subharmonic functions @xmath289 vanish in @xmath275 , and since @xmath243 is harmonic in @xmath290 , we obtain that @xmath291 in @xmath83 .
but due to compactness of trace mapping , we have @xmath292 which contradicts to @xmath243 vanishing in @xmath83 .
thus holds for any @xmath293 .
moreover we can write @xmath294 yielding @xmath295 then using the last part of theorem [ thm : almgren ] , we complete the proof of the lemma .
[ pf : ncontact ] for a fixed @xmath296 consider a functional @xmath297 then , since for a fixed @xmath296 the functional above is continuous and that @xmath298 is nondecreasing in @xmath26 , we obtain the upper semicontinuity of the functional @xmath299 on @xmath24 .
more precisely , we have @xmath300 now , for a contact point @xmath301 we have a sequence of free boundary points @xmath302 converging to @xmath303 . now , near @xmath304 , the minimizer @xmath22 solves the signorini problem and therefore we have @xmath305 and thus , using the upper semicontinuity , we conclude that @xmath306 [ pf : noncontact ] since @xmath303 is not a contact point , we know that there exists a positive @xmath307 such that @xmath22 is harmonic in @xmath308 .
let @xmath243 be a blowup of @xmath22 at @xmath303 : @xmath309 we know that @xmath243 is homogeneous of degree @xmath310 , meaning @xmath311 for @xmath296 and @xmath312 .
we also know that @xmath243 is harmonic in @xmath313 and @xmath243 is nonnegative in @xmath314 .
next , for @xmath36 , define @xmath315 it is easy to see that @xmath316 is homogeneous of degree @xmath317 and @xmath318 thus , the set of possible values of @xmath41 includes @xmath319 .
we want to show that those are the only possible values of @xmath41 .
this fact will follow from the expansion of harmonic functions in slit domains , recently established in ( * ? ? ?
* theorem 3.1 ) .
the latter theorem implies that for any @xmath320 there exists a polynomial @xmath321 of degree @xmath322 such that @xmath323 solely from the fact that @xmath243 is harmonic in @xmath324 , vanishes continuously on @xmath58 and is even in @xmath50 .
taking @xmath325 and using that @xmath243 is homogeneous of degree @xmath41 , we obtain that @xmath326 for a homogeneous polynomial @xmath327 of degree @xmath328 .
thus , @xmath329 for some @xmath36 .
the proof is complete . | we study the signorini problem near a fixed boundary , where the solution is `` clamped down '' or `` glued . ''
we show that in general the solutions are at least @xmath0 regular and that this regularity is sharp .
we prove that near the actual points of contact of the free boundary with the fixed one the blowup solutions must have homogeneity @xmath1 , while at the non - contact points the homogeneity must take one of the values : @xmath2 , @xmath3 , , @xmath4 , . |
we will now derive bounds upon the secret key rate for the 8 members of the gaussian family of cvqkd protocols . in the prepare and measure setting ( p&m ) protocols correspond to alice sending either squeezed @xcite or coherent states @xcite and bob measuring with either homodyne or heterodyne @xcite detection .
each permutation can in turn be used to have bob try and guess alice s encoding , called direct reconciliation @xcite ( dr ) , or alice trying to guess bob s measurement @xcite , called reverse reconciliation ( rr ) .
one can also consider entanglement based ( eb ) schemes in which alice distributes one arm of a two - mode squeezed vacuum to bob @xcite . in fact these methods can be seen as equivalent since if alice heterodyne detects one arm of a squeezed vacuum this corresponds to sending coherent states , while a homodyne detection corresponds to squeezed state protocols @xcite . in the following
, we will calculate the key rates assuming all parties have chosen to measure in the @xmath26 basis . in general
the total key rate will be the average of the this quantity and the analogous expression with @xmath26 and @xmath54 labels permuted .
first we turn to the protocols where both alice and bob homodyne detect . here
alice and bob randomly choose whether they measure the @xmath26 of @xmath54 basis and only keep those times where they agree .
we work in the asymptotic regime @xmath55 where the optimal attacks are known to be the collective attacks . considering the case where bob makes an @xmath26 measurement ,
we can write the secret key rate for the dr protocol as @xcite , @xmath56 where @xmath57 denotes the classical mutual information between alice and bob , with @xmath58 being the continuous shannon entropy of the measurement strings and @xmath59 is the holevo bound . substituting ( [ iab ] ) and ( [ ie ] ) in ( [ k ] ) and comparing with the continuous conditional von neumann entropy @xmath60 we have , @xmath61 which is what one would expect from the devetak - winter relations @xcite . using the entropic style uncertainty relations allows us to bound the eavesdroppers information on the relevant observable . substituting ( [ uc2 ] ) from the main text and recalling that @xmath62 we can write , @xmath63 thus we have bounded the secret key by an expression that depends only upon the conditional shannon entropies and the complementarity of the measurements , both of which are directly accessible to alice and bob .
furthermore one can show via a differential calculation that for any probability distribution @xmath35 the corresponding shannon entropy is maximised for a gaussian distribution of the same variance .
in other words , alice and bob can bound their secret key rate for this protocol by measuring bob s conditional variance . substituting the shannon entropy for a gaussian distribution @xmath64 and @xmath65 we arrive at a final expression for the key rate of @xmath66 the rr expression
is obtained by simply permuting the labels of alice and bob .
the fact that this result can be pessimistic is immediately apparent if we consider the predictions for a perfect channel . under the security proofs presented in @xcite
if alice and bob share an pure epr state ( real or effective ) with variance @xmath67 then the key rate is always positive provided the squeezing parameter is non - zero ( @xmath68 ) . on the other hand
the above result is only positive for @xmath69 which implies a squeezing parameter of @xmath70 or about 1.3 db .
this result was also calculated in @xcite , however as mentioned before the proof relies on the assumption of the applicability of the entropic uncertainty relation to infinite dimensional hilbert spaces .
furthermore ferenczi argues that since for coherent states the directly measured conditional variances are always greater than 1 the above procedure would never predict a positive key rate for coherent state protocols .
however this conclusion comes from a mistaken application of the key rate formulae as we now demonstrate .
consider a dr coherent state protocol , which in the entanglement based picture involves alice making a heterodyne detection upon her arm of an epr pair .
thus she first mixes her mode with vacuum resulting in two modes @xmath71 and @xmath72 upon which she measured @xmath26 and @xmath54 respectively .
bob is still making a homodyne detection , randomly switching between the quadratures .
we will consider the case where bob measures @xmath26 , with the other case following straightforwardly .
the dr key rate is then given by , @xmath73 after alice s projective measurement upon @xmath72 the state @xmath74 is pure and we can again apply the entropic uncertainty relation to write , @xmath75 now this formula might pose a problem , in that we do not measure @xmath54 upon mode @xmath71 , but this can be circumvented if we trust the devices , specifically the beamsplitter , in alice s station for we then have @xmath76 which is measured .
we therefore have , @xmath77 for reverse reconciliation the key rate is given by , @xmath78 once again , the quantity @xmath79 is not measured directly , but provided we trust alice s beamsplitter we have @xmath80 allowing us to evaluate the bound .
these protocols are essentially mirror images coherent state homodyne schemes since , in the eb picture , we now have alice making homodyne measurements and bob making heterodyne measurements .
we now have alice swapping between @xmath26 and @xmath54 measurements while bob splits up his mode measuring @xmath26 upon @xmath81 and @xmath54 upon @xmath82 .
the dr key rate is given by , @xmath83 where we will need to trust the beamsplitter in bob s station to obtain @xmath84 from the directly measured conditional variance .
the rr key rate is given by , @xmath85 where we have again used the form of bob s beamsplitter to write @xmath86 .
the final protocols involve bob making a heterodyne measurement upon coherent states , or alternatively both parties making heterodyne measurements upon an two - mode squeezed vacuum .
thus there are now four modes involved @xmath71 and @xmath81 upon which @xmath26 is measured and @xmath72 and @xmath82 upon which @xmath54 is measured
. we can consider the @xmath26 and @xmath54 channels separately .
note that this is actually an underestimation of the key rate as it essentially allowing eve to devote all her resources to estimating the @xmath26 and @xmath54 measurements separately whereas she must in fact estimate both simultaneously .
the dr key rate for @xmath26 is given by , @xmath87 provided we trust the beamsplitter in bob s station we can calculate @xmath88 .
the rr key rate is given by , @xmath89 provided we trust the beamsplitter in alice and bob s station we can calculate @xmath90 .
another important benefit of utilising entropic uncertainty relations in qkd proofs is that they lend themselves towards one - sided device independent ( 1sdi ) protocols @xcite . for 1sdi - qkd protocols only one of alice or bob
is untrusted and the security is now related to the recently classified steering inequalities @xcite .
the 1sdi nature of entropic proofs is manifest in expressions like ( [ kh ] ) in that one can immediately see it depends only upon measuring a known observable upon one side .
more precisely , for a term like @xmath91 we must be sure that the observable upon alice s side is indeed an @xmath26 measurement but we need make no assumptions about what bob is actually measuring , we simply evaluate the conditional variance .
we note that the proof provided in @xcite is also 1sdi , being independent of alice s devices .
however this device independence does not necessarily extend to the protocols involving heterodyne detection or coherent states ( recall that this is alternatively a heterodyne detection on alice s side in the entanglement based picture ) .
this is essentially because the proof to derive rates such as ( [ kcoh ] ) depends upon characterising the beamsplitter on the supposedly untrusted side thereby invalidating the device independence .
nonetheless the remaining protocols , with the heterodyne detection taking place in the trusted station , are still implementable with high efficiency sources and detection opening the way to long range 1sdi - cvqkd protocols with current technology .
this means that for real epr states both dr and rr are secure provided all parties are homodyning , while bob may safely heterodyne for an rr protocol and alice may heterodyne for a dr protocol .
finally , for dr protocols where alice , who controls the source , is trusted we may also safely make the equivalence between p&m and eb schemes . remarkably , this means that for direct reconciliation is is possible to generate 1sdi key using only coherent states .
we summarise which of the 16 possible protocols are 1sdi in table 1 in the main text . | by applying recently developed entropic uncertainty relations in the continuous variable regime we derive bounds upon the secret key rate of gaussian modulated continuous variable quantum key distribution protocols in the limit of long key length . for several protocols the bounds obtained in this manner can be shown to be one sided device independent , including a protocol that uses only coherent states . though the relevant uncertainty relation is not tight , and
neither are the subsequent key rates , we find that one - sided device independent schemes are experimentally achievable with existing technology for transmission over realistic channels .
quantum systems have information theoretic properties that include both extensions and limitations with respect to their classical counterparts . on the one hand quantum systems
can exhibit entangled correlations whose strength exceeds all classical limitations @xcite , while on the other hand the uncertainty between complementary observables must obey fundamental quantum relations @xcite .
by considering entropic versions of these uncertainty relations @xcite an intimate connection between entanglement and uncertainty has recently been derived @xcite . in principle
this has been know since the seminal work of einstein , podolsky and rosen ( epr ) @xcite but was only recently quantified to derive an entropic uncertainty relation including the possibility of quantum correlations @xcite . beyond fundamental interest , ref .
@xcite demonstrated the usefulness of such relations by compactly re - deriving the secret key rate for the bb84 discrete variable quantum key distribution ( dvqkd ) protocol .
the derivation of similar relations for the smooth min- and max- entropies @xcite , which are appropriate in the non - asymptotic setting , eventually resulted in finite - size bounds for dvqkd that are both general and tight @xcite .
furthermore , security proofs of device independent ( di ) qkd protocols , in which stronger subsets of quantum correlations are used to remove the requirement for the experimental devices themselves to be trusted @xcite , have also proved amenable to these entropic relations @xcite .
finally , entropic relations also naturally lend themselves to situations where only one party s devices are untrusted .
this possibility , first noted in @xcite was subsequently used to prove the security of an experimentally challenging , but feasible , proposal for one - sided device independent ( 1sdi ) dvqkd protocols which are strongly related to the concept of epr steering @xcite .
an extension ref.@xcite to the infinite dimensional hilbert spaces relevant for cv - qkd @xcite has also been applied to propose a discretised 1sdi - cvqkd protocol that accounted for finite - size effects @xcite and a demonstration of scheme independent of bob s devices only , has recently been carried out @xcite . in this letter
we pursue the study of device independence for cvqkd and show that the hardware advantages that drastically improve the ease of detection - loophole free steering tests @xcite , also render 1sdi - cvqkd practical in the near term .
firstly we briefly review progress in entropic uncertainty relations and find culminating in a recently derived uncertainty relation for continuously valued measurements on infinite - dimensional systems , e.g. position and momentum measurements of a harmonic oscillator @xcite .
this relation was independently found by @xcite although the latter derivation is incomplete . applying this relation to the entire family of gaussian cvqkd protocols
we find that the consequent bounds are quite loose with respect to existing results .
however , analysing these bounds from the perspective of one - sided device independence , we find that several of the protocols considered are in fact inherently 1sdi , with the key rate displaying an elegant connection to the relevant steering parameter .
an analysis of their performance over realistic channels shows that 1sdi - cvqkd protocols are feasible , reasonably robust to losses , and hence more practical than their discrete variable counterparts .
we also demonstrate the surprising result that 1sdi - cvqkd is possible , albeit over a more limited range , using only coherent states . _ entropic uncertainty relations for continuous variable systems. _ originally , entropic uncertainty relations were derived assuming one starts without any additional information or with only classical information describing the system in question , i.e. the density matrix @xcite .
the more general case , allowing for quantum side information , was derived in @xcite though only for finite dimensional hilbert spaces and observables with a discrete spectrum , and proceeds as follows . consider a pair of observables @xmath0 with a complementarity @xmath1 where @xmath2 are the eigenvectors of the observables .
these are to be measured on a state @xmath3 which is potentially entangled with another state , @xmath4 , leading to the the following relation for the uncertainty in the pair of observables including access to @xmath4 @xcite , @xmath5 here @xmath6 where @xmath7 is the conditional von neumann entropy of the state @xmath8 whereas @xmath9 is the conditional von neumann entropy of the _ observable _ @xmath10 measured on @xmath3 given by , @xmath11 with @xmath12 the classical shannon entropy . preempting applications to qkd
, one can also consider that the state @xmath8 could have suffered some decoherence which is purified by an environment , or eavesdropper , such that @xmath13 .
using the purity of the overall state ( i.e. @xmath14 ) one can recast ( [ uc1 ] ) to find @xcite , @xmath15 these results have since seen several generalisations both to different entropies and to positive - operator values measurements ( povm s ) @xmath16 @xcite for which @xmath17 where @xmath18 is the infinity or operator norm .
however , these results are only valid for measurements with a finite number of discrete outcomes made on states living in a finite - dimensional hilbert space . for the purposes of cvqkd
we will require an uncertainty relation valid for infinite - dimensional hilbert spaces and continuous - valued measurements .
in particular , we are interested in homodyne measurements of the canonically conjugate quadratures @xmath19 , @xmath20 satisfying @xmath21 } = i\hbar$ ] where @xmath22 and @xmath23 and @xmath24 are bosonic creation and annihilation operators .
just such a relation has been recently developed , building an earlier result for discrete variables on infinite dimensional hilbert spaces @xcite which was first extended to countably infinite measurements which could then be applied by to a discretised version of a homodyne detection @xcite . deriving results for continuous spectra by taking infinite precision limits these coarse - grained povm s
had previously been extensively studied for the shannon entropies @xcite , and an analogous procedure for the quantum conditional von neumann entropy was utilised by ferenczi @xcite and berta et al .
@xcite although the former proof is incomplete .
an alternative derivation was also provided by frank and lieb @xcite .
the final result is the following relation for homodyne detection upon infinite dimensional hilbert spaces @xcite , @xmath25 _ application to cvqkd. _ the most common cvqkd protocols are the gaussian state protocols which encode information in the quadratures of the optical field which are described exactly by operators like @xmath26 and @xmath27 described above .
one can use squeezed @xcite or coherent @xcite states , and measure with either homodyne detection , switching between quadratures , or heterodyne detection @xcite where both are measured simultaneously .
one could also use gaussian entangled states @xcite and these approaches have been shown to be equivalent @xcite .
the communicating parties , called alice and bob in this context , can also use either a direct reconciliation ( dr ) scheme where alice sends corrections to bob or a reverse reconciliation ( rr ) @xcite which allows for losses above 50% .
one can also achieve this loss - tolerance via a post - selection protocol which discards some of the key in order to retain a more correlated subset @xcite .
previous work proved the security of cvqkd in the asymptotic limit against individual @xcite and collective @xcite attacks .
finally the proofs were raised to the level of the most general coherent attacks by use of the de finetti theorem @xcite adapted to infinite dimensions @xcite which shows that collective attacks are in fact optimal .
neglecting detector and reconciliation efficiencies one can show that the rr secret key rate is lower bounded by @xcite , @xmath28 where @xmath29 denotes the classical mutual information between alice and bob , with @xmath30 being the continuous shannon entropy of the measurement strings and @xmath31 is the holevo bound . substituting ( [ iab ] ) and ( [ ie ] ) in ( [ k ] ) and comparing with ( [ cond ] )
we have , @xmath32 substituting the uncertainty relation ( [ eve ] ) and recalling that @xmath33 we can write , @xmath34 thus we have bounded the secret key by an expression that depends only upon the conditional shannon entropies and the complementarity of the measurements , both of which are directly accessible to alice and bob .
furthermore one can show via a variational calculation that for any probability distribution @xmath35 the corresponding shannon entropy is maximised for a gaussian distribution of the same variance .
in other words , alice and bob can bound their secret key rate for this protocol by measuring bob s conditional variance . substituting the shannon entropy for a gaussian distribution @xmath36 @xcite and @xmath37 we arrive at a final expression for the key rate of @xmath38 with the dr expression obtained by simply permuting the labels of alice and bob .
this result was also calculated in ref .
@xcite , however the proof there was incomplete as it relied on the assumption of the applicability of the entropic uncertainty relation .
moreover , it was incorrectly concluded that this method would never predict positive key for coherent state or heterodyne protocols .
in fact , the extension to the other gaussian protocols is straightforward and given in the appendix .
for example the key rate for the dr protocol with coherent states and homodyne detection is given by @xmath39 unfortunately these bounds are actually more pessimistic than previous results @xcite as the uncertainty relation becomes quite loose in the relevant case of large losses .
nonetheless for some of the gaussian protocols these bounds can be shown to allow 1sdi - cvqkd for reasonable parameters and it is to this application that we now turn .
_ one - sided device - independent cvqkd. _ an important benefit of utilising entropic uncertainty relations in qkd proofs is that they lend themselves towards one - sided device independent ( 1sdi ) protocols @xcite .
these are relaxed versions of fully di schemes @xcite , in which all devices are untrusted and the security is guaranteed via a detection - loophole - free bell violation @xcite .
the only assumptions that need to be made are the security of the stations , the causal independence of the measurement trials and a trusted source of randomness for choosing measurement settings . for 1sdi - qkd protocols
only one side , alice or bob , is untrusted and the security is now ensured through the recently classified steering inequalities @xcite .
the 1sdi nature of entropic proofs is manifest in expressions like ( [ kh ] ) in that it depends only upon measuring a known observable upon one side .
specifically , in the derivation of ( [ kh ] ) we only need to know that bob is measuring either @xmath40 or @xmath41 in order to apply the entropic uncertainty relation .
although we write expressions @xmath42 , as this is what will be measured in most experiments , alice could be making any measurement and the security would still hold provided the conditional variance was sufficiently small . in @xcite
1sdi - dvqkd protocols are analysed and although within reach of present technology are very challenging , significantly more so than standard dvqkd as one must perform a detection - loophole - free steering test .
this is difficult because on the untrusted side one can not post - select away events where there is no single - photon detection , which necessitates high detection efficiencies @xcite .
this is in stark contrast to the cv case , where steering is demonstrated via homodyne measurements which are unconditional and thus immune to detection loopholes .
in fact it has since been shown @xcite that cv steering is equivalent to the so - called einstein - podolsky - rosen violation @xcite which has been experimentally feasible for over 20 years @xcite and now sees very strong violations @xcite . for the gaussian cv states relevant to qkd , steering is demonstrated by a violation of the condition on the conditional variances , @xmath43 for alice steering bob @xcite . comparison with eq .
( [ kv ] ) shows that for the rr key rate @xmath44 if and only if @xmath45 , with the corresponding relation between the dr and @xmath46 following straightforwardly . in other words , the condition for positive one - sided device independent
key is more stringent than epr steering , similarly to the case for dvqkd @xcite .
thus for epr states and homodyne detection any positive key is by definition 1sdi , independent of alice for rr and bob for dr .
however this device independence does not necessarily extend to the protocols involving heterodyne detection this is essentially because the proof to derive rates such as ( [ kcoh ] ) depends upon characterising the beamsplitter on the supposedly untrusted side thereby invalidating the device - independence .
alternatively , recall that a steering demonstration requires a measurement choice by the untrusted party which does nt take place if they heterodyne detect .
nonetheless the remaining protocols , with the heterodyne detection taking place in the trusted station , are still implementable with high efficiency sources and detection opening the way to 1sdi - cvqkd protocols with current technology .
this means that for real epr states both dr and rr are secure provided all parties are homodyning , while bob may safely heterodyne for an rr protocol and alice may heterodyne for a dr protocol .
finally , for dr protocols where alice , who controls the source , is trusted we may also safely make the equivalence between prepare and measure ( p&m ) and entanglement - based ( eb ) schemes .
remarkably , this means that for direct reconciliation it is possible to generate 1sdi key using only coherent states .
we summarise which of the 16 possible protocols are 1sdi in table.[default ] .
.the 6 out of 16 cvqkd protocols that are 1sdi .
the subscript indicates who s devices are untrusted , e.g. @xmath47 indicates a protocol is independent of bob s devices .
the coloured line segments identify each protocol with it s loss and noise tolerance in fig.[kcohrr ] , with some of the protocols being degenerate between the p&m and eb versions .
[ default ] [ cols="^,^,^,^,^,^,^ " , ] to evaluate performance we consider a gaussian channel , an excellent model for real fibre optic cable , characterised by a transmission @xmath48 and an excess noise parameter @xmath49 given in units of shot noise .
the noise parameter can be thought of as the noise input to the channel in the sense that a pure state with variance of unity would have a variance @xmath50 after the channel .
neglecting imperfections such as detector and reconciliation efficiency this is the chief factor limiting range . in fig.[kcohrr ] we plot the secure regions for all 4 distinct 1sdi protocols ( there are two redundancies between p&m and eb schemes ) .
the best performing scheme is the rr epr scheme where both parties homodyne . in the limit of low excess noise
this scheme is secure for up to 73% loss .
for very low noises the next best scheme is the rr protocol where bob heterodyne detects but for higher noises the dr protocol with both parties homodyning ( or alternatively where alice sends squeezed states ) performs better . finally , although the dr coherent state scheme performs the poorest it is still secure up to around 33% loss .
these results show that 1sdi - cvqkd is surprisingly robust to decoherence . and excess noise @xmath49 .
for each protocol secure communication is possible for all channels above the corresponding line . reverse and direct reconciliation schemes are plotted in red and blue respectively while a dashed line indicates a heterodyne detection by the trusted party . ] to give a rough idea of the maximum range of these protocols we consider a standard optical fibre for which the transmission is related to the distance @xmath51 in kilometres via @xmath52 .
we choose a low , but achievable excess noise of @xmath53 as measured in @xcite and find that for this kind of channel the homodyne - homodyne rr epr protocol is by far the best performing allowing for 1sdi security out to almost 30 km . _
conclusions. _ we have applied an entropic uncertainty relation that is valid for continuous measurements and infinite dimensional hilbert spaces to the gaussian cvqkd protocols and demonstrated that in the limit of infinitely long key one can derive bounds for several protocols that are one - sided device independent . remarkably we find that even protocols based on sending coherent states can be made independent of bob s detectors ( usually the target of side channel attacks ) for low - loss and low - noise channels .
the protocols involving epr states and homodyne detection were substantially more robust allowing independence from alice s devices for up to 42% pure loss and independence of bob s devices for up to 73% loss pure loss .
our key rates also make explicit the connection to the relevant steering parameter , confirming the link between the asymmetric non - locality of epr steering and the asymmetric performance of 1sdi - cvqkd protocols using different reconciliation directions .
several comments on extensions and directions for future work are in order , beginning with the prospect for extending this security proof to include finite size effects and comparison with the results in refs .
@xcite . there ,
the authors follow a similar program of applying entropic uncertainty relations , in this case to the smooth min - entropies , allowing them to account for all finite size effects while providing an extremely general proof , but only for dr homodyne protocols and are limited to short distances ( 5 km or less ) even with extremely high levels of squeezing . very recently ,
an extension to rr homodyne protocols , for both the asymptotic and finite - size regimes , secure up to 15 km has also appeared @xcite .
it appears very promising then , that these techniques could be adapted to prove the finite - size security of all the other protocols presented here .
a second avenue would be to revisit the restrictions , or lack thereof , made about the eavesdropper including physical assumptions about the quantum memory available to eve @xcite , which has already seen application in di - dvqkd @xcite .
another candidate to further extend the range of these protocols would be the noiseless linear amplifier @xcite which has already been proposed for application to fully di - dvqkd @xcite as well as bell tests @xcite .
even more appealing may be the measurement based versions of these amplification schemes @xcite that have recently been experimentally demonstrated @xcite in light of these results it appears that several 1sdi - cvqkd protocols are within the reach of current technology and multiple possibilities exist to extend the secure range of such schemes to long distances . _
acknowledgements. _ the authors would like to thank d.a .
evans , m.j .
hall , c. branciard , e.g. cavalcanti , t. symul , h.m .
chrzanowski , f. furrer , m. tomamichel and s.m .
assad for helpful discussions .
this research was supported by the arc centre of excellence ce110001027 .
73 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , ,
, * * , ( ) . ,
* * , ( ) . , , , * * , ( ) . , * * , ( ) . , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , ,
* * , ( ) .
, , , * * , ( ) . ,
( ) . , , , , , * * , ( ) . , , , , , , , * * , ( ) . ,
( ) . , , , , , , , , ( ) . , , , , , , , * * , ( )
.
, , , , , , , , , , , * * , ( ) . , , , , , , , ,
* * , ( ) . , , , * * , ( ) . , , , , * * , ( ) .
, , , , , , , * * , ( ) . , , , , , ( ) , . , * * , ( 2013 ) . ,
( ) . , , , , * * , ( ) . , , , ,
* * , ( ) . , , , * * , ( ) .
, , , ( ) . , * * , ( ) . , , , * * , ( ) . , , , , ,
* * , ( ) . , , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , , * * , ( ) . , * * , ( ) . , , , , , * * , ( ) . , * * , ( ) . , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , , , , * * , ( ) . , * * , ( ) . , , , * * , ( ) . ,
* * , ( ) . , * * , ( ) .
, , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , * * , ( ) . , , , , * * , ( ) . , , , , , * * , ( ) . , , ,
* * , ( ) . , , ,
, * * , ( ) . ,
* * , ( ) . , , , , * * , ( ) . , p. ( ) .
, , , , , * * , ( ) . , , , * * , ( ) .
, , , , * * , ( ) . , , , , ,
* * , ( ) . , , , , , , , , * * , ( ) . |
gauge / string duality@xcite provides us new insights into gauge theories in strong coupling regime . according to the gauge / string duality , the dual string theory , which corresponds to conformal gauge theories(e.g . , the @xmath6 super yang millis theory ) , is embedded in @xmath7 space with the metric @xmath8 where @xmath9 is the ten - dimensional metric and @xmath10 is the mostly plus flat space metric .
here we use @xmath11 as indices in ten dimensions , @xmath12 as indices in @xmath13 and @xmath14 @xmath15 as those in four dimensional flat space which lives on the boundary of the @xmath13 space .
@xmath16 , which is the curvature radius of the @xmath13 space , is also equal to the radius of the five - sphere @xmath17 .
it is given by the duality @xmath18 where the string coupling @xmath19 and the string length @xmath20 are given by @xmath21 and @xmath22 with @xmath23 being the regge slope parameter , respectively .
the thooft coupling is defined as @xmath24 .
one can easily see that the large thooft coupling limit is equivalent to the limit @xmath25 . in the limit @xmath26 and @xmath25
, the string theory can be approximated by supergravity .
then the duality reduces to correspondence between gauge theories at large thooft coupling and supergravity , one can investigate the nonperturbative properties of gauge theories at large thooft coupling by studying the corresponding supergravity theory .
there are also some interesting connections between the type @xmath27 superstring theory and the @xmath6 super yang millis ( sym ) theory .
first , the @xmath28 @xmath29 symmetry of the @xmath6 sym is the @xmath30 isometry of @xmath17 .
furthermore , the @xmath31 conformal symmetry of the gauge theory is the isometry of @xmath13 . in addition
, there is an implication that the radial direction ( @xmath32 ) in @xmath13 can be identified with the energy scale in four dimensional sym theory , namely , @xmath33 .
there have been substantial progresses in studying strong coupling gauge theories especially in terms of deep inelastic scattering .
a few years ago , polchinski and strassler@xcite studied the deep inelastic scattering on hadrons by using gauge / string duality where the usual structure functions @xmath34 and @xmath35 are calculated for both spinless and spin-@xmath0 hadrons when bjorken-@xmath2 is finite ( @xmath3 ) where supergravity approximation is valid .
the spinless hadron and spin-@xmath0 hadron correspond to supergravity modes of dilaton and dilatino , respectively .
furthermore , they also investigated the case at small-@xmath2 where the pomeron contribution with a trajectory of @xmath36 was found . since an infrared cutoff @xmath37 is introduced in order to generate confinement , the model
is then called hard wall model .
there are also some earlier studies@xcite on high energy scattering in gaug / string duality .
there have been a lot of further developments along this direction@xcite . a saturation picture based on deep inelastic scattering in ads
/ cft is developed@xcite afterwards and recently reviewed in ref .
in addition , the deep inelastic scattering off the finite temperature plasma in gauge / string duality is recently studied in refs .
@xcite .
our main object in this paper is to extend the calculation of deep inelastic scattering on a spin-@xmath0 fermion in hard wall model , and compute the parity violating structure function @xmath38 as well as the polarized structure functions @xmath39 ,
@xmath5 , @xmath40 , @xmath41 and @xmath42 . among these five polarized structure functions , @xmath40 , @xmath41 and @xmath42
are parity violating structure functions . type @xmath27 superstring theory , which lives in a ten - dimensional space ( e.g. , @xmath43 ) , is a parity - violating theory .
it contains massless left - handed majorana - weyl gravitinos and massless right - handed majorana - weyl dilationos .
the gravitino , which is a spin-@xmath44 fermion , is the superpartner of the graviton .
likewise , the dilatino , which is a spin-@xmath0 fermion , is the superpartner of the dilaton .
it has been proved that type @xmath27 superstring theory is anomaly free@xcite in terms of local ( gauge ) symmetries . here
we expect that the currents in the dual gauge theory is conserved at finite-@xmath2 as we will show later in the paper . in this paper
, we focus on the spin-@xmath0 dilatino and calculate its structure functions as well as form factors . in order to study the polarized structure functions and form factors , we follow the set - up in ref .
@xcite and assume the dilatino has a small mass @xmath45 which eventually can be related to the cutoff scale @xmath37 .
this paper is organized as follows . in section
[ deft ] , we provides the definitions for various structure functions as well as kinematic variables .
in section [ hard ] , we calculate the expectation value of the @xmath29-currents in our gedanken experiment of polarized deep inelastic scattering from gauge / string duality .
this eventually leads to the structure functions at finite @xmath2 .
the section [ disc ] is devoted to the discussions and comments on the structure functions and their sum rules . in section [ elas ] ,
we focus on the elastic scattering and derive the form factors for the spin-@xmath0 hadron .
finally in section [ conc ] , we summarize our results .
the hadronic tensor @xmath46 is defined as @xmath47 | p , \mathcal{q } , s \rangle \ , , \label{dis3}\ ] ] with @xmath48 being the incident current .
the hadronic tensor @xmath49 can be split as @xmath50 according to lorentz and cp invariance , the symmetrical and antisymmetrical parts can be expressed in terms of 8 independent structure functions as@xcite , @xmath51\nonumber\\ & & - \frac{1}{p{\cdot}q}\left ( p_\mu - \frac{p{\cdot}q}{q^2 } \ , q_\mu \right)\left ( p_\nu - \frac{p{\cdot}q}{q^2 } \ , q_\nu \right ) \left[f_2(x , q^2)+\frac{m s\cdot q}{p\cdot q}g_4(x , q^2)\right]\nonumber\\ & & -\frac{m}{2p\cdot q}\left[\left ( p_\mu - \frac{p{\cdot}q } { q^2}q_\mu\right)\left(s_\nu - \frac{s{\cdot}q}{p{\cdot}q } \ ,
p_\nu\right ) + \left ( p_\nu - \frac{p{\cdot}q } { q^2}q_\nu\right)\left(s_\mu - \frac{s{\cdot}q}{p{\cdot}q } \ , p_\mu\right)\right]g_3(x , q^2)\nonumber\\ w_{\mu\nu}^\mathrm{(a ) } & = & -\frac{m \ , \varepsilon_{\mu\nu\rho\sigma } \ , q^\rho}{p{\cdot}q } \left\ { s^\sigma \ , g_1(x , q^2 ) + \left [ s^\sigma - \frac{s{\cdot}q}{p{\cdot}q } \ , p^\sigma \right ] g_2(x , q^2 ) \right\ } -\frac{\varepsilon_{\mu\nu\rho\sigma}q^\rho p^\sigma}{2p{\cdot}q } f_3(x , q^2 ) , \label{wmunu}\end{aligned}\ ] ] where @xmath45 is the mass of the hadron
, @xmath52 is its polarization , @xmath53 is the momentum carried by the current and @xmath54 is the initial momentum of the hadron ( see fig .
( [ dis ] ) . ) . in deep inelastic scattering ,
we define the kinematic variables as the following : @xmath55 the mass of the intermediate state after the scattering is defined as @xmath56 .
all the structure functions are functions of @xmath2 and @xmath57 .
in the so - called hard wall model , polchinski and strassler impose a confinement scale @xmath37 in the fifth dimension of @xmath13 space . as we will see later in the paper ( see eq .
( [ mas ] ) ) , this scale also provides a mass scale for the hadrons . following polchinski and strassler@xcite
, we perform a gedanken experiment of polarized deep inelastic scattering which occurs between the boundary and the cutoff scale @xmath37 . here
we first summarize their set - up before we extend the calculations to the polarized case .
the incident current is chosen to be the @xmath29-current which couples to the hadron as an isometry of @xmath17 .
according to the ads / cft correspondence , the current excites a nonnormalizable mode of a kaluza - klein gauge field at the minkowski boundary of the @xmath13 space @xmath58 where @xmath59 denotes a killing vector on @xmath17 with @xmath60 being the angular coordinates on @xmath17 .
@xmath61 is the external potential in the gauge theory corresponding to the operator insertion @xmath62 on the boundary of the fifth dimension of the @xmath13 space with the boundary condition @xmath63 this gauge field fluctuation @xmath61 can be viewed as a vector boson field which couples to the @xmath29-current @xmath48 on the minkowski boundary , and then propagates into the bulk as a gravitational wave , and eventually interacts with the supergravity modes of the dilatino or dilaton .
the gauge field satisfies maxwell s equation in the bulk , @xmath64 . with a gauge choice
, one can solve this equation for @xmath65 .
usually people choose the gauge @xmath66 .
however , the problem is easier in the lorenz - like gauge , @xmath67 . with given boundary conditions , one obtains the solution in ref .
@xcite . ] @xmath68 where @xmath69(note that in @xmath70 metric signature , @xmath71 for spacelike current . ) . since @xmath72
, the deep inelastic scattering should be localized around @xmath73 which is far away from the cutoff @xmath74 for hard scattering when @xmath75 .
spin-@xmath0 hadrons correspond to supergravity modes of the dilatino . in the conformal region , one can write the dilatino field as @xmath76 where @xmath77 is an @xmath78 spinor on @xmath13 and @xmath79 is a normalized @xmath80 spinor on @xmath17 .
the wave - function @xmath81 satisfies a five - dimensional dirac equation in eq .
( [ dirac0 ] ) while an @xmath82 is missing in eq .
( [ dirac ] ) .
the detailed derivation is provided above . ]
@xmath83 the solution to this dirac equation is@xcite , @xmath84 u_\sigma\ , \label{dso}\ ] ] where @xmath85 here we define the @xmath86-matrices according to the dirac algebra in the mostly plus metric signature @xmath70 ( see , e.g. , the notation in ref .
@xcite ) @xmath87 which gives an additional factor of @xmath88 in @xmath89 ( @xmath90 ) .
it is straightforward to see that this explains the factor @xmath82 in front of the fermion mass in eq .
( [ dirac ] ) and the @xmath91 is the same as the usual definition .
the dilatino is taken to be in a charge eigenstate with charge @xmath92 under the @xmath93 symmetry , which yields @xmath94 .
this @xmath93 symmetry arises from the @xmath93 subgroup of the @xmath28 @xmath29-symmetry . for the initial hadron , by assuming @xmath95 in the interaction region and expanding the bessel functions in eq .
( [ dso ] ) up to linear term in @xmath45 , one gets @xmath96.\ ] ] @xmath97.\ ] ] in order to obtain polarized contribution of structure function , we have kept the next leading order @xmath45 of initial hadron . the conformal dimension @xmath98 of the state is found to be @xmath99 . for the intermediate hadron , @xmath100 and @xmath101 u_{x\sigma } \ .\ ] ] @xmath102 \ .\ ] ] before getting into the detailed calculation , let us look into the center of mass square of the intermediate states in ten dimension .
it is easy to see that @xmath103 thus we know that @xmath104 when @xmath105 . in this range of @xmath2 ,
only massless string states are the relevant intermediate states produced during the interaction , and supergravity calculation should be valid and reliable to obtain the structure functions .
when @xmath2 gets smaller , the massive string modes are excited and string scattering amplitude should be taken into account .
this has been calculated thoroughly for @xmath34 and @xmath35 .
unfortunately , we leave this part of the calculation for polarized structure functions for future studies .
therefore , it is straightforward to compute the matrix element and obtain @xmath106 where @xmath107 is the polarization of the current @xmath108 , @xmath109 and @xmath110 are the tangent space index , and the vielbein @xmath111 and @xmath112 are given by , @xmath113 here the vielbein is used to make the product lorentz invariant due to the fact that the gamma matrices are defined in the flat spacetime .
it then follows that , @xmath114 after changing variables to @xmath115 , one finds @xmath116 where @xmath117 using the following integral results @xmath118,\end{aligned}\ ] ] where the upper limits are approximately set to be @xmath119 , we have , @xmath120 or its complex conjugate do not change sign due to the fact that @xmath121 is imaginary in the notation that we are working with .
] , @xmath122 with the help of eq .
( [ dirac ] ) , it is easy to see that @xmath123 and @xmath124 as a result of current conservation .
in fact , with the present nlo approximation , we can only show that @xmath125 . nevertheless , we can expand the initial wavefunction up to nnlo ( @xmath126 order ) , and find that @xmath127 contributions are cancelled by nnlo terms in the initial wavefunction . if one continues to do this to higher orders , one can show that the current conservation is true for all orders of @xmath127 .
moreover , using the recursion relations of bessel functions , integrating the @xmath128 integral by parts and requiring the @xmath129 and @xmath130 to be the zeros of bessel functions as we use in the later elastic calculation , one can show that @xmath131 as @xmath132 its imaginary part can be written as @xmath133 in large @xmath57 limit , we approximately write @xmath134 . summing over radial excitations and final state spin , but keeping the initial spin , along with the relation @xmath135 derived from the optical theorem , yields , @xmath136\right\}\\ & = & \pi a_{0}^{\prime } { \mathcal{q}}^{2}(\lambda^{2}/q^{2})^{\tau -1}x^{\tau + 1}(1-x)^{\tau -2 } \nonumber \\ & & \times \left\ { \left ( \eta _ { \mu \nu } -\frac{q_{\mu } q_{\nu } } { q^{2}}% \right ) \left ( \frac{1}{2}+\frac{q\cdot s}{2p\cdot q}m\right ) \right .
\nonumber \\ & & \left .
-\frac{1}{p{\cdot } q}\left ( p_{\mu } -\frac{p{\cdot } q}{q^{2}}% \,q_{\mu } \right ) \left ( p_{\nu } -\frac{p{\cdot } q}{q^{2}}\,q_{\nu } \right ) \left ( 1+\frac{q\cdot s}{p\cdot q}m\right ) \right .
\nonumber \\ & & \left .
-\frac{m}{2p\cdot q}\left [ ( p_{\mu } -\frac{p{\cdot } q}{q^{2}}% \,q_{\mu } ) ( s_{\nu } -\frac{s\cdot q}{q^{2}}q_{\nu } ) + ( p_{\nu } -\frac{p{\cdot } q}{q^{2}}\,q_{\nu } ) ( s_{\mu } -\frac{s\cdot q}{q^{2}}q_{\mu } ) \right ] \right\}\end{aligned}\ ] ] and @xmath137 where @xmath138 . to obtain @xmath139 ,
we have used the identity , @xmath140= q^\mu\epsilon^{\nu\alpha\beta\gamma}p_\alpha q_\beta s_\gamma -q^\nu\epsilon^{\mu\alpha\beta\gamma}p_\alpha q_\beta s_\gamma -q^2\epsilon^{\mu\nu\alpha\beta}p_\alpha s_\beta\end{aligned}\ ] ] compare with eq .
( [ wmunu ] ) , we arrive at the final results , @xmath141 the @xmath34 and @xmath35 are exactly the same as the results found in ref .
@xcite by polchinski and strassler .
the results for @xmath38 and all of the polarized structure functions are new .
these structure functions are essentially calculated from the so - called double trace operators with their twist @xmath142 . in figure .
( [ structure ] ) , we illustrate the @xmath2 dependence of the @xmath39 and @xmath5 structure functions .
the @xmath38 , @xmath40 , @xmath41 and @xmath42 structure functions are just twice of the @xmath39 .
@xmath5 structure functions are especially interesting , it is negative at large @xmath2 region and positive at relatively small @xmath2 region which shares the same feature as seen in the proton @xmath5 experiment data . and @xmath5 structure functions , where @xmath143 and @xmath144.,title="fig:",width=302 ] and @xmath5 structure functions , where @xmath143 and @xmath144.,title="fig:",width=302 ]
in this section , we focus on the interpretation of the structure functions that we obtained from last section using gauge / string duality . we also compare the results with the structure functions obtained in qcd for nucleons ( for a review in qcd , see e.g. , refs .
@xcite . ) . *
since only the linear term in @xmath45 is kept in the initial wavefunction and throughout the calculation , the results shown above are from leading order calculation .
the corrections are of order @xmath145 and @xmath146 . * in qcd
, there is an interesting inequality @xmath147 which is derived from the positivity of the cross section@xcite .
here we see that @xmath148 , and the bound is saturated .
this indicates that initial hadron is completely polarized . in terms of string theory language
, this implies that the struck dilatino just tunnels or shrinks to smaller size of order the inverse momentum transfer during the scattering . as a result ,
the structure function exhibits a power law behavior in terms of the @xmath57 dependence which comes from the tunneling probability@xcite . *
it is now straightforward to compute the moments of all the structure functions when the contributions from @xmath149 are negligible .
typically there are just two different kinds of moments , e.g. , @xmath150 we expect that the moments are correct at least for @xmath151 where the low - x contributions are negligible . * in addition , when one sets @xmath152 for @xmath5 , one finds an interesting sum rule @xmath153 which is completely independent of @xmath154 and @xmath57 . in qcd , this sum rule is known as the burkhardt - cottingham sum rule@xcite in large @xmath155 limit .
however , this sum rule can be invalidated by non - regge divergence at low-@xmath2 .
* now let us take a closer look at the @xmath152 moment of the @xmath39 structure functions @xmath156 for sufficiently large @xmath157 , this integral vanishes .
this contradicts with the naive expectation that @xmath158 should remain finite as @xmath157 since the dilatino has spin-@xmath0 .
+ before we explain this problem , let us review the case of @xmath34 and @xmath35 structure function@xcite .
according to energy momentum conservation , the second moment of @xmath34 and the first moment of @xmath35 should have nonzero limit as @xmath157 .
this is known to be determined by the operator product expansion coefficients of @xmath159 .
however , it is not true for the result that we found above .
this indicates that some contributions to @xmath34 and @xmath35 which peak around @xmath160 are missing in above calculation .
the missing contributions are pomeron exchanges . at large thooft coupling
, the pomeron exchange is a graviton exchange which yields @xmath161 at small-@xmath2 where the correction to the pomeron intercept arises from the curvature of @xmath13 .
the pomeron contribution will survive in the large @xmath57 limit and give us a non - vanishing second moment of @xmath34@xcite .
+ therefore , there should be a similar contribution to @xmath39 at small-@xmath2 .
usually , the physical scattering amplitudes , which can be written in terms of @xmath162 and @xmath163 , have the same leading order @xmath164 singularity .
in other words , @xmath39 should always be less singular than @xmath34 . in terms of the regge theory
, there should be an axial vector regge exchange contribution which yields a singular@xcite @xmath165 with @xmath166 when @xmath2 is extremely small .
this contribution will also survive in large @xmath57 limit and yield a finite first moment .
this may indicate that most of the hardon spin is carried by the small-@xmath2 constituents inside the hadron .
* normally in qcd , @xmath39 structure function contains two parts , namely , the singlet part and the non - singlet part .
the singlet part contains the polarized singlet quark and gluon spin contributions , while the non - singlet part can be cast into the bjorken sum rule .
one can subtract off the singlet part and derive the bjorken sum rule by calculating @xmath167 $ ] at large @xmath155 limit . here
@xmath168 , @xmath169 stand for the @xmath39 structure functions of proton and neutron , respectively . since in our above ads / cft calculation we only use the @xmath93 subgroup of @xmath28 @xmath29-flavor - symmetry group and calculate the contributions from double trace operators , we can not distinguish the singlet part from the non - singlet part .
both parts are not included in the calculation .
however , if one includes the contribution from the axial currents and uses the full @xmath28 group , then one gets an additional flavor factor @xmath170 where @xmath171 are the @xmath28 flavor matrices and the flavor indices @xmath172 are set equal .
it is straightforward to see that this flavor factor also contains both singlet and non - singlet parts .
according to our calculation at finite @xmath2 , both of them are small at large @xmath57 limit .
the detail discussions on the small-@xmath2 limit of the @xmath39 structure function will be available in ref .
@xcite . * the parity violating structure functions @xmath38 ,
@xmath40 , @xmath41 and @xmath42 are as large as the @xmath35 structure function due to the reason that the dilatino is right - handed fermion in massless limit .
they are tightly related to the peculiar wavefunction of the dilatino .
however , we expect that @xmath39 and @xmath5 may exhibit some common features of the polarized structure functions of spin-@xmath0 hadrons in the non - perturbative region when the coupling is large .
in this section , we focus on elastic scattering off a spin-@xmath0 fermion in gauge / string duality in the hard wall model framework . in the case of the elastic scattering ,
the final state is the same as the initial state which allows us to set @xmath173 and @xmath174 .
thus , the only variable is @xmath57 . in ads
/ qcd model , the meson form factors have been extensively studied in refs .
furthermore , the nucleon ( spin-@xmath0 hadron ) form factors are then computed in ref .
@xcite . here in this section
, we would like to follow the formalism that we developed for the deep inelastic scattering , and use it in the elastic scattering , then calculate all possible form factors for spin-@xmath0 hadrons . here in this section
, we need to keep the full dilatino wavefunction since @xmath175 is no longer a large parameter . to compute the form factors , one can first write down the most general definition for elastic form factors @xmath176 where we have used the fact that @xmath177 , @xmath178 , @xmath179 , @xmath180 and @xmath91 form the complete sets of @xmath181 @xmath86 matrices . among all these form factors ,
@xmath182 and @xmath183 are the dirac and pauli form factors , respectively .
they are related to the vector current exchange . @xmath184 and @xmath185 are the axial form factors related to axial vector current .
@xmath186 usually vanishes if the current is conserved .
it is easy to see that in our present framework , the @xmath179 component is missing , and thus the @xmath187 is zero can be obtained with further assumptions ( e.g. , see their erice talk and ref .
@xcite . ) . ] . in ref .
@xcite , where non - vanishing pauli form factor @xmath187 is obtained , a new @xmath179 term has to be introduced into the action . before we calculate the form factors from the current expectation value , let us take a look at how current conservation is satisfied
. the current conservation condition can be written as @xmath188.\end{aligned}\ ] ] using the dirac equation , one can simplify above expression and obtain @xmath189\end{aligned}\ ] ] using the following identities , @xmath190=-x^\nu k_{\nu-1}(x ) , \
, \frac{d}{dx}[x^{\nu } j_\nu(x)]=x^\nu j_{\nu-1}(x),\ , \frac{d}{dx}\left[x^{-\nu}j_\nu(x)\right]=-x^{-\nu}j_{\nu+1}(x)\end{aligned}\ ] ] one can easily show that @xmath191 dz\end{aligned}\ ] ] and eventually @xmath192 this indicates that the current is conserved when @xmath193 where @xmath194 and @xmath195 are @xmath196-th zeros of @xmath197 and @xmath198 , respectively .
this is essentially equivalent to the mass spectrum found in ref .
@xcite by requiring vanishing chiral spinor wavefunction on the hard wall located at @xmath74 . furthermore
, we would like to comment that in large @xmath57 limit , the current conservation is trivially satisfied when one set the upper limit of @xmath199-integral as @xmath119 , where we find @xmath200 where @xmath201 is found to be @xmath202\nonumber \\ & & -\frac{2\tau}{m}\left(\frac{m^2}{q^2}\right)^{\tau } \left . _ { 2}\textrm{f}_{1}\right .
\left(\tau-\frac{1}{2 } , \tau+1 ; 2\tau-1;\frac{-4m^2}{q^2}\right).\end{aligned}\ ] ] using taylor expansions of the hypergeometric functions , one can easily show that @xmath203 . ,
@xmath204 and @xmath205 , where we have normalized @xmath206 .
we also set @xmath144 and @xmath207 or @xmath208 where @xmath209 and @xmath210 are the first zero root of @xmath211 and @xmath212 , respectively .
, title="fig:",width=302 ] , @xmath204 and @xmath205 , where we have normalized @xmath206 .
we also set @xmath144 and @xmath207 or @xmath208 where @xmath209 and @xmath210 are the first zero root of @xmath211 and @xmath212 , respectively
. , title="fig:",width=302 ] , @xmath204 and @xmath205 , where we have normalized @xmath206 .
we also set @xmath144 and @xmath207 or @xmath208 where @xmath209 and @xmath210 are the first zero root of @xmath211 and @xmath212 , respectively . , title="fig:",width=302 ] , @xmath204 and @xmath205 , where we have normalized @xmath206 .
we also set @xmath144 and @xmath207 or @xmath208 where @xmath209 and @xmath210 are the first zero root of @xmath211 and @xmath212 , respectively . , title="fig:",width=302 ] , @xmath204 and @xmath205 , where we have normalized @xmath206 .
we also set @xmath144 and @xmath207 or @xmath208 where @xmath209 and @xmath210 are the first zero root of @xmath211 and @xmath212 , respectively .
, title="fig:",width=302 ] , @xmath204 and @xmath205 , where we have normalized @xmath206 .
we also set @xmath144 and @xmath207 or @xmath208 where @xmath209 and @xmath210 are the first zero root of @xmath211 and @xmath212 , respectively .
, title="fig:",width=302 ] assuming @xmath213 , one can set the upper limit of @xmath214 integral as @xmath119 and thus obtain @xmath215 and @xmath216 at large @xmath57 limit , we find that @xmath217 and @xmath218 . in small @xmath57 limit ,
we expand the bessel functions @xmath219 up to @xmath220 but neglect @xmath57 terms .
it is then straightforward to evaluate the @xmath221 integral which yields @xmath222\nonumber\\ & & + 2|c^{\prime}|^2\left(\frac{m}{2\lambda}\right)^{2\tau-1}\frac{q^2}{m^2}\ln\left(\frac{q}{\lambda}\right ) \frac{\ _ 2f_3(\tau-\frac{3}{2},\tau;\tau-1,\tau+1,2\tau-3;-m^2/\lambda^2)}{2\tau\gamma(\tau-1)^2}\nonumber\\ & & + \frac{1}{2}|c^{\prime}|^2 \left(\frac{m}{2\lambda}\right)^{2\tau+1}\frac{q^2}{m^2}\ln\left(\frac{q}{\lambda}\right ) \frac{\ _ 2f_3(\tau-\frac{1}{2},\tau+1;\tau,\tau+2,2\tau-1;-m^2/\lambda^2)}{2(\tau+1)\gamma(\tau)^2 } \\
\mathcal{f}_2\left(q^2\right)&= & 0 \quad \textrm{and } \quad \mathcal{f}_3\left(q^2\right)=0\end{aligned}\ ] ] and @xmath223 together with @xmath224 in the end , one can use numerical methods and evaluate all these form factors with chosen @xmath154 and ratio @xmath129 , then plot them in terms of functions of @xmath175(see figure .
( [ formfactors ] ) ) . according to the power counting rule , we set @xmath225 for now .
it is easy to see that above form factors give rise to logarithmic divergent charge radii for the charged hadron .
this is peculiar in the hard wall model and will be cured in our following - up phenomenological studies@xcite . besides
, we also compare our results of @xmath226 with experimental data , which are shown in figure .
( [ formfit ] ) . in the hard wall model .
the data points are taken from @xcite , width=302 ]
using gauge / string duality , we have calculated the structure functions as well as the form factors of a spin-@xmath0 hadron . especially the polarized structure functions and parity violating structure functions are new .
we find that the burkhardt - cottingham sum rule is also true in our present calculation when the small-@xmath2 contribution to @xmath5 is negligible .
however , the situation for the @xmath39 structure function is more subtle and complicated .
we conjecture that there should be an axial regge contribution to @xmath39 at small-@xmath2 which may indicate that most of the hadron spin is carried by small-@xmath2 partons .
the phenomenological application of above calculation is very appealing and will be available soon@xcite .
we acknowledge inspiring discussions with s. brodsky , v. koch , y. kovchegov , a. mueller , j. qiu , g. teramond and f. yuan .
we would like to thank y. hatta for enormous communication and stimulating discussions during the preparation of this paper . j.g . acknowledges financial support by the national natural science foundation of china under grant no.10525523 .
is supported by the director , office of energy research , office of high energy and nuclear physics , divisions of nuclear physics , of the u.s .
department of energy under contract no .
de - ac02 - 05ch11231 .
e. witten , adv .
* 2 * , 253 ( 1998 ) [ arxiv : hep - th/9802150 ] .
s. s. gubser , i. r. klebanov and a. m. polyakov , phys .
b * 428 * , 105 ( 1998 ) [ arxiv : hep - th/9802109 ] .
j. polchinski and m. j. strassler , phys .
lett . * 88 * , 031601 ( 2002 ) [ arxiv : hep - th/0109174 ] .
j. polchinski and m. j. strassler , jhep * 0305 * , 012 ( 2003 ) [ arxiv : hep - th/0209211 ] .
r. c. brower and c. i. tan , nucl .
b * 662 * , 393 ( 2003 ) [ arxiv : hep - th/0207144 ] .
h. boschi - filho and n. r. f. braga , phys .
b * 560 * , 232 ( 2003 ) [ arxiv : hep - th/0207071 ] .
r. c. brower , m. j. strassler and c. i. tan , arxiv:0710.4378 [ hep - th ] .
r. c. brower , j. polchinski , m. j. strassler and c. i. tan , jhep * 0712 * , 005 ( 2007 ) [ arxiv : hep - th/0603115 ] .
c. a. ballon bayona , h. boschi - filho and n. r. f. braga , jhep * 0809 * , 114 ( 2008 ) [ arxiv:0807.1917 [ hep - th ] ] . c. a. ballon bayona , h. boschi - filho and n. r. f. braga , jhep * 0810 * , 088 ( 2008 ) [ arxiv:0712.3530 [ hep - th ] ] . c. a. ballon bayona , h. boschi - filho and n. r. f. braga , jhep * 0803 * , 064 ( 2008 ) [ arxiv:0711.0221 [ hep - th ] ] .
b. pire , c. roiesnel , l. szymanowski and s. wallon , phys .
b * 670 * , 84 ( 2008 ) [ arxiv:0805.4346 [ hep - ph ] ] .
l. cornalba and m. s. costa , phys .
d * 78 * , 096010 ( 2008 ) [ arxiv:0804.1562 [ hep - ph ] ] .
y. hatta , e. iancu and a. h. mueller , jhep * 0801 * , 026 ( 2008 ) [ arxiv:0710.2148 [ hep - th ] ] .
y. hatta , prog .
* 174 * , 298 ( 2008 ) .
y. hatta , e. iancu and a. h. mueller , jhep * 0801 * , 063 ( 2008 ) [ arxiv:0710.5297 [ hep - th ] ] .
b. lampe and e. reya , phys .
rept . * 332 * , 1 ( 2000 ) [ arxiv : hep - ph/9810270 ] .
w. mueck and k. s. viswanathan , phys .
d * 58 * , 106006 ( 1998 ) [ arxiv : hep - th/9805145 ] .
s. weinberg , `` the quantum theory of fields .
vol . 1 : foundations , '' _ cambridge , uk : univ .
( 1995 ) 609 p _ g. f. de teramond and s. j. brodsky , phys .
lett . *
102 * , 081601 ( 2009 ) [ arxiv:0809.4899 [ hep - ph ] ] . s. j. brodsky and g. f. de teramond , phys .
d * 78 * , 025032 ( 2008 ) [ arxiv:0804.0452 [ hep - ph ] ] .
s. j. brodsky and g. f. de teramond , phys .
d * 77 * , 056007 ( 2008 ) [ arxiv:0707.3859 [ hep - ph ] ] .
s. j. brodsky and g. f. de teramond , phys .
* 96 * , 201601 ( 2006 ) [ arxiv : hep - ph/0602252 ] .
s. j. brodsky and g. f. de teramond , phys .
b * 582 * , 211 ( 2004 ) [ arxiv : hep - th/0310227 ] .
h. r. grigoryan and a. v. radyushkin , phys .
d * 76 * , 115007 ( 2007 ) [ arxiv:0709.0500 [ hep - ph ] ] .
h. r. grigoryan and a. v. radyushkin , phys .
d * 76 * , 095007 ( 2007 ) [ arxiv:0706.1543 [ hep - ph ] ] .
h. r. grigoryan and a. v. radyushkin , phys .
b * 650 * , 421 ( 2007 ) [ arxiv : hep - ph/0703069 ] .
z. abidin and c. e. carlson , phys .
d * 77 * , 115021 ( 2008 ) [ arxiv:0804.0214 [ hep - ph ] ]
. z. abidin and c. e. carlson , phys .
d * 77 * , 095007 ( 2008 ) [ arxiv:0801.3839 [ hep - ph ] ] .
d. k. hong , m. rho , h. u. yee and p. yi , phys .
d * 77 * , 014030 ( 2008 ) [ arxiv:0710.4615 [ hep - ph ] ] . | in this paper , we investigate deep inelastic and elastic scattering on a polarized spin-@xmath0 hadron using gauge / string duality .
this spin-@xmath0 hadron corresponds to a supergravity mode of the dilatino .
the polarized deep inelastic structure functions are computed in supergravity approximation at large t hooft coupling @xmath1 and finite @xmath2 with @xmath3 .
furthermore , we discuss the moments of all structure functions , and propose an interesting sum rule @xmath4 for @xmath5 structure function which is known as the burkhardt - cottingham sum rule in qcd . in the end ,
the elastic scattering is studied and elastic form factors of the spin-@xmath0 hadron are calculated within the same framework . |
after the symmetry breaking theory of landau @xcite , many thought science has reached a comprehensive understanding of different phases of matter . however , a new concept emerged in condense matter physics by quantum hall effect @xcite , quantum spin fluids @xcite and superconductors @xcite which was called topological order .
topological order is a different phase of matter which can not understood by landau theory so that one can find different topological phases with the same symmetries @xcite .
+ the ground state of a quantum system with topological order exhibits a long - range entanglement @xcite .
specifically , the topological entanglement entropy has been defined as a non - local order parameter of topological order @xcite so that it can characterize a quantum phase transition from a topological phase @xcite .
another interesting property of a topological phase is the behavior of the wilson loop where the expectation value of the wilson loop is related to perimeter of the loop in a topological phase , while it is related to area of the loop in a non - topological phase @xcite .
+ long - range entanglement of the ground state of a topological matter also leads to some interesting and exotic properties which have not been seen in symmetry breaking phases .
robust degeneracy of the ground state and exotic statistics of excitations of the system , which are called anyons , are two important characteristics of the topological phases @xcite .
braiding anyons leads to an arbitrary phase factor on wave function of the system in the abelian models and a unitary operator in the non - abelian models @xcite .
the robustness of degeneracy of the ground state and braiding operators against local perturbations is a key property of the topological phases which has converted them to important candidates for fault - tolerant quantum computation @xcite .
+ in spite of many improvements on defining different characteristics of a topological phase , recognizing different topological classes in different quantum systems @xcite is still an open problem which has received much deal of interest . since
topological order is a kind of quantum order , it is clear that the same quantum phases have also the same topological phases .
consequently , considering quantum phase transition between different topological models is a useful approach for classification of topological models .
specifically , there are some recent papers which show a quantum phase transition can well reveal different topological properties of two different topological phases @xcite .
+ this thought that a quantum phase transition reveals necessarily different topological classes of the quantum phases is challenged specifically by 1d quantum models . it has been shown that all 1d quantum phases belong to a trivial topological phase @xcite so that one can find a quantum phase transition between two different 1d quantum models which are topologically trivial @xcite .
in fact , since topological signature of topological phases such as topological entanglement entropy or expectation value of the wilson loop operators can not be defined for a 1d model , topological order of 1d models only defines as a symmetry protected topological phase @xcite .
+ moreover 1d quantum models which are in the same topological phases , there are also two - dimensional ( 2d ) quantum models which belong to a universal topological phase . a well - known example of universal topological phases has been seen in the toric code model ( tc ) with @xmath0 topological order @xcite and the color code model ( cc ) with @xmath1 topological order @xcite .
since the tc and cc model have different gauge symmetries , degeneracies of the ground state of them are different .
while the tc model provides a four - fold degenerate ground state which can be used as a robust quantum memory , degeneracy of the ground state of the cc model is sixteen - fold .
furthermore , unlike the tc model , there is also possibility of applying unitary clifford group in a topological way @xcite .
recently , there is also an experimental realization of the color code on trapped - ion qubits @xcite .
+ although different degeneracies of the cc model and the tc model show that they are in two different quantum phases , other topological characteristics of both models are completely the same . especially , it is possible to show a cc model and two copies of the tc model are in the same quantum phases @xcite . in @xcite ,
the authors have emphasized that since there is an adiabatic evolution without quantum phase transition between a cc model and two copies of the tc model , they have the same topological phases .
this result shows that different quantum phases of the tc and the cc model are completely related to their different gauge symmetries and has not a topological nature .
+ by above facts , there is a good chance that one considers how the same topological characteristics of the cc and the tc model reflects in the properties of a quantum phase transition between them . to this end , in this paper we consider a cc model on a hexagonal lattice besides a single tc model instead of two copies of it on the same lattice .
we show the universal topological phase of the cc and the tc model reflects in properties of the quantum phase transition where we show that the quantum phase transition has a 1d nature .
our result is also another proof for the universal topological phase of the cc model and the tc model .
we emphasize that our proof has a recent and important point where generally shows how the universal topological phase of two 2d topological models can be revealed even in presence of a quantum phase transition .
+ to deriving results , we define a specific version of the tc model on the hexagonal lattice .
we show that a model hamiltonian as interpolation of the cc and the tc model is mapped to many copies of 1d ising models in transverse field which belong to the 2d classical ising universality class .
the mapping is based on re - writing the hamiltonian of the model in a new complete basis .
our method is equivalent with unitary transformation on the hamiltonian which does not change spectrum of the model .
we also study the behavior of wilson loops and explicitly show that the expectation value of the wilson loops in our model captures the non - topological nature of the phase transition .
+ from another point of view , we find the point of the phase transition of our model which is the same as 1d ising model in transverse field at @xmath2 @xcite where @xmath3 and @xmath4 are the couplings of the tc and the cc models , respectively .
it shows when we add the tc hamiltonian as a small perturbation against the cc model , such a perturbation can not ever change the quantum phase of the cc model .
the quantum phase transition occurs only when a considerable perturbation @xmath5 is applied .
such a result show that unlike the small robustness of the cc model against local perturbations such as magnetic field or ising interaction @xcite , it is strongly robust against a topological perturbation like the tc model . +
in section ( [ se1 ] ) , we review different properties of the topological phases of the tc and the cc model . in section ( [ se3 ] ) , we present our model as a interpolation of the cc and the tc model . we define both topological models on a hexagonal lattice and qualitatively show how a quantum phase transition happens . in section ( [ se4 ] ) , we use a mathematical method to map our model to 1d ising models in transverse field which have a well - known phase transition point .
finally , in section ( [ se5 ] ) we consider the behavior of wilson loop operators by a perturbative approach to show the quantum phase transition in our model has not a topological nature .
in this section we review the topological structure of the tc and the cc models .
we specifically emphasize on different topological degeneracies and different topological structures of them by representation of the ground states as a loop - condensate states . for more details , there is also a comparative study of these models which has done in @xcite .
+ a tc model is ordinarily defined as the ground state of a model hamiltonian on an oriented graph where qubits live on edges of the graph .
two commutative operators @xmath6 and @xmath7 are defined corresponding to each plaquette and vertex of the graph in the following form , see figure ( [ toric ] , left ) for a square lattice : ( @xmath8 ) operator on each qubit . by applying a sequence of @xmath9 ( @xmath8 ) operators on qubits it is possible to move a charge ( flux ) anyon on the lattice.,width=377,height=170 ]
@xmath10 where @xmath11 refers to qubits around of a plaquette and @xmath12 refers to qubits around of a vertex and @xmath13 are the pauli operators .
the hamiltonian of the model is as summation of these operators on all plaquettes and vertices of the graph as the following form : @xmath14 by the fact that the vertex and plaquette operators commute with each other , it is simple to show the following state is a ground state of hamiltonian ( [ e1 ] ) : @xmath15 where the state @xmath16 is eigenstate of the operator @xmath8 corresponding to eigenvalue @xmath17 and we ignore the normalization factor for this state . + there is also a simple representation for the state ( [ e2 ] ) which helps for better understanding of the topological order of such a state . to this end , let us span the product of operators @xmath18 on all plaquettes in the relation ( [ e2 ] ) as the following : @xmath19 where the right hand of this relation is the summation of all possible products of the plaquette operators . since each plaquette operator can be interpreted as a loop operator on the lattice , right hand of the relation ( [ e3 ] )
is the summation of all possible loop operators .
+ finally , since @xmath20 where the state @xmath21 is the eigenstate of the operator @xmath8 corresponding to eigenvalue @xmath22 , the ground state of the tc model ( [ e2 ] ) can be interpreted as uniform superposition of all loop constructions of qubits @xmath23 in the sea of qubits @xmath24 which is called the loop condensation .
+ such a state has a topological order which leads to a robust degeneracy in the ground state of the model when we define the model hamiltonian on a torus .
in fact , on a torus topology , there are also non - contractible loop operators in the form of : @xmath25 where @xmath26 refers to the qubits which live on a non - contractible loop around the torus in two different directions @xmath27 , see figure ( [ toric ] , left ) .
therefore , four the following quantum states are the degenerate ground states of hamiltonian ( [ e1 ] ) : @xmath28 where the indices @xmath29 refer to four different quantum states . + topological order of the tc model are understood by three important properties , robust degeneracy , non - local order parameter and anyonic excitations . + the robust degeneracy is due to this fact that four degenerate ground states ( [ e4 ] ) can not be converted to each other by any local parameter so the degeneracy is robust against each local perturbation .
+ another property is that any local operator can not distinguish four degenerate ground states .
in fact , there are two non - local operators which have different expectation values in different degenerate ground states .
such operators are defined in the following form : @xmath30 where @xmath26 refers to the qubits which live on a non - contractible loop around the torus in two different directions @xmath27 , see figure ( [ toric ] , left ) .
the expectation values of these operators in each one of the four degenerate ground states are as follows : @xmath31 @xmath32 @xmath33 @xmath34 therefore these two non - local operators can distinguish the different ground states .
+ finally , excitations of the tc model are quasi - particles with anyonic statistics .
an excitation is generated by applying the pauli operators @xmath8 or @xmath9 on a qubit of the lattice , see figure([toric ] , right ) .
an operator @xmath8 on a qubit does not commute with two plaquette operators which are shared in that qubit and it is interpreted as two flux anyons @xmath35 in two corresponded plaquettes .
also an operator @xmath9 does not commute with two neighbor vertex operators and it is interpreted as two charge anyons @xmath36 in two corresponded vertices . by applying a string of @xmath8 ( @xmath9 ) operators , a flux ( charge ) anyon moves on plaquettes ( vertices ) of the lattice , see figure ( [ toric ] , right ) .
it is simple to show that if a charge anyon winds around a flux anyon , it leads to a minus sign on wave function .
such a factor shows that charge and flux anyons are not fermions or bosons .
the cc model is another well - known kind of the topological lattice models with @xmath37 topological order .
although topological properties of this model seems similar to the tc , additional freedom degree of color in this model leads to some important differences with the tc .
specifically the cc is more efficient than the tc for computational tasks . for examples in the cc on an hexagonal lattice
, all the clifford operators can be applied in a topological way while it is not possible in the tc model @xcite .
+ the cc model can be defined on three - colorable lattices which technically called ` colexes ' and can be generalized to arbitrary dimensions @xcite . as an example we consider a hexagonal lattice where qubits live on vertices of the lattice , see figure ( [ hex ] ) . corresponding to each hexagonal plaquette of the lattice which
is denoted by symbol `` h '' , we define two operators in the following form : @xmath38 where @xmath39 refers to all qubits belonging to the plaquette @xmath40 . the hamiltonian of the model is defined as @xmath41 since all plaquette operators commute with each other , the ground state of this hamiltonian is in the following simple form : @xmath42 ) . to this end , let us span product of operators @xmath43 in the relation ( [ c1 ] ) as the following form : @xmath44 and @xmath45 are attached to each hexagonal plaquette .
edges of the lattice are colored by three different colors corresponding to three different colors of the hexagonal plaquettes .
there are twelve non - trivial loop operators corresponding to two directions on torus and three different colors of the plaquettes .
+ right : corresponding to each hexagonal plaquette of the lattice , there is a triangular plaquette where two qubits live on each edge . in this way
, there are three triangular lattices with three different colors .
since there are two triangular plaquettes with different colors corresponding to each hexagonal plaquette , only two colors are enough for re - presentation of plaquette operators on the triangular lattices .
, width=377,height=226 ] we color all plaquettes of a hexagonal lattice by three different colors , red , blue , green so that any two neighbor plaquettes of the lattice are not in the same color , see figure ( [ hex ] , left ) .
we also color all the edges of the hexagonal lattice with three colors so that each two plaquettes with the same color connect together with an edge with the same color .
each hexagonal plaquette of the hexagonal lattice can also be considered as a triangle plaquette of a triangular lattice where two qubits of each hexagonal plaquette live on each edge of this triangular plaquette , see figure ( [ hex ] , right ) . in this way ,
the edges of the lattice insert in three categories corresponding to each color and we can draw a triangular lattice corresponding to each category of colored edges , see figure([hex ] , right ) .
+ finally , we interpret each plaquette operator of the cc model by a triangular colored loop on one of the three triangular lattices .
by such a interpretation , the summation in the relation ( [ c2 ] ) is regarded as superposition of all possible loop operators with three different colors on the triangular lattices . a closer look to
the hexagonal lattice shows that a plaquette of the initial hexagonal lattice corresponds to two different triangles of two different colored triangular lattices , see figure ( [ hex ] , right ) .
therefore , we have this freedom to select one of the colors for each hexagonal plaquette . by such a freedom ,
it is simple to show that the relation ( [ c2 ] ) can be interpreted as superposition of all loop operators with only two different colors .
therefore , it is well understood that the topological order in the cc is similar to the tc with additional freedom degree of color .
such a order is called @xmath1 topological order versus the @xmath0 topological order for the tc .
+ the topological order in the cc also leads to some important properties similar to the tc such as robust degeneracy , non - local order parameter and anyonic excitations . as a sample ,
if we insert the hexagonal lattice on a torus , there will be many non - contractible loop operators which lead to degeneracy of the ground states .
these loop operators are defined similar to the tc model with a difference that there are three kinds of loop operators corresponding to each color , see figure ( [ hex ] , left ) , in the following form : @xmath46 @xmath47 @xmath48 where @xmath49 refers to the qubits which live on a non - contractible loop around the torus on a red ( green , blue ) triangular lattice in two different directions @xmath27 .
since three above operators are not independent so that the product of them as @xmath50 is equal to a product of plaquette operators , we can generate the sixteen degenerate ground states of the model by applying only two colored loop operators as the following form : @xmath51 where the indices @xmath52
refer to sixteen ground states of the model .
similar to tc model , there are also four non - local order parameters as the following form which can characterize different ground states of the cc model , see figure ( [ hex ] , left ) .
@xmath53 @xmath54 where we apply operators @xmath8 on non - contractible loops with two different colors .
in this section , we consider both the tc and the cc model on the same lattice and define a new hamiltonian as the following form : @xmath55 where @xmath56 and @xmath57 are hamiltonians of the tc and the cc model , respectively .
such a hamiltonian is defined on the qubits which live on the vertices of a hexagonal lattice .
+ definition of the cc on such a lattice is the same as we explained in the previous section .
since qubits live on the vertices of this lattice we can not define an ordinary tc model on such a lattice .
however , there is a simple way to present a tc model on such lattice . to this end
, we divide each hexagonal plaquette to two trapezoid - shaped parts as it is shown in figure([c - t ] ) .
then we paint all trapezoids as chess - pattern with dark and light colors . in this way
, we can use a rotated version of the kitaev model @xcite where we relate an operator @xmath58 to each light plaquette and an operator @xmath59 to each dark plaquette , see figure([c - t ] ) .
it is very simple to show that such a model is exactly the same tc model .
specifically , one can check that the ground state of this model is : @xmath60 are related to light plaquettes and the operators @xmath7 are related to dark plaquettes .
two non - contractible loops which are topologically different in the cc model convert together with product of operators @xmath7 or @xmath6 which are involved by the non - contractible loops .
it show that non - contractible loops with different colors are topologically the same in the tc model . , width=377,height=188 ] such a state is the same as equation ( [ e2 ] ) in the definition of the tc model . +
after definition of the tc and the cc model on the same hexagonal lattice , we are ready to study the properties of the hamiltonian ( [ in1 ] ) . on the one hand
, in limit of @xmath61 , we have a tc model which have a four - fold degeneracy . on the other hand , in limit of @xmath62
, we have a cc model which have a sixteen - fold degeneracy .
hence , by tuning of the coupling @xmath3 from zero to infinity we expect to see a quantum phase transition .
+ since non - contractible loop operators generate degeneracy in the cc and the tc models , it is useful to consider how the subspace of the ground states in the cc model changes by adding the tc model . to this end , consider non - trivial loop operators in the tc and the cc on the hexagonal lattice . as it is shown in figure([c - t ] ) , we consider four non - trivial loop operators @xmath63 , @xmath64 , @xmath65 and @xmath66 for the cc model and two non - trivial loop operators @xmath67 and @xmath68 for the tc model .
it is possible to describe sixteen - fold degenerate subspace of the cc by the projectors which are constructed by non - trivial loop operators as the following form : @xmath69 where @xmath52 .
let us compare above subspace with degenerate subspace of the tc model .
as it is shown in figure ( [ c - t ] ) , four non - trivial operators of the cc model on the hexagonal lattice can also be considered non - trivial operators of the tc model .
but there is a difference that two operators @xmath63 and @xmath65 are in the same homology class in the tc so that they convert to each other by applying product of plaquette operators @xmath7 which are involved by the two non - trivial loops .
the same situation is for operators @xmath64 and @xmath66 where they convert to each other by applying product of plaquette operators @xmath6 which are involved by two non - trivial loops @xmath63 and @xmath65 . in this way , by perturbing the cc model by the tc perturbation , homology classes of the non - trivial loops change so that two non - trivial loops with different colors belong to the same homology class and a quantum phase transition occurs . + it is also interesting if we explain equivalency of non - trivial operators in the tc model in a anyonic picture .
in fact in the cc model two operators @xmath64 and @xmath66 can be interpreted as generating of two charge anyons @xmath70 and @xmath71 which turn around torus and annihilate again . in the cc model , these two charge anyons are not equivalent and they can not fuse together .
consequently they generate new degenerate states while in the tc model , there is only one charge anyon so that @xmath70 and @xmath71 are equivalent and they can fuse together .
+ such an interpretation of the quantum phase transition emphasizes on the role of color in the cc model so that we can claim that the cc model is topologically equivalent with two tc models with two different colors .
this result is explicitly derived in @xcite where authors showed that the cc model is local equivalent with two copies of the tc model . in the next section ,
we show this result by explicit analysis of the quantum phase transition where we show the quantum phase transition has not a topological nature . as another point , it is also useful to emphasize on the role of different gauge symmetries of the tc and the cc model in the quantum phase transition .
in fact , the tc is related to the universality class of the 2d ising model ( classical ) with @xmath0 symmetry , while the cc is related to the universality class of the 2d three - body ising model with @xmath37 symmetry @xcite .
therefore , it is expected that such a quantum phase transition can be characterized as a symmetry breaking process and has not a topological nature . +
as it was explained in the previous section , according to @xcite , because of equivalency of a cc model with two copies of tc model , both models have a universal topological phase . in this section ,
we want to consider properties of the phase transition between the cc and the tc model to understand how the universal topological phase of them reveals in the point of quantum phase transition .
we rewrite hamiltonian eq.([in1 ] ) in a new basis and show that it converts to many copies of 1d ising model in transverse field which belongs to 2d ising universality class .
finally , we conclude that the universal topological phase of those models reflects in 1d nature of the phase transition .
+ consider a set of quantum states as the following form : @xmath72 is the ground state of the cc model .
using @xmath73 in above defenition is necessary to find the effect of operators of the cc model on the state ( [ s1 ] ) .
the values of @xmath74 for indices refer to different eigenstates of this basis .
such a basis has been constructed by the projector operators related to loop operators of the tc model .
it is simple to check that such states generate a complete basis .
in fact , the indices @xmath75 characterize topological degenerate subspace of the tc model and @xmath76 s ( @xmath77 s ) refer to binary variables which correspond to each light ( dark ) plaquette of the lattice , we call them the virtual spins which are equivalent with anyonic excitations of the tc model . in other words , the values @xmath78 and @xmath79 for a virtual spin correspond to presence or absence of an anyon in the related plaquettes . + it is clear that the operators @xmath6 and @xmath7 in this new basis have a simple form . since @xmath80 , we conclude that @xmath81 .
we can also represent the state @xmath82 as @xmath83 where @xmath84 and @xmath85 are eigenstates of operator @xmath9 of the virtual spins which live in the center of each light and dark plaquette of the lattice , respectively .
by such a definition , since @xmath86 , it is clear that the operator @xmath6 plays role of the pauli operator @xmath87 on a virtual spin corresponding to light plaquette @xmath88 .
+ there is similar situation for operators @xmath7 . since @xmath89
, we will have @xmath90 .
consequently , the operator @xmath7 also plays role of the pauli operator @xmath91 on a virtual spin corresponding to dark plaquette @xmath92 .
therefore , the tc model in the new basis is written as the following form : @xmath93 ( @xmath45 ) does not commute with two light ( dark ) plaquette operator @xmath94 ( @xmath95 ) and @xmath96 ( @xmath97 ) and it is equivalent with operator @xmath98 ( @xmath99 ) on virtual spins in the center of the light ( dark ) plaquettes .
similarly , the operator @xmath100 ( @xmath101 ) is equivalent with operator @xmath102 ( @xmath103 ) on virtual spins living in the center of light ( dark ) plaquettes which are denoted by green circles ( red statrs ) .
right : by applying all hexagonal operators in the new basis , we will have @xmath104 copies of 1d ising model corresponding to each row of the lattice.,width=377,height=170 ] it is clear that the operators of the tc model do not ever change indexes @xmath75 in the basis ( [ s1 ] ) .
in other words , we have four degenerate subspaces where there are the same forms for the hamiltonian of the tc model .
interesting point is that there is the same situation in the hamiltonian of the cc model .
it is simple to show that the non - trivial operators @xmath105 and @xmath106 commute with all operators @xmath107 and @xmath45 in the hamiltonian of the cc model .
consequently , the hamiltonian of the cc model has also the same form in the four degenerate states of the tc model so that the hamiltonian of the cc can not generate any transition between those states .
+ by attention to above argument , we consider just one of the degenerate subspaces of the tc model for rewriting hamiltonian ( [ in1 ] ) . finally , we consider following states as new basis : latexmath:[\[\label{s2 } we are ready to find the new form of the cc model in above basis . in the cc model
we have two operators @xmath107 and @xmath45 corresponding to each hexagon of the hexagonal lattice .
we should find the effect of such operators on the states ( [ s2 ] ) .
+ to this end , consider a hexagon @xmath40 of the hexagonal lattice as it has been shown in figure([ising ] ) .
the corresponded hexagonal operator @xmath107 involves six qubits which are common with few dark and light plaquettes of the tc model .
it is clear that the operator @xmath107 commutes with all operators @xmath7 on dark plaquettes and all operators @xmath6 on light plaquettes except of two light - plaquette operators @xmath94 and @xmath96 which have only one joint qubit with the hexagon @xmath40 , see figure ( [ ising ] , left ) .
therefore , we have @xmath109 . since @xmath110 , we conclude that the effect of operator @xmath107 on the state ( [ s2 ] ) leads to rising binary variables @xmath111 and @xmath112 .
such a operation is the same as operator @xmath98 on the basis of virtual spins in ( [ s2 ] ) .
+ the situation is the same for operators @xmath45 .
this operator does not commute with two dark plaquettes @xmath95 and @xmath97 of the tc model , see figure ( [ ising ] , left ) .
therefore , the effect of operator @xmath45 is the same as operator @xmath99 on the basis ( [ s2 ] ) .
+ in the next step , consider another hexagonal plaquette in the neighbor of the plaquette @xmath40 , we denote it by @xmath113 , see figure([ising ] , left ) .
similar to plaquette @xmath40 , operator @xmath100 dose not commute with two light - plaquette operator @xmath96 and @xmath114 so that it is equal to a operator @xmath102 on virtual spins in ( [ s2 ] ) .
if we repeat this work for other plaquettes of the lattice which are in the same row with h and h , we will have an ising model as @xmath115 on virtual spins which live in light - plaquettes of the corresponded row of the lattice , see figure ( [ ising ] , right ) .
the same situation is for operator @xmath116 where it is equal to operator @xmath103 on virtual spins in the dark plaquettes of corresponded row of the lattice , see figure ( [ ising ] , right ) .
+ finally , we can find the effect of operators @xmath107 and @xmath45 associating with all plaquettes of the hexagonal lattice as @xmath117 on the basis ( [ s2 ] ) . as it is shown in figure([ising ] , right ) , the hexagonal lattice insert on a 2d torus where there are @xmath118 rows so that each row of the lattice contains of @xmath118 plaquettes .
if we divide all plaquettes of the lattice into @xmath118 parts which correspond to rows of the lattice , we can write the hamiltonian of the cc model in the following form : @xmath119 where @xmath120 refers to plaquettes which belong to i th row of the lattice .
as it is already discussed , summation of operators @xmath107 ( @xmath45 ) on a row of the lattice in the basis ( [ s2 ] ) is equal to the ising model on virtual spins in light - plaquettes ( dark - plaquettes ) of that row of the lattice .
therefore , the cc model in equation ( [ s4 ] ) is converted to @xmath104 one dimensional ising model on dark and light virtual spins of each row of the lattice in the following form : @xmath121 where @xmath122(@xmath123 ) refers to dark plaquettes ( light plaquettes ) of the row `` r '' . +
after rewriting both the tc model and the cc model in the basis ( [ s2 ] ) according to relations ( [ s3 ] ) and ( [ s5 ] ) , the hamiltonian of the our model ( [ in1 ] ) is converted to the following model in the basis ( [ s2 ] ) : @xmath124 where there are @xmath104 1d ising models in transverse field which have a well - known phase transition point at @xmath125 which belong to the universality class of 2d classical ising model .
+ it is interesting to explicitly consider the ground state in two limits @xmath126 and @xmath127 according to relation([s6 ] ) .
if @xmath126 we will have an ising model on all rows of the lattice and the ground state will be as product of @xmath128 or @xmath129 for each row of the lattice . in the following
we briefly explain how one of the ground states of tc model converts to four ground states of the cc model in effect of the quantum phase transition .
+ consider a state @xmath130 on one of the virtual spins in the light or dark plaquette @xmath131 .
the state @xmath132 in the word of the basis ( [ s2 ] ) is equal to : @xmath133 where @xmath134 refers to all plaquettes expect of the plaquette @xmath131 . on the other hand ,
the state @xmath135 in the word of the basis ( [ s2 ] ) is equal to : @xmath136 in this way , the state @xmath137 will be equal to : @xmath138 therefore , it is simple to show that a state as @xmath128 on whole virtual spins in the lattice which is one of the ground states of the model is equal to : @xmath139 . by attention to relations ( [ w1 ] ) and ( [ w2 ] ) such a state in the basis ( [ s2 ] ) is equal to : @xmath140 this relation shows that if we change a row of the virtual spins on light plaquettes from @xmath141 to @xmath142 it is equal to applying a row of corresponded operators @xmath6 on the state @xmath143 in ( [ w3 ] ) . as it is shown in figure ( [ c - t ] ) , a product of operators @xmath6 on a row of the lattice is equal to product of non - trivial operators @xmath64 and @xmath66 in the cc model . while @xmath64 and @xmath66 are topologically equivalent in the tc model , they are different in the cc model .
therefore , one can easily show that such a operation generates another ground state of the cc model .
+ the same process occurs for the product of @xmath7 on a row of the lattice where it is equal to product of non - trivial operators @xmath63 and @xmath65 in the cc model and generates third ground state .
finally , forth ground state of the cc model is generated by applying a product of @xmath7 and @xmath6 on a row of dark and light plaquttes of the lattice .
therefore , one of the ground states of the tc model splits to four quantum states of the cc model in effect of the quantum phase transition .
+ in this way , we showed that the quantum phase transition between the cc and the tc model is equivalent with the phase transition in 1d ising model in transverse field .
such a result is important in the view point of universal topological phase for the cc and the tc model .
it is well - known that a 1d quantum model has not a topological order @xcite .
in fact for a 1d model , one can not define any topological signature such as wilson loop or topological entanglement entropy .
consequently , our result shows that topological difference of the cc and the tc model has a 1d nature and therefore they are in the same topological phase .
+ the point of phase transition at @xmath125 also shows that the cc model is strongly robust against the tc model .
in fact a tc model with a small coupling @xmath144 can not lead to a quantum phase transition .
while the robustness of the cc against local perturbations such as magnetic field and ising interactions is small @xcite , our result needs to a proper interpretation .
it seems that since the topological phase of the tc model is the same as the cc model , the quantum phase transition in our model only relates to lifting degeneracy of the ground state .
therefore , much robustness of the cc model against the tc model relates to the much robustness of the topological degeneracy of the ground state .
+ in the next subsection
we approximately study the behavior of wilson loop in our model .
we show that the wilson loop does not behave as expected in a topological phase .
in this section , we consider behavior of the wilson loop operators in our model .
it is well - known that behavior of a wilson loop in a perturbed topological phase is as @xmath145 where @xmath146 is the perimeter of the wilson loop .
such a behavior for the wilson loop is a signature of the topological phase @xcite . + we consider a wilson loop operator in the tc model as @xmath147 where @xmath148 is a loop on the chess - patterned hexagonal lattice , see figure ( [ wilson ] ) . such a operator is equal to product of operators @xmath6 which are involved by the wilson loop .
one the one hand , it is clear that the expectation value of such an operator is equal to 1 in the ground state of the tc model . on the other hand , the expectation value of this wilson loop is equal to zero in the ground state of the cc model .
therefore , the expectation value of the wilson loop changes from 1 to 0 when we increase coupling of the cc model from zero to infinity . + which is denoted by red circles .
the wilson loop operators commute with all hexagonal plaquette operators except of the operators which are near of two left and right edges of the wilson loop and are denoted by celtic cross . , width=226,height=226 ] consider the case that @xmath149 , the ground state of our model in such a limit can be found by a perturbation method . to this end
we write a perturbed quantum state as the following form : @xmath150 is the hamiltonian of the cc model and @xmath151 is the normalized ground state of the pure tc model . in the next step
we have to calculate the expectation value of the wilson loop operator into the above state as @xmath152 .
we approximate the ground state ( ( [ ap1 ] ) in first order of @xmath153 as the following form : @xmath154 , we will have : @xmath155 where @xmath156 is the number of hexagonal plaquettes of the lattice . on the other hand , by the fact that @xmath157 we will have : @xmath158 one can easily check that the wilson loop operator @xmath159 commute with all hexagonal plaquettes operators @xmath107 and @xmath45 except of operators @xmath107 which cross two right and left edges of the wilson loop , see figure ( [ wilson ] ) . by this and
the fact that @xmath157 , we conclude that : @xmath160 where @xmath161 refers to the hexagonal operators which commute with the wilson loop and @xmath162 refers to hexagonal operators which cross two right and left edges of the wilson loop .
if we denote the length of two left and right edges of the wilson loop by @xmath163 , we will have : @xmath164 @xmath165 finally the expectation value of the wilson loop operators is calculated as : @xmath166 @xmath167 since @xmath163 is the length of only two edges of the wilson loop , the result shows that the expectation value of the wilson loop does not behave as @xmath168 that we expect for a topological phase .
in fact all wilson loops with different perimeters but the same @xmath163 close to zero with a similar rate by the cc perturbation .
such a behavior shows that in effect of quantum phase transition from the tc to the cc models the topological phase does not change .
understanding how quantum phase transitions between different quantum models can characterize different topological classes seems a challenging task which needs to more exploration in future . in this paper , by a specific example , we considered how topological classes of different quantum phases reflect in the properties of a quantum phase transition between them . to this end , we considered a quantum phase transition between two well - known topological models , the cc and the tc model .
we showed that a interpolation of the cc model and the tc model is mapped to 1d quantum ising model in a transverse field .
the 1d nature of the phase transition revealed the same topological nature of the cc model and the tc model .
in fact , since different quantum phases of both models was related to their different gauge symmetries , it made the possible interpolating line joining them a non - topological flow . to be more explicit
, we also calculated the expectation value of the wilson loop .
the result showed that behavior of the wilson loop captures the same topological phase of the cc and the tc models .
finally , since 1d ising model in transverse field is a well - known model with an exact solution , we could excactly find the point of the phase transition .
the result showed that the cc model is strongly robust against the tc model .
99 l. d. landau , phys .
z. sowjetunion.11 , 26 ( 1937 ) , goldenfeld n. , `` lectures on phase transitions and criti- cal phenomena '' , ( perseus books publishing , l.l.c .
, mas- sachusetts , 1992 ) .
tsui , h.l .
stormer , and a.c .
lett , 48 1559 ( 1982 ) n. read and b. chakraborty , phys . rev .
b 40 , 7133 ( 1989 ) .
r. moessner and s. l. sondhi , phys .
86 , 1881 ( 2001 ) .
e. ardonne , p. fendley , and e. fradkin , ann .
310 , 493 ( 2004 ) .
d. s. rokhsar and s. a. kivelson , phys .
61 , 2376 ( 1988 ) x .- g .
wen , phys . rev .
b 40 , 7387 ( 1989 ) .
b 5 , 1641 ( 1991 ) . t. h. hansson , v. oganesyan , and s. l. sondhi , ann.phys .
313 , 497 ( 2004 ) .
wen , adv .
phys.44 , 405 ( 1995 ) .
chen , z .- c .
gu , and x .-
wen , phys .
b 82 , 155138 ( 2010 ) .
a. kitaev and j. preskill , phys .
96 , 110404 ( 2006 ) .
m. levin and x .- g .
wen , phys .
96 , 110405 ( 2006 ) .
a. hamma , w. zhang , s. haas , and d. a. lidar , phys .
b 77 , 155111 ( 2008 ) .
xiao - gang wen , isrn condensed matter physics , vol .
2013 , ( 2013 ) .
x. g. wen and q. niu , phys . rev .
b 41 , 9377 ( 1990 ) .
wen , quantum field theory of many - body systems : from the origin of sound to an origin of light and elec- trons ( oxford univ .
press , new york , 2004 ) j. m. leinaas and j. myrheim , nuovo cimento soc .
b 37 , 1 ( 1977 ) .
f. wilczek , phys .
48 , 1144 ( 1982 ) .
e. dennis , a. kitaev , a. landahl , j. preskill , j. math .
43 , 4452 - 4505 ( 2002 ) c. nayak , s. h. simon , a. stern , m. freedman , and s. das sarma rev .
mod . phys .
80 , 1083 , ( 2008 ) a. y. kitaev , ann .
( n. y. ) 303 , 2 ( 2003 ) .
v. kalmeyer and r. b. laughlin , phys .
59 , 2095 ( 1987 ) .
wen , f. wilczek , and a. zee , phys . rev .
b 39 , 11413 ( 1989 ) n. read and s. sachdev , phys .
66 , 1773 ( 1991 ) .
t. senthil and m. p. a. fisher , phys .
b 62 , 7850 ( 2000 ) .
s. sachdev and k. park , ann .
298 , 58 ( 2002 ) .
l. balents , m. p. a. fisher , and s. m. girvin , phys .
b 65 ( 2002 ) . s. c. morampudi , c. von keyserlingk , f. pollmann , phys .
b 90 , 035117 ( 2014 ) v. karimipour , l. memarzadeh , p. zarkeshia n , phys .
a 87 , 032322 ( 2013 ) .
xie chen , zheng - cheng gu , xiao - gang wen , phys .
b 83 , 035107 ( 2011 ) w. son , l. amico , r. fazio , a. hamma , s. pascazio , v. vedral , europhys .
95 , 50001 ( 2011 ) v. lahtinen , e. ardonne , phys .
115 , 237203 ( 2015 ) h. bombin , m.a .
martin - delgado , phys.rev.lett .
97 180501 ( 2006 ) d. nigg , m. mller , e. a. martinez , p. schindler , m. hennrich , t. monz , m. a. martin - delgado , r. blatt , science vol .
345 , issue 6194 , pp .
302 - 305 ( 2014 ) h. bombin , g. duclos - cianci , d. poulin , new j. phys . 14 073048 ( 2012 ) a. kubica , b. yoshida and f. pastawski , new j. phys .
17 083026 ( 2015 ) p. pfeuty , ann .
57 , 79 ( 1970 ) s. dusuel , m. kamfor , r. orus , k. p. schmidt , j. vidal , phys .
rev . lett .
106 , 107203 ( 2011 ) m. h. zarei , phys .
a. 91 , 022319 ( 2015 ) s.s .
jahromi , m. kargarian , s farhad masoudi , and k.p .
s chmidt , phys .
b 87 , 094413 ( 2013 ) .
a. langari , a. mohammad - aghaei , and r. haghshenas , phys . rev .
b 91 , 024415 ( 2015 ) saeed s. jahromi , s farhad masoudi , mehdi kargarian , kai phillip schmidt , phys . rev .
b 88 , 214411 ( 2013 ) h. bombin and m. a. martin - delgado , phys .
a 76 , 012305(2007 ) h. bombin and m. a. martin - delgado , phys .
b 75 , 075103 ( 2007 ) helmut g. katzgraber , h. bombin , and m. a. martin - delgado , phys . rev .
103 , 090501 | although the topological order is known as a quantum order in quantum many - body systems , it seems that there is not a one - to - one correspondence between topological phases and quantum phases . as a well - known example
, it has been shown that all one - dimensional ( 1d ) quantum phases are topologically trivial@xcite . by such facts
, it seems a challenging task to understand when a quantum phase transition between different topological models necessarily reveals different topological classes of them . in this paper
, we make an attempt to consider this problem by studying a phase transition between two different quantum phases which have a universal topological phase .
we define a hamiltonian as interpolation of the toric code model with @xmath0 topological order and the color code model with @xmath1 topological order on a hexagonal lattice .
we show such a model is exactly mapped to many copies of 1d quantum ising model in transverse field by rewriting the hamiltonian in a new complete basis .
consequently , we show that the universal topological phase of the color code model and the toric code model reflects in the 1d nature of the phase transition .
we also consider the expectation value of wilson loops by a perturbative calculation and show that behavior of the wilson loop captures the non - topological nature of the quantum phase transition .
the result on the point of phase transition also show that the color code model is strongly robust against the toric code model . * a quantum phase transition from @xmath1 to @xmath0 topological order * + mohammad hossein zarei + physics department , college of sciences , shiraz university , shiraz 71454 , iran |
Subsets and Splits