article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
recent progress of controlled femtosecond pulses has advanced greatly the technology during the last few years @xcite .
a new field of molecular optics has emerged where lasers are used to manipulate the internal and external degrees of freedom of molecules , to deflect beams of molecules , to control molecular dynamics , and to align molecules @xcite .
many investigations have focused on the vibrational motion of diatomic molecules .
the single bond between the atoms acts as a spring and supports harmonic oscillations for small amplitudes , but the bond can break ( dissociate ) when stretched too much
. these phenomena usually occur at time scales between few picoseconds and few hundred femtoseconds . with ultrashort pulses one can now prepare a molecular wave packet and
probe its evolution and observe molecular reactions in this time domain .
successful experiments have been performed on several molecules @xcite .
the most convenient model for studying the vibrational motion of diatomic molecule is the morse potential , which is an exactly solvable system @xcite .
coherent superposition of several vibrational levels of the molecule creates the wave packet which , due to quantum interference , shows revival and fractional revivals @xcite in their time evolution .
fractional revivals are associated with superpositions of separated wave packets ( for example , the so - called schrdinger cat states ) , which manifest clear quantum interference effects and nonclassical features , which can be well visualized in the phase space of the vibrational motion .
although the experimental observation of small quantum interference structures is very challenging , it has already been visualized in pico - meter scale @xcite .
a number of different phase space distribution functions have been introduced @xcite and investigated over the years , and among these the wigner distribution @xcite is particularly useful , because its negativity yields an indication of nonclassical behavior @xcite .
zurek @xcite first showed that this negativity reveals the existence of the smallest structures in phase space i.e. , the sub - planck scale structures .
one may expect that heisenberg s uncertainty principle implies that structures on scales smaller than the planck constant have no observable consequence , while instead zurek @xcite showed that these highly nonclassical structures are expected to be particularly sensitive to decoherence . through a short walk in controversy ,
recently sub - planck scale structures draw considerable attention and have been found by others in different situations @xcite .
decoherence due to the coupling to an external environment is the main responsible for the disappearance of nonclassical manifestations of quantum states and it is considered one of the mechanisms through which the classical world at the macroscopic level emerges from the quantum substrate @xcite .
decoherence on the molecular vibrational degree of freedom is due to the coupling between vibrational and rotational modes @xcite and also to the coupling with the photonic and phononic degrees of freedom @xcite .
the latter are associated with a super - ohmic environment describable in terms of a continuous set of bosonic modes and in this paper we shall focus on the effect of this source of decoherence . to be more specific , we will study the sub - planck scale structures in the morse system and the effect of decoherence on these structures in molecular wave packets .
we shall determine the master equation describing the reduced dynamics of the wave packet and analyze the robustness of the sub - planck structures against decoherence .
the paper is organized as follows . in sec .
ii we give a brief overview of the morse potential and its coherent states , while in sec .
iii we derive the master equation in the case of the coupling with a bosonic environment at thermal equilibrium . in sec .
iv we study the effect of decoherence on the wigner function at the sub - planck level and the sensitivity to decoherence of these structures is analyzed .
finally , we conclude in sec . v.
vibrational dynamics of diatomic molecules are well described by morse potential @xcite .
it can be described as @xmath0 where @xmath1 is the equilibrium value of the inter - nuclear distance @xmath2 and @xmath3 is a range parameter .
@xmath4 is the dissociation energy , which has been extensively studied in a wider context of this model @xcite . defining @xmath5 where @xmath6 is the reduced mass of the vibrational motion ,
the eigenfunctions of the morse potential can be written as @xmath7 where @xmath8 , @xmath9 , and @xmath10 $ ] , with @xmath11 $ ] denoting the integer part of @xmath12 , so that the total number of bound states is @xmath13 $ ] . the parameters @xmath14 and @xmath15 satisfy the constraint condition @xmath16 .
note that @xmath14 is potential dependent and @xmath15 is related to energy @xmath17 and , by definition , @xmath18 . in eq .
( [ eigenstate ] ) , @xmath19 is the associated laguerre polynomial and @xmath20 is the normalization constant : @xmath21^{1/2}.\ ] ] revival and fractional revivals appear during the time evolution of a suitably prepared wave packet and are well studied in the literature @xcite . here we study the effect of decoherence on the motion of a molecular wave packet through its sub - planck scale structures .
there structures are found at one eighth of the fractional revival time in the wigner phase space distribution @xcite .
the initial wave packet is taken here as su@xmath22 coherent state ( cs ) of this potential @xcite , which is obtained upon applying the displacement operator on the ground state .
the cs is given by @xmath23^{1/2 } \eta^{k}|k , s\rangle .
\label{cs}\end{aligned}\ ] ] where @xmath24 $ ] correspond to the bound states of the morse potential and @xmath25 $ ] are the appropriate scattering states .
the parameter @xmath26 is associated with the `` amplitude '' of the cs and possesses the same phase of the displacement amplitude @xmath27 , while its modulus is given by the relation @xmath28 @xcite . in our numerical analysis
, we will always consider low energy coherent states well below the dissociation limit so that only the bound states of the morse potential can be used as basis set .
as described in sec . i , we now investigate the effect of the decoherence of an external phononic or photonic environment on the sub - planck scale structure .
therefore , the total model hamiltonian is @xcite @xmath29 where @xmath30 is the morse hamiltonian of the vibrational mode , @xmath31 is the environment hamiltonian described by a set of independent bosonic modes @xmath32 and @xmath33 is the interaction between the morse particle and the environment , which we choose of the following form ( see also @xcite ) @xmath34 where @xmath35 are coupling constants . this choice corresponds to assume the rotating wave approximation ( rwa ) in the interaction with the environment so that we neglect counter - rotating terms , while the operator @xmath36 is a generic operator of the vibrational mode , whose specific form depends on the considered environment . using standard techniques @xcite ,
one gets in the usual born - markovian approximation , the following master equation for the reduced density operator of the morse oscillator @xmath12 , @xmath37 \label{meq}\\ & + & \left[\hat{o}_{2 } \rho \hat{o}^{\dagger}+\hat{o } \rho \hat{o}_{2}^{\dagger } - \hat{o}^{\dagger } \hat{o}_{2}\rho -\rho \hat{o}_{2}^{\dagger}\hat{o}\right ] \nonumber \\ & + & \left[\hat{o}_{1}^{\dagger}\rho \hat{o}+\hat{o}^{\dagger } \rho \hat{o}_{1 } - \hat{o}\hat{o}_{1}^{\dagger } \rho -\rho \hat{o}_{1}\hat{o}^{\dagger}\right ] , \nonumber\end{aligned}\ ] ] where the operators @xmath38 @xmath39 are new operators of the vibrational mode corresponding to `` modifications '' of the operator associated to the absorption from the environment ( @xmath40 ) or emission into the environment ( @xmath41 ) of vibrational quanta .
this fact is easily understood if we look at their expression in the energy eigenbasis @xmath42 used in eq .
( [ cs ] ) .
in fact , one has @xmath43 where @xmath44 . \label{elemat2}\end{aligned}\ ] ] the quantities @xmath45 are the matrix elements of @xmath36 , @xmath46 is the density of states at the transition frequency between two energy levels , @xmath47 , and @xmath48^{-1}$ ] is the mean thermal number of environmental excitations , being the latter at equilibrium at temperature @xmath49 .
the appearance of these two operators is a direct consequence of the nonlinearity of the molecular vibrational motion .
in fact , in the linear case the transition frequencies @xmath50 do not depend on @xmath51 and @xmath52 and therefore @xmath40 and @xmath41 becomes proportional to @xmath36 . as a consequence ,
master equation ( [ meq ] ) become identical to the master equation of a harmonic oscillator in a thermal environment in the rwa @xcite .
we now solve the master equation ( [ meq ] ) for the specific case of the hi molecule and we adopt the wigner function picture in order to look at the effects of decoherence on sub - planck scale structures in phase space .
the wigner distribution is defined ( @xmath53 ) by @xmath54 and well describes the nonclassical interference effects associated with the time evolution of a wave packet in the nonlinear potential of the morse oscillator .
this fact is visible , for example , in fig .
[ catsub ] , which shows the time evolution of an initial cs wave packet in phase space at two different fractional revival times in the absence of decoherence .
we have considered a hi molecule , which has @xmath55 bound states , with @xmath56 , reduced mass @xmath57 a.u .
, @xmath58 a.u . , and @xmath59 a.u .
we have assumed here ( and also in the following ) that the initial wave packet is well below the dissociation limit , so that it involves only the lower levels of hi molecule ( the energy distribution is peaked around the @xmath60 vibrational level ) .
figure [ catsub](a ) shows the vibrational cat state after one fourth of the fractional revival time . here , the revival time is @xmath61 a.u . due to the anharmonicity of the system
, one can notice the different squeezing effects in the two separated css forming the cat state .
the number of ripples in the interference region increases for increasing mean energy of the initial cs .
the sub - planck scale structures appear in the interference region at one eighth of the fractional revival time [ fig .
[ catsub](b ) ] , where one has a coherent superposition of four well distinct states , forming a so - called compass state @xcite . for this reason we shall focus our attention on the effect of decoherence at this fractional revival time . in the case of a molecular vibration ,
a bosonic environment well describes either the coupling via the dipole interaction with the outside electromagnetic field or , in the case of a molecule immersed in a liquid or gas , the coupling with the acoustic modes of the solvent . in both cases
the operator @xmath36 is connected with the position operator of our morse oscillator .
in fact , @xmath36 describes the upper triangular part ( in the energy basis representation ) of the dipole moment operator of the molecule in the electromagnetic case and of the vibrational coordinate @xmath62 in the acoustic phonon bath case .
however the two situations are analogous because the dipole moment operator is proportional to @xmath62 .
both environments are super - ohmic , that is , we have @xmath63 with @xmath64 characterizing the strength of the system - environment coupling .
the physical meaning of the parameter @xmath64 can be seen from the fact that the master equation ( [ meq ] ) implies that the relaxation rate from level @xmath65 to level @xmath66 , @xmath67 at zero temperature is given by @xmath68 where @xmath69 is the corresponding matrix element of the position operator between the two vibrational levels .
here we have chosen the coupling constant @xmath64 such that the ratio @xmath70 ranges from @xmath71 to @xmath72 .
these values correspond to reasonable values of the coupling constant @xmath64 ; in fact , considering a typical electric dipole moment of a diatomic molecule one gets @xmath73
m@xmath74 s@xmath75 @xmath76 for the case of an electromagnetic environment .
instead , considering a dilute solvent one gets @xmath77 m@xmath74 s@xmath75 @xmath78 a.u . for the case of a phononic environment .
one should note that position variable in our study is a dimensionless quantity ( scaled by @xmath79 ) .
hence , in the case of an electromagnetic environment , the order of @xmath64 in our case would be @xmath80 a.u . ,
which is consistent with our study ( see fig . [ peaks ] ) . with the coupling parameter @xmath64 ( in unit of @xmath81 a.u .
the environment temperature is @xmath82.,width=307 ] moreover , the temperature of the environment is kept fixed at @xmath83 .
figure [ 3ddeco ] shows the wigner distribution at one eighth of the fractional revival time for different values of the coupling with the bosonic environment .
figure [ 3ddeco](a ) refers to no decoherence @xmath84 ) and therefore corresponds to fig .
[ catsub](b ) . figure [ 3ddeco](b )
instead corresponds to @xmath85 a.u . and fig .
[ 3ddeco](c ) corresponds to a stronger decoherence , @xmath86 a.u .
one can clearly see that by increasing the coupling with the bosonic environment , the interference region is more and more affected . with the coupling parameter @xmath64 ( in unit of @xmath81 a.u . ) .
dots are the numerical data from our analysis .
it satisfies an exponential law ( solid line ) @xmath87 , with @xmath88 and @xmath89 .
the environment temperature is @xmath90.,width=307 ] as for the harmonic oscillator case @xcite , decoherence affects the structure as a whole ; also here for the morse oscillator , the sub - planck structures due to quantum interference are more affected than the individual isolated coherent state components .
in fact , a distinct difference can be observed in the decay rate of the amplitude of the sub - planck scale structures and of the individual css .
this is quantitatively shown in fig . [ peaks ] , where these decay rates are plotted versus the decoherence strength @xmath64 .
we consider the left and right peaks of the css at @xmath91 and a negative peak at @xmath92 and @xmath93 , appearing in the sub - planck interference region .
the plot shows that the sub - planck scale structure , i.e. , the central interference patterns ( dashed line in fig . [ peaks ] ) , disappears earlier compared to the individual css , as it happens in the harmonic case .
it is possible to see that the decay of the amplitude of the sub - planck structure follows very well an exponential law as a function of the decoherence strength @xmath64 , as expected in usual bosonic environments @xcite .
we find that a linear exponential function @xmath87 well fits with our results , with @xmath88 and @xmath89 .
figure [ exponential ] shows how the rate of amplitude damping of the chosen negative interference region ( i.e. , sub - planck region ) matches with the exponential form .
it is now worth seeing the effect of environment temperature on decoherence for a fixed value of the coupling constant @xmath64 .
owing to eqs .
( [ elemat1 ] ) and ( [ elemat2 ] ) , one expects a bose - einstein dependence on temperature of the decay of the interference structures associated with sub - planck structures , @xmath94\}$ ] , where @xmath95 corresponds to an effective transition temperature below which the discrete structure of the energy levels of the morse oscillator starts to manifests itself .
this is confirmed by fig .
[ temperature ] , where the numerical results for the value of the negative peak are plotted versus temperature . the data , corresponding to @xmath85 a.u .
, are well fitted by the above curve and the optimal fitting parameters are @xmath96 , @xmath97 and @xmath98 .
the data follow an exponential decay for @xmath99 , while the deviation from the exponential law ( associated with the bose - einstein distribution dependence ) is clearly visible only at very low temperatures , @xmath100 , in the magnified view in the inset of fig .
[ temperature ] .
so far we have been studying the decoherence effect on the sub - planck scale structures at @xmath101 fractional revival time .
hence , it is a natural question to ask what happens at larger times when one can also obtain sub - planck scale structures in the interference region of four - way break up of a coherent state .
thus , we extend our study to the four way break - up or the decoherence through sub - planck scale regions at @xmath102 and @xmath103 fractional revival times .
one expects a larger influence of decoherence on the sub - planck structures for increasing times and this is confirmed by fig . [ 3eight5 ] .
interference fringes in phase space are still visible at @xmath102 fractional revival time , while in fig .
[ 3eight5](b ) , corresponding to @xmath103 fractional revival time , one can see that the sub - planck structures completely disappear due to the larger decoherence effect , whereas the individual coherent states remain almost intact .
the environment temperature is kept constant at @xmath104 , the coupling constant being @xmath105 a.u .
we have investigated the time evolution of a coherent state wave packet in the morse potential under the influence of a bosonic environment describing either photonic or phononic excitations .
we have studied the effect of decoherence on the sub - planck structures in phase space by looking at the evolution of the wigner distribution . as it happens for the harmonic case ,
sub - planck scale structures come out as the most sensitive to decoherence .
a quantitative analysis provides an exponential decay of the amplitude of the quantum interference structures as a function of the coupling with the environment , in agreement with usual predictions @xcite .
influence of the environment temperature on the decoherence is also shown quantitatively .
this is according to the bose - distribution law .
longer time effect on the decoherence is shown for providing another way to see the sensitiveness of sub - planck scale structures compare to their original counterparts .
e. a. shapiro , m. spanner , and m. y. ivanov , phys .
rev . lett . * 91 * , 237901 ( 2003 ) ; k. f. lee , d. m. villeneuve , p. b. corkum , and e. a. shapiro , phys . rev
93 * , 233601 ( 2004 ) ; d. babikov , j. chem
. phys . * 121 * , 7577 ( 2004 ) ; w. schleich and j. a. wheeler , nature ( london ) * 326 * , 574 ( 1987 ) ; w. schleich and j. a. wheeler , j. opt . soc . am .
b * 4 * , 1715 ( 1987 ) ; w. schleich , d. f. walls , and j. a. wheeler , phys .
a * 38 * , 1177 ( 1988 ) ; w. p. schleich , _ quantum optics in phase space _
( wiley - vch , berlin , 2001 ) , and references therein . | we study the effect of decoherence on the sub - planck scale structures of the vibrational wave packet of a molecule .
the time evolution of these wave packets is investigated under the influence of a photonic or phononic environment .
we determine the master equation describing the reduced dynamics of the wave - packet and analyze the sensitivity of the sub - planck structures against decoherence in the case of a hydrogen iodide ( hi ) molecule . |
in the last decade surface enhanced raman scattering ( sers)@xcite has made rapid progress towards applications . with a sensitivity rivaling fluorescence in some cases , and a much higher structural specificity , sers is a highly attractive technique , being developed simultaneously with the field of plasmonics .
progress towards different uses of sers in practical applications has been steady .
new substrates including : arrays of inverted pyramids@xcite , silver pillar and torroid arrays@xcite , adaptive silver films@xcite , and metallic nano - shells@xcite , have been demonstrated .
in addition , many molecules relevant to a myriad of applications such as glucose@xcite , proteins@xcite , dna@xcite , a wide range of medicinal drugs@xcite , and substances for forensic science@xcite , have been characterized . on the other hand ,
the understanding of some fundamental aspects of the phenomenon is still incomplete , and in some cases controversial .
although a full understanding is sometimes not essential for the development of applications , there can be no denying the fact that better comprehension is desirable .
optical pumping of vibrational modes was first suggested in 1996 based on the observed dependence of the anti - stokes / stokes ( as / s ) ratio with incident laser power@xcite , and it was suggested that this could be a possible tool to estimate sers cross sections . however , the interpretation of the experimental results have been the subject of considerable debate in the literature , with many authors simply denying its existence and attributing the experimental observations to either laser - heating , resonance effects , or combinations thereof@xcite . over a series of previous papers
we have suggested the investigation of the as / s - ratio as a function of temperature @xmath0 as an alternative method for the observation of pumping@xcite . in a recent paper , we provided definitive evidence for the existence of vibrational pumping under sers conditions thereby bringing to conclusion one aspect of this complex problem@xcite .
here we address another aspect , namely the problems associated with estimating the sers cross section using this method , and we extend our analysis to include different analytes and laser excitations .
we also highlight several outstanding issues which were not considered previously@xcite .
obtaining good estimates of sers cross sections is in general an extremely hard problem and has been a longstanding ambition of the sers community for reasons that go from the purely applied ( quantification of signals ) to the more fundamental ( comparisons of theoretical electromagnetic enhancement factors with experiment ) .
any method that can produce a standard protocol for the estimation of sers cross sections is , accordingly , of great interest and an effort to understand its principles and limitations is required .
before we go into the details , we briefly review the necessary concepts for understanding as / s - ratios and vibrational pumping in sers .
the method we are going to present has not been widely used and therefore we review the historical background and also its peculiarities and limitations .
the discussion has some natural overlap with our previous papers@xcite but it is presented here for the sake of completeness , the convenience of the reader , and future reference for forthcoming work in progress .
this will inevitably result in a somewhat lengthy introductory section where the principles of the method are laid down .
the next sections are fully devoted to this and are followed by a section at the end with a practical experimental demonstration of sers cross section determination for several standard probes .
let us consider the different contributions to the population of a single vibrational level at temperature @xmath1 during a sers measurement ; we can identify two main contributions : @xmath2 the laser itself which pumps vibrations through raman processes with a rate proportional to its intensity @xmath3 and to the raman - stokes cross section @xmath4 , and @xmath5 thermal excitation and relaxation .
vibrations remain in the level with a finite lifetime @xmath6 , which encompasses all possible relaxation mechanisms such as intramolecular vibrational relaxation ivr ( anharmonic processes ) or external relaxation mechanisms .
there are other possible secondary mechanisms , such as relaxation through an anti - stokes raman processes , or excitation to higher vibrational levels .
it is possible to write rate equations for the detailed dynamics of vibration populations in such a system @xcite . in the regime of weak pumping ( which is the only case considered here ) , where the vibrational population remains small , @xmath7 , it is sufficient to consider only the two main mechanisms , and the rate equation for @xmath8
can then be written as : @xmath9 where @xmath10 is the raman - stokes cross section , @xmath11 the intensity ( power per surface area ) of the laser , @xmath12 the energy of an exciting photon ( @xmath13 is the number of incident photon per unit time and surface area ) , and @xmath14 the energy of the vibration .
the first term on the right is the number of vibrations per unit time being pumped into the level by the action of the laser , while the second and third terms are the contributions of thermal excitation and population relaxation , respectively . in the steady state @xmath15 .
generally , @xmath10 is very small , so when @xmath11 is small the pumping contribution is negligible and the vibrational population is dominated by thermal effects ; i.e. @xmath16 is given by a boltzmann factor .
when @xmath11 is increased ( so that there is a significant pumping contribution to the population of the level ) @xmath8 becomes : @xmath17 where the pumping term is clearly distinguished from the thermal contribution .
in addition , for an ensemble of @xmath18 molecules , the stokes - raman signal is given by @xmath19 while the anti - stokes signal is given by @xmath20 , leading to : @xmath21 moreover , taking the ratio of the anti - stokes to the stokes intensities @xmath22 , we have : @xmath23 where @xmath24 is the _ asymmetry factor _ ( extensively discussed in refs .
@xmath24 includes not only any possible difference between anti - stokes and stokes cross sections arising from resonance effects ( due either to plasmon resonances or to resonant raman scattering for a resonant analyte ) , but also the standard wavelength dependence of raman processes @xmath25 .
@xmath26 can be expressed as : @xmath27 .
\label{pumpratio}\ ] ] this simple model shows that the stokes - raman signal always remains linearly dependent on @xmath11 .
the anti - stokes signal also shows a linear dependence with power when pumping is negligible .
but when pumping dominates over thermal effects , then @xmath28 varies quadratically with power while the as / s - ratio is linearly dependent on @xmath11 .
the original work by kneipp et .
al.@xcite used rhodamine 6 g ( rh6 g ) and crystal violet ( cv ) ( two commonly used sers active dyes ) under @xmath29 excitation at room temperature ( rt ) . the argument for pumping
was primarily based on two observations : ( i ) the as / s - ratios were shown to be larger than expected for a boltzmann factor , and ( ii ) the _ power dependence _ of the signal at rt ( where the thermal contribution is likely to dominate , or be important ) was quadratic .
this observation of a non - linear dependence of the anti - stokes intensities with power ( resulting in a linear dependence in the as / s - ratio ) was thought to provide strong evidence for vibrational pumping in sers .
further measurements on carbon nanotubes@xcite and a dna - base@xcite supported these initial results .
the arguments against this original interpretation are summarized in the following subsection .
haslett et .
al.@xcite were the first to seriously question the existence of vibrational pumping as revealed in the original studies .
extensive measurements were made under similar conditions with a number of both resonant and non - resonant molecules .
they observed an anomalous ratio for all the resonant molecules tested which was independent of power until photo - bleaching occurred .
no anomaly was observed in the case of the non - resonant molecules .
it was concluded that the anomalous ratios observed in the original papers were the result of `` hidden '' resonances rather than pumping@xcite .
al.@xcite also concluded that the observed anomalous ratio could be explained by resonances .
these resonances are accounted for by the asymmetry factor @xmath24 in the model of the previous section . in the absence of pumping ,
the as / s - ratio is predicted to be : @xmath30 it is clear that the @xmath24 factor can indeed result in anomalous ratios ( different from the boltzmann factor ) , even in the absence of pumping .
furthermore , it was suggested that the observed power dependence was not the effect of pumping but rather laser heating .
the effect of heating on @xmath26 in the absence of pumping can be simply understood by including it in eq .
( [ ratio - corrected ] ) .
the real temperature of the probed molecule , @xmath31 , may be different to the nominal temperature @xmath1 because of heating effects , either in the sers substrate or in the molecule itself @xcite . by writing @xmath32 and assuming that @xmath33 we can obtain the corresponding ratio , @xmath34 , by expanding eq .
[ ratio - corrected ] as : @xmath35 this expression first shows that heating can also result in an anomalous as / s ratio , even when @xmath36 .
moreover , it can have an important impact on the power dependence of @xmath26 , because @xmath37 should increase with @xmath11 .
we can assume in a first approximation that @xmath38 , since heat diffusion and transfer are linear problems .
the above expression shows that laser heating should then result in an exponential increase of @xmath34 with @xmath11 .
however , in many cases of interest , the argument in the second exponential is small compared to 1 , and eq .
( [ ratio - x ] ) can then be further expanded to give : @xmath39 a linear dependence of @xmath26 with @xmath11 ( or equivalently a quadratic dependence of @xmath28 with @xmath11 ) _ can then equally be the result of conventional heating effects or pumping_. such an observation is therefore insufficient to demonstrate the presence or not of vibrational pumping .
we also note that heating effects can also explain the mode - dependent behavior observed in the original work @xcite , as the argument of the second exponential in eq .
( [ ratio - x ] ) depends on the mode energy @xmath14 .
most studies have concentrated on finding evidence for sers pumping at room temperature ( rt ) , in particular by studying the power dependence of the as / s ratio .
it is true that the huge enhancements in sers conditions greatly increases the contribution from pumping , but may also contribute to an increased heating ( in particular directly in the probe molecules ) . at room temperature ( rt ) and above , the thermal contribution to @xmath26 is relatively large and in many cases dominates , such that @xmath40 .
the study of vibrational pumping at rt therefore involves measuring small departures from an already existing ( large ) thermal population .
moreover , as the previous discussion has shown it is also extremely difficult to distinguish the relative contributions of pumping and heating at rt .
a more practical approach is to study the temperature dependence of the as / s ratios , and in particular the low - temperature regime where thermal effects are expected to be completely negligible compared to any other mechanisms , and in particular pumping .
the temperature dependence can be directly studied using the model presented so far .
figure [ schmratios ] summarizes the different scenarios for as / s - ratios as a function of temperature .
the solid lines show the variation of the ratio with @xmath1 when pumping is absent ; i.e. an exponential decrease as @xmath1 decreases , following the boltzmann factor @xmath41 .
the dashed lines show the case when pumping is present . at high @xmath1 s
the ratio is approximately the same as if pumping were absent .
this is the _ thermally - dominated regime_. as @xmath1 is decreased , there is a cross - over to a _ pumping - dominated regime _ , where the ratio reaches a plateau .
we refer to the cross - over temperature between these two regimes as t@xmath42 ; indicated by arrows in fig .
[ schmratios ] .
the cross - over occurs when pumping and thermal terms in eq .
( [ pumpratio ] ) are comparable , which occurs for : @xmath43^{-1}\ ] ] the cross - over occurs at a larger temperature for higher powers @xmath11 , for which pumping is stronger , but also for higher energy peaks ( with a larger @xmath14 ) , for which thermal excitation is weaker . in the pumping - dominated regime ,
the as / s - ratio is constant and equal to @xmath44 .
@xmath45 can be estimated to a high degree of accuracy for a given experimental setup .
if @xmath46 ( no asymmetry between @xmath47 and @xmath48 ) we can therefore deduce the product @xmath49 from the plateau in the as / s - ratio below @xmath50 .
in general , and in particular under sers conditions , @xmath51 , but its value can be determined from the exponential dependence in the thermally - dominated regime ( above @xmath50 ) . in practice , a fit of the experimental data over the whole temperature range with two parameters enables us to determine both @xmath24 and @xmath52 @xcite . nevertheless , the problem of estimating @xmath6 still needs to be circumvented to determine the sers cross - section @xmath10 itself . by reducing @xmath1 , the contribution from pumping becomes dominant , making measurements much easier , and more reliable .
it also simplifies greatly their interpretation .
but the main advantage is that it enables one to unambiguously rule out heating effects as an alternative explanation . as discussed , the appearance of plateaus for @xmath53 in fig .
[ schmratios ] is a clear characterization of the pumping - dominated regime .
note that it does not mean that heating is absent , but simply that its effect on as / s - ratios is negligible ( relative to the pumping contribution ) .
any heating effects may possibly affect the value of @xmath50 , but as long as a plateau is observed , vibrational pumping must occur .
moreover , the value of @xmath26 in the plateau region ( from which we will infer sers cross - sections ) is independent of heating .
studies of the temperature dependence of @xmath26 therefore provide a more conclusive demonstration of sers pumping .
however , the technique is not without its drawbacks . in order to extract sers cross - sections for the method discussed here
there are several issues that need special attention as described below .
several issues need to be considered in the practical determination of sers cross sections via vibrational pumping .
not all of them can be resolved satisfactorily and we have to include some _ ad - hoc _ approximations in order to obtain estimates of the cross sections .
while there is nothing intrinsically wrong with this , it is necessary to be aware of the range of validity of the estimates in order to be able to comprehend the current limitations of the method as proposed .
it is worth mentioning here that in most situations in raman ( and in sers ) we talk plainly about the cross section @xmath54 while in reality there are at least three possible cross sections we can refer to : @xmath2 radiative , @xmath5 non - radiative , and @xmath55 total . in the case of vibrational pumping , for example
, the rate at which vibrations are pumped into the molecule is proportional to the _ total _ cross section . but some of the re - emitted photons might not be observable in the far - field because they are absorbed by plasmon resonances in the vicinity of the probe .
this is what makes the distinction between _ radiative _ and _ non - radiative _ cross sections .
this could lead potentially to inconsistencies among cross sections for different modes if they happen at frequencies where the non - radiative contributions are different .
we will not insist on this distinction in the following , but it is worth keeping in mind that as far as the cross sections is concerned what we measure is the `` radiative '' part of an effect ( pumping ) which depends on the `` total '' ( radiative plus non - radiative ) cross section .
this distinction could under certain circumstances be important .
we summarize in this subsection the different aspects of the problem of lifetime estimation for the purpose of obtaining sers cross sections . * ultimately , the observation of a cross - over from a thermally - dominated to a pumping - dominated as / s - ratio leads to an estimation of both the asymmetry factor @xmath24 and the product @xmath56 .
an estimation of @xmath47 itself requires the knowledge of @xmath6 , as pointed out before .
the latter can not be directly obtained , in general , from the sers spectrum potentially leading to a serious limitation of this technique . in the first report of cross section estimation via vibrational pumping@xcite a lifetime of @xmath57
was simply estimated without any reference to the experimental data . in order to improve upon this
we need to make a series of assumptions .
the validity of these assumptions needs to be evaluated on a case - by - case basis and the results obtained for the cross section have to be interpreted within these approximations / assumptions .
let us first review the problems associated with trying to extract lifetimes from the sers spectra .
* raman peaks in molecules have lineshapes which are seldom pure lorentzians and have several contributions to their natural widths ; a subject of a longstanding history in spectroscopy@xcite .
the observed linewidth has typically contributions from : @xmath2 groups of raman modes piled up together in narrow energy regions ; @xmath5 inhomogeneous broadening , @xmath55 phase coherence relaxation , and @xmath58 population relaxation .
the last two are also known as the _ off - diagonal _ ( dephasing ) and _ diagonal _ ( population ) relaxation times in density matrix formalism@xcite , respectively , and are part of the _ homogeneous broadening _ of the peak .
these relaxation times are different from mode to mode and can not be easily separated from the observed linewidth in a plain sers spectrum .
the _ population relaxation _ is the one we need ( @xmath6 ) for the estimation of @xmath10 . *
the presence of multiple modes contributing to a peak is a drawback that can be overcome in many situations . in the case of dyes
it is not unusual for several raman active modes , which are closely spaced in energy and have different cross - sections , to contribute to the observed peaks .
this will naturally result in distortions of the lineshape and problems with any estimate of the lifetime based on the width of the peak .
this can be overcome by the use of well characterized vibrations which are relatively isolated from the others and even ( if possible ) complementing the information by density functional theory ( dft ) calculations of the raman spectra@xcite to support the selection of those modes that are the best candidates for a reliable estimation of @xmath6 . *
if the broadening is homogeneous@xcite , the only remaining contributions to the linewidth are the diagonal and off - diagonal relaxation times .
molecular vibrations have typically very strong anharmonic couplings with other vibrations in the molecule and contributions to inhomogeneous broadenings are mostly negligible . even in sers experiments where the single molecule limit is approached@xcite ( which would be more sensitive to inhomogeneous broadening ) the full width at half maximum ( fwhm ) , @xmath59 , of the peaks does not change by more than @xmath60 @xmath61 in peaks with a typical @xmath59 of @xmath62 @xmath61 . the validity of this assumption ( whether inhomogeneous broadening is important or not ) has to be assessed obviously in each specific situation . *
the separation between diagonal and off - diagonal relaxation times can not be achieved by a simple sers spectrum .
it is always possible to invoke a different type of spectroscopy ( time resolved)@xcite , but this may not be feasible in most situations .
one simple possibility ( as done in ref .
@xcite ) is to estimate the population relaxation by using the fwhm @xmath63 of the peaks via @xmath64 ; i.e. essentially ignoring the contribution of the dephasing ( off - diagonal ) relaxation to the width .
this will produce an underestimation of @xmath6 and an overestimation of @xmath47 .
in addition , this typically produces cross - sections for different peaks which are not entirely consistent with the relative ( integrated ) intensities among peaks ; thus showing an intrinsic inconsistency in the estimation of the @xmath6 s ( which can be of the order of a factor of 2 to 10 for the examples we examined ) .
it is generally difficult to make a simple estimation that will be valid for all modes .
it is quite clear that for the cross sections to be meaningful ( and if they are not affected by different amounts of non - radiative processes ) the relative cross sections among peaks must be in accordance with the relative integrated intensities observed in a normal sers spectrum .
* population relaxation is caused mainly by the anharmonic coupling to the `` thermal bath '' .
this bath includes all other modes in the molecule as well as those in the solvent or substrate . generally speaking
, higher energy modes have shorter lifetimes due to a greater number of possible decay pathways .
intramolecular anharmonic decays are energy conserving , meaning that high energy modes have more possibilities for a decay than low energy ones .
population relaxation is also the only dynamical process that contributes to the vibrational linewidth in the limit @xmath65 . *
it is important to realize that , in principle , _ if we find a single peak in the spectrum where the population relaxation can be gained directly from the fwhm via _
@xmath64 , _ then all of the cross sections of the other modes follow immediately through the relative integrated intensities of the peaks_. when lacking a time resolved experiment to isolate the contributions properly , _ the problem comes down to a judicious choice of a mode where the width is dominated by relaxation population_. a compromise then is to extract the values of the cross sections by means of the width of the highest ( isolated ) raman active mode .
a rule of thumb is that the lower the temperature and the higher the energy of the mode the more the linewidth will be dominated by population relaxation .
we can then use one mode to estimate its @xmath47 , and all the others follow automatically from this assumption .
the cross sections are now consistent by construction and the method provides , in fact , the means to estimate the relative lifetimes of the modes .
we call this procedure ( i.e. estimating the @xmath6 of the raman active mode with the largest raman shift , and making the other cross - sections consistent with this value through the relative integrated intensities ) the _ corrected lifetime method _
( clm ) . *
if we are comparing different substrates with the same analyte we could also directly compare the values of @xmath52 without any further assumption . if there is no reason to believe that relaxation population is different between the two substrates , _ the ratio of @xmath52 for two different substrates provides a direct comparison of sers cross sections_. * problems with measuring the asymmetry factor `` @xmath24 '' can be as important as those to obtain a reliable lifetime .
heating in the thermally - dominated regime can be a _
major _ problem . as demonstrated in sec .
ii b , heating can be in part a source for a larger than normal @xmath24 , if it is not properly identified .
* ideally , the best estimation of @xmath24 can come from measurements at rt with long integration times and very low powers .
the obtained @xmath24 s should be more or less the same for a given analyte in the presence of the same metal ( they are mainly related to the local interaction of the dye with the metal@xcite ) . *
the product @xmath66 is measured with high accuracy in the pumping dominated `` plateau '' at low temperatures .
accordingly , the best use of the technique is to use one well - characterized analyte ( without photo - bleaching if possible ) as a means to compare the relative sers cross - sections of the dye on different substrates of the same metal , by simple comparison of the plateaus .
the product @xmath66 is perhaps the best `` objective '' comparison between the cross sections of two substrates without any assumption .
however , it is also important to note that the cross sections we are comparing are not the average cross sections but rather a biased average towards the highest cross sections in the sample , as explained in the next subsection .
there are additional complications with the estimation of the cross section .
it is generally accepted that sers signals are dominated by the presence of `` hot - spots '' or places with high local enhancements .
the formulae presented above made implicitly the assumption that the cross section is the same for all @xmath18 molecules in the sample .
the question we want to address now is : do we obtain an estimate of the average cross section with this method ?
we shall show in what follows that what we actually obtain from the experiment is an estimate which is heavily biased towards the sites with the highest enhancements ; i.e. the method provides an estimate of the enhancements in hot - spots .
this is of course an advantage and a disadvantage at the same time .
we show this explicitly in this section .
we consider a number of molecules ( @xmath67 from 1 to @xmath18 ) in the sample .
each molecule can have a different anti - stokes , ( @xmath68 ) and stokes ( @xmath69 ) sers cross sections ( according to the asymmetry factor @xmath70 for that molecule ) .
we assume for simplicity that the incident power @xmath11 is the same for all .
the rate equation for the average phonon population @xmath71 of a vibrational mode of _ one molecule _ , @xmath67 , and its stationary state solution are given by expressions similar to eqs .
( [ rateeqn ] ) and ( [ nsteady ] ) : following the other formulae in the previous sections , the stokes signal for this molecule is simply @xmath72 , while the anti - stokes signal is given by : @xmath73 .
the anti - stokes / stokes ratio _ for this molecule _ , @xmath74 , is given by an expression similar to eq .
( [ pumpratio ] ) .
if all the molecules experienced the same enhancements , then the total stokes and anti - stokes intensities would be @xmath75 and @xmath76 , and the ratio would be @xmath77 .
however , if the molecules have different cross sections ( sers enhancements ) , which is a much more realistic assumption , we then have : @xmath78 and @xmath79 all the averages here @xmath80 are the usual statistical ensemble averages .
the as / s - ratio is then : @xmath81.\ ] ] comparing this expression with that obtained previously in eq .
( [ pumpratio ] ) , what we called @xmath24 is now replaced by : @xmath82 note that this is not exactly the average of the @xmath83 s , i.e. @xmath84 . moreover , what we measure ( instead of @xmath85 ) is now : @xmath86 to understand the meaning of @xmath87 ( @xmath88 for ensemble ) , we need to understand the sources of non - uniformity .
the main variation is in the cross - sections , because of the large differences in sers enhancements at different locations on the sers substrate .
we typically have values which are , say , @xmath89 larger at hot - spots than in other places .
however , for a given molecule , the ratio @xmath90 should not vary as much ( at least not by more than one order of magnitude ) . in fact , this ratio can in a first approximation be taken as a constant , which is an intrinsic property of the adsorbed molecule / metal complex@xcite . assuming that @xmath91 is the same for all molecules
, we simply get that @xmath92 , i.e. the @xmath24 we measure is the correct one . moreover , using @xmath93 , we can express @xmath87 as a function of @xmath94 only as : @xmath95 if the distribution of @xmath10 s were fairly uniform ( for example a gaussian around an average value with a small standard deviation ) then @xmath87 would be a good estimate of the average of the distribution ( slightly overestimated ) .
but this is far from reality in sers conditions .
a more realistic situation is to have a small number of molecules at a hot - spot ( hs ) , and a large number of molecules at non - hs positions . for
the sake of argument , let us demonstrate the effect on this average by considering @xmath96 molecules with one molecule at a hs and 999 molecules at places with much lower enhancements .
the expression given above would give : @xmath97 if there is a difference of a few orders of magnitude between the cross sections @xmath98 and @xmath99 , the numerator is then easily dominated by the one molecule at the hs ( due to the square ) . for the numerator ,
it is not as clear - cut , it all depends on how much stronger the hs is and how many molecules are at non - hs positions . but overall , the denominator is likely to be dominated by @xmath98 , possibly slightly larger if the second contribution is not negligible .
this shows that @xmath100 , or perhaps slightly smaller .
a more quantitative argument is only possible if we had a more realistic distribution of sers enhancements for a given substrate , but the qualitative conclusion is that _ pumping experiments provide a good lower estimate of the cross - section of the few molecules experiencing the highest enhancements_. the last aspect we want to briefly touch upon in this analysis of cross section estimations via vibrational pumping is photo - bleaching
. the exact mechanisms of photo - bleaching under sers conditions are still poorly understood and would deserve a full study in itself .
photo - bleaching is particularly important in the method we are describing here because molecules sited at hs s (
i.e. those exposed to the largest enhancements ) are affected to a greater degree . at high power densities ,
for example , the population of molecules at hs s will be bleached at a faster rate than the average and this will change the measured cross section . studying @xmath26 under photo - bleaching conditions could , in principle , tell us something about this distribution .
there is robust experimental evidence that photo - bleaching can complicate the analysis , produce experimental artifacts , and -to a large extent- even decide the answer we obtain from a specific experiment .
this is added to several experimental complications in as / s - ratios which include the fact that sometimes ( depending on the dispersion of the spectrometer ) the anti - stokes and stokes sides can not be measured simultaneously .
this implies a delay between the two measurements and therefore a difference in exposure times to the laser . by choosing a certain power level we are , in a way , selecting the population of hs
s that we are going to measure .
we shall come back to this problem in the discussion of the results .
a rule of thumb is that larger laser spots with low power densities ( but sufficient to produce detectable pumping ) , short integration times , and non - resonant ( with the dye ) laser excitation , are in general preferable for a more reliable estimate of sers cross sections .
all possible measures should be taken ( including the type of scanning used ) to address and minimize the undesirable effects of photo - bleaching .
we turn now to an experimental demonstration of the principles underlined above .
sers measurements have been performed on substrates formed by dried ag colloids on silicon .
the colloids were prepared using the standard lee and meisel technique@xcite .
sers active samples were prepared by mixing the colloids with a @xmath101 kcl solution in equal amounts .
the analytes were then added to give a concentration of @xmath102 in each case . a small amount of this solution
was then dried on to a silicon substrate .
the investigated analytes are rhodamine 6 g ( rh6 g ) , crystal violet ( cv ) , and 3,3-diethyloxadicarbocyanine ( dodc ) .
figure [ analytes ] shows the basic sers spectra of the analytes .
we shall show the explicit temperature dependencies of the as / s - ratios for only a few modes ( labeled in fig .
[ analytes ] ) while the tables show additional data for other modes .
samples were mounted in a closed - cycle he - cryostat ( cti - cryogenics ) with temperature control in the range @xmath103 .
raman measurements were performed using several laser lines of a kr@xmath104 and ar@xmath104-ion lasers which were focused to a 20 @xmath105 m diameter spot .
the signal was collected using a high - numerical aperture photographic zoom lens ( canon , @xmath10610 magnification ) onto the entrance slit of a high - dispersion double - additive u1000 jobin - yvon spectrometer coupled to a liquid n@xmath107-cooled ccd detector .
peaks were analyzed using standard voigt functions with subtracted backgrounds .
the variation in the as / s - ratio with @xmath1 for each mode was then fitted to eq .
[ pumpratio ] . for this purpose , it is convenient to modify the expression to : @xmath108 , \label{lnpumpratio}\ ] ] where @xmath109 , and @xmath110 which are the ( dimensionless ) fitted parameters .
the @xmath6 s of the peaks ( necessary to obtain the cross sections ) are estimated in two different ways following the prescriptions in sec .
iii b. as discussed earlier , it is of critical importance that the sample remains stable over the entire length of the experiment .
a range of integration times were used depending on the specific sample and measurement to achieve a good signal to noise ratio for different peaks , but power densities were kept to a minimum compatible with the observation of pumping and exposure to the beam was minimized as much as possible in between changes of @xmath1 .
we used @xmath111 spread over the 20 @xmath105 m diameter spot .
as the rate of photo - bleaching may change with @xmath1@xcite a suitable power density to ensure reasonable stability must be decided at rt , where the effect is greatest .
the relatively large spot and low power densities have the triple advantage of ( @xmath67 ) reducing photo - bleaching to a negligible level , ( @xmath112 ) reducing any indirect laser heating effects and ( @xmath113 ) improving the averaging over cluster geometries . still , even under low power density excitation
, the choice of laser line can have an important effect on the measurement .
an explicit example is shown in fig .
[ photobleach ] for rh6 g under @xmath114 excitation .
as explained in the caption , this observation is compatible with a bleaching of the hs s with largest cross sections .
laser lines that are strongly resonant with the probe molecule@xcite should be avoided so that the photo - stability of the analyte and the accumulated exposure to the laser do not become an issue in the interpretation and analysis of the data .
for that reason , we only concentrate hereafter on results obtained with the @xmath115 and @xmath116 lines of a kr@xmath117-laser .
these two lines are close enough to the visible ( to profit fully from sers enhancements ) but , by the same token , they do not produce excessive photo - bleaching for the dyes under consideration here .
having all the theoretical tools from the previous sections , we can now easily scan through several experimental results that demonstrate the method in practice .
figure [ tdeps ] shows the as / s - ratios ( ln(@xmath26 ) ) vs. @xmath1 , as measured using the @xmath118 line , for four modes of each of the investigated analytes .
the solid lines represent the best fit to the experimental values using eq .
[ lnpumpratio ] . a series of small imperfections can be seen in the data for some modes , but overall the behavior is very well represented by eq .
[ lnpumpratio ] , with rh6 g being the best example .
the arrows indicate the cross - over points @xmath119 between the thermally - dominated and pumping - dominated regimes for all the investigated modes .
@xmath50 occurs at higher @xmath1 s for higher energy modes , as expected .
the fact that pumping is observed for each of the modes indicates that the effect is fairly general and should be observable for a large number of analytes for as long as @xmath10 is high enough to make pumping observable at moderate power densities@xcite . in order to transform these data into estimates of the cross sections we need to address the different methods of estimating @xmath6 .
table [ tab : fitsigmas ] shows the experimental results for rh6 g taken with the @xmath116 line . for the sake of comparison
, we display the cross section by using two different estimations of the @xmath6 s .
this gives a better idea of how reliable these numbers are and how differences among them should be interpreted .
table [ tab : fitsigmas ] shows both : @xmath2 the extraction of @xmath6 directly from the widths via @xmath120 ( this normally renders cross sections which are not fully consistent with the relative integrated intensities of the peaks ) , and @xmath5 the extraction of @xmath6 from the highest raman active mode and the subsequent adjustment of the other cross - sections to agree with the relative integrated intensities ( as described in sec .
iii b ) . a column with relative intensities and cross - sections among modes
is also provided for completeness .
the anti - stokes cross sections are obtained through the stokes ones via the fitted value of the asymmetry parameter @xmath24 from eq .
[ lnpumpratio ] .
these results prove the point that the estimation of the lifetime is a crucial step in this method to transform the experimental values into an estimation of the cross sections themselves .
the values should be interpreted within this assumption , and their validity should be assessed on a case - by - case basis .
the preferred ( more consistent ) method is of course the one that ensures the relative cross sections agree with the relative intensities of peaks ( clm in sec .
iii b ) .
finally , we can discuss the effect of estimating the cross sections for two different excitation wavelengths .
the experiment is somewhat restricted in the choice of excitations ( to be compatible with photo - stability ) but still provides a hint of resonance contributions to the cross sections with some limitations .
table [ tab : fitsigmasall ] shows the cross - sections obtained for rh6 g , dodc , and cv for 647 and @xmath116 excitations for different modes .
all the cross sections were obtained by the clm analysis .
it is worth noting that in this latter method we only need ( in principle ) one temperature dependent as / s - ratio for the highest raman active mode ( judged to be dominated by population relaxation ) .
this together with the plain sers spectrum of the dye is enough to estimate all the cross sections of the modes via the relative integrated intensities .
the results will be consistent ( by construction ) but not necessarily accurate ( depending on the one mode that has been chosen to estimate the initial cross section ) .
the main conclusions of these experiments can be summarized as follows : @xmath2 rh6 g shows larger cross - sections at 647 than at @xmath116 .
this is consistent with a resonant increase known to exist in rh6 g towards the yellow part of the spectrum@xcite .
@xmath5 on the contrary , both dodc and cv experience larger cross - sections at @xmath121 , consistent with the maximum of the absorption of these dyes being at longer wavelengths with respect to rh6 g and , therefore , having a much more red - shifted resonant interaction than rh6g@xcite .
@xmath55 although there is some variations from mode to mode , it is generally true that cv experiences the largest cross - sections in this near - ir region .
the numbers in tables [ tab : fitsigmas ] and [ tab : fitsigmasall ] deserve a special comment regarding their interpretation . for a start , the main source of error in these numbers ( as pointed out in the caption of table [ tab : fitsigmas ] ) is _ systematic _ rather than statistical , and comes from the different methods used in the estimation of @xmath6 .
error bars for these systematic errors are not easy to obtain from the data , for they depend on the relative contribution to the fwhm of population relaxation and dephasing , which is unknown and changes from mode to mode .
the number of significant figures is decided then by the range in which these numbers span for a given group of modes and laser excitation .
it only makes sense to compare statistical errors ( from the propagation of errors in the parameters of the fits and the experimental magnitudes ) for two cross sections with the same systematic errors .
this is the situation , for example , of the comparison between the method presented in this paper and the power - dependence method ( quadratic dependence of the @xmath28 signal as a function of power ) as reported in ref .
@xcite . in this latter case
the same systematic error coming from the estimation of @xmath6 affects both results and , hence , the statistical error bars are meaningful to claim consistency or inconsistency between the two values . in any other case ,
the overriding source of error in these estimates of the cross sections is systematic .
the order of magnitude of these numbers deserves a special comment too .
according to the results in sec .
iii d , we expect these numbers to represent something close to the maximum available enhancement in these samples .
in addition , the fact that @xmath6 could be in general underestimated by any of the methods presented here leads automatically to an overestimation of @xmath54 .
last but not least , the values of @xmath48 are obtained from those of @xmath47 via the asymmetry factor @xmath24 . as a consequence of all these , we expect the values to be : @xmath2 at the higher - end of what is known for hs s in sers for @xmath47 and , @xmath5 with a larger error for the anti - stokes values which use the estimation of the asymmetry parameter @xmath122 two times ; once to obtain @xmath47 itself through the fit of @xmath123 vs. @xmath1 , and twice to obtain @xmath48 from @xmath47 . indeed ,
table [ tab : fitsigmasall ] shows that cross - sections are at the higher - end of what is commonly accepted for sers enhancements .
@xmath48 is bound to be less exact than @xmath47 because it depends crucially on the asymmetry parameter and its associated errors ( treated in sec .
the most consistent set of values are obtained for @xmath47 using the clm - analysis ( which should be the preferred option ) and are all of the order of @xmath124 . in exceptional cases ,
cross sections of the order of @xmath125 are obtained , which are still expected if we take into account that the estimation is for hs s and that the values are possibly overestimated by the choice of @xmath6 .
unrealistically large values of @xmath47 beyond that should be interpreted as a limitation of the technique , in general , and should be tracked down to a problem in the assignment of @xmath6 for that mode .
even for these cases , it is still possible to compare @xmath126 ( or @xmath127 ) for the same modes in different substrates providing a relative comparison of cross section performances at hs s without any additional assumptions .
this comparison is independent of the number of molecules involved in the signal , due to the self - normalizing nature of the method .
table [ tab : fitsigmasall ] summarizes , therefore , a typical example where the performance of different dyes can be tested through appropriate estimates of the cross sections derived from the method presented in this paper .
we have presented an in - depth discussion of the state - of - the - art of sers cross section estimation via vibrational pumping and have given several experimental examples of it . despite the approximations needed and the intrinsic problems of the technique , it is still an excellent tool ( and in many cases possibly the only tool ) to extract an estimation of the sers cross - sections in situations where the number of molecules and hot - spots present in the sample are not known or are difficult to estimate .
the _ self - normalizing _ nature of the method ( with respect to the number of molecules involved ) is certainly a major advantage .
in addition , we feel that these latter experimental and theoretical developments have moved forward to establish vibrational pumping as a real phenomenon in sers after almost a decade of controversy about its very existence .
one outstanding issue that surely deserves further investigation is the effect of photo - bleaching under sers conditions ( in particular in hs ) . as we showed in sec .
iii d , the fact that this technique produces an estimate of the cross sections which is heavily biased towards hs s provides a unique opportunity to explore their characteristics .
this is particularly true in the comparison among different substrates ( in which the photo - stability of the analyte plays a special role ) .
a forthcoming study on the topic is in preparation and will be published elsewhere .
pge and lfc acknowledge support by epsrc ( uk ) under grant gr / t06124 .
rcm acknowledges partial support from the national physical laboratory ( uk ) and the hospitality of the macdiarmid institute at victoria university ( new zealand ) where the measurements have been performed .
special thanks are given to dominik hangleiter ( wellington college , nz ) for taking the data in fig .
m. moskovits , rev .
phys . * 57 * , 783 ( 1985 ) .
n. m. b. perney , j. j. baumberg , m. e. zoorob , m. d. b. charlton , s. mahnkopf , and c. m. netti , optics express * 14 * , 847 ( 2006 ) .
m. green and f. m. liu , j. phys .
b * 107 * , 13015 ( 2003 ) . v. p. drachev , m. d. thoreson , v. nashine , e. n. khaliullin , d. ben - amotz , v. j. davisson , and v. m. shalaev , j. raman spect . * 36 * , 648 ( 2005 )
. j. b. jackson , s. l. westcott , l. r. hirsch , j. l. west , and n. j. halas , appl .
* 82 * , 257 ( 2003 ) .
o. lyandres , n. c. shah , c. r. yonzon , j. t. walsh , m. r. glucksberg , and r. p. van duyne , anal . chem . * 77 * , 6134 ( 2005 ) .
m. larsson and j. lindgren , j. raman spectrosc .
* 36 * , 394 ( 2005 ) .
a. rasmussen , and v. deckert , j. raman spectrosc . * 37 * , 311 ( 2006 ) . s. cinta - pinzaru , n. peica , b. kustner , s. schlucker , m. schmitt , t. frosch , j. h. faber , g. bringmann , and j. popp , j. raman spectrosc . * 37 * , 326 ( 2006 ) .
j. binoy , i. h. joe , v. s. jayakumar , o. f. nielsen , and j. aubard , laser phys . lett . * 2 * , 544 ( 2005 ) .
a. v. szeghalmi , l. leopold , s. pinzaru , v. chis , i. silaghi - dumitrescu , m. schmitt , j. popp , and w. kiefer , biopolymers * 78 * , 298 ( 2005 ) .
j. m. sylvia , j. a. janni , j. d. klein , and k. m. spencer , anal .
chem . * 72 * , 5834 , ( 2000 ) .
k. kneipp , y. wang , h. kneipp , i. itzkan , r. r. dasari , and m. s. feld , phys .
* 76 * , 2444 ( 1996 ) .
a. g. brolo , a. c. sanderson , and a. p. smith , phys .
b * 69 * , 045424 ( 2004 ) . t. l. haslett , l. tay , and m. moskovits , j. chem
. phys . * 113 * , 1641 ( 2000 ) .
e. c. le ru and p. g. etchegoin , faraday discussions * 132 * ( in press ) .
p. v. teredesai , a. k. sood , a. govindaraj , and c.n.r .
rao , appl .
sci . * 182 * , 196 ( 2001 ) .
r. c. maher , l. f. cohen , e. c. le ru , and p. g. etchegoin , faraday discussions * 132 * ( in press ) .
r. c. maher , l. f. cohen , j. c. gallop , e. c. le ru , and p. g. etchegoin , j. phys .
b * 110 * , 6797 ( 2006 ) .
r. c. maher , p. g. etchegoin , e. c. le ru , and l. f. cohen , j. phys .
b ( in press ) .
k. kneipp , h. kneipp , v. b. kartha , r. manoharan , g. deinum , i. itzkan , r. r. dasari , and m. s. feld , phys .
e * 57 * , r6281 ( 1998 ) .
r. c. maher , j. hou , l. f. cohen , f. m. liu , n. green , r. j. c. brown , m. j. t. milton , e. c. le ru , j. m. hadfield , j. e. harvery , and p. g. etchegoin , j. chem . phys . * 123 * , 084702 ( 2005 ) .
a. laubereau and w. kaiser , rev .
phys . * 50 * , 607 ( 1978 ) .
h. watanabe , n. hayazawa , y. inouye , and s. kawata , j. phys .
b * 109 * , 5012 ( 2005 ) .
s. naumov , s. kapoor , s. thomas , s. venkateswaran , and t. mukherjee , j. mol . structure ( theochem ) * 685 * , 127 ( 2204 ) .
w. demtrder , _ laser spectroscopy _ ( springer verlag , berlin , 2003 ) .
e. c. le ru , m. meyer , and p. g. etchegoin , j. phys . chem .
b * 110 * , 1944 ( 2006 ) . p.c .
lee , d. meisel , j. phys .
chem . * 86 * , 3391 ( 1982 ) .
n. p. w. pieczonka and r. f. aroca , chem .
phys . chem . * 6 * , 2473 ( 2005 ) . measurements on adenine with these lasers showed no signs of optical pumping suggesting that the cross - section is small relative to the analytes used here .
mode & @xmath10 ( @xmath128 ) & @xmath48 ( @xmath128 ) & b & @xmath10 ( @xmath128 ) & @xmath48 ( @xmath128 ) & b + ( @xmath61)&(@xmath129 ) & ( @xmath129 ) & ( @xmath130 ) & ( @xmath129 ) & ( @xmath129 ) & ( @xmath130 ) + + 610 & 20 & 112 & 44.8 & 1.8 & 7 & 25.6 + 780 & 25 & 136 & 25.9 & 2.2 & 11 & 6.85 + 1360 & 2 & 463 & 1.26 & 2.1 & 58 & 1.32 + 1510 & 4 & 499 & 6.38 & 3.7 & 114 & 2.13 + 1650 & 14 & 232 & 22.5 & 1.2 & 95 & 2.11 + + 875 & 1.8 & 26 & 11.1 & 0.90 & 33 & 3.04 + 925 & 1.7 & 34 & 8.40 & 0.94 & 33 & 3.34 + 970 & 1.7 & 34 & 7.32 & 1.14 & 55 & 2.17 + 1128 & 2.5 & 51 & 4.93 & 1.74 & 143 & 8.23 + 1375 & 1.3 & 25 & 1.21 & 1.01 & 132 & 0.30 + + 760 & 1.4 & 1.8 & 93.5 & 0.2 & 0.5 & 53.6 + 805 & 10.9 & 14.5 & 74.7 & 3.4 & 6.7 & 28.6 + 910 & 3.9 & 6.5 & 124 & 1.1 & 3.9 & 6.82 + 1175 & 19.5 & 30.7 & 18.0 & 4.3 & 31.8 & 5.62 + 1292 & 3.7 & 6.3 & 8.24 & 1.6 & 4.8 & 3.94 + 1620 & 6.4 & 14.4 & 9.99 & 2.0 & 32.7 & 2.26 + | we present an in - depth analysis of the experimental estimation of cross sections in surface enhanced raman scattering ( sers ) by vibrational pumping .
the paper highlights the advantages and disadvantages of the technique , pinpoints the main aspects and limitations , and provides the underlying physical concepts to interpret the experimental results .
examples for several commonly used sers probes are given , and a discussion on future possible developments is also presented . |
quantum spaces are revealing rich geometrical structures and are the subject of intense research activities . in this paper
we present a construction of monopoles on the quantum projective plane @xmath1 . by a monopole
we mean a line bundle over @xmath1 , that is to say a ` rank 1 ' ( in a sense to be made precise ) finitely generated projective module over the coordinate algebra @xmath2 , endowed with a connection having anti - selfdual curvature ; these are described in sect .
[ se : mono ] .
necessary for the anti - selfduality was a differential calculus and a hodge star operators on forms . in sect .
[ se : diff ] we give a full differential @xmath3-calculus on @xmath1 with an hodge star operator which satisfies all the required properties . both the k - theory and the k - homology groups of the quantum projective plane
are known : @xmath4 while @xmath5 .
thus , a finitely generated projective module over @xmath2 is uniquely identified by three integers .
hermitian vector bundles over a homogeneous space actually the corresponding modules of sections can be equivalently described as vector valued functions over the total space which are equivariant for a suitable action of the structure group : that is to say , one thinks of sections as equivariant maps. this construction still makes sense for quantum homogeneous spaces . in sect .
[ se : hb ] we describe general hermitian ` vector bundles ' on @xmath1 as modules of equivariant elements ; we then specialize to line bundles , for which we construct explicit projections ( they are all finitely generated projective modules ) .
the generators of the k - homology are obtained in sect .
[ se : top ] where we also compute their pairings with the line bundle projections ; the resulting numbers are interpreted as the rank ( always 1 for line bundles ) and 1st and 2nd chern numbers of the underlying ` vector bundles ' . in particular , we identify three natural generators for the k - theory of @xmath1 and three dual generators for its k - homology . classically , topological invariants are computed by integrating powers of the curvature of a connection , and the result depends only on the class of the vector bundle , and not on the chosen connection . in order to integrate the curvature of a connection on the quantum projective space
@xmath1 one needs ` twisted integrals ' ; moreover , one does not gets integers any longer but rather @xmath0-analogues of integers . in sect .
[ se : qinv ] we give some general result on equivariant k - theory and k - homology and corresponding chern - connes characters , and then focus on @xmath1 with equivariance under the action of the symmetry algebra @xmath6 . in particular , we construct twisted cocycles that , when paired with the line bundle projections , give @xmath0-analogues of monopole and instanton numbers of the line bundle . as a corollary , we obtain that when the deformation parameter @xmath0 is transcendental the equivariant @xmath7-group has ( at least ) a countable number of generators , @xmath8 . after this introduction , we start in sect . [ se : qqp ] with some basic results on the geometry of the quantum projective plane @xmath1 and its symmetry algebra @xmath6 . one of our motivation for the present work was to study quantum field theories on noncommutative spaces .
the construction of monopoles on @xmath1 that we present here ( instanton configurations on @xmath1 will be reported in @xcite ) is a step in this direction . in this spirit , in sect .
[ se : gl ] we study gauged laplacian operators acting on modules of sections of monopole bundles .
their complete diagonalization is made possible by the presence of many symmetries . from the point of view of physics
, such an operator would describe ` excitations moving on the quantum projective space ' in the field of a magnetic monopole . in the limit @xmath9 it provides a model of quantum hall effect on the projective plane .
* notations . *
+ throughout this paper the real deformation parameter will be taken to be @xmath10 .
the symbol @xmath11 $ ] denotes the @xmath0-analogue of @xmath12 , @xmath13:=\frac{q^x - q^{-x}}{q - q^{-1 } } \ ; ; \ ] ] for @xmath14 a positive integer the @xmath0-factorial is @xmath15!:=[n][n-1]\ldots[2][1]$ ] , with @xmath16!:=1 $ ] and the @xmath0-binomial is given by @xmath17{\smash[t]{\begin{array}{c}n \\ m\end{array}}}\!\!\right]$}}:=\frac{[n]!}{[m]![n - m]!}\;.\ ] ] for convenience we define the @xmath0-trinomial coefficient as @xmath18!=q^{-(jk+kl+lj)}\frac{[j+k+l]!}{[j]![k]![l ] ! } \;.\ ] ] we shall use sweedler notation for coproducts @xmath19 with a sum understood . finally , all algebras will be unital @xmath3-algebras over @xmath20 , and their representations will be implicitly assumed to be unital @xmath3-representations .
we recall from @xcite some results on the geometry of the quantum projective plane . + let @xmath6 denote the hopf @xmath3-algebra generated by elements @xmath21 , with @xmath3-algebra structure @xmath22 and @xmath23 , and relations @xmath24=0\;,\qquad [ e_i , f_i]=\frac{k_i^2-k_i^{-2}}{q - q^{-1}}\ ; , \qquad [ e_i , f_j]=0 \quad\mathrm{if}\;i\neq j \ ; , \\ k_ie_ik_i^{-1}=qe_i\ ; , \qquad k_ie_jk_i^{-1}=q^{-1/2}e_j \quad\mathrm{if}\;i\neq j \ ; , \end{gathered}\ ] ] and ( serre relations ) @xmath25 additional relations for the @xmath26 s are obtained from the above ones by taking their @xmath3-conjugated .
our @xmath6 is the ` compact ' real form of the hopf algebra denoted @xmath27 in sect .
6.1.2 of @xcite . with the @xmath0-commutator defined as @xmath28_q:=ab - q^{-1}ba \;,\ ] ] relations can be rewritten in the form @xmath29_q]_q=0\,$ ] or @xmath30_q , e_i]_q=0\,$ ] .
+ counit , antipode and coproduct are given by @xmath31 for @xmath32 .
an easy check on generators gives for the square of the antipode : @xmath33 for obvious reasons we denote @xmath34 the hopf @xmath3-subalgebra of @xmath6 generated by the elements @xmath35 , while @xmath36 denotes the hopf @xmath3-subalgebra generated by @xmath34 and @xmath37 and @xmath38 .
we shall also use the notation @xmath39 for the hopf @xmath3-subalgebra generated by @xmath40 and @xmath41 .
we need finite - dimensional highest weight irreducible representations for which the @xmath42 s are positive operators ; these are all the finite - dimensional irreducible representations that are well defined for @xmath43 . for @xmath6
these representation are labelled by a pair of non - negative integers @xmath44 .
basis vectors of the representation @xmath45 carry a multi - index @xmath46 satisfying the constraints @xmath47 the representation @xmath45 is conveniently described as follows .
we let @xmath48 denote its highest weight vector , i.e. @xmath49 and @xmath50 for @xmath32 , and we let @xmath51 be the element @xmath52!}{\textrm{\footnotesize$\left[\!\!\rule{0pt}{14pt } \smash[b]{\smash[t]{\begin{array}{c}\!n_1-j_1\ !
\\ k\end{array}}}\!\!\right]$ } } f_1^{\frac{1}{2}(j_1+j_2)-m+k}[f_2,f_1]_q^{n_1-j_1-k}f_2^{j_2+k } \,,\!\ ] ] with @xmath53}\sqrt{\frac{[\tfrac{j_1+j_2}{2}+m ] ! } { [ \tfrac{j_1+j_2}{2}-m]!}\,\frac{[n_2-j_2]![j_1]!}{[n_1-j_1]![j_2]!}\ , \frac{[n_1+j_2 + 1]![n_2+j_1 + 1]!}{[n_1]![n_2]![n_1+n_2 + 1]!}}\;.\ ] ] then , the vector space @xmath54 underlying the representation is spanned by the vectors @xmath55 . using the commutation rules of @xmath6
one proves that in this basis the representation is given by ( cf .
2 ) ) @xmath56[\tfrac{1}{2 } ( j_1+j_2)+m+1 ] } \,{\left|j_1,j_2,m+1\right > } \;,\\
e_2{\left|j_1,j_2,m\right > } & : = \sqrt{[\tfrac{1}{2 } ( j_1+j_2)-m+1]}\,a_{j_1,j_2}{\left|j_1 + 1,j_2,m-{\smash[t]{\smash[b]{\tfrac{1}{2}}}}\right > } \\ &
\qquad\qquad\qquad + \sqrt{[\tfrac{1}{2 } ( j_1+j_2)+m]}\,b_{j_1,j_2}{\left|j_1,j_2 - 1,m-{\smash[t]{\smash[b]{\tfrac{1}{2}}}}\right>}\;,\end{aligned}\ ] ] with @xmath26 the transpose of @xmath57 , and with coefficients given by @xmath58[n_2+j_1 + 2][j_1 + 1 ] } { [ j_1+j_2 + 1][j_1+j_2 + 2 ] } } \;,\\[5pt ] b_{j_1,j_2}&:=\begin{cases } \sqrt{\dfrac{[n_1+j_2 + 1][n_2-j_2 + 1][j_2]}{[j_1+j_2][j_1+j_2 + 1 ] } } \quad & \mathrm{if}\;j_1+j_2\neq 0\ ; , \\ 1 & \mathrm{if}\;j_1+j_2=0\;. \end{cases}\end{aligned}\ ] ] it is a @xmath3-representation for the inner product @xmath59 . with our notation , the highest weight vector is @xmath60 .
the representation @xmath61 , has matrix elements @xmath62 .
since @xmath37 commutes with all elements of @xmath34 , any irreducible representation @xmath63 of @xmath36 is the product of a representation of @xmath34 ( these are labelled by the ` spin ' @xmath64 with @xmath65 ) , and a representation of charge @xmath66 of the hopf @xmath3-subalgebra @xmath39 generated by @xmath40 , that is @xmath67 . restricting the representation @xmath68 of @xmath6 to @xmath36
we get : @xmath69 and thus the representations @xmath63 of @xmath36 appearing as components in at least one representation of @xmath6 are only those for which @xmath70 , or equivalently @xmath71 ( since @xmath72 ) . for later use ( in sect .
[ se : gl ] ) we need the casimir operator ; it is the operator given ( in a slightly enlarged algebra , cf .
@xcite ) by @xmath73_q[e_1,e_2]_q+qh^{-1}[f_1,f_2]_q[e_2,e_1]_q \ ; , \end{gathered}\ ] ] with @xmath74 . in the representation
@xmath68 it has spectrum @xmath75 ^ 2 + [ \tfrac{1}{3}(2n_1+n_2)+1]^2+[\tfrac{1}{3}(n_1 + 2n_2)+1]^2\;.\ ] ] + the deformation @xmath77 of the hopf @xmath3-algebra of representative functions of @xmath76 is given in @xcite ( see also @xcite , sect .
9.2 ) . as a @xmath3-algebra it is generated by elements @xmath78 , with @xmath79 ,
having commutation relations @xmath80&=0 \ ; , & [ u^i_k , u^j_l]&=(q - q^{-1})u^i_lu^j_k \ ; , & & \forall\;i < j,\;k < l\;. \end{aligned}\ ] ] there is also a cubic relation @xmath81 with the sum over all permutations @xmath82 of the three elements @xmath83 and @xmath84 is the length of @xmath82 .
the @xmath3-structure is given by @xmath85 with @xmath86 and @xmath87 , as ordered sets .
thus for example @xmath88 .
coproduct , counit and antipode are the standard ones : @xmath89 there is a non - degenerate dual pairing ( cf .
@xcite , sect .
9.4 ) @xmath90 which allows one to define left @xmath91 and right @xmath92 canonical actions of @xmath6 on @xmath77 , @xmath93 by using the counit on these equations one gets that @xmath94 for all @xmath95 and @xmath96 .
also , it is known that the left ( resp .
right ) canonical action is dual to the right ( resp
. left ) regular action .
for the case at hand this is the statement that @xmath97 for all @xmath98 and @xmath96 . on generators
the actions are given by @xmath99 by the peter - weyl theorem @xcite a linear basis @xmath100 of @xmath77 is defined implicitly by @xmath101{\smash[b]{t^{n_1,n_2}_{\underline{i},\underline{j}}}}\right>}= \rho^{n_1,n_2}_{\underline{i},\underline{j}}(h ) \;,\qquad\forall\;h\in{\mathcal{u}_q(\mathfrak{su}(3))}\;.\ ] ] from the definition it follows that @xmath102 and @xmath103 ( that is , @xmath104 is a unitary matrix ) .
also , this matrix transforms according to the representation @xmath68 under the left / right canonical action : @xmath105 for all @xmath106 .
for @xmath107 the elements @xmath108 are just the generators @xmath78 ( properly reordered ) .
let us describe them in general .
we have @xmath109 where @xmath110 are given in . by the poincar - birkhoff - witt theorem ( * ? ? ?
@xmath111 ) , the vector space @xmath6 is spanned by elements of the form @xmath112 , where @xmath113 is a product of @xmath57 s and @xmath114 is a product of @xmath26 s . since @xmath115 , it follows that @xmath116 unless @xmath113 and @xmath114 have degree zero . for @xmath117 one
gets @xmath118 , since @xmath50 .
let us define @xmath119 . from the explicit formul it follows that @xmath120 .
thus , by using , @xmath121 unless @xmath117 .
when @xmath117 , from @xmath122 it derives that @xmath123 .
hence , @xmath124 for all @xmath106 .
but this implies that @xmath125{t^{n_1,n_2}_{\underline{i},\underline{j}}}\right > } & = \rho^{n_1,n_2}_{\underline{i},\underline{j}}(h ) = { \left<\underline{i}|\smash[b]{h|\underline{j}}\right > } = { \left<\,\uparrow\!|\smash[b]{(x_{\underline{i}}^{n_1,n_2})^*h x_{\underline{j}}^{n_1,n_2}|\!\uparrow\,}\right > } \\ & = \rho^{n_1,n_2}_{\uparrow,\uparrow}((x_{\underline{i}}^{n_1,n_2})^*h x_{\underline{j}}^{n_1,n_2 } ) = { \left<(\smash[b]{x_{\underline{i}}^{n_1,n_2})^*h x_{\underline{j}}^{n_1,n_2},u^{n_1,n_2}_{\uparrow,\uparrow}}\right > } \\ & = { \left < h\smash[b]{,x_{\underline{j}}^{n_1,n_2}{\triangleright}u^{n_1,n_2}_{\uparrow,\uparrow}{\triangleleft}(x_{\underline{i}}^{n_1,n_2})^*}\right > } \;,\end{aligned}\ ] ] using the identity .
thus @xmath126 , which is just ( [ eq : eqdef ] ) . + the most natural way to arrive to @xmath1 is via the @xmath127-sphere @xmath128
. we shall therefore start from the algebra of coordinate functions on the latter , defined as @xmath129 and , as such , it is the @xmath3-subalgebra of @xmath77 generated by elements @xmath130 of the last ` row ' . in @xcite
it is proved to be isomorphic , through the identification @xmath131 , to the abstract @xmath3-algebra with generators @xmath132 , @xmath133 , and relations : @xmath134=0 \;,\qquad [ z_2^*,z_2]=(1-q^2)z_1z_1^ * \;,\qquad [ z_3^*,z_3]=(1-q^2)(z_1z_1^*+z_2z_2^ * ) \ ; , \\
\rule{0pt}{16pt } z_1z_1^*+z_2z_2^*+z_3z_3^*=1 \;.\end{gathered}\ ] ] these relations will be useful later on .
the @xmath3-algebra @xmath2 of coordinate functions on the quantum projective plane @xmath1 is the fixed point subalgebra of @xmath77 for the right action of @xmath36 , @xmath135 clearly , both @xmath136 and @xmath2 are left @xmath6-module @xmath3-algebras .
the @xmath3-algebra @xmath2 is generated by elements @xmath137 and from the relations of @xmath136 one gets analogous commutation relations for @xmath2 : @xmath138 with @xmath139 .
the elements @xmath140 are the matrix entries of a projection @xmath141 , that is @xmath142 .
this projection has @xmath0-trace : @xmath143 this projection gives the ` tautological line bundle ' on @xmath1 ; general line bundles will be discussed in sect .
[ se : lb ] below .
the two equalities in give algebra inclusions @xmath144 and @xmath145 .
these could be seen as ` noncommutative principal bundles ' with ` structure hopf algebra ' @xmath146 and @xmath147 respectively . indeed , the action of @xmath36 on @xmath77 dualizes to a coaction of @xmath146 for which @xmath2 is the algebra of coinvariants .
analogously , the action of @xmath39 on @xmath136 dualizes to a coaction of @xmath148 for which again @xmath2 is the algebra of coinvariants .
these are noncommutative analogues of the classical @xmath149-principal bundle @xmath150 and @xmath147-principal bundle @xmath151 .
on the manifold @xmath1 we shall select suitable ` monopole ' bundles . these will come as associated bundles to the principal fibrations on @xmath1 mentioned in the previous section .
we start with the general construction of associated bundles : the starting idea is to define their modules of ` sections ' as equivariant vector valued functions on @xmath152 .
+ with any @xmath14-dimensional @xmath3-representation @xmath153 , one associates an @xmath2-bimodule of equivariant elements : @xmath154 where @xmath155 is a column vector and row by column multiplication is implied .
it is easy to see that @xmath156 is also a left @xmath157-module . in particular
, @xmath158 is the module associated to the trivial representation given by the counit .
a natural @xmath2-valued hermitian structure on @xmath156 is defined by @xmath159 where @xmath160 is the conjugate transpose of @xmath161 and again one multiplies row by column . indeed , if @xmath162 , for @xmath163 , and denoting @xmath164 , we have that @xmath165 thus @xmath166 as it should be .
we think of each @xmath156 endowed with this hermitian structure as the module of sections of an equivariant ` hermitian vector bundle ' over @xmath1 .
[ innprod ] composing the hermitian structure with ( the restriction to @xmath2 of ) the haar functional @xmath167 , we get a non - degenerate @xmath20-valued inner product on @xmath156 , @xmath168 .
this will be used later on in sect .
[ se : hsci ] when defining a hodge operator on forms and gauged laplacian operators on modules of sections .
clearly @xmath169 for any pair of representations @xmath170 .
we have also an inclusion @xmath171 , where @xmath172 is the hopf tensor product of the two representations .
indeed , for @xmath173 and @xmath174 , the product @xmath175 satisfies @xmath176 where we used anti - comultiplicativity of the antipode , @xmath177 thus , @xmath178 .
the inclusion is an isometry for the natural inner product @xmath179 on @xmath180 . for the representations discussed at the end of sect .
[ se:2.1 ] , @xmath181 , with @xmath65 and @xmath71 , we denote @xmath182 .
all these @xmath183 are finitely - generated and projective as one sided ( i.e. both as left or right ) @xmath2-modules .
now , since there always exists an irreducible representation @xmath68 of @xmath6 containing @xmath63 as a summand , and since @xmath184 is a direct summand ( as a bimodule ) in the corresponding bimodule @xmath185 , to establish the statement it is enough to show that all the @xmath185 are free both as left and right ( but not as bimodules ! ) @xmath2-modules .
but this is easy ; indeed , as left @xmath2-modules , we have the following isomorphism @xmath186 where @xmath187 are the elements defined in .
the maps @xmath188 and @xmath189 are well defined from ( [ eq : azt ] ) ; and clearly , they are left @xmath2-module maps . that they are one the inverse of the other follows from the unitarity of the matrix @xmath104 . similarly , as right @xmath2-modules we have the following isomorphism @xmath190 + in the construction of the vector bundles over @xmath1 we have in particular @xmath191 .
moreover @xmath192 are ` sections ' of ` line bundles ' ; before we discuss them in more details in this section , we note that they can equivalently be obtained out of the sphere @xmath128 : @xmath193 we are ready to compute projections @xmath194 for the modules @xmath192 , thus realizing them as @xmath195 as right modules , and @xmath196 as left modules ( notice the ` @xmath197 ' sign ; and remember that @xmath66 labels representations ) , for a suitable integer @xmath198 .
for @xmath199 we have @xmath200!(z_1^jz_2^kz_3^l)(z_1^jz_2^kz_3^l)^*=1 \;,\\
\sum\nolimits_{j+k+l = n}q^{2(j - l)}[j , k , l]!(z_1^jz_2^kz_3^l)^*(z_1^jz_2^kz_3^l)=q^{-2n } \;,\end{gathered}\ ] ] where @xmath201!$ ] are the @xmath0-trinomial coefficients in ( [ eq : q3coef ] ) .
we seek coefficients @xmath202 such that @xmath203 . from the algebraic identity
@xmath204 we get the recursive equations @xmath205 where @xmath206 and with ` initial datum ' @xmath207 .
thus , @xmath202 are the vertices of a @xmath0-tartaglia tetrahedron . since @xmath208=q^{-k - l}[j]+q^{j - l}[k]+q^{j+k}[l ] \;,\ ] ] one verifies that @xmath209!$ ] is a solution by plugging it into ( [ eq : tetra ] ) . similarly , since @xmath210
, we have the algebraic identity @xmath211 that gives the recursive equations on the coefficients @xmath212 with ` initial datum ' @xmath213 .
the solution is @xmath214!$ ] , as one can check by using the identity @xmath215=q^{k+l}[j]+q^{l - j}[k]+q^{-j - k}[l]$ ] .
[ lemma : linebundles ] define @xmath216!}\,z_1^jz_2^kz_3^l\ ; , & & \textup{if}\;n\geq 0\ ; \quad\textup{and with}\quad \ ; j+k+l = n\,,\\ ( \psi_{j , k , l}^n)^*&:=q^{-n+j - l}\sqrt{[j , k , l]!}\,(z_1^jz_2^kz_3^l)^*\ ; , & & \textup{if}\;n\leq 0\ ; \quad\textup{and with}\quad \ ; j+k+l =- n\,.\end{aligned}\ ] ] let @xmath217 be the column vector with components @xmath218 and @xmath194 be the projection of size @xmath219 given by @xmath220 then one has @xmath221 as right @xmath2-modules and @xmath196 as left @xmath2-modules .
the column vector @xmath217 has entries in number @xmath222 and @xmath223 is a row vector of the same size . by the previous lemma @xmath224 , so @xmath225 is a projection .
next , consider the right @xmath2-module map @xmath226 since each @xmath227 is @xmath34-invariant and @xmath228 , it follows that @xmath229 for all @xmath230 and @xmath231 .
hence , the image of @xmath161 is in @xmath192 .
the inverse @xmath2-module map is @xmath232 thus proving that we have an isomorphism of right @xmath2-modules .
the proof for the left module structure is completely analogous , the two module maps being in this case @xmath233 and @xmath234 .
it turns out that what we just computed are the elements @xmath235 for @xmath236 ( resp .
@xmath237 if @xmath238 ) .
the next proposition yields the exact relation with the matrix entries of @xmath223 .
[ lemma : psit ] it holds that @xmath239 for @xmath199 and @xmath240 the generic label of the representation @xmath241 , definition gives @xmath242!}\sqrt{\frac{[n-2l]!}{[n]!}\,\frac{[l+m]!}{[l - m]!}}\ , f_1^{l - m}f_2^{2l } { \triangleright}z_3^n \,.\ ] ] induction on @xmath66 yields @xmath243z_2z_3^{n-1}\;,\ ] ] and induction on @xmath244 yields @xmath245!}{[n-2l ] ! } \ , z_2^{2l}z_3^{n-2l } \;.\ ] ] similarly , changing the labels @xmath246 and @xmath247 : @xmath248!}{[l+m ] ! } \,z_1^{l - m}z_2^{l+m } \;.\ ] ] thus @xmath249!}{[n-2l]![l+m]![l - m]!}}\ , q^{-\frac{n-1}{2}}q^{-l(2l-1)}q^{-\frac{2l-1}{2}}q^{-\frac{1}{2}(l - m)(l - m-1)}z_1^{l - m}z_2^{l+m } z_3^{n-2l } \\ & = ( \psi^n_{l - m , l+m , n-2l})^ * \;,\end{aligned}\ ] ] which establishes the case @xmath199 .
the proof for the case @xmath238 is similar .
it is computationally useful to introduce the left action @xmath250 , of @xmath6 on @xmath77 , given by @xmath251 ; it satisfies @xmath252 due to the presence of the antipode .
also , it is a unitary action for the inner product on @xmath77 coming from the haar state @xmath253 . the proof is a simple computation : @xmath254 for all @xmath255 and @xmath106 . with this action
the bimodule @xmath156 can be viewed as the set of elements of @xmath256 that are invariant under the action @xmath257 of @xmath163 .
equivalence classes of finitely generated projective ( left or right ) modules over an algebra @xmath258 the algebraic counterpart of vector bundles are elements of the group @xmath259 . equivalence classes of even fredholm modules
the algebraic counterpart of ` fundamental classes ' gives a dual group @xmath260 .
the natural non - degenerate pairing between k - theory and k - homology , which pairs projective modules with even fredholm modules , gives index maps @xmath261 .
for the quantum projective plane , @xmath4 . the result for k - theory can be proved viewing the corresponding @xmath262-algebra as the cuntz krieger algebra of a graph @xcite .
the group @xmath7 is given as the cokernel of the incidence matrix canonically associated with the graph ; the result for k - homology can be proven using the same techniques : the groups @xmath263 is now given as the kernel of the transposed matrix @xcite .
it is worth mentioning that @xmath264 with the group @xmath265 ( resp .
@xmath266 ) given as the kernel ( resp .
the cokernel ) of the incidence matrix ( resp . the transposed matrix ) .
thus a finitely generated projective ( left or right ) module over @xmath2 is uniquely identified by three integers and these are obtained by pairing the corresponding idempotent with the three generators of the k - homology .
this section is devoted to the explicit construction of three fredholm modules that , in the next section will be shown to be generators of the k - homology by pairing them with suitable idempotents .
+ we recall that a @xmath267-summable even fredholm module for the algebra @xmath258 is a triple @xmath268 consisting of a @xmath269-graded hilbert space @xmath270 , a graded representation @xmath271 and an odd operator @xmath272 such that @xmath273[f , a_1]\ldots [ f , a_k]$ ] is of traceclass for any @xmath274 . with a @xmath267-summable even fredholm module there
are canonically associated even cyclic cocycles @xmath275 , for @xmath276 .
the map @xmath277 is given by @xmath278[f , a_1]\ldots [ f , a_{2n } ] \right)\ ] ] where @xmath279 is the grading and the symbol @xmath82 for the representation is understood .
we recall that a cyclic @xmath280-cocycle @xmath281 is a @xmath20-linear map @xmath282 which is cyclic , i.e. it satisfies @xmath283 and is a hochschild coboundary , i.e. it satisfies @xmath284 , with @xmath285 the boundary operator : @xmath286 when applied to an idempotent one gets a pairing @xmath287,[e]\right > } & : = \mathrm{ch}^n_{(\pi,{\mathcal{h}},f ) } ( [ e ] ) = \tfrac{1}{2}(-1)^n{\mathrm{tr}}_{{\mathcal{h}}\otimes{\mathbb{c}}^m}(\gamma f[f , e]^{2n+1 } ) , \label{eq : kkpair}\end{aligned}\ ] ] where @xmath288 is the size of @xmath113 and matrix multiplication is understood .
the pairing is integer valued , being the index of the fredholm operator @xmath289 from @xmath290 to @xmath291 , with the natural extensions @xmath292 forming an even fredholm module over @xmath293 .
the result of the pairing depends only on the classes of @xmath113 and of the fredholm module ( for details see @xcite ) . to construct three independent fredholm modules we need ( at least ) four representations .
four is exactly the number of _ irreducible _ representations of the algebra @xmath2 , that we describe in the next section .
it is peculiar of the quantum case that one we needs to consider only irreducible representations : at @xmath294 irreducible representation are @xmath295-dimensional and give only one of the generators of the k - homology ( the trivial fredholm module ) .
an additional true ` quantum effect ' is that for @xmath1 one gets three independent @xmath295-summable fredholm modules corresponding to independent traces on @xmath2 .
thus all relevant information leaves in degree zero .
in contrast , for the classical @xmath296 one needs ( there exist ) cohomology classes in degree zero , two and four .
both at the algebraic and at the @xmath262-algebraic levels there is a sequence @xmath297 of @xmath3-algebra morphisms . in reverse order : the empty space , the space with one point , the quantum projective line and the quantum projective plane .
the first map is the quotient of @xmath2 by the ideal generated by @xmath298 and @xmath299 : roughly speaking , by putting @xmath300 the remaining generators satisfy the relations of @xmath301 , the quantum projective line also known as the standard podle sphere ; and any @xmath3-representation of @xmath2 with @xmath298 in the kernel comes from a representation of @xmath301 .
the second map is the further quotient by the ideal generated by @xmath302 and @xmath303 : roughly speaking , in @xmath2 we send @xmath304 and get the algebra @xmath20 ; this is the only non - trivial character @xmath305 of the algebra @xmath2 , the ` classical point ' of @xmath1 .
the last map is simply @xmath306 .
+ for @xmath307 there is only one irreducible representation , the trivial one , not enough to construct a fredholm module ( and indeed , @xmath308 ) . for @xmath309 there is the representation coming from the morphism @xmath310 , and one further faithful irreducible representation given by the identity map @xmath311 .
these two are all - representations of @xmath20 are @xmath295-dimensional ( it is abelian ) . by linearity , they are in @xmath295 to @xmath295 correspondence with maps @xmath312 , with @xmath313 a projection .
hence @xmath314 or @xmath295 , and the corresponding representations are the trivial map @xmath315 , and the identity map @xmath311 . ]
the irreducible @xmath3-representations of @xmath20 , and are enough to construct an even fredholm module , @xmath316{\smash[t]{\begin{array}{ccc}c & 0 \\ 0 & 0\end{array}}}\!\right)$ } } \;,\qquad f_0={\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 1 \\ 1 & 0\end{array}}}\!\right)$ } } \;.\ ] ] which generates @xmath317 . the pairing with @xmath318 ( projections in @xmath319 are equivalent iff
they have the same rank ) is given by the matrix trace : @xmath320)={\mathrm{tr}}(e).\ ] ] the pullback of this fredholm module to @xmath1 just substitutes the representation @xmath321 of @xmath20 with the character @xmath322 of @xmath2 . using the same symbol for the character ,
one gets a map @xmath323)={\mathrm{tr}}\,\chi_0(e ) \;,\ ] ] the geometrical meaning is the following : the rank of a vector bundle is the dimension of the fiber at any point @xmath324 of the space , and this coincides with the trace of the corresponding projection evaluated at @xmath324 .
here we have only one ` classical point ' , and the map in computes the rank of the restriction of the vector bundle to this classical point .
notice that if the module is free , then ( [ eq : rank ] ) is really the rank of the module .
we pass to the next level of the sequence ( [ eq : seq ] ) , that is the algebra @xmath325 .
there are two irreducible @xmath3-representations coming from the map in the sequence the trivial one and the character of the algebra , and one further representation that is faithful and irreducible .
the latter is the restriction of the representation @xmath326 given by [ eq : chione ] @xmath327 a potential fredholm module is given by @xmath328{\smash[t]{\begin{array}{ccc}\chi_1(a ) & 0 \\ 0 & \chi_0(a ) \,{\textup{id}}_{\ell^2({\mathbb{n } } ) } \end{array}}}\!\right)$ } } \;,\qquad f_1={\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 1 \\ 1 & 0\end{array}}}\!\right)$ } } \;.\ ] ] however , this is _ not _ a fredholm module over @xmath136 , since @xmath329=\bigl\{\chi_1(z_3)-\chi_0(z_3)\bigr\}{\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}1 & 0 \\ 0 & 1\end{array}}}\!\right)$}}\ ] ] is not compact . on the other hand ,
this _ is _ a fredholm module when restricted to @xmath2 ( or @xmath325 ) .
indeed , due to the tracial relation @xmath330 , a complete set of generators for @xmath2 is made of @xmath331 and their adjoints .
on these generators the representation @xmath332 is [ eq : chionep ] @xmath333 and @xmath334 is of trace class for all @xmath335 . in particular
, this means that the fredholm module is @xmath295-summable .
the associated character is @xmath336)={\mathrm{tr}}_{\ell^2({\mathbb{n}})\otimes{\mathbb{c}}^m}\,(\chi_1-\chi_0)(e ) \;,\ ] ] where @xmath288 is the size of the matrix @xmath113 .
the value in ( [ eq : charge ] ) depends only on the restriction of the ` vector bundle ' to the subspace @xmath301 , and will be called for this reason the _
1st chern number _
( or also the _ monopole charge _ ) .
+ besides the representations ( and the fredholm modules ) coming from the sequence ( [ eq : seq ] ) , the algebra @xmath2 has a further irreducible representation and a further fredholm module , which is independent from the previous two .
the representation is faithful and comes from the representation @xmath337 given by [ eq : chitwo ] @xmath338 the construction of the fredholm module is a bit involved .
let us use the labels @xmath339 and @xmath340 . a new basis for the hilbert space
is then given by @xmath341 with @xmath65 and @xmath342 .
in this basis the representation reads @xmath343{\smash[b]{\tfrac{1}{2}}}},m-{\smash[t]{\smash[b]{\tfrac{1}{2}}}}\right > } \;,\\ \chi_2(z_3){\left|\ell , m\right > } & = \sqrt{1-q^{2(\ell+m+1)}}{\left|\ell+{\smash[t]{\smash[b]{\tfrac{1}{2}}}},m+{\smash[t]{\smash[b]{\tfrac{1}{2}}}}\right > } \;.\end{aligned}\ ] ] for the third fredholm module we take as hilbert space @xmath344 two copies of the linear span of orthonormal vectors @xmath341 , with @xmath65 and @xmath345 .
the grading @xmath346 and the operator @xmath347 are the obvious ones .
it remains to describe the representation @xmath348 . as subrepresentation @xmath349
we choose @xmath350 one checks that modulo traceclass operators : @xmath351 we define the subrepresentation @xmath352 by adding multiplicities to @xmath332 . on the generators : @xmath353 on each invariant subspace with a fixed @xmath64 , putting @xmath354
one recovers the representation @xmath332 .
since @xmath355 is finite , on the subspace @xmath356 the operators @xmath357 and @xmath358 are trace class , and so @xmath359 is of trace class as well for all @xmath360 : the fredholm module is @xmath295-summable with corresponding character @xmath361)={\mathrm{tr}}_{{\mathcal{h}}_2\otimes{\mathbb{c}}^m}\,(\pi_+-\pi_-)(e ) \;,\ ] ] where @xmath288 is the size of the matrix @xmath113 .
the above replaces the 2nd chern class of the module . working with generators and relations
as done in @xcite for quantum spheres , it is not difficult to prove that any irreducible @xmath3-representation of @xmath2 is equivalent to one of the representations described above . for completeness
, we give the proof in app .
[ app - ir ] . by iterating the construction , for any positive integer @xmath14 one obtains the @xmath362 irreducible representations of the quantum projective spaces @xmath363 ( only one of these is faithful ,
the others coming from the morphism @xmath364 ) , and the corresponding @xmath365 fredholm modules ; details will be reported elsewhere @xcite . + we know from sect .
[ se : lb ] that the bimodule @xmath366 is isomorphic to @xmath367 as left module and to @xmath368 as right module , with @xmath222 and @xmath225 given by prop .
[ lemma : linebundles ] .
we next compute rank , and 1st and 2nd chern numbers of @xmath194 .
we focus the discussion on @xmath199 , the case @xmath238 being similar .
since @xmath369 , the rank given by ( [ eq : rank ] ) results in @xmath370)=1 \;,\ ] ] thus justifying the name ` line bundles ' for the virtual bundles underlying the modules @xmath192 .
the same result is valid when @xmath238 .
next , we compute @xmath371!}{[j]![k]![l ] ! } ( z_3^l)^*(z_2^k)^*(z_1^j)^*z_1^jz_2^kz_3^l\ ] ] and @xmath372!}{[k]![l ] ! } \chi_1(z_3^*)^l\chi_1(z_2^*)^k\chi_1(z_2)^k\chi_1(z_3)^l \;,\\ ~\\ \chi_0\left ( { \mathrm{tr}}_{{\mathbb{c}}^{d_n}}(p_n ) \right ) & = \chi_0(z_3^*)^n\chi_0(z_3)^n=1 \;.\end{aligned}\ ] ] the use of ( [ eq : one ] ) and ( [ eq : two ] ) leads to @xmath373{{\mathrm{tr}}_{{\mathbb{c}}^{d_n}}}(p_n)\big|n\right > } \\ & = \sum\nolimits_{k+l = n}q^{-kl}\frac{[n]!}{[k]![l ] ! } q^{2k(n+l)}(1-q^{2(n+1)})(1-q^{2(n+2)})\ldots(1-q^{2(n+l ) } ) \\ & = \sum\nolimits_{k=0}^nq^{2kn}+o(q)= \begin{cases } \frac{1-q^{2n(n+1)}}{1-q^{2n}}+o(q)=1+o(q ) & \mathrm{if}\;n>0\;,\\ n+1 + o(q ) & \mathrm{if}\;n=0\;. \end{cases}\end{aligned}\ ] ] as usual we can compute the index for @xmath374 ( see @xcite ) ; we get @xmath375 the very same result holds for @xmath238 .
so , the 1st chern number is @xmath376)=n.\ ] ] it remains to compute the last chern number .
the representations @xmath377 in the third fredholm module are the restriction of ( homonymous ) representations of @xmath136 given as follows . with @xmath305 and @xmath378 the representation described in the previous section
, one has that @xmath379 if @xmath380 , and @xmath381 if @xmath356 . on the other hand , for any @xmath382 it holds that @xmath383 the use of yields : @xmath384!}{[r]![s ] ! }
q^{2r(\ell+m+s)}\prod\nolimits_{k=1}^s(1-q^{2(\ell+m+k ) } ) \\ & = \sum\nolimits_{r=0}^n \ , q^{2r(\ell+m)}\{1+o(q)\}=1+n\delta_{\ell ,-
m}+o(q ) \;,\\\end{aligned}\ ] ] for the representation @xmath352 ; for the representation @xmath349 one gets instead @xmath385 if @xmath386 , and @xmath387!}{[i]![j]![k ] ! }
q^{2i(2\ell+j+k)}q^{2j(\ell+m+k ) } \times \\ & \quad\quad\quad\quad\quad \times\ ; \prod\nolimits_{r=1}^j(1-q^{2(\ell - m+r ) } ) \prod\nolimits_{s=1}^k(1-q^{2(\ell+m+s ) } ) \\ & = \sum\nolimits_{i+j+k = n } \ , q^{4i\ell+2j(\ell+m)}+o(q ) \\ & = 1+n\delta_{\ell ,- m}+\tfrac{1}{2}n(n+1)\delta_{\ell,0}+o(q),\end{aligned}\ ] ] if @xmath388 . in the computation
we have used @xmath389!}{[i]![j]![k]!}=q^{-(ij+jk+ki)}\bigl(1+o(q)\bigr)$ ] .
putting things together results in @xmath390 again , we compute in the limit for @xmath374 to get @xmath391 ) & = \lim_{q\to 0^+ } \sum\nolimits_{\ell\in\frac{1}{2}{\mathbb{n}}}\sum\nolimits_{m=-\ell}^\ell \tfrac{1}{2}n(n+1)\delta_{\ell,0}+o(q ) \\ & = \tfrac{1}{2}n(n+1 ) \;.\end{aligned}\ ] ] the same formula is valid for @xmath238 .
we summarize the results in a proposition .
[ prop : same ] for any @xmath392 , the ( right ) module @xmath192 has ` rank ' @xmath370 ) = 1 \ ; , \ ] ] ` 1st chern number ' @xmath376 ) = n \ ; , \ ] ] and ` 2nd chern number ' @xmath393 ) = \tfrac{1}{2}n(n+1 ) \ ; .\ ] ] a complete set of generators @xmath394 of @xmath395 is given by @xmath396 $ ] the class of the rank one free ( left or right ) @xmath2-module @xmath397 , @xmath398 the class of the left module @xmath399 ( or equivalently of the right module @xmath400 ) , @xmath401 the class of the left module @xmath400 ( or equivalently of the right module @xmath399 ) .
thus , in terms of left modules , @xmath398 is the class of the tautological bundle and @xmath401 the class of the dual vector bundle .
a complete set of generators of @xmath402 is given by the classes of the fredholm modules @xmath403 , @xmath404 , given in sect . [ se : top ] .
we have already mentioned that @xmath405 .
a set of generators of the abelian group @xmath406 is the same as a basis of @xmath406 as a @xmath407-module .
suppose we have three elements @xmath408 and three elements @xmath409 in the dual space of @xmath407-linear maps ; call @xmath410 the matrix with elements @xmath411 .
assume that @xmath412 and that the inverse @xmath413 of @xmath414 is an element of @xmath415 .
then for any linear map @xmath416 the difference @xmath417 vanishes on all the @xmath418 s and , by the linear independence over @xmath407 of the @xmath418 s , we deduce that any element @xmath419 is a sum @xmath420 with integer coefficients @xmath421 .
hence , the @xmath422 s are a basis of @xmath423 . similarly for any @xmath424
the difference @xmath425 is in the kernel of all @xmath426 , meaning that the @xmath418 s are a basis of @xmath406 over @xmath407 .
now , let @xmath427 and @xmath428\in k^0({\mathcal{a}(\mathbb{c}\mathrm{p}^2_q)})$ ] , @xmath133 , be the classes in the corollary . by prop .
[ prop : same ] the matrix @xmath414 is given by @xmath429{\smash[t]{\begin{array}{ccc}1 & 1 & 1 \\ 0 & -1 & 1 \\ 0 & 0 & 1\end{array}}}\!\right)$ } } \;.\ ] ] it is invertible in @xmath415 with inverse @xmath430{\smash[t]{\begin{array}{ccc}1 & 1 & -2 \\ 0 & -1 & 1 \\ 0 & 0 & 1\end{array}}}\!\right)$ } } \;.\ ] ] this proves that we have generators of k - theory and k - homology .
on the line bundles described previously we aim to define and study ( anti-)selfdual connections . to this end
we need the full machinery of a differential calculus on @xmath1 and additional ingredients like a hodge star operator .
we start with forms .
+ in @xcite we studied the antiholomorphic part of the differential calculus on @xmath1 .
antiholomorphic forms were defined as @xmath431 we complete the calculus presently .
each bimodule of forms @xmath432 will be identified with a suitable bimodule @xmath156 of equivariant elements as described in the previous section .
that is to say , @xmath433 for a ( not necessarily irreducible ) representation @xmath434 of the hopf algebra @xmath36 on @xmath435 .
the representations @xmath436 are known from @xcite and the @xmath437 are obtained by conjugation .
the dimension of @xmath438 does not depend on @xmath0 , and for @xmath294 gives the rank @xmath439 of @xmath432 .
there is a unique sub - representation of the hopf tensor product @xmath440 with the prescribed dimension : this allows one to identify @xmath438 .
the representations relevant for the calculus are listed in the following figure : [ cols="^,^,^ " , ] in the diamond on the right , in the position @xmath441 , we give the values of spin and charge @xmath442 of the representation @xmath438 . for later use , we list the representations @xmath443 and @xmath444 explicitly : @xmath445{\smash[t]{\begin{array}{ccc}q^{\frac{1}{2 } } & 0 \\ 0 & q^{-\frac{1}{2}}\end{array}}}\!\right)$ } } \ ; , & \sigma_{\frac{1}{2},n}(e_1 ) & = { \textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 1 \\ 0 & 0\end{array}}}\!\right)$ } } \ ; , & \sigma_{\frac{1}{2},n}(f_1 ) & = { \textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 \\ 1 & 0\end{array}}}\!\right)$ } } \;,\\ \rule{0pt}{30pt } \sigma_{1,n}(k_1 ) & = { \textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}q & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & q^{-1}\end{array}}}\!\right)$ } } \ ; , & \sigma_{1,n}(e_1 ) & = [ 2]^{\frac{1}{2}}{\textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 1 & 0 \\ 0 & 0 & 1 \\ 0 & 0 & 0\end{array}}}\!\right)$ } } \ ; , & \sigma_{1,n}(f_1 ) & = [ 2]^{\frac{1}{2}}{\textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & 0 \\ 1 & 0 & 0 \\ 0 & 1 & 0\end{array}}}\!\right)$ } } \;,\end{aligned}\ ] ] and furthermore , @xmath446 ( resp .
@xmath447 ) is @xmath448 times the identity matrix .
+ we first make @xmath449 a bi - graded associative algebra .
let @xmath450 , and suppose we have a bi - graded associative left @xmath36-covariant product on @xmath451 , denoted @xmath452 .
for @xmath453 and @xmath454 , with @xmath455 and @xmath456 , @xmath457 , define @xmath458 using left @xmath36-covariance of the product on @xmath451 and the fact that @xmath77 is a @xmath36-bimodule algebra so that @xmath459 , we get @xmath460 if @xmath461 is invariant ( i.e. it belongs to @xmath462 , this can be simplified and becomes @xmath463 if also @xmath464 is invariant ( i.e. it belongs to @xmath465 ) , we get @xmath466 thus , @xmath452 defines a bilinear map @xmath467 .
its associativity follows from associativity of both the products in @xmath77 and @xmath451 .
the datum @xmath468 is automatically a left @xmath157-module algebra , since left and right canonical actions commute and @xmath77 is a left @xmath157-module algebra .
this means that @xmath469 for all @xmath106 and @xmath470 .
all we need is then a graded associative left @xmath36-covariant product on @xmath451 . for all @xmath471 , we shall now construct a left @xmath36-module map @xmath472 which is unique up to some normalization constants .
when @xmath294 , one can fix the normalization , up to some phase factors and angles , by requiring that these maps are partial isometries . requiring that vectors with real components form a subalgebra ( so that real forms are an algebra ) , the phases must be @xmath473 ; the remaining angles and signs are then fixed by the requirement of associativity and graded commutativity of the product . for @xmath474 , partial isometries do not give an associative product .
we determine in app .
[ app : b ] the most general value of the normalization constants in order to have a left @xmath36-covariant product on @xmath451 which is i ) associative , ii ) graded commutative for @xmath294 , and iii ) it sends real vectors into real vectors .
here we just present the result .
[ pr : const ] a left @xmath36-covariant graded associative product @xmath452 on @xmath451 , sending real vectors to real vectors and graded commutative for @xmath294 , is given by @xmath475c_1}\mu_2(v , w)^t-\frac{c_0}{[2]c_2}vw_4 \;,\\
v^{1,0}\times v^{1,0 } & \to v^{2,0}\ ; , & v{\wedge_q\mkern-1mu}w & : = c_4\mu_0(v , w)^t\;,\\ v^{1,0}\times v^{0,1 } & \to v^{1,1}\ ; , & v{\wedge_q\mkern-1mu}w & : = \bigl(-q^{\frac{1}{2}s}c_1\mu_1(v , w),q^{-\frac{3}{2}s}c_2\mu_0(v , w)\bigr)^t \;,\\ v^{1,0}\times v^{1,2 } & \to v^{2,2}\ ; , & v{\wedge_q\mkern-1mu}w & : = \frac{c_3c_4}{c_0}\mu_0(v , w)^t\;,\\ v^{1,0}\times v^{1,1 } & \to v^{2,1}\ ; , & v{\wedge_q\mkern-1mu}w & : = -q^{-\frac{1}{2}s}\frac{c_4}{[2]c_1}\mu_2(v , w)^t - q^{\frac{3}{2}s}\frac{c_4}{[2]c_2}vw_4 \;,\\ v^{1,2}\times v^{1,0 } & \to v^{2,2}\ ; , & v{\wedge_q\mkern-1mu}w & : = \frac{c_3c_4}{c_0}\mu_0(v , w)^t\;,\\ v^{2,1}\times v^{0,1 } & \to v^{2,2}\ ; , & v{\wedge_q\mkern-1mu}w & : = c_3\mu_0(v , w)^t\;,\\ v^{1,1}\times v^{0,1 } & \to v^{1,2}\ ; , & v{\wedge_q\mkern-1mu}w & : = -q^{-\frac{1}{2}s}\frac{c_0}{[2]c_1}\mu_3(v , w)^t - q^{\frac{3}{2}s}\frac{c_0}{[2]c_2}v_4w \;,\\ v^{1,1}\times v^{1,0 } & \to v^{2,1}\ ; , & v{\wedge_q\mkern-1mu}w & : = \frac{c_4}{[2]c_1}\mu_3(v , w)^t-\frac{c_4}{[2]c_2}v_4w \;,\\ v^{1,1}\times v^{1,1 } & \to v^{2,2}\ ; , & v{\wedge_q\mkern-1mu}w&:=-q^{-\frac{1}{2}s}\frac{c_3c_4}{[2]|c_1|^2}\mu_4(v , w)-q^{\frac{3}{2}s}\frac{c_3c_4}{[2]|c_2|^2}v_4 w_4 \;,\end{aligned}\ ] ] where the maps @xmath476 s are @xmath477^{-\frac{1}{2}}(q^{\frac{1}{2}}v_1w_2-q^{-\frac{1}{2}}v_2w_1)\;,\\ \mu_1:{\mathbb{r}}^2\times{\mathbb{r}}^2&\to{\mathbb{r}}^3\ ; , & \mu_1(v , w)&:=\bigl(v_1w_1,[2]^{-\frac{1}{2}}(q^{-\frac{1}{2}}v_1w_2+q^{\frac{1}{2}}v_2w_1),v_2w_2\bigr)\;,\\ \mu_2:{\mathbb{r}}^2\times{\mathbb{r}}^3&\to{\mathbb{r}}^2\ ; , & \mu_2(v , w)&:=\bigl(qv_1w_2-q^{-\frac{1}{2}}[2]^{\frac{1}{2}}v_2w_1 , q^{\frac{1}{2}}[2]^{\frac{1}{2}}v_1w_3-q^{-1}v_2w_2\bigr)\;,\\ \mu_3:{\mathbb{r}}^3\times{\mathbb{r}}^2&\to{\mathbb{r}}^2\ ; , & \mu_3(v , w)&:=\bigl(q^{\frac{1}{2}}[2]^{\frac{1}{2}}v_1w_2-q^{-1}v_2w_1 , qv_2w_2-q^{-\frac{1}{2}}[2]^{\frac{1}{2}}v_3w_1\bigr)\;,\\ \mu_4:{\mathbb{r}}^3\times{\mathbb{r}}^3&\to{\mathbb{r}}\ ; , & \mu_4(v , w)&:=qv_1w_3-v_2w_2+q^{-1}v_3w_1 \;.\end{aligned}\ ] ] the parameters @xmath478 and @xmath479 are not fixed for the time being . to get an involution we use the fact that the spin @xmath480 ( resp .
spin @xmath295 ) representation of @xmath34 is quaternionic ( resp .
rephrased in terms of the representations @xmath443 and @xmath444 of @xmath36 we have the following lemma , which takes into account the fact that real / quaternionic structures change sign to @xmath66 .
[ lemma : j ] let @xmath481 be the vector space underlying the representation @xmath63 of @xmath36 .
an antilinear map @xmath482 satisfying @xmath483 and such that @xmath484 for any @xmath163 , is given , for @xmath485 , by @xmath486 for any @xmath487 , @xmath488 and @xmath489 respectively .
moreover , if @xmath490 , the map @xmath491 is a graded involution , i.e. it satisfies @xmath492 and @xmath493 for all @xmath494 and @xmath495 . the property @xmath483 is easily checked in all the above mentioned cases , and the condition @xmath496 is a direct corollary : @xmath497 , and @xmath498 is a sum of spaces @xmath499 having @xmath500 with the same parity of @xmath501 .
since @xmath502 always commutes with @xmath503 , and @xmath504 , the claim for @xmath505 is trivially satisfied .
we have to check for the remaining generators @xmath506 , and in the not - trivial cases @xmath507 . a direct computation yields : @xmath508{\smash[t]{\begin{array}{ccc}0 & -q^{-\frac{1}{2 } } \\ q^{\frac{1}{2 } } & 0\end{array}}}\!\right)$ } } { \textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}q^{\frac{1}{2 } } & 0 \\ 0 & q^{-\frac{1}{2}}\end{array}}}\!\right)$}}{\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & q^{-\frac{1}{2 } } \\ -q^{\frac{1}{2 } } & 0\end{array}}}\!\right)$ } } = { \textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}q^{-\frac{1}{2 } } & 0 \\ 0 & q^{\frac{1}{2}}\end{array}}}\!\right)$ } } \\[5pt ] & = \sigma_{\frac{1}{2},-n}(k_1^{-1 } ) \;,\\[10pt ] j\sigma_{\frac{1}{2},n}(e_1)j^{-1}&={\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & -q^{-\frac{1}{2 } } \\ q^{\frac{1}{2 } } & 0\end{array}}}\!\right)$ } } { \textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 1 \\ 0 & 0\end{array}}}\!\right)$}}{\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & q^{-\frac{1}{2 } } \\ -q^{\frac{1}{2 } } & 0\end{array}}}\!\right)$}}={\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 \\ -q & 0\end{array}}}\!\right)$ } } \\[5pt ] & = -q\sigma_{\frac{1}{2},-n}(f_1 ) \;,\\[5pt ] j\sigma_{\frac{1}{2},n}(f_1)j^{-1}&={\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & -q^{-\frac{1}{2 } } \\ q^{\frac{1}{2 } } & 0\end{array}}}\!\right)$ } } { \textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 \\ 1 & 0\end{array}}}\!\right)$}}{\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & q^{-\frac{1}{2 } } \\ -q^{\frac{1}{2 } } & 0\end{array}}}\!\right)$}}={\textrm{\footnotesize$\left(\!\rule{0pt}{14pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & -q^{-1 } \\ 0 & 0\end{array}}}\!\right)$ } } \\[5pt ] & = -q^{-1}\sigma_{\frac{1}{2},-n}(e_1 ) \;,\\[10pt ] j\sigma_{1,n}(k_1)j^{-1}&={\textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & -q^{-1 } \\ 0 & 1 & 0 \\ -q & 0 & 0\end{array}}}\!\right)$ } } { \textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}q & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & q^{-1}\end{array}}}\!\right)$}}{\textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & -q^{-1 } \\ 0 & 1 & 0 \\ -q & 0 & 0\end{array}}}\!\right)$ } } = { \textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}q^{-1 } & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & q\end{array}}}\!\right)$ } } \\[5pt ] & = \sigma_{1,-n}(k_1^{-1 } ) \;,\\[10pt ] j\sigma_{1,n}(e_1)j^{-1}&=[2]^{\frac{1}{2}}{\textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & -q^{-1 } \\ 0 & 1 & 0 \\ -q & 0 & 0\end{array}}}\!\right)$ } } { \textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 1 & 0 \\ 0 & 0 & 1 \\ 0 & 0 & 0\end{array}}}\!\right)$}}{\textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & -q^{-1 } \\ 0 & 1 & 0 \\ -q & 0 & 0\end{array}}}\!\right)$ } } { \textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & 0 \\ -q & 0 & 0 \\ 0 & -q & 0\end{array}}}\!\right)$ } } \\[5pt ] & = -q\sigma_{1,-n}(f_1 ) \;,\\[10pt ] j\sigma_{1,n}(f_1)j^{-1}&=[2]^{\frac{1}{2}}{\textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & -q^{-1 } \\ 0 & 1 & 0 \\ -q & 0 & 0\end{array}}}\!\right)$ } } { \textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & 0 \\ 1 & 0 & 0 \\ 0 & 1 & 0\end{array}}}\!\right)$}}{\textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & 0 & -q^{-1 } \\ 0 & 1 & 0 \\ -q & 0 & 0\end{array}}}\!\right)$ } } = { \textrm{\footnotesize$\left(\!\rule{0pt}{22pt}\smash[b]{\smash[t]{\begin{array}{ccc}0 & -q^{-1 } & 0 \\ 0 & 0 & -q^{-1 } \\ 0 & 0 & 0\end{array}}}\!\right)$ } } \\[5pt ] & = -q^{-1}\sigma_{1,-n}(e_1 ) \;.\end{aligned}\ ] ] also , by a direct computation one checks that @xmath509 with these , is straightforwardly established . from now on , we take that @xmath510 for the coefficients of prop .
[ pr : const ] .
the composition of the involution on @xmath77 with the @xmath511 in the previous lemma , yields a map @xmath512 sending forms into forms and extending the involution of the algebra @xmath513 .
indeed , with @xmath163 , and @xmath164 , from , we get @xmath514 invariance of @xmath515 gives @xmath516 and in turn , using @xmath517 , @xmath518 thus , the involution maps invariant elements into invariant elements , i.e. forms into forms . as a consequence of
, @xmath519 is a graded @xmath3-algebra : @xmath520 the algebra @xmath521 is generated in degree @xmath295 , that is any form of degree @xmath522 can be written as a sum of products of @xmath295-forms .
we need to show that the maps @xmath523 and @xmath524 are surjective for all @xmath335 .
we give the proof for the first map , the second being analogous .
+ if @xmath525 is a scalar , the claim is clearly true . + if @xmath526 ( resp .
@xmath527 ) the scalar @xmath528^{-\frac{1}{2 } } \;,\qquad \mathrm{resp . } \quad ( 1,0)^t{\wedge_q\mkern-1mu}(0,1)^t = c_3[2]^{-\frac{1}{2 } } \;,\ ] ] is a basis @xmath529 ( resp .
@xmath530 ) , and the map is clearly surjective .
+ if @xmath531 , the map @xmath452 is invertible ; indeed , @xmath532 with @xmath533 the unitary matrix in .
+ finally , if @xmath534 the vectors @xmath535^{-1 } ( 1,0)^t \;,\quad ( 0,1)^t{\wedge_q\mkern-1mu}(0,0,0,1)^t =- c_0c_2^{-1}[2]^{-1 } ( 0,1)^t \;,\ ] ] form a basis of @xmath536 , and the maps surjective .
this concludes the proof .
+ having an ( associative , graded involutive ) algebra of forms , the next steps consist in endowing it with i ) two derivations @xmath537 and @xmath538 giving a double complex , with @xmath539 the total differential ; ii ) a closed integral ; iii ) an hodge star operator .
we postpone to the next section the explicit construction of the exterior differentials and start here with the integral .
recall from remark [ innprod ] that an inner product on forms is given by composing the natural hermitian structure on the module @xmath540 of forms with the restriction to @xmath2 of the haar state @xmath167 ; that is @xmath541 where the sum is over all the homogeneous components @xmath542 of @xmath461 and @xmath464 . both the left canonical action @xmath91 and the left action @xmath543 of @xmath6 on @xmath77 are unitary for this inner product , as well as the action of @xmath2 by left multiplication .
now @xmath544 is a free module of rank one , with basis a central element which we denote @xmath295 ( the volume form ) .
indeed , we call @xmath545 the form with all components equal to zero but for the one in degree @xmath546 , which is @xmath295
. we think of this as the volume form and define an integral by @xmath547 in particular , @xmath548 . if the differentials @xmath537 and @xmath538 are given via the ( right ) action of elements of @xmath6 which are in the kernel of the counit @xmath549 , the integral is automatically closed , i.e. @xmath550 a simple consequence of the invariance of the haar state : @xmath551 . using the hermitian structure on @xmath552 , given by @xmath553
the hodge star operator is defined on _ real _ forms via the usual requirement that @xmath554 .
this can be extended to complex forms both linearly ( as e.g. in @xcite ) or antilinearly ( as e.g. in @xcite ) . in the context of solutions of the yang - mills equations , a curvature ,
being the square of a connection , is always real and it does nt matter which extension we choose . we choose the former .
recalling that the hermitian structure @xmath555 is linear in the second entry and antilinear in the first , the hodge star on complex forms is the linear operator @xmath556 defined by @xmath557 applying the haar states to both sides of previous equation we get the usual relation @xmath558 with a @xmath3-calculus , a closed integral and the graded leibniz rule for the differential , the equality ( [ eq:4.3 ] ) implies that @xmath559 which becomes @xmath560 if @xmath561 . to obtain this last property , which is automatic when @xmath294 , one needs suitable constraints on the parameters @xmath562 s in prop .
[ pr : const ] .
[ lemma:4.3 ] on any form @xmath461 one has @xmath563 if @xmath564^{-\frac{1}{4}}\sqrt{|c_0| } \;,\qquad c_2=\pm q^{\frac{3}{4}s}[2]^{-\frac{1}{4}}\sqrt{|c_0| } \;,\qquad c_3=\pm [ 2]^{\frac{1}{2 } } \;,\ ] ] with arbitrary signs .
for this choice of parameters , on ( anti-)holomorphic @xmath565-forms the hodge star is the identity , and on @xmath566 forms it is the linear map @xmath567 if @xmath568 and @xmath464 are homogeneous with different degree , both sides of ( [ eq:4.4 ] ) are zero .
it is then enough to consider the case @xmath569 , @xmath570 . from the definition of the involution on forms , for the possible values of the labels
, one gets @xmath571^{-\frac{1}{2 } } c_3\ , \omega^\dag \cdot \omega ' \;,\\ ( i , j)=(1,1):\quad \omega^*\ ! & { \wedge_q\mkern-1mu}\omega'= \frac{c_3c_4}{[2 ] } \bigl(-q^{-\frac{1}{2}s}|c_1|^{-2}w^\dag , q^{\frac{3}{2}s}|c_2|^{-2}w_4^\dag\bigr ) \omega ' \;.\end{aligned}\ ] ] condition is satisfied if @xmath572^{-\frac{1}{2}}c_3\omega_{j , i } & \mathrm{if}\;(i , j)=(0,1),(1,0),(1,2),(2,1)\ ; , \\
\frac{1}{[2]}c_3c_0 ( -q^{-\frac{1}{2}s}|c_1|^{-2}w , q^{\frac{3}{2}s}|c_2|^{-2}w_4 ) & \mathrm{if}\;(i , j)=(1,1)\;. \end{cases}\ ] ] the square of @xmath573 on @xmath461 is verified to be @xmath574 if the @xmath562 are those given in the statement of the proposition . with these , is immediately checked . with the previous lemma ,
all parameters are fixed , but for some arbitrary signs and a global rescaling^{\frac{1}{2}}$ ] . in the notations of @xcite
we would have @xmath575^{-\frac{1}{2}}$ ] . ]
encoded in @xmath576 . on the other hand ,
fixing the sign of @xmath577 corresponds to fixing an orientation , as flipping the the orientation results in exchanging selfdual with anti - selfdual forms . from now on we assume that @xmath578 , so that by selfdual @xmath566-forms are of the type @xmath579 , and anti - self dual ones are of the type @xmath580 .
the hodge star operator is an isometry .
notice that since @xmath581 , by graded involutivity of the conjugation of forms we have @xmath582 where we used the fact that @xmath461 and @xmath464 have the same degree , and @xmath583 and @xmath584 have the same parity .
there is an equivalent definition of the hodge star operator .
firstly , one can define the `` exterior multiplication '' @xmath585 and the dual `` contraction '' @xmath586 by the formul : @xmath587 then , from and one has @xmath588 which , from the non degeneracy of the scalar product , yields @xmath589 for any @xmath590 .
+ we are left with the definition of the exterior derivative @xmath591 . as we recall in app .
[ ap : cal ] , in order to have a real differential calculus for @xmath592 one needs a derivation , that is a map @xmath593 obeying the leibniz rule and such that @xmath594 and @xmath595 . then the exterior derivative @xmath591 is extended uniquely to forms of higher degree . if @xmath596 , this is equivalent to write @xmath597 with two derivations @xmath598 and @xmath599 , such that @xmath600 ; again both @xmath537 and @xmath538 are extended uniquely to forms of higher degree .
we write @xmath601 , with @xmath602 and @xmath603 the basis vectors of @xmath604 . then
, we set @xmath605 and determine the conditions on the elements @xmath606 s that yield a @xmath537 mapping @xmath441-forms , that is any @xmath607 , to @xmath608-forms ; namely we impose that @xmath609 .
since the assignment @xmath610 is a representation , for all @xmath611 one gets that @xmath612 , with the right adjoint action given by @xmath613 in turn , left covariance of the wedge product yields @xmath614 for any @xmath163 .
then , for @xmath615 an element which is invariant for the tensor product of the right adjoint action with @xmath616 , i.e. for @xmath615 such that @xmath617 we conclude that @xmath618 this means that @xmath537 maps invariant elements of @xmath619 into invariant elements of @xmath620 , that is forms to forms . since both @xmath621 and @xmath438 are @xmath3-representation , taking the adjoint of equation yields @xmath622 which means that @xmath623 maps @xmath441-forms to @xmath624-forms . similarly ,
by replacing @xmath604 with the representation @xmath625 , an invariant element @xmath626 of @xmath627 would give an operator @xmath628 mapping @xmath441-forms to @xmath629-forms and the adjoint @xmath630 mapping @xmath441-forms to @xmath631-forms .
on @xmath1 the only elements of @xmath6 which act on the right as derivations of @xmath2 are the operators @xmath632_q$ ] and @xmath633_q$ ] .
these must be applied in such a way to get an element of @xmath634 .
one defines two exterior derivations @xmath598 and @xmath599 by @xmath635_q}}a)^t \ ; , \qquad { \bar{\partial}}a:=({\mathcal{l}_{k_1k_2[f_2,f_1]_q}}a,{\mathcal{l}_{k_2f_2}}a)^t \;,\ ] ] for all @xmath360 .
the map @xmath636 satisfies the conditions @xmath637 and @xmath638 .
the elements @xmath639 and @xmath640 given by @xmath641_q)^t\;,\qquad y:=(k_1k_2[f_2,f_1]_q , k_2f_2)^t \;,\ ] ] are invariant , hence by the above discussion the maps @xmath642 and @xmath643 send @xmath432 into @xmath644 , resp .
the wedge product on zero forms is diagonal multiplication and on functions these maps become ( proportional to ) ( [ eq : ddb ] ) . since @xmath646 is anticomultiplicative and @xmath647 , from the covariance of the right canonical action we get that for all @xmath255 @xmath648 if @xmath649 are indeed in @xmath2 , right @xmath36-invariance gives @xmath650 and @xmath651 , so @xmath537 and @xmath538 are derivations on @xmath2 .
reality is a simple check . from @xmath652
it follows that @xmath653_q+q(k_1k_2)^{-1}[f_1,f_2]_q,(k_2-k_2^{-1})f_2\bigr ) \\ & = a{\triangleleft}\bigl((q - q^{-1})f_1f_2,0\bigr ) \\ & = 0 \;.\end{aligned}\ ] ] for the generators @xmath654 of @xmath2 one computes that @xmath655 and shows that @xmath656 are a generating family for @xmath657 as a left @xmath2-module .
indeed , for any @xmath658 the coefficients @xmath659 are right @xmath36-invariant , i.e. @xmath660 .
since @xmath661 the following algebraic identities @xmath662 hold . assembling all together , for any @xmath295-form @xmath461 one
finally gets @xmath663 proving that the vectors @xmath656 are a generating family for @xmath634 as a left @xmath2-module .
this concludes the proof . using the properties @xmath664 , @xmath665 and @xmath666 , for any element @xmath360 , simple manipulations in yields @xmath667 also , modulo a proportionality constant the operator @xmath538 coincides with the one of @xcite .
we close this section with few remarks on invariant ( anti-)selfdual @xmath565-forms that we shall use later on in the paper when dealing with monopole connections .
since @xmath668 , @xmath669 , and @xmath670 , by the harmonic decomposition @xmath671 a @xmath565-form @xmath461 is invariant iff @xmath672 and it has the form @xmath673 with @xmath674 . by such
a @xmath565-form is anti - selfdual , for the choice of orientation ( @xmath578 ) made above .
as mentioned in the introduction , by a monopole we mean a line bundle over @xmath1 , that is to say a ` rank 1 ' finitely generated projective module over the coordinate algebra @xmath675 , endowed with a connection having anti - selfdual curvature . in this section we present some of these connections . in sect .
[ se : lb ] we have described at length the isomorphism @xmath676 as left @xmath2-modules and the isomorphism @xmath677 as right @xmath2-modules .
any of the two isomorphisms may be used to transport the grassmannian connection on the bimodule @xmath192 .
we use the second one because it is notationally simpler .
recall that the isomorphism @xmath678 is given @xmath679 , with inverse @xmath680 , @xmath681 ; and @xmath217 is the column vector with components @xmath218 given in prop .
[ lemma : linebundles ] , @xmath219 and @xmath225 .
+ the grassmannian connection on the right @xmath2-module @xmath682 , with respect to the differential calculus of sect . [ se : diff ] is the map @xmath683 for its curvature we get @xmath684 were we used that @xmath685 for any @xmath686 and the identity @xmath687 and so @xmath688 , both valid for any idempotent @xmath113 ( and any differential calculus ) .
when transported to equivariant maps , the connection @xmath689 is readily found to be given by @xmath690 and the curvature @xmath691 becomes the operator of left wedge multiplication by the @xmath565-form , still denote @xmath692 , given by @xmath693 [ incon ] the connection @xmath694 is left @xmath6-invariant . from prop .
[ lemma : psit ] we deduce @xmath695 in turn , for any @xmath696 , ( [ eq : azt ] ) yields @xmath697 thus , for any @xmath696 , @xmath698 this concludes the proof . the connection being invariant from lemma [ incon ] , its curvature @xmath699 is invariant as well .
then , from the discussion at the end of sect .
[ se : diff ] , the two - form @xmath700 is necessarily of the type @xmath701 for some @xmath702 .
hence by it is anti - selfdual .
the connection @xmath694 is anti - selfdual , that is to say , its curvature is a ( constant ) anti - selfdual two - form : @xmath703 for its use later on we compute the constant @xmath704 for @xmath199 ; a similar computation being possible for @xmath238 as well . by construction @xmath705 since @xmath706 , it also holds that @xmath707 using the condition @xmath224 and covariance of the action one deduces that @xmath708 these equations allow one to compute @xmath709 by ( [ eq : azt ] ) we have @xmath710^{-\frac{1}{2 } } \sum\nolimits_{\underline{i } } q(t^{0,n}_{\uparrow,\underline{i}}{\triangleleft}e_2)(t^{0,n}_{\uparrow,\underline{i}}{\triangleleft}e_2)^ * + q^{-1}(t^{0,n}_{\uparrow,\underline{i}}{\triangleleft}e_2e_1)(t^{0,n}_{\uparrow,\underline{i}}{\triangleleft}e_2e_1)^ * \\ & = -c_2q^{-\frac{3}{2}s-3}[2]^{-\frac{1}{2 } } \sum\nolimits_{\underline{i},\underline{j},\underline{k } } \bigl\{q \rho^{0,n}_{\uparrow,\underline{j}}(e_2)\rho^{0,n}_{\underline{k},\uparrow}(f_2 ) \\ & \qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad + q^{-1}\rho^{0,n}_{\uparrow,\underline{j}}(e_2e_1)\rho^{0,n}_{\underline{k},\uparrow}(f_1f_2 ) \bigr\}t^{0,n}_{\underline{j},\underline{i}}(t^{0,n}_{\underline{k},\underline{i}})^ * \\ & = -c_2q^{-\frac{3}{2}s-3}[2]^{-\frac{1}{2 } } \sum\nolimits_{\underline{j},\underline{k } } \bigl\{q\rho^{0,n}_{\uparrow,\underline{j}}(e_2)\rho^{0,n}_{\underline{k},\uparrow}(f_2 ) + q^{-1}\rho^{0,n}_{\uparrow,\underline{j}}(e_2e_1)\rho^{0,n}_{\underline{k},\uparrow}(f_1f_2 ) \bigr\}\delta_{\underline{j},\underline{k } } \\ & = -c_2q^{-\frac{3}{2}s-3}[2]^{-\frac{1}{2 } } \rho^{0,n}_{\uparrow,\uparrow}(qe_2f_2+q^{-1}e_2e_1f_1f_2 ) \\ & = -c_2q^{-\frac{3}{2}s-3}[2]^{-\frac{1}{2 } } { \left<0,0,0|qe_2f_2+q^{-1}e_2e_1f_1f_2|0,0,0\right > } \;.\end{aligned}\ ] ] with @xmath711 the highest weight vector of the representation @xmath712 . using the vanishing @xmath713
, one gets that @xmath714f_2|0,0,0\right>}= { \left<0,0,0|e_2\smash[t]{\smash[b]{\tfrac{k_1 ^ 2-k_1^{-2}}{q - q^{-1}}}}f_2|0,0,0\right > } \\ & = { \left<0,0,0|e_2f_2\smash[t]{\smash[b]{\tfrac{qk_1 ^ 2-q^{-1}k_1^{-2}}{q - q^{-1}}}}|0,0,0\right>}= { \left<0,0,0|e_2f_2|0,0,0\right > } \;,\end{aligned}\ ] ] giving @xmath715^{\frac{1}{2}}q^n{\left<0,0,0|e_2f_2|0,0,0\right > } $ ] .
+ since @xmath716}{\left|0,1,\smash[t]{\smash[b]{\frac{1}{2}}}\right>}$ ] we finally get @xmath717^{\frac{1}{2}}q^n[n]$ ] . hence @xmath718 \ , \nabla_1 ^ 2 \;,\ ] ] where @xmath719 is an irrelevant normalization constant .
+ with monopoles connection on the modules @xmath192 we can study the corresponding gauged laplacian operator acting on @xmath192 .
such an operator describes ` excitations moving on the quantum projective space ' in the field of a magnetic monopole and in the limit @xmath9 provides a model of quantum hall effect on the projective plane .
next , we relate the gauged laplacian operator to the casimir operator in . now , when acting from the right on @xmath192 , a straightforward computation leads to @xmath747 ^ 2+[\tfrac{1}{3}n+1]^2+[\tfrac{2}{3}n+1]^2 + ( q^{\frac{1}{3}n + 1}+q^{-\frac{1}{3}n -1})f_2e_2 \\
+ f_2 f_1 \left ( q^{-\frac{1}{3}n+1 } [ e_1,e_2]_q - q^{\frac{n}{3 } } [ e_2,e_1]_q \right ) \bigr\ } \;,\end{gathered}\ ] ] which , using again reduces to @xmath748 ^ 2+[\tfrac{1}{3}n+1]^2+[\tfrac{2}{3}n+1]^2 + ( q^{\frac{1}{3}n}+q^{-\frac{1}{3}n } ) \bigl ( [ 2 ] f_2e_2 - f_2 f_1 e_2 e_1 \bigr ) \bigr\ } \;.\end{gathered}\ ] ] the generator @xmath37 of the ` structure algebra ' @xmath39 act on @xmath192 as @xmath749 .
then , a comparison of and yields the following proposition .
the gauged laplacian is related to the casimir operator by @xmath750\ , \frac { l - l^{- 1 } } { q - q^{-1 } } \right ) \\
= \mathcal{c}_q - ( { q - q^{-1}})^{-1 } \left ( ( l^{\frac{1}{3 } } - l^{-\frac{1}{3 } } ) ^2 + ( q l^{\frac{1}{3 } } - q^{-1 } l^{-\frac{1}{3 } } ) ^2 + ( q l^{\frac{2}{3 } } - q^{-1 } l^{-\frac{2}{3 } } ) ^2 \right ) \;.\end{gathered}\ ] ] or @xmath751[n ] \bigr ) = \mathcal{c}_q-[\tfrac{1}{3}n]^2-[\tfrac{1}{3}n+1]^2-[\tfrac{2}{3}n+1]^2 \;.\ ] ] the diagonalization of the gauged laplacian is made simple by the observation that for the casimir left and right action is the same : @xmath752 for all @xmath753 , and by the fact that with respect to the left action of @xmath6 there are decompositions : @xmath754 the eigenvalues of the gauged laplacian @xmath755 are given by @xmath756[n+n+2]+[2][n ] & & \mathrm{if}\;n\geq 0\;,\\ \lambda_{n , n}&= ( 1+q^{-3})[n+2][n - n]+[2][n ] & & \mathrm{if}\;n\leq 0\;,\end{aligned}\ ] ] with @xmath757 .
using the above harmonic decomposition of @xmath192 and the spectrum ( [ eq : spcq ] ) of the casimir operator , one gets @xmath758
^ 2+[n+\tfrac{2}{3}|n|+1]^2 -[\tfrac{1}{3}n+1]^2-[\tfrac{2}{3}n+1]^2\big ) + [ 2][n ] \;.\ ] ] with @xmath757 and for any @xmath392 .
this expression can be simplified using the identity
@xmath759 ^ 2-[b]^2=[a][a+2b]$ ] . for @xmath199
this becomes @xmath760\big([n+\tfrac{2}{3}n+2]+[n+\tfrac{4}{3}n+2]\big)+[2][n ] \;,\ ] ] that with a further simplification is the claimed expression in the proposition .
one proceeds similarly for the case @xmath238 .
it is worth stressing that the spectrum of @xmath755 is not symmetric under the exchange @xmath761 , not even when exchanging in addition the parameter as @xmath762 .
a similar phenomenon was already observed in @xcite for gauged laplacians on the standard podle sphere ; the latter can be though of as the quantum projective line @xmath301 .
classically , topological invariants are computed by integrating powers of the curvature of a connection , the result being independent of the particular connection . on the other hand , in order to integrate the curvature of a connection on the quantum projective space @xmath1 one needs ` twisted integrals ' ; the results are not integers any longer but rather @xmath0-analogues , as we shall see in sect .
[ se : qtinvcp2 ] .
we start with some general result on equivariant k - theory and k - homology and corresponding chern - connes characters .
+ classically , the equivariant topological @xmath7-group of a manifold is the grothendieck group of the abelian monoid whose elements are equivalence classes of equivariant vector bundles .
it has an algebraic version that can be generalized to noncommutative algebras .
one has a bialgebra @xmath763 and a left @xmath763-module algebra @xmath258 .
equivariant vector bundles are replaced by one sided ( left , say ) @xmath764-modules that are finitely generated and projective as left @xmath258-modules ; these will be simply called ` equivariant projective modules ' .
any such a module is given by a pair @xmath765 , where @xmath113 is an @xmath766 idempotent with entries in @xmath258 , and is a representation and the following compatibility requirement is satisfied ( see e.g. ( * ? ? ?
2 ) ) , @xmath767 with ` @xmath768 ' denoting transposition .
the corresponding module @xmath769 is made of elements @xmath770 in the range of the idempotent , @xmath771 , with left - module structure given by @xmath772 an equivalence between any two equivariant modules is simply an invertible left @xmath764-module map between them ; @xmath773 will denote the abelian monoid whose elements are equivalence classes of equivariant projective left modules with operation the direct sum , as usual .
the equivariant k - theory group @xmath774 is the grothendieck group of the abelian monoid @xmath773 .
the equivalence of equivariant projective modules can be rephrased in terms of idempotents .
what follows is a direct extension of well known results @xcite .
we give the proof for completeness .
[ lemma : ku ] two equivariant projective modules @xmath769 and @xmath775 are equivalent iff @xmath776 and @xmath777 for some @xmath778 and @xmath779 satisfying the equivariance conditions @xmath780 let @xmath781 be the generic element of @xmath782 .
if @xmath783 is a left @xmath258-module map , then @xmath784 , so the map is uniquely determined by the its value on rows of the idempotent @xmath113 , and similarly for @xmath785 .
we call @xmath778 ( resp .
@xmath779 ) the matrix with entries @xmath786 ( resp . @xmath787 ) . since @xmath82 maps into the range of @xmath788 ( resp .
@xmath785 maps into the range of @xmath113 ) , we have the conditions @xmath789 that in term of row vectors become @xmath790 next , we apply @xmath785 to the equation on the left , @xmath82 to the one on the right , and use @xmath258-linearity . since @xmath791 is the @xmath792-th component of image , via @xmath82 , of the @xmath793-th row of @xmath113 , and @xmath785 is the inverse map , we have @xmath794 and similarly for @xmath785 . thus : @xmath795 from the definition of @xmath796 , the equivariance properties of @xmath113 , and the fact that @xmath82 is an @xmath763-module map , we get @xmath797 which is the first equivariance condition in .
similarly one proves the equivariance condition for @xmath161 . with this , the ` only if ' part is proved .
next , assume that @xmath776 and @xmath777 for some @xmath796 and @xmath161 satisfying above . then , @xmath798 which means @xmath113 that satisfies ( [ eq : cov ] ) . similarly for @xmath788 .
we define @xmath783 and @xmath799 via the formul @xmath800 and @xmath801 ; we need to show that ( i ) the maps are well defined , ( ii ) they are one the inverse of the other , ( iii ) they are left @xmath764-module maps . point ( iii ) is a consequence of the fact that left and right multiplication commute , and of the equivariance conditions for @xmath796 and @xmath161 .
point ( ii ) follows from the identity @xmath802 ( resp .
@xmath803 ) , and the fact that @xmath525 is in the range of @xmath113 ( resp .
@xmath788 ) .
finally if @xmath804 , we have @xmath805 i.e. @xmath806 , and similarly for @xmath785 : the maps are well defined .
there is a natural map from equivariant k - theory to equivariant cyclic homology given for instance in @xcite .
we adapt that construction to our situation .
one start with @xmath807 , the collection of @xmath20-linear maps from @xmath763 to @xmath808 , and defines operations @xmath809 , for @xmath810 , @xmath811 and an operation @xmath812 , @xmath813 they are the face operators and the cyclic operator of equivariant cyclic homology as we are going to show . the maps @xmath814 make up a presimplicial module
one checks that @xmath815 for all @xmath816 , so that @xmath817 is a boundary operator @xcite .
the hopf algebra @xmath763 acts on @xmath808 via the rule @xmath818 and @xmath819 will denote the collection of elements @xmath820 which are ` equivariant ' , meaning that @xmath821 for all @xmath822 .
next , one establishes that the operators @xmath814 commute with the action of @xmath763 ( a not completely trivial task for the last one @xmath823 ) , and it makes sense to consider the complex of equivariant maps . the cyclic operator @xmath824 commutes with the action of @xmath763 thus it descends to an operator on @xmath819 as well .
finally , with @xmath825 it holds that @xmath826 , which says that the boundary operator @xmath827 maps @xmath828 into @xmath829 .
the homology of this last complex is called the ` @xmath763-equivariant cyclic homology ' of @xmath258 with corresponding homology groups usually denoted @xmath830 . next , for @xmath831 a representation as above ,
consider the set @xmath832 this is a subalgebra of @xmath833 ; indeed given any two of its elements @xmath834 one has : @xmath835 moreover , @xmath836-equivariant @xmath766 idempotents as in are elements of @xmath837 . due to the definition of @xmath837 there exists a map @xmath838 given by @xmath839 where @xmath840 denotes composition of the tensor product over @xmath20 with matrix multiplication .
also , @xmath841 which amounts to say that @xmath842 transform the ordinary cyclic operator for the algebra @xmath837 into the ` @xmath763-equivariant ' cyclic operator for @xmath258 . since @xmath843 ,
the map @xmath842 is a morphism of differential complexes , mapping the complex of the cyclic homology of @xmath837 to the complex of the @xmath763-equivariant cyclic homology of @xmath258 .
this construction is completely analogous to the ` non - equivariant ' case , cf .
1.2.3 ) . at this point
, one can repeat verbatim the proof of thm 8.3.2 in @xcite , replacing the ring @xmath844 there , with @xmath837 ( which is still a matrix ring ) and replacing the generalized trace map there , with @xmath842 , to prove the following theorem .
[ thm ] a map @xmath845 is defined by @xmath846 we give in app .
[ app : alt ] an alternative , more explicit proof for the cases @xmath847 .
we stress that what we denote here @xmath848 and call cyclic homology is connes first version of cyclic homology , i.e. the homology of connes complex denoted @xmath849 in @xcite .
modulo a normalization , the cycle @xmath850 is the usual chern - connes character in cyclic homology ( and in fact , no @xmath836 s in the formul ) . on the other hand , @xmath851 is what we are about to use for @xmath1 . in general
, one fixes a group - like element @xmath852 calling @xmath738 the corresponding automorphism of @xmath258 , @xmath853 for all @xmath854 .
then , one pairs @xmath855 with the hochschild cohomology of @xmath258 with coefficients in @xmath856 ; the latter is @xmath258 itself with bimodule structure @xmath857 .
indeed , the pairing @xmath858 defined by @xmath859 when used to compute the dual @xmath860 of the face operators introduced in , yields the formul : @xmath861 these are just the face operators of the hochschild cohomology @xmath862 of @xmath258 with coefficients in @xmath856 ( cf .
thus , the pairing in descends to a pairing @xmath863 + as mentioned , for @xmath1 we take @xmath864 the element implementing the square of the antipode ( cf . ) .
now , the haar state of @xmath77 , satisfies ( cf .
( 11.26 ) and eq . ( 11.36 ) in @xcite ) @xmath865 which for @xmath866 results in @xmath867 this just means that the restriction of the haar state of @xmath77 to @xmath2 is the representative of a class in @xmath868 .
an additional zero cocycle is given by the restriction of the counit @xmath549 of @xmath152 , which on @xmath1 yields the ` classical point ' , that is the character @xmath305 in . on the other hand , with the integral defined in by using the haar state as well , an element @xmath869\in h^4({\mathcal{a}(\mathbb{c}\mathrm{p}^2_q)},{_\eta}{\mathcal{a}(\mathbb{c}\mathrm{p}^2_q)})$ ] is constructed as @xmath870
let us check that it is a cocycle .
leibniz rule gives @xmath871 which is zero by the modular property of the haar state .
a @xmath565-cocycle can be defined in a similar way .
recall that elements of @xmath872 have the form @xmath873 , with @xmath874 .
let @xmath875 be the projection onto the second component @xmath876 , and extend it to a projection @xmath877 by setting @xmath878 if @xmath879 or @xmath880 .
the map @xmath82 is an @xmath2-bimodule map .
then , the map @xmath881 is the representative of a class @xmath882\in h^2({\mathcal{a}(\mathbb{c}\mathrm{p}^2_q)},{_\eta}{\mathcal{a}(\mathbb{c}\mathrm{p}^2_q)})$ ] . indeed , by the leibniz rule , @xmath883 being @xmath82 a bimodule map we get in turn @xmath884 which is zero by the modular property of the haar state .
both classes @xmath869 $ ] and @xmath882 $ ] will be proven to be not trivial by pairing them with the monopole projections .
firstly , the monopole projections @xmath885 in are equivariant with respect to the representation @xmath886 of @xmath6 defined as @xmath887 by prop .
[ lemma : psit ] , @xmath888 for all @xmath106 . using @xmath889
, we get @xmath890 , and the equivariance of @xmath194 follows . from the construction of the previous section , we can pair @xmath891 with @xmath869 $ ] and @xmath892 with @xmath882 $ ] .
the pairing of @xmath882 $ ] with @xmath892 gives @xmath893 since @xmath217 are ` functions ' on the total space of the bundle , we can not move them inside @xmath82 ( which is an @xmath2-bimodule map , not an @xmath77-bimodule map ) . nevertheless , with a little abuse of notations
the form @xmath894 is a constant , and @xmath895 with this , and using @xmath896 , we come to the final formula @xmath897 in , we have already shown that @xmath898 \,\nabla_1 ^ 2 $ ] with @xmath899 .
thus the corresponding quantum characteristic class is proportional to @xmath900 $ ] : @xmath901 \,.\ ] ] at @xmath294 the integral of the curvature is ( modulo a global normalization constant ) the _ monopole number _ of the bundle ; it is the same as the first chern number . as for the pairing of @xmath891 with @xmath902 , @xmath903 using the modular properties of the haar state , which means @xmath904 and is valid for any row vector @xmath905 with entries in @xmath77
, we get : @xmath906 we need to compute the top form @xmath907 . from the identity it follows that @xmath908 , and @xmath909 leading to @xmath910 using @xmath898 \,\nabla_1 ^ 2 $ ] the corresponding quantum characteristic class
is found to be proportional to @xmath911 ^ 2 $ ] : @xmath912 ^ 2 \;.\ ] ] at @xmath294 , the integral of the square of the curvature is ( modulo a global normalization constant ) the _ instanton number _ of the bundle .
the pairing of a projection with the third fredholm module as in sect .
[ se : inpm ] does not give the ` classical ' instanton number , that is @xmath913 , but rather the 2nd chern number which is combination of the instanton number and of the monopole number . by pairing @xmath914 with the haar state , and using its modular property ,
one gets @xmath915 on the other hand , the pairing with the classical point @xmath305 yields @xmath916 with @xmath917 the highest weight of the representation @xmath712 if @xmath199 or @xmath918 if @xmath238 . as a byproduct
, we see that if @xmath0 is transcendental , the equivariant @xmath7-group has ( at least ) a countable number of generators : @xmath8 , i.e. all @xmath919 $ ] are independent .
indeed , were the classes @xmath919 $ ] not independent , there would exist a sequence @xmath920 of integers all zero but for finitely many such that @xmath921 , and @xmath0 would be the root of a non - zero polynomial with integer coefficients .
the results above are analogous to the ones for the standard podle sphere ( cf .
prop . 5.1 and 5.2 in @xcite ) .
in fact they are instances of the general fact ( cf .
@xmath922 ) ) that the equivariant @xmath7 group is a free abelian group and generators are , for the present case , in bijection with equivalence classes of irreducible corepresentations of @xmath36 .
on a four - dimensional manifold ( anti)self - dual connections are stationary points ( usually minima ) of the yang - mills action functional , i.e. they are solutions of the corresponding equations of motion . in dimension
greater than four , ` generalized ' instantons can be defined as solutions of hermitian yang - mills equations .
on @xmath923 a basic instanton solution is associated to the canonical ( universal ) connection on the stiefel bundle @xmath924 .
the extension of this construction to _ quantum _ complex projective spaces using the differential calculus in @xcite will be explored in future works , and it should help , in particular , to understand how to generalize hermitian yang - mills equations to noncommutative spaces .
in this appendix , we determine the most general value of the normalization constants in order to have a left @xmath36-covariant product on @xmath451 which is i ) associative , ii ) graded commutative for @xmath294 , and iii ) it sends real vectors into real vectors .
indeed , as a way of illustration , let us start by considering the cases @xmath925 and @xmath926 . as vector spaces @xmath927 .
for @xmath928 we order the components of @xmath929 as : @xmath930 .
a unitary equivalence @xmath533 between @xmath931 and @xmath932 is given by @xmath933^{-\frac{1}{2 } } & q^{\frac{1}{2}}[2]^{-\frac{1}{2 } } & 0 \\ 0 & 0 & 0 & 1 \\ 0 & q^{\frac{1}{2}}[2]^{-\frac{1}{2 } } & -q^{-\frac{1}{2}}[2]^{-\frac{1}{2 } } & 0 \end{array}\!\right)}\ ] ] it is easy to check that @xmath934 for all @xmath163 by doing it explicitly on all generators . for @xmath935
this is trivial . for @xmath936 , and omitting the representation symbols , we have @xmath937^{\frac{1}{2}}v_2w_2 \\ 0 \\ 0 \end{array}\!\right ) } \;,\ ] ] and @xmath938^{\frac{1}{2 } } { \left(\!\begin{array}{cccc}0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0\end{array}\!\right)}u(v\otimes w)= { \left(\!\begin{array}{cccc}0 & q^{-\frac{1}{2 } } & q^{\frac{1}{2 } } & 0 \\ 0 & 0 & 0 & { } [ 2]^{\frac{1}{2 } } \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0\end{array}\!\right ) } ( v\otimes w ) \;.\ ] ] thus @xmath939
. then the statement holds for @xmath940 since @xmath941 .
the map @xmath942 is then @xmath943 , where @xmath944 are arbitrary for the time being .
when @xmath945 we would get partial isometries ; by composing @xmath533 with the orthogonal projection onto the last component , we would get a partial isometry from @xmath926 .
the general situation is listed in the following proposition .
[ wedpro ] the most general left @xmath36-covariant graded product @xmath452 on @xmath451 , sending real vectors to real vectors , is given by @xmath946^{-\frac{1}{2}}\mu_2(v , w)^t+{c_0 ^ 2}vw_4 \;,\\ v^{1,0}\times v^{1,0 } & \to v^{2,0}\ ; , & v{\wedge_q\mkern-1mu}w & : = { c_4}\mu_0(v , w)^t\;,\\ v^{1,0}\times v^{0,1 } & \to v^{1,1}\ ; , & v{\wedge_q\mkern-1mu}w & : = \bigl(-{d_0}\mu_1(v , w ) , \,{d_1}\mu_0(v , w)\bigr)^t \;,\\ v^{1,0}\times v^{1,2 } & \to v^{2,2}\ ; , & v{\wedge_q\mkern-1mu}w & : = { d_2}\mu_0(v , w)^t\;,\\ v^{1,0}\times v^{1,1 } & \to v^{2,1}\ ; , & v{\wedge_q\mkern-1mu}w & : = { c_1 ^ 1}[3]^{-\frac{1}{2}}\mu_2(v , w)^t+{c_1 ^ 2}vw_4 \;,\\ v^{1,2}\times v^{1,0 } & \to v^{2,2}\ ; , & v{\wedge_q\mkern-1mu}w & : = { d_3}\mu_0(v , w)^t\;,\\ v^{2,1}\times v^{0,1 } & \to v^{2,2}\ ; , & v{\wedge_q\mkern-1mu}w & : = { d_4}\mu_0(v , w)^t\;,\\ v^{1,1}\times v^{0,1 } & \to v^{1,2}\ ; , & v{\wedge_q\mkern-1mu}w & : = -{c_2 ^ 1}[3]^{-\frac{1}{2}}\mu_3(v , w)^t+{c_2 ^ 2}v_4w \;,\\ v^{1,1}\times v^{1,0 } & \to v^{2,1}\ ; , & v{\wedge_q\mkern-1mu}w & : = -{c_3 ^ 1}[3]^{-\frac{1}{2}}\mu_3(v , w)^t+{c_3 ^ 2}v_4w \;,\\ v^{1,1}\times v^{1,1 } & \to v^{2,2}\ ; , & v{\wedge_q\mkern-1mu}w&:={c_4 ^ 1}[3]^{-\frac{1}{2}}\mu_4(v , w)+{c_4 ^ 2}v_4 w_4 \;. \end{aligned}\ ] ] the coefficients @xmath947 and @xmath948 are arbitrary for the time being ; and the maps @xmath476 s are @xmath477^{-\frac{1}{2}}(q^{\frac{1}{2}}v_1w_2-q^{-\frac{1}{2}}v_2w_1)\;,\\ \mu_1:{\mathbb{r}}^2\times{\mathbb{r}}^2&\to{\mathbb{r}}^3\ ; , & \mu_1(v , w)&:=\bigl(v_1w_1,[2]^{-\frac{1}{2}}(q^{-\frac{1}{2}}v_1w_2+q^{\frac{1}{2}}v_2w_1 ) , \,v_2w_2\bigr)\;,\\ \mu_2:{\mathbb{r}}^2\times{\mathbb{r}}^3&\to{\mathbb{r}}^2\ ; , & \mu_2(v , w)&:=\bigl(qv_1w_2-q^{-\frac{1}{2}}[2]^{\frac{1}{2}}v_2w_1 , \
, q^{\frac{1}{2}}[2]^{\frac{1}{2}}v_1w_3-q^{-1}v_2w_2\bigr)\;,\\ \mu_3:{\mathbb{r}}^3\times{\mathbb{r}}^2&\to{\mathbb{r}}^2\ ; , & \mu_3(v , w)&:=\bigl(q^{\frac{1}{2}}[2]^{\frac{1}{2}}v_1w_2-q^{-1}v_2w_1 , \
, qv_2w_2-q^{-\frac{1}{2}}[2]^{\frac{1}{2}}v_3w_1\bigr)\;,\\ \mu_4:{\mathbb{r}}^3\times{\mathbb{r}}^3&\to{\mathbb{r}}\ ; , & \mu_4(v , w)&:=qv_1w_3-v_2w_2+q^{-1}v_3w_1 \;.\end{aligned}\ ] ] on the other hand , there is no need to to specify the multiplication rule by elements in @xmath949 , being scalars . when all the coefficients @xmath562 s and @xmath950 s are equal to @xmath473 and @xmath951 , the maps in the proposition are partial isometries . in general ,
not all the choices give an associative product .
associativity fixes the value of the parameters @xmath950 s and @xmath952 s .
we have [ eq : dueto ] @xmath953 \ , \mu_0(v , v')v''+v\,\mu_0(v',v '' ) \label{eq : duetod}\;,\\ \mu_3(\mu_1(v , v ' ) , v'')&=\mu_0(v , v')v''+[2 ] \,v\,\mu_0(v',v '' ) \label{eq : duetoe}\;,\end{aligned}\ ] ] for all @xmath954 and @xmath955 . by direct computation .
both sides of ( [ eq : duetoa ] ) are equal to @xmath956^{-\frac{1}{2}}(q^{-\frac{1}{2}}v_1v'_2+q^{\frac{1}{2}}v_2v'_1)w_2+q^{-1}v_2v'_2w_1 \;,\ ] ] both sides of ( [ eq : duetob ] ) are equal to @xmath957^{-\frac{1}{2}}v_1w_2v'_2-v_1w_3v'_1-v_2w_1v'_2+q^{-\frac{3}{2}}[2]^{-\frac{1}{2}}v_2w_2v'_1 \;,\ ] ] both sides of ( [ eq : duetoc ] ) are equal to @xmath958^{-\frac{1}{2}}v_2(q^{-\frac{1}{2}}w_1v'_2+q^{\frac{1}{2}}w_2v'_1)+q^{-1}v_3w_1v'_1 \;,\ ] ] both sides of ( [ eq : duetod ] ) are equal to @xmath959^{-\frac{1}{2}}v_1v'_1v''_2 + q^{\frac{3}{2}}[2]^{-\frac{1}{2}}v_1v'_2v''_1 -q^{-\frac{1}{2}}[2]^{\frac{1}{2}}v_2v'_1v''_1 } { \rule{0pt}{18pt } q^{\frac{1}{2}}[2]^{\frac{1}{2}}v_1v'_2v''_2 -q^{-\frac{3}{2}}[2]^{-\frac{1}{2}}v_2v'_1v''_2 -q^{-\frac{1}{2}}[2]^{-\frac{1}{2}}v_2v'_2v''_1 } \;,\ ] ] both sides of ( [ eq : duetoe ] ) are equal to @xmath959^{\frac{1}{2}}v_1v'_1v''_2 -q^{-\frac{3}{2}}[2]^{-\frac{1}{2}}v_1v'_2v''_1 -q^{-\frac{1}{2}}[2]^{-\frac{1}{2}}v_2v'_1v''_1 } { \rule{0pt}{18pt } q^{\frac{1}{2}}[2]^{-\frac{1}{2}}v_1v'_2v''_2 + q^{\frac{3}{2}}[2]^{-\frac{1}{2}}v_2v'_1v''_2 -q^{-\frac{1}{2}}[2]^{\frac{1}{2}}v_2v'_2v''_1 } \;.\ ] ] the map @xmath452 in prop .
[ wedpro ] is a graded associative product on @xmath451 if [ subeq : coef ] @xmath960}}{[2]}{\frac{c_0}{c_1 } } \ ; , & { c^1_3}&={s_1}q^{-\frac{1}{2}{s_2}}\frac{\sqrt{[3]}}{[2]}{\frac{c_0}{c_1 } } \ ; , \label{eq : coefb } \\ { c^2_1}&=-\frac{1}{[2]}{\frac{c_0}{c_2 } } \ ; , & { c^2_3}&=-{s_1}q^{\frac{3}{2}{s_2}}\frac{1}{[2]}{\frac{c_0}{c_2 } } \ ; , \label{eq : coefc } \\
{ c^1_2}&=-{s_1}q^{-\frac{1}{2}{s_2}}\frac{\sqrt{[3]}}{[2]}{\frac{c_4}{c_1 } } \ ; , & { c^1_4}&=-\frac{\sqrt{[3]}}{[2]}{\frac{c_4}{c_1 } } \ ; , \label{eq : coefd } \\ { c^2_2}&=-{s_1}q^{\frac{3}{2}{s_2}}\frac{1}{[2]}{\frac{c_4}{c_2 } } \ ; , & { c^2_4}&=-\frac{1}{[2]}{\frac{c_4}{c_2 } } \ ; , \label{eq : coefe } \\ { c^1_5}&=-{s_1}q^{-\frac{1}{2}{s_2}}\frac{\sqrt{[3]}}{[2]}{\frac{c_3c_4}{|c_1|^2 } } \ ; , & { c^2_5}&=-{s_1}q^{\frac{3}{2}{s_2}}\frac{1}{[2]}{\frac{c_3c_4}{|c_2|^2 } } \ ; , \label{eq : coeff } \\ { d_2 } & = { \frac{c_3c_4}{c_0 } } \ ; , & { d_3 } & = { \frac{c_3c_4}{c_0 } } \ ; , \label{eq : coefg } \\ { d_4 } & = { c_3 } \ ; , \label{eq : coefh } \;,\end{aligned}\ ] ] where @xmath961 are arbitrary signs .
the algebra is graded commutative in the @xmath962 limit iff @xmath963 .
we seek solutions with all parameters different from zero the product is graded by construction .
we impose the condition : @xmath964 if one of the three vectors @xmath965 is a scalar , i.e. an element of @xmath966 , @xmath529 , @xmath967 or @xmath530 , the equality follows by the bilinearity of @xmath452 .
the non - trivial cases are when all three vectors are not scalars .
if the total degree is greater than @xmath546 , we get @xmath968 .
the remaining non - trivial cases are firstly @xmath969 and ( two ) permutations : @xmath970^{-\frac{1}{2}}\mu_2(v,\mu_1(v',v '' ) ) + { c_2}{c^2_1}v\mu_0(v',v '' ) \;,\\[8pt ] ( v{\wedge_q\mkern-1mu}v''){\wedge_q\mkern-1mu}v ' & = -{c_1}{c^1_3}[3]^{-\frac{1}{2}}\mu_3(\mu_1(v , v''),v ' ) + { c_2}{c^2_3}\mu_0(v , v'')v ' \;,\\ v{\wedge_q\mkern-1mu}(v''{\wedge_q\mkern-1mu}v ' ) & = -{d_0}{c^1_1}[3]^{-\frac{1}{2}}\mu_2(v,\mu_1(v'',v ' ) ) + { d_1}{c^2_1}v\mu_0(v'',v ' ) \;,\\[8pt ] ( v''{\wedge_q\mkern-1mu}v){\wedge_q\mkern-1mu}v ' & = { d_0}{c^1_3}[3]^{-\frac{1}{2}}\mu_3(\mu_1(v'',v),v ' ) + { d_1}{c^2_3}\mu_0(v'',v)v ' \;,\\ v''{\wedge_q\mkern-1mu}(v{\wedge_q\mkern-1mu}v ' ) & = { c_0}v''\mu_0(v , v ' ) \;.\end{aligned}\ ] ] using ( [ eq : duetod]-[eq : duetoe ] ) one sees that associativity is just the conditions ( [ eq : coefa]-[eq : coefc ] ) respectively .
+ then , @xmath971 and ( two ) permutations : @xmath972^{-\frac{1}{2}}\mu_2(v,\mu_1(v',v''))+{d_1c^2_2}v\mu_0(v',v '' ) \ ; , \\[8pt ] ( v{\wedge_q\mkern-1mu}v''){\wedge_q\mkern-1mu}v ' & = { d_0c^1_4}[3]^{-\frac{1}{2}}\mu_3(\mu_1(v , v''),v')+{d_1c^2_4}\mu_0(v , v'')v ' \ ; , \\ v{\wedge_q\mkern-1mu}(v''{\wedge_q\mkern-1mu}v ' ) & = { c_1}{c^1_2}[3]^{-\frac{1}{2}}\mu_2(v,\mu_1(v'',v'))+{c_2}{c^2_2}v\mu_0(v'',v ' ) \;,\\[8pt ] ( v''{\wedge_q\mkern-1mu}v){\wedge_q\mkern-1mu}v'&=-{c_1}{c^1_4}[3]^{-\frac{1}{2}}\mu_3(\mu_1(v'',v),v')+{c_2}{c^2_4}\mu_0(v'',v)v ' \;,\\ v''{\wedge_q\mkern-1mu}(v{\wedge_q\mkern-1mu}v')&={c_4}v''\mu_0(v , v ' ) \;.\end{aligned}\ ] ] using again ( [ eq : duetod]-[eq : duetoe ] ) one sees that associativity is just the condition ( [ eq : coefd]-[eq : coefe ] ) respectively .
+ finally , @xmath973 and ( five ) permutations : @xmath974^{-\frac{1}{2}}\mu_4(\mu_1(v , v'),v '' ) \;,\\
v{\wedge_q\mkern-1mu}\{v'{\wedge_q\mkern-1mu}v''\}&={c_3}{c^2_2}\mu_0(v , v'w_4)+{c_3}{c^1_2}[3]^{-\frac{1}{2}}\mu_0(v,\mu_2(v',v '' ) ) \;,\\[8pt ] \{v{\wedge_q\mkern-1mu}v''\}{\wedge_q\mkern-1mu}v'&={c^2_1d_3}\mu_0(vw_4,v')+{d_3c^1_1}[3]^{-\frac{1}{2}}\mu_0(\mu_2(v , v''),v ' ) \;,\\
v{\wedge_q\mkern-1mu}\{v''{\wedge_q\mkern-1mu}v'\}&={c_3}{c^2_4}\mu_0(v , w_4v')-{c_3}{c^1_4}[3]^{-\frac{1}{2}}\mu_0(v,\mu_3(v'',v ' ) ) \;,\\[8pt ] \{v'{\wedge_q\mkern-1mu}v\}{\wedge_q\mkern-1mu}v''&={d_1c^2_5}\mu_0(v',v)w_4-{d_0c^1_5}[3]^{-\frac{1}{2}}\mu_4(\mu_1(v',v),v '' ) \;,\\ v'{\wedge_q\mkern-1mu}\{v{\wedge_q\mkern-1mu}v''\}&={d_2c^2_1}\mu_0(v',vw_4)+{d_2c^1_1}[3]^{-\frac{1}{2}}\mu_0(v',\mu_2(v , v '' ) ) \;,\\[8pt ] \{v'{\wedge_q\mkern-1mu}v''\}{\wedge_q\mkern-1mu}v&={d_4c^2_2}\mu_0(v'w_4,v)+{d_4c^1_2}[3]^{-\frac{1}{2}}\mu_0(\mu_2(v',v''),v ) \;,\\
v'{\wedge_q\mkern-1mu}\{v''{\wedge_q\mkern-1mu}v\}&={d_2c^2_3}\mu_0(v',w_4v)-{d_2c^1_3}[3]^{-\frac{1}{2}}\mu_0(v',\mu_3(v'',v ) ) \;,\\[8pt ] \{v''{\wedge_q\mkern-1mu}v\}{\wedge_q\mkern-1mu}v'&={d_3c^2_3}\mu_0(w_4v , v')-{d_3c^1_3}[3]^{-\frac{1}{2}}\mu_0(\mu_3(v'',v),v ' ) \;,\\ v''{\wedge_q\mkern-1mu}\{v{\wedge_q\mkern-1mu}v'\}&={c_2}{c^2_5}w_4\mu_0(v , v')+{c_1}{c^1_5}[3]^{-\frac{1}{2}}\mu_4(v'',\mu_1(v , v ' ) ) \;,\\[8pt ] \{v''{\wedge_q\mkern-1mu}v'\}{\wedge_q\mkern-1mu}v&={d_4c^2_4}\mu_0(w_4v',v)-{d_4c^1_4}[3]^{-\frac{1}{2}}\mu_0(\mu_3(v'',v'),v ) \;,\\
v''{\wedge_q\mkern-1mu}\{v'{\wedge_q\mkern-1mu}v\}&={d_1c^2_5}w_4\mu_0(v',v)-{d_0c^1_5}[3]^{-\frac{1}{2}}\mu_4(v'',\mu_1(v',v ) ) \;.\end{aligned}\ ] ] using ( [ eq : dueto ] ) , associativity for the first couple is shown to be equivalent to ( [ eq : coeff ] ) , for the second couple to the second condition in ( [ eq : coefg ] ) , for the third couple to the first condition in ( [ eq : coefg ] ) , for the fourth couple to ( [ eq : coefh ] ) .
for the last two couples the associativity is automatically satisfied .
graded anticommutativity in the @xmath962 limit is a simple check , based on the observation that for @xmath294 : @xmath975 is antisymmetric , @xmath976 is symmetric , @xmath977 , and @xmath978 is symmetric .
this concludes the proof .
in this appendix we review the material that leads to the construction of the differential calculus @xmath979 on @xmath1 , in particular the fact that it is enough to define things for one - forms and then extend them in a natural and unique ( by universality ) way .
recall that a first order differential calculus on a unital @xmath3-algebra @xmath258 is a pair @xmath980 , where @xmath981 is an @xmath258-bimodule giving the space of one - forms and @xmath982 is the exterior differential a linear map satisfying the leibniz rule , @xmath983 we also assumes that @xmath984 and that we are dealing with a @xmath3-calculus with the @xmath3-structure on @xmath985 given ( uniquely ) by @xmath986 for all @xmath987 . to go beyond one
forms a differential graded @xmath3-algebra @xmath988 is defined as follows .
the set @xmath989 is a graded algebra , with @xmath990 and graded product denoted @xmath991 .
universal _ differential calculus of @xmath258 is obtained when @xmath452 is the tensor product , @xmath985 is the kernel of the multiplication map @xmath992 and @xmath993 . with a general associative ( unital )
graded multiplication @xmath452 , the calculus will always be a quotient of the universal calculus by a differential ideal .
element in @xmath994 ( that is _ @xmath995-forms _ ) are ( sums of ) products of @xmath995 1-forms : @xmath996 .
the differential is extended to @xmath994 by requiring that its square vanishes , @xmath997 , and that it is a graded derivation of degree 1 , that is @xmath998 and @xmath999 for any two forms @xmath461 and @xmath464 . by universality
, these two properties uniquely identify @xmath591 .
it is the map given on @xmath295-forms by @xmath1000 on any product of @xmath995 @xmath295-forms by @xmath1001 and extended by linearity .
finally , a @xmath3-structure on a differential calculus is given by a graded involution which anti - commutes with the differential . given a @xmath3-structure on @xmath258 , a @xmath3-structure on @xmath1002 is uniquely defined firstly on @xmath295-forms by @xmath1003 , and then by induction on @xmath995-forms by @xmath1004 where @xmath161 is a @xmath295-form and @xmath525 is a @xmath1005 form .
with this , one easily check , again by induction , that @xmath1006 for any form @xmath461 . an equivalent way to give a @xmath3-differential calculus @xmath1002 is via a differential double complex @xmath1007 . that is to say , starting with a bigradation @xmath1008 with a bigraded product ,
the following statements are equivalent : * there is a graded derivation @xmath1009 satisfying @xmath1010 * there are two graded derivations @xmath1011 and @xmath1012 , satisfying @xmath1013
we give an alternative proof that the map @xmath845 given by @xmath1014 is well - defined .
this proof is valid for @xmath847 and for it we use the results of lemma [ lemma : ku ] stating the equivalence of equivariant projective modules in terms of idempotents .
thus , we aim at proving that for any @xmath1015\in k_0^{{\mathcal{u}}}({\mathcal{a}})$ ] , the expression in ( [ eq : eccc ] ) is an equivariant cyclic @xmath14-cycle , and that if @xmath765 is in the same class of @xmath1016 , the difference @xmath1017 is a boundary .
in fact , when @xmath14 is odd ( [ eq : eccc ] ) @xmath1018 and @xmath1019 reduces in @xmath828 to the trivial equivariant cyclic cocycle @xmath1020 . for @xmath14
even , one finds that @xmath1021 which vanishes in @xmath828 since @xmath1022 is now odd and the class of @xmath1023 is zero as before .
thus , @xmath1019 is an equivariant cyclic cocycle . for the last part of the proof ,
@xmath1024 is the boundary of the function @xmath1025 @xmath1026 is the boundary ( modulo the image of @xmath1027 ) of @xmath1028 and @xmath1029 is the boundary ( modulo the image of @xmath1030 ) of @xmath1031 this concludes the proof . in general , for an arbitrary high @xmath14 , it is not an easy task to write @xmath1017 explicitly as the boundary of something else .
in this appendix we prove the result mentioned at the end of sect . [ se : inpm ] , namely that any non - trivial irreducible @xmath3-representation ( irrep for short ) of @xmath2 ( by _ bounded _ operators on a hilbert space ) is unitarily equivalent to one of the representations @xmath305 , @xmath332 or @xmath378 described in the paper .
[ lemma : app1 ] let @xmath1032 be a @xmath3-representation such that @xmath1033
. then there exists @xmath1034 , @xmath133 , such that @xmath1035 is positive and @xmath1036 with this , a @xmath3-representation @xmath1037 is defined by @xmath1038 and it satisfies @xmath1039 .
moreover , @xmath1040 , so that the restriction of @xmath1041 to @xmath2 is exactly @xmath82 .
the @xmath3-representation @xmath1041 is irreducible if @xmath82 is irreducible .
recall that the generators @xmath140 of @xmath2 satisfy the commutation relations in sect .
[ sec:2.3 ] , the quadratic relations @xmath1042 , @xmath133 , and the real structure is given @xmath1043 .
they are related to the generators @xmath227 of @xmath136 by @xmath654 .
since @xmath82 is a @xmath3-representation , @xmath1044 is a sum of positive operators and then it is positive as well .
the positive operators @xmath1045 are well defined and are mutually commuting since @xmath1046 are mutually commuting .
we have @xmath1047 , for all @xmath133 .
using the commutation rules between @xmath140 and @xmath1048 yields @xmath1049 by solving these equations we get @xmath1050 and a third relation for @xmath1051 that we do nt need .
similarly , using the commutation rule @xmath1052=(1-q^2)(p_{21}p_{13}+p_{22}p_{23})$ ] , the relation @xmath1053 can be rewritten as @xmath1054 writing @xmath1055 for the polar decomposition of a bounded operator @xmath1056 , for the polar decomposition of @xmath1057 , @xmath1058 , from we get @xmath1059 .
therefore @xmath1060 .
applying @xmath82 to both sides of the equation yields @xmath1061 where we used the fact that @xmath1062 and @xmath1063 commute , and so do @xmath1064 and @xmath1065 . for any bounded operator @xmath1056 , it holds that @xmath1066 . being @xmath1064 positive with @xmath1067
, we have @xmath1068 . since the range of @xmath1064 is dense in @xmath1069 , we can simplify the @xmath1064 factor on the right of both sides of , and since @xmath1067 we can simplify the @xmath1064 on the left too .
we get @xmath1070 .
next , we show that @xmath1074 defines a @xmath3-representation of @xmath136 , i.e. that the @xmath1075 s satisfy the commutation rules of the algebra @xmath136 .
then the rest of the proof is straightforward : if the relations of @xmath136 are satisfied , some algebraic manipulation immediately gives from that @xmath1076 with @xmath1077 by hyphothesis .
also , since @xmath1078 , if the latter space is dense in @xmath1069 , so must be the former : so if @xmath82 is irreducible , @xmath1041 is irreducible too . among the defining relations of @xmath1 we have @xmath1079 for @xmath1058 .
applying @xmath82 to both sides and using it becomes @xmath1080 .
this reduces to @xmath1081 , since @xmath1082 allowing to simplify the factor @xmath1035 on the left and get .
being @xmath1035 positive it can be diagonalized , and by previous relation @xmath1075 intertwines the eigenspace of @xmath1083 with eigenvalue @xmath1084 with the eigenspace of @xmath1083 with eigenvalue @xmath1085 , i.e. it maps the eigenvalue @xmath1086 of @xmath1035 to @xmath1087 ( @xmath1088 is excluded by the positivity of @xmath1035 ) .
this proves @xmath1089 since @xmath1090 , by conjugating we get @xmath1091 .
the defining relation @xmath1092 yields @xmath1093 , using and . again
@xmath1083 can be simplified having kernel @xmath1094 and so @xmath1095 .
similarly from @xmath1096 we deduce @xmath1097 and by conjugation @xmath1098 . from @xmath1090
we get the relation @xmath1099=0 $ ] ; from @xmath1100 we get @xmath1101=(1-q^2)z_1z_1^*$ ] , from @xmath1102 we get @xmath1103=(1-q^2)(z_1z_1^*+z_2z_2^*)$ ] . using previous formulas for @xmath1101 $ ] and @xmath1103 $ ] , the tracial relation @xmath1104 gives the spherical relation @xmath1105 . with this ,
all the defining relations of @xmath128 are satisfied and the proof is complete .
it is in general not true that a representation on a hilbert space of a subalgebra of a given algebra can be extended to a representation of the full algebra on _ the same _ hilbert space : one needs at least to extend the hilbert space .
the point here is that we can extend the representation from @xmath2 to @xmath136 without enlarging the hilbert space . by previous lemma any irrep @xmath1032 with @xmath1033
is the restriction of an irrep @xmath1037 with @xmath1106 . on the other hand
any irrep @xmath1032 with @xmath1107 for all @xmath133 is the pullback of an irrep of the standard podle sphere , the @xmath3-algebra morphism @xmath1108 being the map @xmath1109 where @xmath1110 and @xmath1111 are the original generators of podle @xcite .
next lemma shows that only these two cases are possible .
suppose there is a vector @xmath161 such that @xmath1112 .
the element @xmath1062 commutes with the generators @xmath140 for all @xmath1113 , while @xmath1079 and @xmath1114 for all @xmath1115 .
this means that for any polynomial @xmath1116 in the generators @xmath1117 and by linearity @xmath1118 for any @xmath360 .
we conclude that the kernel of @xmath1119 carries a subrepresentation of @xmath82 , but @xmath82 is irreducible thus we have only two cases : @xmath1033 or @xmath1120 . in the latter case
@xmath1121 a sum of positive operators @xmath1122 is zero iff each one is zero , and for a bounded operator @xmath1056 the condition @xmath1123 implies @xmath1124 .
thus @xmath1120 implies that @xmath1125 as well .
we need to recall the representation theory of quantum spheres .
irreps of quantum orthogonal spheres @xmath1126 are classified in @xcite and in particular we are interested in @xmath136 , while the case of the standard podle sphere @xmath301 is in @xcite .
the generators @xmath227 of @xmath136 are related to the @xmath1127 s used in @xcite by @xmath1128 and by the replacement @xmath1129 .
we collect some results of @xcite into the following proposition , adapted to our notations .
[ prop : app ] any non - trivial irrep of @xmath325 is unitarily equivalent to one of the following two representations ( cf .
4.i ) , with parameter @xmath1130 ) .
the first is one - dimensional , @xmath1131 ( and @xmath1132 ) , the second is just the map @xmath332 in .
any irrep of @xmath136 with @xmath1133 not in the kernel is unitarily equivalent to one of the following family , parametrized by @xmath1134 ( cf .
@xcite , eq . ( 3.10 ) , @xmath1135 ) : @xmath1136 note that @xmath1137 .
for @xmath1138 we get the representation @xmath1139 in .
notice that restricted to @xmath2 all the representations @xmath1140 are unitarily equivalent .
indeed , let @xmath533 be the unitary transformation @xmath1141 ; one easily checks that @xmath1142 for all @xmath360 ( indeed it is enough to prove it for @xmath1143 ) . by lemma [ lemma : app1 ] and lemma [ lemma : app2 ]
any non - trivial irrep of @xmath2 is either the pullback of one of @xmath325 and then unitarily equivalent to @xmath305 or @xmath332 or it is the restriction of an irrep @xmath1041 of @xmath136 such that @xmath1039 and by prop .
[ prop : app ] it is unitarily equivalent to ( the restriction of ) @xmath378 .
j. cuntz , on the homotopy groups for the space of endomorphisms of a @xmath262-algebra ( with applications to topological markov chains ) , in : _ operator algebras and group representations _ , pitman , london , 1984 , pp . | we present several results on the geometry of the quantum projective plane .
they include : explicit generators for the k - theory and the k - homology ; a real calculus with a hodge star operator ; anti - selfdual connections on line bundles with explicit computation of the corresponding ` classical ' characteristic classes ( via fredholm modules ) ; complete diagonalization of gauged laplacians on these line bundles ; ` quantum ' characteristic classes via equivariant k - theory and @xmath0-indices . |
the ` supersoft ' x - ray sources form a new and exciting but inhomogeneous class of x - ray objects . though a few members of the class were known before rosat , it is this observatory which has characterised the class as objects with temperatures of 15 80 ev and luminosities approaching the eddington limit for a solar mass object , i.e. @xmath2 @xmath3 erg s@xmath1 .
the model which best fits the observations has a white dwarf steadily or cyclically burning hydrogen - rich material accreted onto its surface at @xmath4 m@xmath5 yr@xmath1 ( van den heuvel et al .
@xcite ) .
van den heuvel et al .
( @xcite ) theorise that this material may be supplied by a near main - sequence a / f companion star , more massive than the white dwarf and in a close binary orbit , which is overflowing its roche lobe .
such systems are known as close binary supersoft sources ( cbss ) , for further discussion of the nature of the class see kahabka & van den heuvel ( @xcite ) .
35 supersoft x - ray sources have so far been identified within our galaxy , m31 , the local group galaxy ngc 55 , the lmc & smc ( listed in kahabka & van den heuvel @xcite ) . of these systems
( hereafter rx j0019 ) is the only northern hemisphere cbss identified , so far , in our galaxy .
rx j0019 is a close binary with a 15@xmath685 period and both short - term and long - term optical variability , with orbital variability of @xmath7 ( beuermann et al .
@xcite , hereafter b95 ; greiner & wenzel @xcite ) .
ubvri photometry at photometric maximum and minimum reveal a very blue continuum , in common with other cbss .
the optical and uv spectra show balmer @xmath8 and he ii lines , especially @xmath91640,4686 , in emission with enhanced blue wings , but no he i lines ( b95 ) .
balmer @xmath10 and @xmath11 lines exhibited p - cygni type profiles .
.spectroscopic and photometric jhk magnitudes and spectral indices .
jhk fluxes were obtained from spectra by fitting the continua with low order polynomials and letting the flux be the value of the fit at 1.25 , 1.6 and 2.2 @xmath0 m .
the photometric phases were calculated using the ephemeris of will & barwig ( @xcite ) .
the spectral index , @xmath8 , was calculated for spectra dereddened by e(b v)=0.1 . [
cols="^,^,^,^,^,^,^,^,^ " , ] no photospheric absorption lines are seen in the infrared spectrum taken at minimum light , indicating that either the underlying spectral type of the donor is not later than f , we are seeing the system at low inclination and irradiation is supressing absorption , or the accretion disc is dominating the emission . through the optical and ir to at least 2.4
@xmath0 m the orbital modulation of flux appears to be nearly constant , implying that they are dominated by the same component .
the continuum is flatter than that expected for a stellar photosphere , suggesting the accretion disc is dominant . given the variability of the source over a variety of timescales ( greiner & wenzel @xcite ) , simultaneous multiwavelength observations of rx j0019 would be necessary to properly determine the relative contribution of light sources , and consequently place constraints on the spectral class of the donor , inclination of and distance to the system .
one explanation of the satellite lines to p@xmath11 , p@xmath10 and
br@xmath11 is that they are the infrared signature of jets , similar to the doppler shifted components of the balmer @xmath8 , @xmath10 and he ii @xmath124686 lines of another supersoft source , ( hereafter rx j0513 , crampton et al .
@xcite ; southwell et al .
@xcite ) , in which @xmath13 km s@xmath1 . with this interpretation in mind
we fitted each of these three composite lines profiles with three gaussians .
the details of these gaussians are given in table [ tab : jets ] .
the mean separation of the satellite lines from the core is @xmath14 km s@xmath1 , and their mean fwhm is @xmath15 km s@xmath1 . assuming the jet interpretation is correct , and the jets are perpendicular to an accretion disc which lies in the plane of the binary , the true velocity of the outflow is simply a function of the inclination angle ( @xmath16 km s@xmath1 . ) .
following shahbaz et al .
( @xcite ) we can also estimate the opening angle of the jet , @xmath17 , where @xmath18 is the fwhm of the satellite lines and @xmath19 the observed velocity .
meyer - hofmeister et al .
( @xcite ) model orbital modulation of the system with a best - fit inclination of 56@xmath20 , corresponding to a outflow velocity of @xmath21 km s@xmath1 and an opening angle of 28@xmath20 .
hachisu & kato ( @xcite ) model outbursts of the source with an inclination of 45@xmath20 , corresponding to a velocity of @xmath22 km s@xmath1 and a very large opening angle of 45@xmath20 .
an inclination angle of around 75@xmath20 would be required to derive the same velocity as the jet in rx j0513 , and would imply a much more collimated jet ( @xmath23 ) . alternatively the lower jet velocity could be intrinsic and reflect the lower escape velocity required from a lower mass white dwarf . indeed hachisu & kato ( @xcite ) argue for a mass of around 0.6 m@xmath24 for the white dwarf in rx j0019 , around half that estimated for rx j0513 ( southwell et al .
@xcite ) .
_ note added in manuscript _ just before submission we became aware of a related work by becker et al .
( @xcite ) .
they have found transient satellite lines around hydrogen and he ii lines in optical spectra of rx j0019 .
they also interpret the satellite lines to be the signature of jets and find @xmath25 800 km s@xmath1 , consistent with this work .
we thank the staff responsible for the + ukirt service program for their assistance in obtaining these observations .
the ukirt is operated by the observatories on behalf of pparc .
the data were reduced using figaro software and computers provided by starlink .
thanks to t. r. marsh for the use of his molly software for spectral analysis .
hq acknowledges receipt of a pparc studentship .
rpf was supported during the period of this research initially by astron grant 781 - 76 - 017 and subsequently by ec marie curie fellowship erbfmbict 972436 .
we thank guy pooley for making the radio observations for us , and acknowledge useful discussions with andre van teeseling , peter kahabka , rob hynes and martin still .
becker c. m. , remillard r. a. , rappaport s. a. , mcclintock j. e. 1998 , apj , submitted beuermann k. et al .
1995 , a&a , 294 , l1 cardelli j. a. , clayton g. c. , mathis j. s. 1989 , apj , 345 , 245 crampton d. et al .
1996 , apj , 456 , 320 fender r.p .
, bell burnell s.j . , 1996 , in : ` rontgensstrahlung from the universe ' , eds .
zimmerman , j.e .
trumper , h. yorke . ,
mpe report 263 , p135 greiner j. , wenzel w. 1995 , a&a , 294 , l5 hachisu i. & kato m. 1998 , apj , submitted horne k. 1986 , pasp , 98 , 609 kahabka p. , van den heuvel e. p. j. 1997 , ara&a , 35 , 69 meyer - hofmeister e. , schandl s. , meyer f. 1997 , a&a , 321 , 245 shahbaz t. , livio m. , southwell k.a . , charles p.a .
1997 , apj , 484 , l59 southwell k. a. et al .
1996 , apj , 470 , 1065 van den heuvel e. p. j. , bhattacharya d. , nomoto k. , rappaport s. a. 1992 , a&a , 262 , 97 will t. , barwig h. 1996 , in `` supersoft x - ray sources '' , ed .
greiner j. , 99 | we present jhk photometry and spectroscopy of rx j0019.8 + 2156 .
the spectrum appears to be dominated by the accretion disc to at least 2.4 @xmath0 m , over any other source of emission .
we find paschen , brackett and he ii lines strongly in emission , but no he i. there are satellite lines approximately 850 km s@xmath1 either side of the strongest , unblended hydrogen lines .
these satellite lines may be the spectral signature of jets from the accretion disc .
accretion , accretion disks ism : jets and outflows stars : emission - line x - rays : stars infrared : stars stars : individual : rx j0019.8 + 2156 |
cataclysmic variables ( cvs ) are short period binary systems in which a roche lobe filling low mass main sequence secondary transfers mass to a white dwarf ( wd ) primary .
the transferred matter accretes onto the wd either through a disc or a stream and slowly builds up a hydrogen rich surface layer on the wd . with continuing accretion the pressure at the bottom of this layer increases , and
hydrogen burning eventually starts .
the thermodynamic conditions at ignition determine how the burning proceeds ( e.g. fujimoto 1982 ) .
if the degeneracy is very high , a thermonuclear runaway occurs , leading to a violent outburst terminated by the ejection of all or most of the accumulated envelope .
classical novae are thought to be objects undergoing such an outburst ( cf .
livio 1994 for a recent review ) .
ignition at moderate or weak degeneracy causes strong or weak h shell flashes , whereas stable or stationary hydrogen burning requires fairly high accretion rates @xmath5yr which are not expected to occur in cvs .
mass transfer in cvs is driven by orbital angular momentum losses which generally shrink the binary and maintain the semi
detached state .
the observed properties of short period cvs below the cv period gap ( orbital period @xmath6 h ) are consistent with gravitational wave radiation as the only driving mechanism .
a much stronger angular momentum loss , usually assumed to be magnetic stellar wind braking , is needed for systems above the gap ( @xmath7 h ) .
the assumption that magnetic braking ceases to be effective once the secondary becomes fully convective in turn provides a natural explanation for the period gap as a period regime where the systems are detached and therefore unobservable ( spruit & ritter 1983 , rappaport et al .
the resulting typical mass transfer rate @xmath8 in cvs is @xmath9yr below the gap and @xmath10yr above the gap ( see e.g. king 1988 , kolb 1996 , for reviews ) . for negligible wind losses from the system
the accretion rate is essentially the same as the transfer rate . with the above typical values the h ignition on the wd turns out
to be degenerate enough to cause more or less violent outbursts ( e.g. prialnik & kovetz 1995 ) . as the mass to be accumulated before ignition is very small ( @xmath11 ) the outbursts recur on a time @xmath12 much shorter than the mass transfer timescale which determines the long term evolution .
studies of the secular evolution of cvs make use of this fact and replace a sequence of nova outbursts with given recurrence time @xmath13 and ejected envelope mass @xmath14 by a continuous isotropic wind loss from the wd at a constant rate @xmath15 .
such a procedure obviously neglects any effect that nova outbursts may have on the long term evolution .
these are in particular + ( i ) the evolution of the system is not continuous but characterized by sudden changes of the orbital parameters , causing the mass transfer rate to fluctuate around the continuous wind average value ( see sect . 2 ) .
it is not a priori clear if the continuous wind average properly describes the system s evolution close to mass transfer instability .
+ ( ii ) at visual maximum ( and the following decline ) nova envelopes have pseudo photospheric radii of typical giants , i.e. much larger than the orbital separation .
therefore the secondary is engulfed in this envelope and possibly interacting with it .
drag forces on the secondary moving within the envelope can lead to frictional angular momentum loss ( faml ) from the orbit , and accretion of envelope material onto the secondary could increase its photospheric metal abundances ( pollution ) .
+ ( iii ) the h burning hot wd is extremely luminous ( @xmath16 , compared with accretion luminosities @xmath17 ) for as long as a few years .
this might drive additional mass loss from the secondary star . in this paper we will focus on the effects of mass loss discontinuities and faml .
we neglect irradiation as it does not last long enough to influence the long term evolution .
pollution is expected to be important only for metal poor secondaries ( stehle 1993 ) ; we neglect it altogether . in sect .
2 we formally derive the continuous wind average and consider the stability of mass transfer in the presence of nova discontinuities analytically , with faml of arbitrary strength .
we review previous studies on faml and follow livio et al .
@xcite to derive a simple quantitative model for faml in sect .
3 . using this description
we perform numerical calculations of the long term evolution of cvs with various strengths of faml , both for sequences of nova outbursts and the continuous wind average .
results of such computations where the faml strength and the ejected mass per outburst have been varied systematically are shown in sect .
sequences with faml parameters taken from the consistent set of nova models by prialnik & kovetz @xcite are shown at the end of sect .
section 5 discusses our results .
we begin by investigating how mass transfer discontinuities induced by nova outbursts affect mass transfer stability . introducting the well known conservative mass transfer stability criterion
we develop a formalism to extend its applicability to the discontinuous case .
the strength of faml enters as a free parameter . following ritter @xcite the mass transfer rate @xmath8 in a cv
can be approximated by @xmath18 ( note that @xmath8 is always positive ) . here
both @xmath19 @xmath20 and the ratio @xmath21 of the photospheric pressure scale height @xmath22 to the donor s radius @xmath23 are roughly constant for the secondaries under consideration .
the secondary s roche radius @xmath24 can be written as a fraction @xmath25 of the orbital distance @xmath26 , @xmath27 @xmath25 depends only on the mass ratio @xmath28 and is for @xmath29 given by @xmath30 to better than @xmath31 ( paczyski 1971 ) . to assess how @xmath8 changes with time and to find a stationary value for @xmath8 ( where @xmath32 ) we have to consider both @xmath33 and @xmath34 . from the total orbital angular momentum @xmath35 , @xmath36 and the derivative @xmath37 ( @xmath38 ) we find with ( [ eq-2.1.3 ] ) @xmath39 where @xmath40 , @xmath41 and @xmath42 denote the primary mass , donor mass and total mass , respectively . in ( [ eq-2.1.5 ] )
we formally separate `` mass loss related '' angular momentum loss @xmath43 , caused by mass which leaves the binary and carries a certain specific angular momentum @xmath44 ( quantified by the dimensionless parameter @xmath45 ) , i.e.@xmath46 and `` systemic '' angular momentum loss @xmath47 which operates without ( noticeable ) mass loss , e.g. gravitational wave radiation and magnetic braking . by combining the mass changing terms ( [ eq-2.1.5 ] )
is usually rewritten as @xmath48 thus defining the mass radius exponent @xmath49 of the secondary s roche radius . in the simple case of conservative mass transfer where @xmath50 constant , i.e. @xmath51 , we find from ( [ eq-2.1.5 ] ) @xmath52 in the more general case of an isotropic wind loss at a rate @xmath53 ,
i.e. @xmath54 , we obtain @xmath55 an expression especially useful when @xmath56 and @xmath45 are constant .
figure [ fig : zetaroche ] depicts @xmath57 and @xmath49 with @xmath58 and either @xmath59 or @xmath60 as a function of @xmath41 for @xmath61 .
similarly , it is standard practice to decompose the secondary s radius change into the adiabatic response @xmath62 and the thermal relaxation @xmath63 , @xmath64 fig .
[ fig : zetastar ] shows the adiabatic mass radius exponent @xmath65 as a function of stellar mass for low
mass zams secondaries ( hjellming 1989 ) .
fully convective stars and stars with deep convective envelopes ( @xmath66 ) have @xmath67 .
using these definitions the time derivative of @xmath8 according to ( [ eq-2.1.2 ] ) becomes @xmath68 \ : , \label{eq-2.1.9}\ ] ] with @xmath69 and @xmath70 the binary attempts to settle at the stationary mass transfer rate @xmath71 ( @xmath32 for @xmath72 ) .
the stationary rate is _ stable _ if @xmath73 at @xmath72 , i.e. if the system opposes any instantaneous perturbation in @xmath8 .
this translates into @xmath74 , hence the familiar criterion @xmath75 a similar stability criterion , @xmath76 , against thermal timescale mass transfer can be derived ( see e.g.ritter 1996 ) , where @xmath77 is the thermal equilibrium mass radius exponent ( @xmath78 for low mass main sequence stars , see fig .
[ fig : zetastar ] ) .
the differential equation ( [ eq-2.1.9 ] ) governs the evolution of the mass transfer rate as a function of time @xmath79 . for constant @xmath80 and @xmath81
the solution of ( [ eq-2.1.9 ] ) is @xmath82 where @xmath83 denotes the initial value of @xmath8 .
hence any changes of the transfer rate proceed on the characteristic timescale @xmath84 , i.e. @xmath85 is a small fraction @xmath86 of the systemic angular momentum loss timescale @xmath87 . in reality
both @xmath80 and @xmath81 change with time , typically on a the secular timescale @xmath88 , but for most practical cases where we investigate mass transfer stability against short timescale perturbations @xmath80 and @xmath81 can be considered as constant .
formally ( [ eq-2.2.1 ] ) is a solution of ( [ eq-2.1.9 ] ) even if @xmath89 ( i.e. @xmath90 ) . in this case , assuming dynamical stability ( @xmath74 ) , @xmath91 is negative and no longer a stationary value for the mass transfer rate .
rather ( [ eq-2.2.1 ] ) shows that in this case @xmath8 decreases exponentially , i.e. the system detaches ( note that @xmath92 is a physical mass transfer rate and as such always positive ) .
if finally the mass transfer is unstable ( @xmath93 ) then we see from ( [ eq-2.1.9 ] ) that unless @xmath80 has a large negative value the transfer rate grows .
the growth timescale is initially @xmath85 , but becomes shorter and shorter with further increasing @xmath8 .
so far we considered only continuous ( though not necessarily stationary ) mass transfer .
events that change orbital parameters on a timescale much shorter than the characteristic time @xmath94 yr for reestablishing the local stationary value @xmath91 may be regarded as discontinuous and instantaneous .
nova outbursts certainly belong to this category ; the nova envelope expands beyond the orbit within less than a few @xmath95 days after ignition and returns within @xmath96 years ( actually these are upper limits for the slowest novae ) . in the following we
apply the above formalism separately to the outburst phase and the inter
outburst phase , then combine them to describe the full nova cycle . as a result of the outburst the envelope mass @xmath14
is ejected .
we expect that the specific angular momentum carried away by @xmath14 is higher than the specific orbital angular momentum of the wd due to dynamical friction of the secondary orbiting within the envelope ( faml ) .
hence we write for the discontinuous change of the orbital angular momentum @xmath97 where @xmath98 accounts for the specific orbital angular momentum @xmath99 of the wd ( represented as a point mass ) .
the free parameter @xmath100 measures the strength of faml and will be estimated in terms of a simple model for the frictional processes in sect . 3 below . correspondingly , using ( [ eq-2.1.5 ] ) with @xmath101 and @xmath102 ( i.e. neglecting the small amount of mass accreted onto the secondary as well as any mass transfer during this short phase )
, the change @xmath103 of the roche radius is @xmath104 after the outburst mass transfer continues .
once the mass accreted onto the wd exceeds the critical ignition mass @xmath105 the next outburst occurs . for simplicity
we assume _ conservative _ mass transfer during the inter
outburst phase .
according to ( [ eq-2.1.7 ] ) the total change of the roche radius in the inter
outburst phase is @xmath106 where @xmath107 is the total change of the orbital a.m. due to systemic losses between outbursts . for the study of the long term evolution of cvs it is convenient to replace the sequence of nova cycles with mass transfer discontinuities by a mean evolution where the mass @xmath14 is regarded as being lost continuously at a constant rate over the cycle in form of an isotropic stellar wind carrying the specific orbital angular momentum @xmath108 .
if @xmath109 denotes the corresponding roche lobe index then we have from ( [ eq-2.1.7 ] ) @xmath110 again with @xmath111 because we neglect any change of @xmath112 during outburst .
as the total change of the roche radius over a complete cycle is @xmath113 , comparison with ( [ eq-2.3.2 ] ) and ( [ eq-2.3.4 ] ) gives @xmath114 this can be written as @xmath115 here @xmath116 specifies the change of the wd mass _ during outburst _ in units of the mass lost from @xmath41 _ during the inter
outburst phase_. equation ( [ eq-2.3.7 ] ) in fact is equivalent to ( [ eq-2.1.14 ] ) with @xmath117 and @xmath118 ; @xmath119 determines if the wd mass will grow or shrink in the long term evolution .
the generalization of ( [ eq-2.3.7 ] ) which allows for non conservative mass transfer between outbursts and a change of @xmath41 during outburst is given in the appendix . as was shown in sect .
2.1 mass transfer in the fictious mean evolution which mimics the effect of nova outbursts by a continuous wind loss is dynamically stable if @xmath120 clearly , in the case of discontinuous mass transfer with nova cycles the stability considerations leading to this criterion are no longer applicable , as the system never settles at a stationary transfer rate . after an outburst
the transfer rate @xmath8 follows the solution ( [ eq-2.2.1 ] ) and approaches the conservative stationary rate @xmath121 ( where @xmath122 is @xmath123 according to ( [ eq-2.1.11 ] ) with @xmath57 ) . at the next outburst
@xmath8 changes discontinuously from the pre
outburst value @xmath124 to the ( new ) post outburst value @xmath125 , i.e. increases ( or drops ) by a factor @xmath126 \ : , \label{eq-2.4.2}\ ] ] see eqs .
( [ eq-2.1.2 ] ) and ( [ eq-2.3.2 ] ) . in the following we
consider mass transfer stability in the presence of nova cycles and derive a generalized stability criterion that reduces indeed to the simple form ( [ eq-2.4.1 ] ) . in our approach
we compare the mass transfer rate @xmath127 and @xmath128 _ immediately before _
subsequent outbursts @xmath129 and @xmath130 in a sequence of cycles with constant @xmath131 , @xmath14 , @xmath122 and @xmath80 .
( [ eq-2.2.1 ] ) with @xmath132 describes @xmath8 as a function of time between outbursts . at time
@xmath133 when the outburst @xmath130 ignites , the mass transferred since the last outburst is just @xmath131 , i.e. @xmath134 this can be solved for the outburst recurrence time @xmath13 ( cycle time ) @xmath135 \ : .
\label{eq-2.4.3}\ ] ] inserting @xmath13 for @xmath79 in ( [ eq-2.2.1 ] ) gives the new pre
outburst value @xmath136 , @xmath137 \\
\nonumber + x^i_{\rm pre } \exp \left ( -{\cal b}^f { \delta m_{\rm ig}}\right ) \ : , \end{aligned}\ ] ] where we used ( [ eq-2.4.2 ] ) to replace @xmath138 by @xmath139 , ( [ eq-2.3.6 ] ) and the definition of @xmath140 analogous to @xmath122 , i.e. @xmath141 .
hence the relative change of the pre
outburst value is @xmath142 - \left [ 1- \exp \left ( -{\cal b}^f{\delta m_{\rm ig}}\right ) \right ] } \\
\nonumber \lefteqn { \simeq { \delta m_{\rm ig}}\left ( \frac{{\cal a}}{x^i } - { \cal b}^f \right ) } \end{aligned}\ ] ] ( the subscripts for all @xmath8 have been dropped as they all read `` pre '' )
. we will refer to @xmath143 as the growth function . for @xmath144 the mass transfer rate
_ grows _ from outburst to outburst , for @xmath145 it decreases , on a timescale @xmath146 .
formally , @xmath147 for @xmath148 , with @xmath149 @xmath150 represents the stationary _ pre outburst _ mass transfer rate if it is positive , i.e. if @xmath80 and @xmath140 have the same sign . for it to be stable we require @xmath151 at @xmath152 so that the system evolves back to @xmath150 after a perturbation increases / decreases @xmath8 .
this is the case if , and only if , @xmath153 .
then the stationary solution exists only if @xmath154 , in all other cases the system either detaches , or the transfer rate grows unlimited .
figure [ fig : fhat ] summarizes schematically the formal functional dependency of @xmath143 on @xmath155 for various combinations of the signs of @xmath80 and @xmath140 ( see table [ tab : fhat ] ) .
hence the generalized stability criterion for discontinuous nova
amplified mass transfer is @xmath156 , which is indeed equivalent to ( [ eq-2.4.1 ] ) . remarkably , this condition is independent of the sign of @xmath122 , i.e. the system can evolve in a stable manner with formally dynamically unstable conservative mass transfer between subsequent outbursts .
.branches of possible solutions of eq .
( [ eq-2.4.5 ] ) as shown in fig .
[ fig : fhat ] . [ cols="^,^,^,^ " , ] the analysis of the system s behaviour once it violates condition ( [ eq-2.4.1 ] ) is not straightforward .
the main problem is that @xmath80 is expected to change its sign ( from positive to negative ) at about the same time as @xmath140 .
this can be seen from an estimate of the thermal relaxation term @xmath157 if the real evolution is replaced by the continuous wind average : stehle et al .
@xcite have shown that the secular evolution rapidly converges to an attracting evolutionary path characterized by @xmath158 whatever the initial configuration of the system .
assuming stationarity , equating ( [ eq-2.4.8 ] ) with ( [ eq-2.1.7 ] ) , and ( [ eq-2.1.8 ] ) with ( [ eq-2.1.7 ] ) gives after elimination of @xmath159 @xmath160 as an estimate for @xmath80 .
hence @xmath153 during a phase of stable , stationary mass transfer , but @xmath161 once @xmath162 .
in practice this means that when a stable system , characterized by the full line in fig .
[ fig : fhat ] , approaches instability , the curve becomes flatter and flatter with @xmath150 almost constant , and then inverts to the ( unstable ) long dashed line .
in the previous section we derived analytic expressions describing stationary mass transfer and mass transfer stability in the presence of nova cycles . a non zero faml effect was allowed , and the strength of faml was treated as a free parameter @xmath100 . to proceed further and underline the above findings with numerical examples we quantify @xmath100 by adopting a model for the frictional processes leading to faml .
macdonald @xcite discussed for the first time the possibility that an interaction between the secondary and the extended envelope might tap energy from the orbit and thereby reduce the nova decay time .
he determined the angular momentum transferred from the orbit to the envelope by describing the nova envelope as a polytrope with index 3 at rest , and restricting the combined nuclear , internal and frictional energy generation to the eddington luminosity ( macdonald 1986 ) .
shara et al .
@xcite and livio et al .
@xcite used a more direct way , based on bondi hoyle accretion , to estimate the transfer of angular momentum .
their approach ( which we adopt below ) offers a very simple analytic treatment , as the explicit expression for faml has only one free parameter , the envelope expansion velocity @xmath0 , measuring the strength of faml .
more recently , kato & hachisu @xcite have again confirmed their earlier result that faml has only minor influence on the decay phase of nova outbursts .
this can be understood in terms of the high expansion velocities in their models and will be discussed in sect . 5 . on the other hand ,
lloyd et al .
@xcite showed that common enevlope evolution can contribute to the shaping of the nova remnant . in bondi
hoyle accretion linear momentum is transferred through the drag force @xmath163 where @xmath164 is the gas stream velocity at infinity , @xmath165 the accretion radius ( see below ) and @xmath166 the density of the accreted medium .
the dimensionless drag coefficient @xmath167 varies with mach number @xmath168 and specific heat ratio @xmath169 and is of order unity .
equation ( [ eq : fdrag ] ) is strictly valid for the highly supersonic case , but also applicable for @xmath170 if an adequate interpolation for the accretion radius is used , e.g. @xmath171 ( shima et al . 1985 ) ; @xmath172 is the local sound speed .
detailed 3dim .
hydrodynamical calculations summarized in ruffert @xcite confirm that ( [ eq : fdrag ] ) describes the subsonic case as well when @xmath173 is replaced by the geometrical cross section .
but these simulations also reveal additional complexities not represented in the simple form ( [ eq : fdrag ] ) , e.g. the influence of the accretor s size , even in the supersonic case .
most important , the drag coefficient can deviate significantly from unity .
kley et al .
@xcite point out that radiation pressure may reduce @xmath167 by a factor of @xmath174 because it increases the sound speed and thus lowers @xmath168 to the subsonic case . given these uncertainties , and
the fact that idealized bondi hoyle accretion is merely a rough approximation to the situation of a secondary orbiting in an expanding nova envelope , the quantitative expressions for faml derived below are order of magnitude estimates only .
we now consider the faml situation in a simplified geometry ( fig .
[ fig : sketch ] ) : the wd is at rest in the centre of a spherically symmetrical expanding nova envelope . then the secondary s orbital velocity @xmath175 ( typically @xmath176 km / s ) is perpendicular to the wind expanding with @xmath0 .
hence the velocity of the accretion flow with respect to the secondary is @xmath177 as @xmath178 the ratio of accretion radius ( [ eq : racc ] ) to stellar radius @xmath23 is @xmath179 ( where @xmath180 is the quadratic mean of @xmath181 , @xmath0 and @xmath172 ) , suggesting that @xmath182 for most cases . setting @xmath183 and @xmath184 we obtain for the loss of angular momentum due to friction from the tangential component of the drag force ( [ eq : fdrag ] ) @xmath185 using the continuity equation @xmath186 to eliminate @xmath166 , ( [ eq : vsec ] ) and ( [ eq-2.1.3 ] ) then give @xmath187^{\rm 2 } \ : \frac{v_{\rm rel}}{v_{\rm exp } } \ : \dot{m}_{\rm 1 } \ : \sqrt{g\:m\:a } .
\label{eq : jdotfaml2}\ ] ] if most of the mass and angular momentum can be assumed to be lost at roughly constant rate and velocity ( livio et al .
1991 ) we finally obtain the total faml , integrated over the outburst , @xmath188 where @xmath189^{\rm 2 } , \label{eq : aofq}\ ] ] @xmath190 and @xmath191 comparison with ( [ eq-2.3.1 ] ) then shows that @xmath192 i.e. the mean specific a.m. @xmath193 of the ejected material is @xmath194 the functional dependence on @xmath195 in eq .
( [ eq : k1depend ] ) , shown in fig .
[ fig : j_valid ] ( top panel ) , leads to the unphysical limit @xmath196 for a very slow envelope expansion @xmath197 .
this is mainly due to the neglect of the envelope spin up which would effectively reduce the relative velocity ( livio et al .
1991 ) .
the secondary can not spin the envelope up to speeds faster than corotation , so @xmath193 is certainly smaller than @xmath198 ( where @xmath199 is the binary s angular velocity ) . if only a certain fraction of the envelope is spun
up to corotation this upper limit is correspondingly smaller , e.g. @xmath200 if only the torus traced by the secondary s orbital motion corotates .
equation ( [ eq : nufaml ] ) specifies our model of faml in terms of a single parameter , @xmath201 ( or @xmath0 ) and determines the value of @xmath109 according to eq .
( [ eq-2.3.7 ] ) .
we use ( [ eq : jmax ] ) as a physical upper limit on the effect of faml . to illustrate the effect of faml in the parametrization ( [ eq : nufaml ] ) we consider the resulting relative change @xmath202 of a system parameter @xmath203 as a result of the nova outburst . for the orbital distance
we obtain , similar to ( [ eq-2.3.2 ] ) , @xmath204 which translates into @xmath205 for the orbital period @xmath206 . the approximate expressions in ( [ eq : daa ] ) and ( [ eq : dpp ] ) make use of the weak dependence on @xmath207 ; @xmath208 is @xmath209 . for a given ejection mass and white dwarf mass
, @xmath210 critically determines the magnitude of the orbital change . from ( [ eq-2.4.2 ] ) and ( [ eq-2.3.2 ] )
the change of the mass transfer rate @xmath211 is @xmath212 where we used @xmath213 to obtain the last line .
figure [ fig : m_jumps ] illustrates eq . ( [ eq : dmm ] ) . depending on the strength of faml and the mass ratio
the changes in system parameters can be either positive or negative .
the critical @xmath214 where the amplitudes vanish are shown in fig .
[ fig : k1crit ] . the critical lines for @xmath206 and
@xmath8 almost coincide and separate the two different outburst types , those which increase the mass transfer rate _ and _ decrease the orbital period , and those which decrease the transfer rate _ and _ increase the period . to consider the effect faml has on the continuous wind average evolution we first note that faml is a consequential angular momentum loss ( caml ; cf .
king & kolb 1995 ) as @xmath215 is proportional to the mass transfer rate , see ( [ eq : deltafaml ] ) with @xmath216 ( where @xmath8 denotes the continuous wind average mass transfer rate ) . as such
faml amplifies the transfer rate which would be driven by systemic angular momentum losses alone in the absence of faml .
an analytic estimate of this amplification factor , the ratio @xmath217 of the mass transfer rate with and without faml , can be obtained if we assume that the system follows the corresponding uniform evolutionary track found by stehle et al .
@xcite also in the case with faml .
this is a reasonable assumption as long as the amplification is moderate and the resulting mass transfer timescale longer than the thermal time of the secondary .
the radius reaction of the secondary along its track is given by ( [ eq-2.4.8 ] ) . hence using @xmath218 and ( [ eq-2.1.7 ] ) we have @xmath219 here @xmath220 const . and
@xmath49 is from ( [ eq-2.1.14 ] ) with @xmath58 , @xmath59 .
we emphasize that the ratio @xmath221 does not depend on the mass @xmath14 ejected per outburst , rather only on @xmath201 and the mass ratio @xmath207 .
[ fig : factor_fix ] shows @xmath221 as function of @xmath222 for various values of @xmath201 .
obviously faml amplification is always small for large @xmath207 ( small @xmath223 ) . for a given @xmath201
it increases with decreasing @xmath207 and formally goes to infinity where the system approaches thermal instability ( @xmath224 ) . the full line in fig .
[ fig : factor_fix ] indicates the upper limit for @xmath221 if the specific angular momentum the ejected envelope can carry is limited by @xmath225 ( see sect .
here we apply the faml description introduced in the previous section and test the analytical considerations of sect . 2 and 3 explicitly with numerical sequences of the secular evolution of cvs .
the above faml model contains @xmath201 , @xmath131 and @xmath14 as free parameters .
unfortunately both observations and theoretical models describing the outburst itself yield often contradicting and inconsistent estimates for these parameters .
therefore we restrict the following investigation to a simple parameter study to obtain a systematic picture of the effect of faml with a given strength on the secular evolution .
in particular we present calculations with constant @xmath131 ( @xmath226 ) and @xmath201 , for various values of @xmath131 and @xmath201 . to model the binary evolution numerically
we describe the secondary star by either full stellar models using mazzitelli s stellar evolution code ( e.g. mazzitelli 1989 ) , or by a simplified bipolytrope structure using the generalized bipolytrope code ( kolb & ritter 1992 ) .
both codes have been modified to include faml and to resolve individual nova outbursts .
the bipolytrope code is about 4 orders of magnitude faster than the full code and allows one to compute the secular evolution for several gyr with every single nova outburst fully resolved .
several sequences calculated with the full code serve as an independent check of the simplified description . despite a careful calibration of the bipolytrope model to full stellar models , allowing a quantitative match of the results to usually better than 10% , there are well known limitations of the simplified description ( see kolb & ritter 1992 ) , e.g. the increasing deviation of @xmath65 from the values of actual stars for @xmath227 .
but these effects are negligible for the purpose of this paper . in all the examples shown below
we compute magnetic braking according to verbunt & zwaan @xcite with the calibration parameter set to unity .
we have performed calculations with the following two modes : 1 .
fully resolved : mass transfer is conservative until the mass accreted on the wd exceeds the ignition mass . in the time step immediately thereafter the ejecta mass and the angular momentum ( [ eq : deltafaml ] ) is removed from the system .
thus fully resolved evolutions are discontinuous and show repeating nova cycles .
continuous wind average : mass loss from the system is continuous at a rate equal to the transfer rate , carrying the _ increased _ specific angular momentum according to ( [ eq : jej ] ) ; hence the secular evolution is continuous .
we generated short stretches of fully resolved evolutionary sequences with full stellar models by switching into the fully resolved mode in the middle of a continuous wind average calculation for a typical cv above the period gap .
the mode switch was made from a model well established on the uniform evolutionary track characterized by ( [ eq-2.4.8 ] ) .
[ fig : down ] shows for the case of weak faml with @xmath228 a sawtooth ( or shark fin ) like modulation of the mass transfer rate over the orbital period ( full line ) .
the system evolves from right to left ; the crosses mark intervals of 1000 yr . as expected from fig .
[ fig : k1crit ] the outburst results in a _ longer _ period and a _ lower _ mass transfer rate with this choice of parameters . in the case of strong faml with @xmath229 ( fig .
[ fig : up ] ) the outburst leads to a _
shorter _ period and a _ higher _ mass transfer rate . the approximate time evolution according to eq . ( [ eq-2.2.1 ] ) , with @xmath230 and @xmath122 evaluated from the full sequence at the post outburst position , marked with a filled diamond , is shown as a dotted curve and matches the full sequence very well .
the dash dotted line in figs . [ fig : down ] & [ fig : up ] indicates the level of the local stationary conservative mass transfer rate as given by eq .
( [ eq-2.1.15 ] ) with @xmath231 .
note that there is a difference between this value and the mass transfer rate of a completely conservative evolution with the same initial values ( but no outbursts ) as the system here does _ not _ evolve conservatively on average but loses mass together with angular momentum . the averaged stationary mass transfer rate which includes faml with @xmath109 according to eq .
( [ eq-2.3.7 ] ) is shown as a dashed line .
this value is identical to the mass transfer rate calculated in the averaged mode and to the time average of the full evolution .
the system approaches the conservative value between outbursts and oscillates around the average transfer rate .
note that if , beginning with negligible faml , the strength of faml is continuously increased , the outburst amplitudes will decrease until the stationary ( local ) conservative transfer rate and the continuous wind average ( which grows with faml ) are equal .
further increase of faml will now lead to growing amplitudes , but with opposite signs . fig .
[ fig : mdotpfaml ] compares the continuous wind average evolution for a reference system ( @xmath232 ; @xmath233 at turn on ; @xmath234 ) with strong faml and without faml , computed with full stellar models . as expected from fig .
[ fig : factor_fix ] the mass transfer rate with faml is significantly larger than in the case without only at long periods , where the system is close to ( but not beyond ) thermal instability . due to the higher mass transfer rate the secondary in the faml sequence
is driven further out of thermal equlibrium and therfore larger at the upper edge of the period gap .
at the same time the secondary s mass is smaller upon entering the detached phase .
both effects , well known from systematic studies of cv evolution ( e.g. kolb & ritter 1992 ) , cause a wider period gap by both increasing the period at the upper edge and decreasing the period at the lower edge of the gap .
the analytical considerations of sect .
3.3 are confirmed by a parameter study performed with the bipolytrope code .
the evolutionary sequences with all outbursts resolved depicted in fig .
[ fig : satz1 ] summarize the results .
the reference system is shown in the left column , a less massive system in the right column .
the strength of faml increases from top to bottom , with consequences for the outburst amplitudes as discussed in sect . 3.2 and in the previous subsection .
note that the sawtooths here _ do not represent _ single outburst cycles , but arise because we connect the pre and post outburst values for every n@xmath235 ( @xmath236 or @xmath237 ) outburst only .
the advantage of this display method is that both magnitude and sign of the outburst amplitude ( the vertical flanks ) can be followed through the evolution . in reality
the system undergoes many individual cycles like those in figs .
[ fig : down ] & [ fig : up ] between subsequent outbursts depicted in fig .
[ fig : satz1 ] .
the two evolutionary tracks plotted in each panel differ only in the ignition mass @xmath131 ( here @xmath238 ) .
the thick line corresponds to @xmath239 , the thin line to @xmath240 . in the scale
used the low ignition mass sequences are practically indistinguishable from the continuous wind average evolution ( remember that both oscillate around the _ same average value _ , cf . sect . [ sec-4.5 ] ) .
inspection of fig .
[ fig : m_jumps ] shows that the growth of amplitudes towards shortest periods is mainly due to the increase of @xmath207 . only in the high faml ( low @xmath0 ) case , i.e. in the very steep region of eq .
( [ eq : k1depend ] ) , there is an additional contribution from the increasing orbital velocity ( hence decreasing @xmath195 ) at short @xmath206 . to illustrate and test the analytical mass transfer stability considerations of sects .
2.3 and 2.4 explicitly we consider the secular evolution of a system which runs into an instability , i.e. violates the formal stability criterion ( [ eq-2.4.1 ] ) at some point of the evolution .
to establish this situation we artificially increase the strength of faml along a fully resolved evolutionary sequence ( computed with the bipoltrope code , with constant @xmath241 ) , starting from a small value at the onset of mass transfer . as a consequence @xmath109 becomes larger than @xmath65 at some intermediate secondary mass .
in particular , we adopt @xmath242 as the functional form of the envelope expansion velocity .
( @xmath41 will be large enough to avoid negative @xmath0 ) .
the corresponding @xmath109 with @xmath100 from ( [ eq : nufaml ] ) becomes larger than @xmath65 at @xmath243 , and the mass transfer rate indeed begins to grow without limit at about this mass well above the period gap ( fig . [
fig : runaway ] ) .
a closer look at the final phase immediately before the runaway ( fig .
[ fig : runzoom ] ) shows that the system evolves beyond the formally unstable point apparently unaffected and enters the runaway mass transfer phase only @xmath244 later .
in fact this is just as we would expect if the evolution were replaced by the corresponding continuous wind average . in this case
( [ eq-2.1.9 ] ) with @xmath245 describes the variation of the ( average ) mass transfer rate @xmath8 .
the characteristic timescale on which @xmath8 changes is @xmath246 , which becomes very large when @xmath247 , i.e. around @xmath248 at the instability point ; note that @xmath230 is also small at this point , see ( [ eq-2.4.9 ] ) .
proceeding further increases the strength of faml even more and makes @xmath140 more negative , hence the timescale @xmath249 progressively shorter . the time delay until the runaway finally begins is essentially determined by the crossing angle between the functions @xmath250 and @xmath251 and the value of @xmath8 at this point .
figure [ fig : frun ] shows the growth function @xmath252 defined in ( [ eq-2.4.5 ] ) for selected models immediately before the runaway for the evolutionary sequence depicted in fig .
[ fig : runzoom ] .
the different curves are computed with data from the run taken at the times as labelled ( cf .
[ fig : runzoom ] ) . as discussed in sect .
[ sec-2.4 ] the change of @xmath252 from a stable curve with negative slope to an unstable curve with positive slope occurs when @xmath248 .
the sequence shown in figs [ fig : runaway]-[fig : frun ] is typical of the behaviour of evolutionary sequences with nova cycles and very strong faml .
it convincingly demonstrates that the faml
amplified istropic wind average properly describes the effect of a sequence of nova outbursts , even in the most extreme case when the system approaches a faml induced instability .
a more systematic investigation of the significance of this instability for the secular evolution of cvs will be published elsewhere ( kolb et al .
, in preparation ) .
the faml parameters @xmath201 , @xmath14 and @xmath131 are certainly not constant along the evolution but depend on the actual state of the outbursting system .
the three governing parameters which crucially determine the outburst characteristics of thermonuclear runaway ( tnr ) models for classical novae are the wd mass @xmath40 , the ( mean ) accretion rate @xmath8 and the wd temperature @xmath253 ( e.g. shara 1989 ) . both the mass and
the temperature of the wd change only slowly during the binary evolution , thus the dominant dependence of @xmath0 is from the ( mean ) mass transfer rate .
as faml itself drives the mass transfer this could lead to an interesting feedback and possibly self amplification in the system , a question considered in a separate paper ( kolb et al .
, in preparation ) .
although published tnr models mostly give values for @xmath131 and estimates for @xmath14 , few or no data are available on @xmath201 .
unfortunately this is also true for the most complete and consistent set of published nova models ( prialnik & kovetz 1995 ) which we chose to use as input to the faml description derived in sect .
[ sec-3.1 ] .
these authors tabulate only the average expansion velocity @xmath254 , the time average of the _ maximum _ velocity in the flow during the whole mass loss phase . in the absence of more relevant velocity data we simply set @xmath255 , but we are aware that this choice represents an upper limit for @xmath0 ( which has to be taken at the secondary s position ) .
two exemplary relations @xmath256 are shown in fig . [ fig : vav ] , for a low and a high mass wd , complemented by large values ( @xmath257 km / sec , hence negligible faml ) when no envelope ejection was found in the hydrodynamical simulations .
the evolutionary sequences obtained with @xmath258 ( using the velocity data for nova models with a @xmath259 wd ) and with @xmath260 are shown in figs .
[ fig : dina_high ] & [ fig : dina_low ] .
the changes of @xmath261 during the whole evolution are very small ( @xmath262 ) . as in fig .
[ fig : satz1 ] the sawtooths result from connecting only every @xmath263 outburst . for comparison
a standard continuous wind average evolution without faml but otherwise identical parameters has been overlaid in thick linestyle . as expected from the high expansion velocities faml
has very little effect on the global evolution .
by choice we underestimated the strength of faml in the sequences shown above .
but although the expansion velocity at the location of the secondary is indeed smaller than @xmath254 in the models by prialnik & kovetz @xcite they currently appear to be too large to cause a faml effect with significance for the long term evolution of cvs ( cf .
also discussion in next section ) .
perhaps the most interesting feature of the sequences presented in this paragraph is the large outburst amplitude of the mass transfer rate in low
mass wd ( @xmath264 ) systems below the period gap .
the reason for this lies mainly in the fact that @xmath14 and @xmath131 to a first approximation scale like @xmath265 ( fujimoto 1982 , @xmath266 is the wd radius ) , i.e. increase with decreasing wd mass ( and to a lesser extent , increase with decreasing @xmath8 ) .
moreover the slow but continuous decrease of the wd mass leads to a further increase of the outburst amplitudes towards shorter orbital period in this sequence ( fig . [
fig : dina_low ] ) . note that this effect adds to the ones observed before ( lowermost panels of fig .
[ fig : satz1 ] ) which were caused by the change of @xmath207 due to decreasing @xmath41 , and to a lesser extent to the slowly increasing orbital velocity @xmath181 , mimicing increasing faml strength .
such amplitudes account for a non
negligible time interval with a secular mean mass transfer rate significantly _ below _ the continuous wind average transfer rate , i.e. below @xmath267/yr , the typical value for non degenerate cvs below the period gap driven by gravitational radiation alone . as an example fig .
[ fig : problow_t ] plots for a model at @xmath268 h from the sequence shown in fig .
[ fig : dina_low ] the time spent with a mass transfer rate _ below _ a value @xmath8 , as a function of @xmath8 .
one nova cycle lasts @xmath269 myr , and the mass transfer rate remains for 0.5 myr after the outburst below @xmath270/yr .
intrinsically , these cvs with low - mass and moderate mass wds form the vast majority ( e.g. de kool 1992 , politano 1996 ) .
thus , ignoring selection effects , we would expect to observe 1 out of 10 - 20 short period cvs with small sub
gr driven mass transfer rate .
this seems to offer an alternative explanation for the low mass transfer rate sproats et al .
@xcite claim to find observationally in so called tremendous outburst amplitude dwarf novae ( toads ) which howell et al .
@xcite interpret as post minimum period systems .
as the secular mean mass transfer rate is predicted to drop substantially ( @xmath271/yr ) cvs which have evolved past the minimum period would be even fainter than the above post outburst cvs , and it is unclear if they are detectable at all .
we note that at least 4 objects in the list of sproats et al .
@xcite have been labelled as novae ( cf .
duerbeck 1987 ) : al com ( orbital period @xmath272 d , outburst suspected 1961 ) , vy aqr ( @xmath273 d , 1907 ) , rz leo ( @xmath274 d , 1918 ) and ss lmi ( no period known ; 1980 ) . with exception of ss lmi
they are now firmly considered to be dwarf novae .
this does not exclude a priori that the first recorded outburst ( e.g. in 1907 for vy aqr ) or even earlier ones were actually novae .
hence it is at least conceivable that the present low mass transfer rate and the toad characteristics might be a consequence of this last nova event .
the three systems with known orbital period are located at the extreme end of fig . 4 of sproats et al .
@xcite , showing large outburst amplitude and low quiescence magnitude . moreover , among the open square symbols indicating dne in their fig . 3 they are also those with lowest period _ and _ absolute magnitude in quiescence ( mass transfer rate @xmath275/yr ) .
the fact that the claimed deviation from the secular mean magnitude is shrinking with the time elapsed since the potential nova outburst ( 1961 - 1918 - 1907 ) is probably just a coincidence .
we caution that it is difficult to decide if the set of models by prialnik & kovetz @xcite properly describes classical nova outbursts on wds in cvs below the period gap .
observationally , only 4 out of 28 classical novae with determined orbital period are below the gap ( e.g. ritter & kolb 1995 ) .
this is in conflict with standard population models of cvs if the prialnik & kovetz @xcite ignition masses are used to predict an observable period distribution for novae ( kolb 1995 ) .
in this paper we considered the effects of nova outbursts on the secular evolution of cvs . as a result of these outbursts
the secular mean mass transfer rate and the orbital period are not continuous functions of time but change essentially discontinuously with every nova outburst by an amount proportional to the ejected envelope mass .
in addition , energy and angular momentum can be removed from the orbit due to dynamical friction of the secondary orbiting in the expanding nova envelope .
the discontinuous evolution with a given strength of frictional angular momentum loss ( faml ) is usually replaced by the corresponding continuous wind average evolution , where the mass and angular momentum loss associated with a nova outburst is assumed to be distributed over the inter
outburst time and to form an isotropic wind from the white dwarf .
we showed analytically that the well known mass transfer stability criterion for the latter case can also be derived from a proper analysis of the real , discontinuous process , for an arbitrary strength of faml amplification .
we specified a quantitative model for faml within the framework of bondi hoyle accretion following livio et al . @xcite . in this model
the strength of faml depends crucially on the expansion velocity @xmath0 of the envelope at the location of the secondary , being the stronger the smaller @xmath0 is .
we expect that the resulting simple one parameter description properly describes the order of magnitude of the faml effect and , more importantly , the differential dependences on fundamental binary parameters .
hence although it is a useful way to study the potential influence of faml systematically , it certainly can not replace a detailed modelling of the frictional processes .
calculations of the long term evolution of cvs verified the validity of the replacement of the discontinuous sequence of nova cycles with the continuous wind average , even for situations close to mass transfer instability , whatever the strength of faml .
the mass transfer rate in the continuous wind average evolution is faml amplified , i.e. by the factor ( [ eq : mdotratio ] ) larger than the transfer rate driven by systemic angular momentum losses alone . for a given cv this factor
is determined by @xmath201 alone , i.e. independent of the ejection mass .
we emphasize that faml only _ amplifies _ the transfer rate caused by systemic losses , it does not _ add _ to them .
hence faml is a particular example of consequential angular momentum losses ( caml ) investigated in detail by king & kolb @xcite . in general
, the faml amplification factor turns out to be large only when the envelope expansion is very slow ( @xmath276 , i.e. @xmath277 km / s ) _ and _ when the system is already close to thermal mass transfer instability ( fig.[fig : factor_fix ] ) .
this latter condition means that e.g. for an evolution with strong faml @xmath278 const .
the averaged mass transfer rate is significantly affected only at long orbital period , @xmath279 h. the magnitude and direction of the outburst amplitudes of the mass transfer rate @xmath8 and the orbital period @xmath206 depend on both the ejection mass @xmath14 and the faml parameter @xmath201 . for weak ( or negligible ) faml the outbursts are towards lower mass transfer rate and longer orbital period , for strong faml towards larger @xmath8 and shorter @xmath206 .
there is an intermediate regime where the outburst amplitudes essentially disappear .
theoretical models for nova outbursts generally find larger expansion velocities ( e.g. prialnik 1986 , prialnik & kovetz 1995 ) , typically @xmath280 .
this is certainly true for the terminal velocities , but in more recent models probably also for the crucial velocity at the secondary s location , i.e. closer to the wd .
kato & hachisu @xcite argue that in the wind mass loss phase of a nova outburst the main acceleration ( at about the sonic point ) takes place at a temperature where the opacity has a maximum .
thus the introduction of the opal opacities had the effect of moving this point closer to the wd .
a comparison of the radial velocity profiles found by kato & hachisu @xcite with those obtained previously ( e.g. prialnik 1986 , kato 1983 ) confirms this . as a result , in the newer models
the velocities at the location of the secondary are already quite close to their terminal values .
this explains why kato & hachisu @xcite find only marginal effects from dynamical friction .
this seems to suggest that the overall influence of faml on the long term evolution of cvs is small .
however , in view of the considerable simplifications of our faml description and the uncertainties of theoretical tnr models for nova outbursts , it is worthwhile to investigate sytematically mass transfer stability with faml , and to consider the role of a feedback between outburst characteristics ( hence faml strength ) and the mean mass transfer rate prior to the outburst .
we will study this in a forthcoming paper ( kolb et al .
, in preparation ) .
to illustrate further the effect faml might have on the secular evolution of cvs we have calculated evolutionary sequences with envelope expansion velocities , ignition and ejection masses taken from the extended set of nova models by prialnik & kovetz @xcite . as expected , the continuous wind average evolution hardly differs from the standard cv evolution without faml . as a consequence of the large ejection mass ( several @xmath281/yr ) the outburst amplitudes
become very large ( a factor @xmath282 ) below the period gap for intermediate mass wds ( @xmath283 ) even more so as the outburst
induced decrease of the mass transfer is largest if faml vanishes .
such systems have a mass transfer rate less than @xmath2/yr for @xmath3 myr after the outburst , and could account for intrinsically faint cvs below the period gap .
we speculated ( sect .
4.5 ) if toads ( e.g. sproats et al . 1996
) could represent such systems .
we finally note that the apparent scatter in observationally derived values for the mass transfer rate of cvs with comparable orbital period , well known since the review of patterson @xcite , is unlikely to be due to outburst amplitudes .
first , we expect from fig .
[ fig : dina_high ] that these amplitudes are small or negligible above the period gap , and second the systems spend most of the time close to the continuous wind average mass transfer rate ( cf .
[ fig : problow_t ] ) .
a more promising explanation for this scatter assumes mass transfer cycles which could be irradiation induced ( e.g. king et al .
1995 , 1996 ) .
[ [ acknowledgements . ] ] acknowledgements .
we thank d. prialnik and a. kovetz for giving access to nova model parameters prior to publication , and m. livio and l. yungelson for providing a copy of an unpublished manuscript on faml .
k.s . would like to thank m. ruffert for many discussions on bondi hoyle accretion .
we thank a. king for improving the language of the manuscript .
k.s . obtained partial financial support from the swiss national science foundation .
theoretical astrophysics research at leicester is supported by a pparc rolling grant .
de kool m. 1992 , a&a , 261 , 188 duerbeck h.w .
1987 , space sci .
rev .. 45 , 1 & 2 fujimoto m.y .
1982 , apj , 257 , 752 hjellming m.s .
1989 , phd thesis , urbana champaign , illinois howell s. , rappaport s. , politano m. 1997 , mnras , 287 , 929 kato m. 1983 , pasj , 35 , 507 kato m. , hachisu i. 1994 , apj , 437 , 802 king a.r
. 1988 , qjras , 29 , 1 king a.r . , kolb u. 1995 , apj , 439 , 330 king a.r . ,
kolb u. , frank j. , ritter h. 1995 , apj , 444 , l37 king a.r . , frank j. , kolb u. , ritter h. 1996 , apj , 467 , 761 kley w. , shankar a. , burkert a. 1995 , a&a , 297 , 739 kolb u. 1995 , in bianchini a. , della valle m. , orio m. , eds , astrophysics and space science library , vol .
cataclysmic variables_. kluwer academic publishers , dordrecht , p. 511
kolb u. 1996 , in evans a. , wood j.h .
, eds , iau coll . 158 , _ cataclysmic variables and related objects_. kluwer academic publishers , dordrecht , p. 433
kolb u. , ritter h. 1992 , a&a , 254 , 213 livio m. 1994 , in nussbaumer h. , orr a. , eds , proc . of the 22nd saas fee advanced course , _ interacting binaries_. springer , berlin , p. 135
livio m. , govarie a. , ritter h. 1991 , a&a , 246 , 84 lloyd h.m .
, obrien t.j .
, bode m.f .
1997 , mnras , 284 , 137 macdonald j. 1980 , mnras , 191 , 933 macdonald j. 1986 , apj , 305 , 251 mazzitelli i. 1989 , apj , 340 , 249 paczyski b. 1971 , ara&a , 9 , 183 patterson j. 1984 , apjs , 54 , 443 politano m. 1996 , apj , 465 , 338 prialnik d. 1986 , apj , 310 , 222 prialnik d. , kovetz a. 1995 , apj , 445 , 789 rappaport s. , verbunt f. , joss p.c . 1983 , apj , 275 , 713 ritter h. 1988 , a&a , 202 , 93 ritter h. 1990 , in cassatella a. , viotti r. , eds , iau colloquium 122 , _ physics of classical novae_. springer , berlin , p. 313
ritter h. 1996 , in wijers r.a.m.j .
, davies m.b . ,
, eds , nato asi , series c , vol .
477 , _ evolutionary processes in binary stars_. kluwer , dordrecht , p. 223
ritter h. , kolb u. 1995 , in lewin w.h.g . , van paradijs j. , van den heuvel e.p.j .
, eds , _ x - ray binaries_. cambridge university press , cambridge , p. 578
ruffert m. 1995 , a&a , 311 , 817 shara m.m .
1989 , pasp , 101 , 5 shara m.m . , livio m. , moffat a.f.j . , orio m. 1986 , apj , 311 , 163 shima e. , matsuda t. , takeda h. , sawada k. 1985 , mnras , 217 , 367 sproats l.n .
, howell s.b . , mason k.o . 1996 , mnras , 282 , 1211 spruit h.c . , ritter h. 1983 , a&a , 124 , 267 stehle r. 1993 , _ diploma thesis _
, ludwig maximilians universitt mnchen stehle r. , ritter h. , kolb u. 1996 , mnras , 279 , 581 verbunt f. , zwaan c. 1981 , a&a , 100 , l7
starting from the fundamental eq .
( [ eq-2.1.5 ] ) one can derive a more general expression for the relative change of @xmath24 during both the outburst and inter outburst phase .
specifically we allow a small fraction @xmath284 of the mass ejected during the outburst to accrete on to the secondary , i.e. @xmath285 .
this gives @xmath286 } \end{aligned}\ ] ] instead of ( [ eq-2.3.2 ] ) , which is obtained in the case @xmath287 . additionally we allow a wind ( or other ) mass loss during the inter outburst accretion phase .
thus we use the general @xmath49 from ( [ eq-2.1.14 ] ) rather than the conservative one to generalize ( [ eq-2.3.4 ] ) .
this results in two additional parameters , @xmath288 and @xmath289 , describing the system s `` wind '' mass and corresponding angular momentum loss during the inter
outburst phase .
hence we get ( using @xmath290 ) @xmath291 \ : , } \end{aligned}\ ] ] which with ( [ eq : a1 ] ) , ( [ eq-2.3.5 ] ) and @xmath292 gives the generalized @xmath109 @xmath293 + { \delta m_{\rm ig}}\left [ \frac{\beta_2 - 2}{q } + \right .
\right . } \\
\nonumber \lefteqn { \left .
\left .
( \beta_2 + 2)(\gamma-1 ) - \frac{\gamma-2}{1+q}(1 + 2\:\nu^w ) \right ] \right ) \ : . }
\end{aligned}\ ] ] six parameters describe the system : @xmath131 and @xmath14 ( ideally taken from nova models ) , @xmath100 and @xmath284 ( requiring the specification of a particular faml model ) , and @xmath294 ( or @xmath288 ) , @xmath289 ( characterizing the wind loss during accretion ) .
the ratio of mass retained by the wd to that lost by the secondary @xmath295 which relates @xmath14 and @xmath131 similar to the simple case @xmath119 in the main body of the paper , now depends on @xmath284 and @xmath294 .
using a weighted @xmath45 to describe the average specific angular momentum gives @xmath296 and inserting these values into eq .
( [ eq-2.1.14 ] ) would also have directly led to eq .
( [ eq : a3 ] ) . | we investigate the effects of discontinuous mass loss in recurrent outburst events on the long term evolution of cataclysmic variables ( cvs ) .
similarly we consider the effects of frictional angular momentum loss ( faml ) , i.e. interaction of the expanding nova envelope with the secondary .
the bondi
hoyle accretion model is used to parameterize faml in terms of the expansion velocity @xmath0 of the nova envelope at the location of the secondary ; we find that small @xmath0 causes strong faml .
numerical calculations of cv evolution over a wide range of parameters demonstrate the equivalence of a discontinuous sequence of nova cycles and the corresponding mean evolution ( replacing envelope ejection by a continuous wind ) , even close to mass transfer instability .
a formal stability analysis of discontinuous mass transfer confirms this , independent of details of the faml model .
faml is a consequential angular momentum loss which amplifies the mass transfer rate driven by systemic angular momentum losses such as magnetic braking .
we show that for a given @xmath0 and white dwarf mass the amplification increases with secondary mass and is significant only close to the largest secondary mass consistent with mass transfer stability .
the amplification factor is independent of the envelope mass ejected during the outburst , whereas the mass transfer amplitude induced by individual nova outbursts is proportional to it . in sequences calculated with nova model parameters taken from prialnik & kovetz @xcite faml amplification
is negligible , but the outburst amplitude in systems below the period gap with a white dwarf mass @xmath1 is larger than a factor of 10 . the mass transfer rate in such systems is smaller than @xmath2/yr for @xmath3 myr ( @xmath4 of the nova cycle ) after the outburst .
this offers an explanation for intrinsically unusually faint cvs below the period gap .
= = = = = = = = # 1 # 1 # 1 # 1 @mathgroup@group @mathgroup@normal@groupeurmn @mathgroup@bold@groupeurbn @mathgroup@group @mathgroup@normal@groupmsamn @mathgroup@bold@groupmsamn = `` 019 = ' ' 016 = `` 040 = ' ' 336 = " 33e = = = = = = = = # 1 # 1 # 1 # 1 = = = = = = = = [ firstpage ] novae , cataclysmic variables binaries : close stars : evolution |
the precise quantization of the hall conductance in the integer quantum hall states is protected by the non - trivial topology of band structures .
this topological property is characterized by the thouless - kohmoto - nightingale - den nijs ( tknn ) number , or the chern number @xcite , which takes non - zero values only when time - reversal symmetry is broken . in recent years
, tremendous progress has been achieved in a new class of topologically non - trivial band insulators in the presence of time - reversal symmetry , which are termed as topological insulators @xcite .
topological insulators exist in both two ( 2d ) and three dimensions ( 3d ) , which are characterized by the @xmath0 topological index .
these topological states have robust gapless helical edge modes with odd number of channels in 2d @xcite , and odd number of surface dirac cones in 3d @xcite .
topological insulators have been experimentally observed in 2d quantum wells through transport measurements @xcite , and also in 3d systems of bi@xmath1sb@xmath2 , bi@xmath3te@xmath4 , bi@xmath3se@xmath4 , and sb@xmath3te@xmath4 through the angle - resolved photo - emission spectroscopy @xcite , and the absence of backscattering in the scanning tunneling spectroscopy @xcite .
interaction effects in topological insulators remain an open question . due to their gapped nature ,
topological insulators remain stable against weak interactions .
however , strong interactions may change their topological properties . for 2d topological insulators
, it has been found that the two - particle correlated backscattering , which is an interaction effect and is allowed in the time - reversal invariant hamiltonian , can gap out the helical edge states by spontaneous developing magnetic ordering under strong repulsive interactions @xcite . in this case , time - reversal symmetry is spontaneously broken along edges , although the bulk remains paramagnetic . at mean - field level
, interaction effects can destabilize the quantum anomalous hall state of the haldane - hubbard model @xcite and the 2d topological insulating state of the kane - mele - hubbard ( kmh ) model @xcite by developing long - range charge density wave and antiferromagnetic orders , respectively @xcite .
interactions can also change the topologically trivial band structures into non - trivial ones at mean - field level by developing bulk order parameters raghu2008,sun2009,zhang_ran2009 , wen2010 . due to the difficulty of analytic studies on strong correlation physics , exact results from numeric simulations are desirable .
recently , an exact diagonalization has been carried on the spinless haldane - hubbard model @xcite . a first order phase transition between quantum anomalous hall insulating state and topologically trivial mott - insulating state
is found .
quantum monte - carlo ( qmc ) simulations play an important role in studying strongly correlated systems @xcite . a major obstacle to apply the qmc to fermion systems
is the notorious sign problem . in the particular method of the determinant qmc
, the 4-fermion interaction terms are decoupled through the hubbard - stratonovich ( hs ) transformation and fermions are able to be integrated out .
the resultant fermion determinant , generally speaking , is not positive - definite , which is the origin of the notorious sign problem .
this problem prevents qmc simulations to achieve a good numerical precision at low temperatures and large sample sizes .
nevertheless , in a number of interacting models , the sign problem disappears . as presented in ref .
wu2005 , these models include the negative-@xmath5 hubbard model , the positive-@xmath5 hubbard model at half - filling and in bipartite lattices , and a class of models whose interactions can be decomposed in a time - reversal invariant way .
we find that the kane - mele model augmented by the hubbard interaction with purely imaginary next - nearest - neighbor hoppings has a particle - hole symmetry .
such a symmetry has interesting consequences such as the absence of edge charge and spin currents , which shows the edge currents are not a reliable criterion for topological properties . more importantly , the particle - hole symmetry ensures the absence of the sign problem in the quantum monte - carlo simulations .
this provides a wonderful opportunity to study interaction effects in topological insulating systems . in this article
, we perform a determinant qmc study on the stability of the topological insulating state of the kmh model with the strong hubbard interaction @xmath5 .
antiferromagnetic long - range - order has been found at large values of @xmath5 .
consequently , the quantum phase diagram of the kmh model can be classified into paramagnetic bulk insulating phases and antiferromagnetic mott insulating phases .
when further consider the stability of helical edges with infinitesimal two - particle backscattering , which is not contained in kmh model but generally allowed by time - reversal symmetry , the paramagnetic bulk insulating phase can be divided into two regimes according to their edge state luttinger parameters @xcite . the topological band insulator with stable helical edges
are stable in the weak interaction regime , while the helical edges become unstable by two - particle correlated backscattering at the intermediate interaction regime .
we have also studied the nature of spin - liquid phase in the pure hubbard model with @xmath6 , showing that it is neither a spontaneous haldane type quantum anomalous hall insulator , nor , a kane - mele type quantum spin hall insulator .
this article is organized as follows . in section [
sect : sign ] , we prove the absence of the sign problem in the kmh model under certain conditions . in section [
sect : qmc ] , we present the simulations on the developing of antiferromagnetic long - range orders in the bulk . in section
[ sect : edge ] , the edge properties are studied including both the edge single particle excitations and the edge spin correlations . in section [
sect : spinliquid ] , we present the simulation of the charge and spin current orders in the pure hubbard model in the honeycomb lattice .
conclusions are given in section [ sect : conclusion ] .
the kane - mele model is a straightforward generalization of the haldane model in the honeycomb lattice @xcite defined as @xmath7 where @xmath8 is the nearest - neighbor ( nn ) hopping integral as scaled to @xmath9 below ; @xmath10 is the next - nearest - neighbor ( nnn ) spin - orbit hopping integral ; @xmath11 is the chemical potential . in the general case of the kane - mele model , the nnn hopping for the spin-@xmath12 electrons
are complex - valued and complex - conjugate to each other . as a special case , the nnn hopping in eq .
[ eq : km ] is purely imaginary .
the hubbard interaction is defined as usual @xmath13 \big[n_{i\downarrow}-\frac{1}{2}\big ] .
\label{eq : hubbard}\end{aligned}\ ] ] in this section , we will present the symmetry properties of eq .
[ eq : km ] and eq .
[ eq : hubbard ] , and prove the absence of the sign problem in the determinant qmc .
[ eq : km ] and eq .
[ eq : hubbard ] has the particle - hole symmetry at @xmath14 as explained below .
we define the transformation as usual @xmath15 under this transformation , a hermitian fermion bilinear operator connecting two sites belonging to two different sublattices transforms as @xmath16 while that connecting two different sites in the same sublattice transforms as @xmath17 the onsite particle density transforms as @xmath18 where no summation over spin - index is assumed in eq .
[ eq : density ] . clearly in eq .
[ eq : km ] , the nn - hopping is real and the nnn - hopping is purely imaginary , thus its band structure is invariant at @xmath14 .
eq : hubbard is obviously invariant .
the particle - hole symmetry also implies that @xmath14 corresponds to half - filling .
an important conclusion based on the particle - hole symmetry is that both charge and spin currents vanish on all the bonds for the kmh model of eq .
[ eq : km ] and eq .
[ eq : hubbard ] at @xmath14 .
this result applies to arbitrary boundary conditions with broken bonds but with the homogeneous on - site potential which maintains the particle - hole symmetry on each site .
the proof is straightforward . through the continuity equation ,
the current operators of each spin component along the nn and nnn bonds are defined as @xmath19 respectively , where no summation over spin - index is assumed . both @xmath20 and @xmath21 are odd under the particle - hole transformation ,
thus they vanish even with the open - boundary condition . by the same reasoning
, the charge current also vanishes in the haldane - hubbard model with the purely imaginary nnn - hoppings and the particle - hole symmetric charge interactions of @xmath22 this result shows that edge charge and spin currents are not good criteria for quantum anomalous hall and topological insulators . in order to have a better understanding on this counter - intuitive result , we have considered the simplest non - interacting haldane model with the purely imaginary nnn hoppings by diagonalization .
there are indeed gapless one - dimensional single particle chiral edge modes clearly seen from the spectra as commonly presented in literatures .
clearly this branch of edge mode contributes to edge currents .
however , we find that the continuous bulk spectra also contribute to edge currents .
perfect cancellation occurs which results in zero current on each bond , including each edge bond , although we know for sure that the band structure is topologically non - trivial . for interacting models , there are no well - defined single particle states .
we can not separate the edge and bulk contributions anymore .
nevertheless , we expect that current correlation functions should exhibit difference between topological insulators and trivial insulators .
another conclusion inferred from the particle - hole symmetry is that the average particle density for each spin component on each site is strictly @xmath23 even when the translational symmetry is broken .
for example , it applies to any disordered pattern of the hopping integrals , as long as the nn hoppings are real and the nnn hoppings are purely imaginary .
edge currents do appear if the particle - hole symmetry is broken .
for example , for the non - interacting haldane model with generally complex - valued nnn hoppings , edge currents appear along open boundaries .
so far we only consider the sharp edges of broken bonds but with homogeneous on - site potential . for edges with the confining single particle potential ,
the particle - hole symmetry is broken which also results in edge currents .
in particular , for a weak linear external potential , the linear response should still give rise to quantized hall conductance in the insulating region . the hubbard model on the honeycomb lattice , which corresponding to the case of @xmath6 of eq .
[ eq : km ] and eq .
[ eq : hubbard ] has been recently simulated at half - filling @xcite .
as hubbard @xmath5 increases from zero to a moderate value and then the strong coupling regime , the ground state emerges from a semi - metal phase , to a new spin - liquid phase and then the antiferromagnetic insulating phase .
below we will prove that the sign problem still vanishes with nonzero values of @xmath10 .
the absence of the sign problem can be proved for both the finite temperature and the zero temperature algorithms for the determinant qmc . in this subsection , we prove this property for the finite temperature method for simplicity , and leave the more lengthy proof for the zero temperature algorithm in appendix [ sect : appendix ] .
we emphasize that the simulations presented in this article are done at the zero temperature . just as the ref.(@xcite )
does , we employ a discrete hs transformation which respects the @xmath24 symmetry for every fixed hs field configuration by decoupling in the density channel .
we rewrite the hubbard interaction and decompose it in the density channel by using imaginary numbers as @xmath25 where the discretized hs fields take values of @xmath26 , @xmath27 ; @xmath28 , and @xmath29 . for the convenience of presentation ,
we prove the absence of the sign problem in the finite temperature formalism with @xmath30 .
the proof for the zero temperature projector algorithm is similar .
the partition function at half - filling reads z & = & _ \ { l } \ {
( tr_p = m^1 e^-_i , jc_i^ k_ij^ c_j e^i _ i _ i , p(l ) ( c_i^c_i- ) ) + & & ( tr_p = m^1e^-_i , jc_i^ k_ij^c_je^i _ i _ i , p(l ) ( c_i^c_i - ) ) _ i , p_i , p(l ) } , where @xmath31 sums over all the configurations of the discrete hs fields @xmath32 and @xmath33 ; @xmath34 and @xmath35 are indices of discretized grids along the spatial and temporal directions , respectively ; @xmath36 takes the trace of the fermion space ; @xmath37 is the discretized time slice which is set to @xmath38 in the simulations in this article ; @xmath39 equals the imaginary time @xmath40 . by using the particle - hole transformation defined in eq .
[ eq : p - h ] , we show that the onsite particle density transforms according to eq .
[ eq : density ] ; the nn - hopping matrix kernel transforms according to eq .
[ eq : p - h - nn ] ; the nnn - hopping matrix kernel transforms according to eq .
[ eq : p - h - nnn ] . when the following two conditions are satisfied , the fermion determinants of two spin components are complex - conjugate to each other , thus the product of them is positive - definite : @xmath41 apparently , eq . [ eq : km ] and eq .
[ eq : hubbard ] satisfy these conditions , and thus are sign problem free .
please note that the kmh mode is sign problem free only when the nnn - hopping is purely imaginary .
generally speaking , the interacting model without the sign problem can have complex - valued hoppings with opposite signs , which still gives rise to opposite chern numbers for the band structures of spin-@xmath42 and @xmath43 , respectively .
however , they are not related by time - reversal symmetry anymore .
the hubbard model in the honeycomb lattice , which corresponds the case of @xmath6 in eq .
[ eq : km ] and eq .
[ eq : hubbard ] , has been simulated in ref .
when @xmath5 increases from zero , the single particle charge gap appears at @xmath44 , while the antiferromagnetic long - rang order emerges at @xmath45 .
the mismatch reveals an exotic spin liquid phase in between . when the intrinsic spin - orbit coupling , _
i.e. _ , the nnn hopping term in eq .
[ eq : km ] , enters , the model describes the topological band insulator .
it already has a band gap even at @xmath46 . as increasing @xmath5 ,
the antiferromagnetic structure factor is still a good quantity to tell when the magnetic long range order appears .
however , the bulk gap is no longer an appropriate quantity to judge a possible transition from the topological band insulator to a antiferromagnetic mott - insulator . here
we use the local single particle gap on edge sites as an indicator of the stability of edge states and topological properties .
we also study the edge effects to antiferromagnetic correlations . in this section
, we will simulate the bulk antiferromagnetic structure factor , and leave the study of edge properties in section [ sect : edge ] .
based on the above proof of the absence of the sign problem , we perform the qmc simulation for the kmh model at _ zero temperature _ by using the projective method @xcite .
we perform measurements from @xmath47 different random number series and each independent measurement has @xmath48 sample sweeps after warming up , the discrete imaginary time step @xmath49 is set to be @xmath38 . in this section , we use periodic boundary conditions for bulk properties calculation , e.g. , the bulk antiferromagnetic structure factor .
the spin - orbit nnn hopping in eq .
1 breaks the @xmath24 symmetry but preserves the conservation of @xmath50 . as a result ,
the antiferromagnetic correlation of @xmath50 should be different from those of @xmath51 and @xmath52 . in the large @xmath5-limit
, the nnn hopping generates an anisotropic exchange as @xmath53 with @xmath54 , which is ferromagnetic in the @xmath55-plane and antiferromagnetic along the @xmath56-direction @xcite . as the combined effect from the nnn anisotropic exchange and nn isotropic antiferromagnetic exchange , the magnetic exchange along the @xmath56-axis is frustrated while those along @xmath57 and @xmath58- axes are not .
thus the neel ordering favors the easy @xmath55-plane .
our qmc simulations have confirmed this picture .
the antiferromagnetic structure factor along the @xmath57-direction ( @xmath59-afsf ) and the @xmath56-direction ( @xmath60-afsf ) are defined as @xmath61 ^{2}|g\rangle , \notag \\
s_{af}^{zz}&=&\frac{1}{n}\langle g|\left [ \sum_{i}(-1)^{i}s_{i}^{z } \right ] ^{2}| g\rangle,\end{aligned}\ ] ] where @xmath62 means average over the ground state ; @xmath63 is the number of sites ; @xmath64 is the size ; @xmath65 takes the values of @xmath66 for the @xmath67 and @xmath68-sublattices , respectively . the comparison between @xmath69 and @xmath70
is plotted in fig .
fig : xxzzafsf , which clearly shows the easy - plane feature .
below we will use the @xmath59-afsf to describe the antiferromagnetic properties , and perform the simulation at @xmath71 with different values of @xmath5 and sample sizes of @xmath72 .
the extrapolation to the thermodynamic limit for different hubbard @xmath5 is plotted in of fig .
fig : xxafsf .
it can be seen that the magnetic long range order emerges at @xmath73 for @xmath71 . in fig .
[ fig : phase_diagram ] we present the qmc simulation on the magnetic phase diagram of the kmh model for in the parameter space of @xmath74 .
the phase boundary separating the af long - range - ordered phase and non - magnetic phases are marked for various values of @xmath10 .
the spin - orbit coupling opens the band gap at the order of @xmath10 , thus the interaction effect @xmath5 becomes important only when @xmath5 is larger than @xmath10 . as a result , the critical value of @xmath75 for the onset of the af phase increases with @xmath10 .
the phase diagram fig .
[ fig : phase_diagram ] exhibits a large regime of non - magnetic insulating state outside the af phase at @xmath76 . at small values of @xmath5 , it should be the @xmath77 topological band insulating phase which is stable against weak interactions . as increasing @xmath5
, it enters the af mott insulating phase at a critical line of @xmath75 . in an updated version of ref .
@xcite , it is found that the spin - liquid phase also extends to a small but finite value of @xmath10 .
however , the nature of this spin - liquid state remains unclear .
the bulk paramagnetic regime actually has rich internal structures . according to the stability of the helical edge states with respect to the two - particle spin - flip backscattering
, this paramagnetic insulating phase is divided into two different regimes with the effective edge luttinger parameter @xmath78 , respectively .
the analysis is presented below in sect .
[ sect : stability ] .
we believe that the edge properties is crucial to expose the topological aspect of the kmh model . in this section , we will show that the antiferromagnetic correlations along the edge become strongly relevant as increasing the hubbard @xmath5 while the bulk remains paramagnetic .
we consider the lattice configuration plotted in fig .
[ fig : open ] with the periodical and open boundary conditions along the @xmath57 and @xmath58-directions , respectively . as proved in sect .
[ sect : sign ] , the edge currents , both for charge and spin , are always zero due to the particle - hole symmetry .
we use another quantity , the local single particle excitation gap on edge sites , to check whether the edges are gapped or gapless .
it can be extracted from the tail of on - site time displaced green s function on the edge @xmath79 , which is defined by @xmath80 where @xmath81 is the many - body ground state .
the dependency of @xmath82 with @xmath83 for the site @xmath34 on the tip of the zig - zag edges are plotted in the inset of fig .
[ fig : edge_gap ] , where the long tail of @xmath84 shows a linear behavior with @xmath83 and the slope measures the excitation gap . here
the lattice has a ribbon geometry with @xmath85 zig - zag rows .
we fix the width of the ribbon @xmath86 and increase its length .
the extrapolations of the edge excitation gaps with @xmath64 are depicted in fig .
[ fig : edge_gap ] with @xmath10 fixed at 0.1 and different values of @xmath87 . clearly increasing @xmath5 significantly reduces the weight of the low energy spectra .
the bosonization analysis of the stability of the helical edge states has been performed in ref .
for the parameter regime of fig .
[ fig : edge_gap ] , the bulk remains paramagnetic , or , time - reversal invariant .
for the current kmh - model , @xmath50 is conserved which prohibits the existence of the two - particle spin - flip scattering term to open the gap .
the luttinger liquid theory of such a helical edge branch , _
i.e. _ , the right and left movers are with opposite spin polarizations , is characterized by only one luttinger parameter @xmath88 , which describes the forward scattering between these two branches .
due to the helical nature of the edge states , the long wavelength charge fluctuations and the @xmath56-component of the spin fluctuations are not independent but are conjugate to each other . both of them are gapless in the thermodynamic limit , and so does the single particle edge excitations .
the onsite imaginary time single - particle green s function decays as @xmath89 with the exponent @xmath90 at @xmath91 , the low energy density of states does not open a full gap but are depleted according to a power - law , and thus exhibit a pseudo - gap behavior .
the non - zero gap values in fig .
[ fig : edge_gap ] may be an artifact of finite size scaling and a result of tunneling between two opposite edges .
a more detailed numerical analysis is needed to further clarify the nature of the single particle excitations .
we further investigate the edge effects to the antiferromagnetic correlations .
we define the antiferromagnetic structure form factor for each zig - zag row parallel to the zig - zag boundary as @xmath92 ^ 2| g\rangle , \label{eq : lines_af}\end{aligned}\ ] ] where @xmath93 is the index of the zig - zag row ; @xmath34 is the site index along the @xmath94-th zig - zag line ; @xmath95 is the number of sites in each row .
the @xmath59-afsf for all the rows are depicted in fig .
[ fig : lines_af ] .
it is interesting to observe that the af correlations are strongest on edges , and become weaker inside the bulk .
this effect is most prominent at small and intermediate values of @xmath5 , because the single particle band gap due to @xmath10 is suppressed around edges , which enhances the interaction effects .
when @xmath96 , the bulk antiferromagnetism develops .
the antiferromagnetic correlations along both the edge and central rows are enhanced by @xmath5 . however , their difference is suppressed due to the disappearance of the helical edge states .
the finite - size scaling of the @xmath59-afsf for the edge rows for different values of @xmath5 are presented in fig . [
fig : edge_af ] . compared with the @xmath59-afsf calculated in the bulk ( fig .
[ fig : xxafsf ] ) , the edge antiferromagnetic correlations are much stronger than those of the bulk .
although the extrapolation to the infinite size in fig .
[ fig : edge_af ] implies a finite value of the neel order of @xmath97 on the edge , we believe that it is an artifact due to the power - law scaling of the af correlations . the 1d nature of the edge states and the conservation of @xmath50 prohibits the true long range neel ordering of @xmath98 but allows the quasi - long - range ordering , which is confirmed in the two - point spin correlations in sect .
sect : stability .
according to the bosonization analysis in ref .
@xcite , the scaling dimension of the @xmath99 neel order of the @xmath55-components is @xmath88 , thus their equal - time correlations decays as @xmath100 . if the condition of the conservation of @xmath50 is released , a time - reversal invariant two - particle correlated spin - flip backscattering term is allowed as @xmath101 at the particle - hole symmetric point of the kmh - model that we are simulating , the above term becomes the umklapp term which conserves the lattice momentum .
such a term reduces the @xmath102 spin symmetry down to @xmath77 .
it has the scaling dimension @xmath103 , and becomes relevant at @xmath104 . in this case
, it opens a gap by developing the long range @xmath99 magnetic ordering of @xmath97 or @xmath52 .
even for the cases that the two - particle spin - flip backscattering are random disordered or at a single site , they still can destabilize the helical edge states at smaller values of the luttinger parameter @xmath88 @xcite . according to the above analysis
, the bulk paramagnetic regime at weak and intermediate coupling strengths should be divided into two regimes . at weak interactions ,
the helical edge states are stable against interaction effects .
the two - particle backscattering terms only have perturbative effects . on the other hand , at intermediate level of interaction strength ,
interaction effects are non - perturbative which breaks time - reversal symmetry along edges and thus destroys the helical edges .
we emphasize that this destabilizing helical edges occurs when the bulk remains paramagnetic and time - reversal invariant . to numerically verify this picture , we present the calculation of the real space equal - time two - point correlations along the zig - zag edge in fig .
fig : edgeaf_power .
since each unit cell contains two non - equivalent sites , we denote the sites on the tips of the edge as @xmath67-sites and the other slightly inner sites as @xmath68-sites .
the correlation functions are defined as @xmath105 where @xmath106 and @xmath107 are along the zig - zag edge .
the simulated results for @xmath71 are plotted at different values of @xmath5 in the bulk paramagnetic regime .
the edge spin correlation exhibits the ferrimagnetic correlations among @xmath67 and @xmath68-sites because the edge breaks the equivalence between @xmath67 and @xmath68-sites .
the magnetic correlations are stronger among the outer @xmath67-sites , and are weaker among the inner @xmath68-sites .
all of these correlations obey the power law and their decay exponents ( @xmath108 ) are fitted .
as further increasing @xmath5 towards to the bulk antiferromagnetic regime , the difference between @xmath109 and @xmath110 correlations become weaker . due to the domination of the magnetic correlation at @xmath67-sites
, we use the decay exponents of @xmath111 to fit the effective luttinger parameter @xmath88 for the helical edge .
the three plots in fig .
[ fig : edgeaf_power ] ( a ) at @xmath112 and @xmath113 gives rises to @xmath114 , and @xmath115 , respectively .
the case of @xmath116 belongs to the topological band insulating phase in which interaction effects are perturbative . for the case of @xmath117
at which the bulk remains non - magnetic , although the edge remains gapless , it is only because the conservation of @xmath50 which is _ not _ an essential symmetry of topological insulators .
as long as the above umklapp term eq .
[ eq:2pctbg ] is introduced , which unfortunately can not be simulated by our qmc method , the gapless helical edge states are destabilized .
we argue that the system enters a new phase with paramagnetic bulk but unstable edges .
the transition point between these two paramagnetic phases at @xmath71 lies at @xmath118 with @xmath119 .
we have calculated the edge spin correlations for other values of spin - orbit coupling and interaction parameters to map the boundary with @xmath120 between two different bulk paramagnetic phases .
the boundary is plotted in fig .
fig : phase_diagram .
as @xmath10 decreases , the dispersion of the edge spectra becomes more flat , and interaction effects go stronger . as a result , the boundary shifts to lower values of @xmath5 .
in particular at @xmath6 , the edge spectra become exactly flat , we expect edge ferromagnetism at infinitesimal @xmath5 due to the density of state divergence .
thus the boundary should pass the origin .
in particular , the edge ferromagnetism of graphene ribbon has been simulated in ref .
since meng _ et al .
_ @xcite claimed the existence of a spin - liquid phase for hubbard model @xmath122 at @xmath123 ( see fig .
[ fig : phase_diagram ] ) , it has attracted considerable interests and debates on the nature of this phase .
one possibility of such a phase is that it could be a relative spin - orbit symmetry breaking phase with a non - trivial mean - field band structure @xcite .
if it is the case , a finite @xmath124 behaves like an external field to pin down the order parameter along the external spin - orbit configuration .
then the semi - metal and spin - liquid phase are indistinguishable at finite @xmath124 . in this section
, we will check the form factor of the such a spin - orbit order parameter between nnn sites at @xmath6 , and find negative results . without loss of generality
, we only consider the horizontal bonds .
we define the positive directions for the nnn horizontal bonds as depicted in fig .
fig : curr_corre .
two different nnn current orders are designed , including the charge flux order and the kane - mele type spin - orbit order , or , equivalently , the spin - current flux order .
their form factors are denoted as @xmath125 and @xmath126 and are defined as @xmath127 where @xmath65 takes the values of @xmath9 or @xmath128 for site @xmath34 in the @xmath67 and @xmath68 sublattices , respectively ; the charge current @xmath129 , and spin current @xmath130 ; @xmath131 is the nnn vector along horizontal direction .
please note that the bond current operator here @xmath132 is different from that in eq .
eq : nnn_curr as @xmath133 where no summation over @xmath134 is assumed .
we have performed the simulation of the nnn charge and spin - current form factors defined in eq .
[ eq : currents ] for the hubbard model at @xmath6 .
the extrapolations of the form factors to the infinite lattice size are depicted in fig .
[ fig : nnn_sf ] .
the curves represents three typical hubbard @xmath5 values @xmath135 and @xmath136 , which fall in semi - metal phase , spin - liquid phase and mott insulating phase , respectively . for all the three parameters ,
both the charge and spin nnn current antiferromagnetic form factors vanish in thermodynamic limit , indicating the absence of the nnn charge and spin - current orders in all these three phases , especially the spin liquid phase .
the nature of this spin - liquid phase , whether it is actually a subtly ordered phase or a genuinely exotic phase with non - trivial topological property , remains an unsolved question .
we have studied the particle - hole symmetry in the kmh model , which results in the absence of the charge and spin currents and the absence of the quantum monte - carlo sign problem .
the determinant qmc simulations have been performed for both the bulk and edge properties .
the bulk antiferromagnetic long range order appears at large values of @xmath5 . with the open boundary condition ,
the antiferromagnetic correlation is strongest along edges .
we also studied the stability of helical edges in paramagnetic insulating phase when turn on infinitesimal two - particle backscattering term , which can be introduced by time - reversal invariant but @xmath50 not conserved interaction terms .
the paramagnetic insulating phase in fig .
[ fig : phase_diagram ] can be classified into two regimes of weak and intermediate interactions , respectively . in the weak interaction regime ,
the helical edge states remain gapless which is robust against the two - particle back - scattering ; in the intermediate interaction regime , the edge states can spontaneously break time - reversal symmetry by developing magnetic ordering along the edge by the two - particle backscattering term .
since this destabilizing helical edges occurs when the bulk remains time - reversal invariant , it is an interesting and open question whether the non - trivial bulk @xmath77-topology is still maintained in this regime .
we also checked that the spin - liquid phase in the hubbard model at @xmath137 in the honeycomb lattice is neither a spontaneously developed haldane - type quantum anomalous hall insulator , nor , the kane - mele type quantum spin hall insulator .
c. w. and d. z. thank j. e. hirsch for his education on the determinate quantum monte - carlo simulations and his encouragement on this work . c. w. is grateful to s. kivelson for suggesting the simulation in section sect : spinliquid , and thanks l. balents , l. fu , t. xiang for helpful discussions . c. w. is supported by aro - w911nf0810291 and sloan research foundation .
d. z and g. m. z. are supported by nsf of china and the national program for basic research of most - china . _ note added _ during the preparation of this manuscript , we learned the work on the qmc simulation on the same kmh model by hohenadler _
et al _ @xcite .
in this appendix , we prove the absence of the sign problem of the kmh model at half - filling for zero temperature determinant qmc , which is essentially an imaginary time projector algorithm .
the explanation to the algorithm can be find in ref .
for readers convenience , we also give a brief introduction below .
the hamiltonian composes of free and interaction parts h = h_t+h_i .
the free part reads h_t=_i , jc_i,^k_i , j^c_j , , [ eq : hk ] where the kinetic energy matrix kernels @xmath138 and @xmath139 of the kane - mele model in eq .
[ eq : hk ] are given in section ii .
they satisfy the relation of eq .
[ eq : nosign ] .
the interaction part is h_i=_i(n_i+n_i-1)^2 .
[ eq : hhint ] the expectation value of a physical observable operator @xmath140 at zero temperature is defined as = = , [ eq : expec ] where @xmath141 is the ground state ; @xmath142 is a projection parameter large enough to ensure the trial wavefunction @xmath143 is projected to the ground state @xmath141 .
the discretized hs transformation of the interaction term eq . [ eq : hhint ] is performed in the density channel as the same as that in eq .
[ eq : hs ] .
the imaginary time propagator , i.e. , the projection operator , is represented as e^-h & = & _ \{l } \
{ u_\{l}(,0 ) _ i , p_i , p(l)e^-i_i , p(l ) } , + u_\{l}(,0)&= & _ p = m^1e^-_i , jc_i^k_ij^c_je^i_ic_i^ _ i , p(l)c_i + & & _ p = m^1e^-_i , jc_i^k_ij^c_je^i _ ic_i^_i , p(l)c_i , + [ eq : imgprop ] where @xmath33 and @xmath32 are the space - time discretized hs fields defined in eq .
[ eq : hs ] with @xmath144 taking values of @xmath145 ; @xmath31 represents the summation over the spatial and temporal configurations of the hs field ; @xmath146 is the propagation operator for the hs configuration @xmath147 .
the trial wavefunction @xmath148 is required to be a slater determinant , which we will specify later .
the ground state @xmath149 can be obtained from applying the imaginary time propagator @xmath150 of eq .
[ eq : imgprop ] on @xmath151 as |_0 = _ \{l } \ { u_\{l}(,0 )
_ i , p_i , p(l)e^-i_i , p(l ) } |_t , we further perform the calculation of eq .
[ eq : expec ] as where @xmath152 is the average value of @xmath153 for the space - time hs configuration @xmath147 defined as _
\{l}= , and @xmath154 is the corresponding probability of the hs field configuration @xmath147 as p_\{l } & = & _ t|u_l(2,0)|_t_i , p_i , p(l)e^-i_i , p(l ) .
+ @xmath155 is defined as @xmath156 .
the summation over the hs configurations @xmath157 can be done by using the monte carlo method .
next we prove the absence of the sign problem for the kmh model with purely imaginary nnn hoppings at half - filling in the zero temperature qmc method , _
i.e. _ , the probability @xmath158 is positive - definite .
we factorize the @xmath159 , where @xmath160 is a slater - determinant state for spin-@xmath42 electrons with the particle number @xmath161 , and similar convention applies for @xmath162 .
then @xmath158 reads as let us perform a particle - hole transformation only to the spin-@xmath43 component c_i^d_i=(-1)^ic_i^ , c_id_i^=(-1)^ic_i , then slater - determinant state @xmath166 changes to another slater - determinant state of holes with the hole number @xmath167 denoted as @xmath168 .
we arrive at p_\{l } & = & _
t^n^| _ p=2m^1e^-_i , jc_i^k_ij^c_j + & & e^i_i_i , p(l ) ( c_i^c_i- ) |_t^n^ + & & _ t^h , n - n^| _ p=2m^1e^-_i , jd_i^k_ij^d_j + & & e^-i_i_i , p(l ) ( d_i^d_i- ) |_t^h , n - n^ + & & _ i , p_i , p(l ) . &
= & _ j=1^n - n^(^q^)_j|0_h , where @xmath169 and @xmath170 are the particle vacuum and hole vacuum states , respectively ; @xmath171 is the number of of lattice sites ; @xmath172 is a @xmath173-dimensional rectangular matrix , and @xmath174 is a @xmath175-dimensional matrix ; @xmath176 and @xmath177 are vector notations for @xmath178 and @xmath179 with @xmath180 to @xmath171 .
the slater - determinant wavefunction has nice properties as e^c^ m c _ j=1^n_p ( c^q)_j|0=_j=1^n_p [ c^e^m q^]_j|0 , and & & 0|_j=1^n_p(c q^)_j e^c^ m c _ j=1^n_p(c^q^)_j|0 + & = & , where @xmath181 is an @xmath182 hermitian matrix , or anti - hermitian matrix . based on these properties
, we have p_\{l } & = & + & & + & & _ i , p_i , p(l ) , where the matrix kernels satisfy @xmath183 and @xmath184 is a purely real diagonal matrix whose @xmath34-th diagonal element reads _
ii= _ i , p(l ) . | we prove that the kane - mele - hubbard model with purely imaginary next - nearest - neighbor hoppings has a particle - hole symmetry at half - filling .
such a symmetry has interesting consequences including the absence of charge and spin currents along open edges , and the absence of the sign problem in the determinant quantum monte - carlo simulations .
consequentially , the interplay between band topology and strong correlations can be studied at high numeric precisions .
the process that the topological band insulator evolves into the antiferromagnetic mott insulator as increasing interaction strength is studied by calculating both the bulk and edge electronic properties . in agreement with previous theory analyses ,
the numeric simulations show that the kane - mele - hubbard model exhibits three phases as increasing correlation effects : the topological band insulating phase with stable helical edges , the bulk paramagnetic phase with unstable edges , and the bulk antiferromagnetic phase . |
over the years , transition metal oxides have represented a traditional platform for strongly correlated electron physics , which has nowadays become a field of its own , encompassing several classes of compounds with one common denominator : the localized character of the low - energy electronic wavefunctions ( with @xmath1 or @xmath2 orbital character ) and the corresponding prominence of coulomb interactions in driving the electronic properties of these materials . in conventional metals or semiconductors , the fermiology is essentially determined by the lowering of the total kinetic energy which becomes possible when a periodic potential supports the delocalization of the local orbitals into extended wavefunctions with a well - defined momentum and a correspondigly homogeneous distribution of the charge density . in correlated electron systems , the on - site coulomb repulsion between two electrons in the same @xmath1 or @xmath2 orbital can overcome the kinetic energy part of the hamiltonian , inducing the electronic system to find new ways to lower its total energy , often by spontaneous breaking of the native symmetries of the lattice ( translational and/or point group symmetry ) .
this tendency leads to the emergence of a rich variety of symmetry - broken electronic phases , and represents a distinctive trademark of strongly correlated systems , spanning across families of compounds otherwise very different from a chemical standpoint @xcite . within the extended class of correlated electron systems ,
copper oxides represent a unique breed due to the mixed character of the electronic bands , frustrated by the conflicting interplay between the o-@xmath3 states promoting itinerancy and electron hopping and the cu-@xmath4 states which hinder charge fluctuations and are conducive to mott - hubbard physics and an insulating , antiferromagnetic ( af ) ground state @xcite .
this delicate balance can be tuned and controlled by carrier doping , resulting in a phase diagram of astonishing richness and complexity , yet to be fully understood @xcite .
antiferromagnetism has been long known as the ground state in charge - transfer insulating undoped copper oxides @xcite , while unconventional superconductivity was discovered in 1986 @xcite . to date , several broken symmetries have been detected in the cuprates , which can be categorized into zero - momentum ( @xmath5 ) and finite - momentum ( @xmath6 ) orders , breaking rotational and translational symmetry , respectively .
these different types of order can involve both the charge sector ( nematic state for @xmath5 , and charge - density - wave for @xmath6 ) and the spin sector ( @xmath5 magnetic order , and spin - density - wave or antiferromagnetism with @xmath6 ) , leading to several proposed forms of intertwined electronic orders @xcite . in this review
we will focus on charge order , which is defined as an electronic phase breaking translational symmetry via a self - organization of the electrons into periodic structures incompatible with the periodicity of the underlying lattice .
the denomination of _ charge order _ , or equivalently _ charge - density - wave _ , has over time embraced a broad phenomenology .
however , charge order was first discovered in the form of stripe order , whose real - space representation is depicted in fig.[intro_fig]a . in this ordered state , the doped holes ( dark circles ) are segregated into unidirectional structures which act as boundaries separating undoped domains characterized by af order .
therefore , stripe order in its generalized meaning is an electronic ground state characterized by a _ combination of magnetic order and charge order with specific geometrical constraints on the ordering wavevectors_. nd@xmath7sr@xmath8cu@xmath9o@xmath10 . *
( _ a _ ) schematic representation of the stripe pattern : circles represent the cu sites in the cuo@xmath9 plane , with arrows denoting the cu spins and gray circles indicating the location of the doped holes .
( _ b _ ) @xmath11 projection of the reciprocal space showing the location of the measured charge and spin satellites ( black circles ) along the @xmath12 and @xmath13 directions , respectively ; white circles represent bragg reflections .
( _ c , d _ ) elastic neutron scattering measurement of the charge order peak at @xmath14 ( _ c _ ) and of the magnetic peaks at @xmath15 and @xmath16 ( _ d _ ) , with scans along the arrows marked in ( _ b _ ) .
readapted from ref .. ] .
[ intro_fig ] historically , the first proof of stripe order came from neutron scattering , a momentum - resolved technique which had been extensively used in the early days of high - temperature superconductors ( htscs ) due to its excellent energy resolution and large magnetic cross section .
the first neutron scattering studies of htscs focused on the doping and temperature dependence of af order in yba@xmath9cu@xmath17o@xmath18 ( ybco ) @xcite and la@xmath19sr@xmath8cuo@xmath10 ( lsco ) @xcite , which was determined to be located at wavevectors @xmath20 , corresponding to @xmath21 in physical units ( typically , @xmath22 ) . also ,
unless otherwise stated , we will refer the wavevectors to the undistorted unit cell , where the * a * and * b * axes ( and correspondingly the reciprocal axes * h * and * k * ) are parallel to the cu - o bond directions with lattice parameters equal to the nearest cu - cu distance . ] . around the same time
, the first experimental hint of stripe order was revealed by the independent neutron scattering measurements of yoshizawa _ et al .
_ @xcite and birgeneau _ et al . _
@xcite , who reported incommensurate af modulations in underdoped la@xmath23sr@xmath24cuo@xmath10 ( 8% hole doping ) and la@xmath25sr@xmath26cuo@xmath10 ( 11% hole doping ) , respectively , suggesting the presence of an electronic phase characterized by the quasi - static ordering of the doped holes . a subsequent study by thurston _
_ @xcite revealed that the incommensurate af order could only be found in the superconducting phase of lsco [ it was later shown to be present also in the insulating phase , albeit with different modulation vector @xcite ] .
the full momentum structure and doping dependence of incommensurate spin order in lsco was later uncovered by cheong _
@xcite , and shown to manifest itself as a set of four magnetic peaks at @xmath27 and @xmath28 ( see diagram in fig.[intro_fig]b ) , with @xmath29 representing the doping - dependent incommensurability .
the experimental observations from neutron scattering spurred an intense theoretical activity , aimed at addressing two primary phenomenological findings : ( i ) the momentum structure of the incommensurate af order and its relationship to the ordering of the doped holes ; and ( ii ) the emergence of incommensurate af order only within the superconducting phase , which pointed at an intimate interplay between short - range magnetism and superconductivity in the cuprates .
the short - range attractive interaction between segregated charges was initially proposed as a pairing mechanism for the superconducting state @xcite , while several numerical studies of the hubbard model near half - filling were performed under different conditions : using mean - field theory without @xcite and with @xcite charge fluctuations ; within hartree - fock approximation @xcite ; and by using exact diagonalization methods @xcite .
all these studies pointed to a symmetry - broken ground state with short - ranged charge inhomogeneities organized into an ordered pattern of unidirectional charged arrays , segregated away from the magnetic domains and acting as boundaries for the latter @xcite .
the stripe structure with holes condensed along rivers of charge - separating antiferromagnetic domains was expected to manifest itself also as a periodic modulation of the charge , with a wavevector twice as long as in the case of the incommensurate af peaks @xcite and tied to the hole concentration ( a counting rule stating that the number of domain line unit cells is equal to the number of carriers . " , from ref . ) .
such a stripe state was found shortly thereafter by tranquada _
_ in the nickelate compound la@xmath9nio@xmath30 @xcite , as evidenced by the presence of charge and magnetic satellite reflections in neutron scattering at the ( collinear ) wavevectors @xmath31 and @xmath32 , respectively , reflecting the fundamental relationship between the spin and charge ordering incommensurability factors : @xmath33 .
these two characteristics ( collinearity and two - fold proportionality of wavevectors ) represent the basic conditions for stripe order , as opposed to other geometries ( such as checkerboard ) .
the search for static stripe order in the isostructural lsco cuprate compound was complicated by the necessity for native fourfold symmetry in the cuo@xmath9 planes to be broken in order to accommodate a stripe state .
a solution was eventually obtained with the addition of rare earth impurities ( nd and eu , in place of la ) distorting the lattice in the otherwise square - symmetric cuo@xmath9 planes and providing an inner strain field stabilizing stripe order .
consequently , in 1995 tranquada _
@xcite used neutron scattering to detect charge and spin order peaks in the same sample of underdoped nd - substituted lsco , la@xmath34nd@xmath35sr@xmath8cu@xmath9o@xmath30 ( lnsco ) at @xmath36 and @xmath37 , as shown in fig.[intro_fig]c and d , respectively .
this finding confirmed the wavevector relationship already proven in the nickelate compound , namely that @xmath38 , as well as the equivalence between the incommensurability and the doping with the former amounting to @xmath39 at the ` magical ' doping level of 12.5% which is another signature of stripe order .
this work represented the first direct evidence of stripe order in the cuprates and , together with closely following studies using again neutron @xcite , as well as x - ray @xcite scattering , initiated the whole experimental field of charge ordering phenomena in htscs .
for more extended reviews of neutron scattering studies in the cuprates , we refer the reader to refs . , while for a comprehensive discussion of the theoretical works on charge and spin order we refer the reader to refs .. towards the end of the 90 s , a series of advancements in scanning tunnelling microscopy ( stm ) methods made it possible to obtain atomically - resolved maps with spectroscopic information on the energy - dependent local density of states ( ldos ) via the measurement of the bias - dependent tip - sample differential tunnelling conductance as a function of the spatial coordinate @xmath40 , @xmath41 .
the atomic resolution enabled by stm setups was key to the detection of density modulations with remarkably short correlation lengths , leading the way to a new era in the study of the nanoscale interplay between different electronic orders and of the resulting granular structure in htscs .
k ) bi@xmath9sr@xmath9cacu@xmath9o@xmath42 ( bi2212 ) .
( _ b _ ) differential tunnelling conductance ( @xmath43 ) map ( integrated between 1 and 12mev ) between an applied magnetic field of 5 and 0 t , showing the appearance of period-4 density modulations within the magnetic vortex cores , and over the same spatial region of ( _ a _ ) .
readapted from ref .. ( _ c _ ) momentum space representation of the periodic structures observed using stm on bi2212 , including the fourier peaks from the atomic lattice ( a ) and the structural supermodulation ( s ) , as well as the charge order peaks ( q ) at @xmath44 and @xmath45 , in reciprocal lattice units .
( _ d _ ) energy - resolved fourier - transformed conductance map along the @xmath46 direction , showing no dependence of the charge order peak position ( q ) on energy , thereby demonstrating the independence of this phenomenon from quasiparticle interference effects .
readapted from ref ..
( _ e - g _ ) conductance maps in na - ccoc at different bias voltages , and ( _ h - j _ ) corresponding 2d fourier transforms , whose linecuts ( _ k - m _ ) along the cu - o bond directions show the charge order spatial frequencies and the lack of energy dispersion in the momentum - resolved structures .
readapted from ref .. ] the new spectroscopic imaging scanning tunnelling microscopy ( si - stm ) capabilities were soon applied to the study of htscs , with the family of choice being that of bi - based cuprates , owing to the presence of natural cleavage planes yielding extended , atomically - flat surfaces that turned out to be ideal for stm studies .
the first compound to be studied was bi@xmath9sr@xmath9cacu@xmath9o@xmath42 ( bi2212 ) , a layered cuprate with a bio sro
cuo@xmath9ca cuo@xmath9sro
bio stacking of crystallographic planes , characterized by a weak , van der waals - type bonding between adjacent bio layers . in 2002 , 7 years after the original discovery of stripe order in the la - based compounds , the stm investigation of bi2212 by hoffman _
_ @xcite revealed the presence charge order in yet another cuprate family .
the authors studied cleaved surfaces of slightly overdoped bi2212 samples , whose topographic stm map is shown in fig.[stm_fig]a .
the corresponding differential conductance map over the same field of view , and upon application of a 5 t magnetic field , is depicted in fig.[stm_fig]b . in the field - induced vortex cores ( darker spots ) , where the superconducting order is suppressed ,
four - fold , bi - directional modulations of the local density of states can be clearly imaged .
this density modulation , with correlation lengths of the order of 30 , provided experimental evidence in support of the expectation for a charge instability @xcite in the doping range where the field - induced period-8 spin density modulation was discovered in lsco shortly before @xcite .
the emergence of the periodic modulations within the vortex cores further suggested a direct competition between charge order and superconductivity , in analogy to the phenomenology in the la - based cuprates @xcite . while similar evidence for charge order was found by howald _
@xcite at low temperature and in the absence of a magnetic field , around the same time the discovery of quasi - particle interference ( qpi ) @xcite in the superconducting state and below an energy scale of the order of the superconducting gap underscored the importance of spectroscopic measurements resolving the observed structure as a function of sample - to - tip bias voltage .
the energy dependence of the charge order signal in fourier space was later explored in the stm study of vershinin _ et al . _ on bi2212 @xcite and hanaguri _ et al . _
@xcite on na - doped ca@xmath9cuo@xmath9cl@xmath9 ( na - ccoc ) .
figure [ stm_fig]c shows the reciprocal space chart of bi2212 obtained by fourier transforming the differential tunnelling conductance map in the pseudogap state ( @xmath47k ) , where qpi effects from superconducting quasiparticles are not active .
the two - dimensional momentum structure of charge order can be also followed as a function of energy ( corresponding to the bias between the stm tip and the underlying surface ) , and fig.[stm_fig]d shows a series of linecuts across the charge ordering wavevector @xmath48r.l.u .
, along the @xmath49 direction , as a function of energy .
a non - dispersive charge order peak ( q ) was found to be present across an extended energy range with almost constant amplitude , thereby proving its independence from the energy - dependent features due to quasi - particle interference .
charge order in highly underdoped ( @xmath50 ) na - ccoc , including non - superconducting samples , was discovered around the same period and shown to similarly exhibit an energy - independent momentum structure across a broad range of energies in the pseudogap state .
panels e - g and h - j in fig.[stm_fig ] show , respectively , the two - dimensional real- and reciprocal- ( fourier - transformed ) maps of the differential tunnelling conductance at 8 , 24 , and 48mv bias .
distinctive features can be identified in the fourier transform maps that correspond to near period-4 electronic modulations along the cu - o bond directions .
the energy - independence of these features appears evident from the linecuts shown in fig.[stm_fig]k - m , demonstrating the presence of a broken symmetry driving static electronic modulations affecting all the electronic states over an extended energy range .
the charge - ordered state in bi2212 and na - ccoc was later shown by kohsaka _
@xcite to break down into bond - centered , locally unidirectional domains , as revealed by tunnelling asymmetry imaging .
these measurements also revealed the prominence of these electronic modulations at large energy scales , of the order of the pseudogap energy , a tendency confirmed in subsequent studies @xcite .
evidence for charge order in single - layered bi - based compounds , bi@xmath51pb@xmath35sr@xmath52la@xmath53cuo@xmath54 ( bi2201 ) , was reported in 2008 by wise _
@xcite over an extended range of dopings , demonstrating that the charge order wavevector evolution is compatible with a possible instability arising from the antinodal region of the fermi surface .
a similar doping dependence was also detected in bi2212 @xcite .
in more recent years , a possible connection between charge order and the pseudogap state has been unveiled in bi2212 using high - temperature stm to correlate the onset of the electronic signal from incipient stripes with the pseudogap temperature @xmath55 @xcite .
these findings were followed by temperature - dependent measurements which show a competition between charge order and superconductivity near the fermi energy @xcite .
even more recently , it was shown that the zero - temperature density - wave order disappears in the vicinity of a putative quantum critical point located at a doping of @xmath56 for bi2201 @xcite and @xmath57 for bi2212 @xcite . throughout the years
, stm and si - stm provided a constant stream of scientific results and valuable insights on the existence , nature , symmetry , and nanoscale structure of charge order in underdoped cuprates ; many of these studies have now been covered by several reviews @xcite .
x - rays have been long used to investigate the inner structure of matter , thanks to the interaction of light with the electronic clouds surrounding the atomic nuclei @xcite .
the role of photon energy has traditionally been secondary with the exception of anomalous x - ray diffraction , a technique which relies on the strongly varying x - ray absorption near the absorption edges of certain elements to simplify the phase retrieval problem in crystallography and provide a full reconstruction of the atomic positions in complex macromolecular systems @xcite .
the first glimpses of resonant x - ray effects were revealed at the beginning of the 1970 s , when de bergevin and brunel @xcite demonstrated that x - rays are also sensitive to the distribution of electronic spins in magnetic materials by detecting antiferromagnetic bragg reflections in nio , thus confirming theoretical predictions by platzman and tzoar @xcite . following these seminal findings
, synchrotron - based x - ray magnetic scattering has been subsequently used as a powerful alternative to neutron scattering on several magnetically - ordered systems @xcite , while around the same time the foundations for a comprehensive theory of resonant x - ray scattering ( rxs ) were laid by blume and coworkers @xcite . since its inception during the 1980 s ,
resonant hard x - ray scattering has developed into an extremely versatile tool @xcite for the selective study of ordering phenomena involving the charge , spin , orbital , and lattice degrees of freedom , often offering a unique perspective on their respective interplay . historically , hard x - ray methods have anticipated soft x - ray methods due to the @xmath58 dependence of the x - ray attenuation length on photon energy @xmath59 ( @xmath60 is the angular frequency of the radiation field ) , which eliminates the need for vacuum - based experimental chambers for energies @xmath61kevm ( 3 mm ) and , as a result , the loss in x - ray flux over 1 m ( representing the typical dimension of a scattering diffractometer ) is @xmath62% ( 100% ) . due to the continuous ,
power - law dependence of x - ray absorption on the photon energy , no precise cutoff can be defined for soft x - rays .
however , conventional cr-@xmath63 x - ray sources exist that operate in air at @xmath64kev , albeit requiring reduced source - to - sample and sample - to - detector distances . in addition
, we note that hard x - rays beamlines still require vacuum conditions along the long pipes transporting the photons from the storage ring to the experimental chamber and from the chamber to the photon detectors . ] .
plane in the non - superconducting cuprate compound sr@xmath65cu@xmath66o@xmath67 measured using soft x - ray scattering ( _ a _ ) off - resonance ( 526ev ) and ( _ b _ ) at the main absorption peak for doped o-@xmath3 holes ( 528.6ev , corresponding to the @xmath68 transition ) . at resonance ( _ b _ ) , a superlattice peak is revealed at @xmath69 , signaling the crystallization of o-@xmath3 holes into a static and periodic pattern ( wigner crystal ) .
readapted from ref .. ( _ c _ ) resonant soft x - ray scattering measurements of the stripe order peak of la@xmath70ba@xmath71cuo@xmath10 [ located at @xmath72 across the o-@xmath73 edge ( @xmath68 , left ) , with highlighted the x - ray transitions into the mobile carrier states ( mobile carrier peak , mcp ) and upper hubbard band ( uhb ) .
( _ d _ ) same as in ( _ c _ ) , but at the cu-@xmath74 ( @xmath75 , right ) absorption edge .
the scattering peak intensity ( red line and markers ) is resonantly enhanced in the vicinity of the features of the absorption spectra ( green line ) corresponding to electronic transitions into the o and cu sites in the cuo@xmath9 planes .
readapted from ref .. ] soft x - ray absorption ( xas ) was first pioneered using conventional x - ray sources and in a transmission geometry on thin films of rare earth metals and oxides , known for the large absorption cross - section and rich multiplet structure characterizing the @xmath76 edges ( @xmath77 transitions ) @xcite .
important advancements in xas came with the use of synchrotron facilities @xcite , which produced a higher x - ray flux and further enabled the control of linear and circular x - ray polarization , making it possible to detect magnetic x - ray dichroism effects @xcite . in the same years
, improved detection schemes were being progressively adopted , such as partial electron yield @xcite , total electron yield @xcite , and fluorescence yield @xcite .
the successes of resonant hard x - ray scattering spurred the development of soft x - ray scattering instruments and methods , which presented clear benefits
access to the electronic states and corresponding degrees of freedom controlling the electronic structure in transition metal and rare earth oxides , greater sensitivity to surface and interface effects , native control of incoming light polarization but also inherent complications limited accessible reciprocal space ( and a consequent lower boundary of @xmath78on the smallest measurable periodicity , at @xmath79kev ) , the incompatibility with crystal - based energy or polarization analyzers , the need for high vacuum environment .
the rise of resonant soft x - ray scattering required special diffractometers capable of simultaneous , in - vacuum control of the sample and detector angles .
the first concept of a two - circle diffractometer ( with independent sample and detector rotations in the scattering plane ) was pioneered already in the late 1980 s @xcite and further advancements were realized throughout the 1990 s and early 2000 s at different facilities including brookhaven national lab ( ref .
and later ref . ) , lure ( ref . ) , the synchrotron radiation center ( ref ) , the european synchrotron radiation facility ( ref . ) , the daresbury laboratory ( refs . ) , bessy ( ref . ) , the swiss light source ( ref . ) , and the canadian light source ( ref . ) .
motivated by the new insights produced by early resonant inelastic scattering studies at the cu-@xmath73 edge ( @xmath80kev ) @xcite , the first resonant scattering measurements on cuprates in the soft x - ray regime were performed in 2002 by abbamonte _
@xcite on thin films of oxygen - doped la@xmath9cuo@xmath10 , at the o-@xmath73 edge ( @xmath81 , @xmath82ev ) and cu-@xmath83 edge ( @xmath84 , @xmath85ev ) .
the sensitivity of these two absorption edges to the doped holes in cuprates had been demonstrated previously @xcite , in particular through the identification of a mobile carrier peak ( mcp ) structure at @xmath86ev , i.e. below the onset of the main o-@xmath73 edge , and representing transitions onto the doped hole electronic states with large spectral weight on the o-@xmath3 orbitals within the cuo@xmath9 planes .
the study by abbamonte _
_ explored momentum space in the ranges @xmath87 ( orthogonal to the cuo@xmath9 planes ) and @xmath88 ( parallel to the cuo@xmath9 planes ) and the evolution of the momentum - dependent interference fringes across the resonances was determined to be indicative of a rounding of the carrier density near the film - substrate interface .
most remarkably , the authors revealed the extent of resonant enhancement at these absorption edges in the cuprates , which amounts to a single doped hole in the cuo@xmath9 planes scattering as strongly as 82 electronic charges , with a resulting magnification of the experimental signal equal to the squared of the scattering amplitude , or @xmath89 .
this can be regarded as one of the key figures of merit for resonant x - ray methods , and represents the foundational mechanism underlying the success of rxs in detecting weak ordering phenomena ( such as charge order ) in the cuprates .
later , abbamonte _
_ et al.__@xcite used rxs to discover and characterize the ordering of doped holes in the non - superconducting cuprate compound sr@xmath65cu@xmath66o@xmath67
. a peak in reciprocal space , representing the signature of a well - defined , periodic modulation of the electronic density along the reciprocal @xmath90 axis and with wavevector @xmath69 , could only be measured at the o-@xmath73 pre - peak resonance at 528.6ev ( fig.[rsxs_origin_fig]b ) while no signal was detected once the photon energy was tuned off - resonance by only a few ev ( fig.[rsxs_origin_fig]a ) .
this work represents the very first rxs evidence of charge order in a copper - oxide compound , corresponding to the crystallization into a wigner crystal phase , manifested as a modulation of the electronic charge triggered by electron - electron coulomb interactions .
this observation demonstrates the power of resonant scattering methods as opposed to conventional diffraction techniques , which , being non - resonant , are largely insensitive to subtle electronic ordering phenomena not involving the lattice degrees of freedom . the first rxs study of charge order in one of the superconducting cuprate families came along in 2005 , when abbamonte _
@xcite investigated stripe order in lbco .
while the momentum structure of the stripe - ordered phases in this compound had been previously studied using neutron scattering @xcite , no direct information was yet available on the role and involvement of the electronic degrees of freedom , due to the fact that neutron scattering predominantly measures periodic distortions in the lattice .
et al . _ found a peak in reciprocal space at @xmath91 , consistent with previous observations @xcite and with a weak dependence on the out - of - plane momentum @xmath90 , thus confirming the prominent two - dimensional nature of the charge ordered state .
once again , this study reaffirmed the fundamental role played by the resonant enhancement : figure[rsxs_origin_fig]c and d show the intensity of the charge order peak ( red markers ) overlaid onto the absorption profiles ( green lines ) for the o-@xmath73 and cu-@xmath74 edges , respectively .
the intensity for charge scattering is nonzero , i.e. detectable on top of the fluorescent background and above noise level , only near the oxygen prepeak ( mcp ) and the cu excitonic resonance .
this work had profound implications for the understanding of charge order in cuprates , since it clarified that stripe order is predominantly configured as a modulation of the electronic density followed by a distortion of the lattice , which however represents a secondary effect , as demonstrated by the absence of any diffracted intensity away from the electronic resonances ( in the soft x - ray range ) . in the following sections we will provide a brief description of the theory of resonant scattering and of the experimental scheme . for a more comprehensive treatise on these topics ,
we refer the reader to refs .
and references therein .
resonant x - ray scattering is a _ photon in
photon out _ technique , where photons get scattered from a material due to the interaction with the electronic clouds . for radiation - matter scattering to occur , the interaction hamiltonian has to contain operator combinations of the kind @xmath92 , where @xmath93 [ @xmath94 is the operator creating [ annihilating ] a photon with wavevector @xmath95 , polarization state @xmath96 , and frequency @xmath97 .
the effective nonrelativistic interaction hamiltonian can be derived from the full electron - matter minimal coupling hamiltonian , and reads : @xmath98}^{2 } + v ( { \mathbf{r}}_{j } , t ) \right\rbrace + { \sum}_{j \neq k } \frac{{e}^{2}}{{\vert{\mathbf{r}}_{j } - { \mathbf{r}}_{k}\vert}^{2 } } + { h}_{\mathrm{em } } \nonumber \\ & = \underbrace{{h}_{\mathrm{el } } + { h}_{\mathrm{em}}}_{{h}_{0 } } + \underbrace{\frac{e}{{m}_{e } c } { \sum}_{j } \mathbf{a } ( { \mathbf{r}}_{j } , t ) \cdot { \mathbf{p}}_{j}}_{{h}_{\mathrm{int}}^{\mathrm{lin } } } + \underbrace{\frac{{e}^{2}}{2{m}_{e } { c}^{2 } } { \sum}_{j } { { \mathbf{a}}^{2 } ( { \mathbf{r}}_{j } , t)}}_{{h}_{\mathrm{int}}^{\mathrm{quad}}}\end{aligned}\ ] ] where @xmath99 and @xmath100 are the fundamental electronic charge and mass , @xmath101 and @xmath102 represent the momentum and position coordinates of the _ j_-th electron respectively , and @xmath103 and @xmath104 are the lattice potential and the coulomb interaction terms , respectively .
@xmath105 represents the vector potential , @xmath106 is the hamiltonian of the electronic system , while @xmath107 $ ] is the hamiltonian of the electromagnetic ( em ) field alone .
the interaction operators @xmath108 and @xmath109 , which are respectively linear and quadratic in the vector potential , couple the electromagnetic field and the electronic degrees of freedom . at this point
, we can use as basis set for the light - matter quantum system the states @xmath110 , where @xmath111 represents the electronic part of the wavefunction ( with eigenvalues @xmath112 , and @xmath100 labeling a generic set of quantum numbers ) while @xmath113 indicates a photon state with photon occupation @xmath114 , corresponding to having @xmath115 photons with wavevector and polarization @xmath116 , @xmath117 photons with @xmath118 , and so on . in this notation
, @xmath119 labels the global set of quantum numbers , i.e. @xmath120 .
note that due to the radiation - matter interaction , the states @xmath121 are not eigenstates of the system , but they can be used as basis set in a perturbative scheme , in which case we define the unperturbed ( i.e. , with the interaction terms turned off ) energy spectrum as @xmath122 . within this framework , a scattering process is defined as a transition from an initial photon state @xmath123 to a final photon state @xmath124 , corresponding to the annihilation of an incoming photon with wavevector and polarization @xmath125 and the concomitant creation of an outgoing photon @xmath126 . to an intermediate state @xmath127 .
the subsequent radiative recombination of the excited electron with the core hole results in the creation of an outgoing ( scattered ) photon .
( _ b _ ) the different photon energy dependence of resonant and non - resonant processes , showing the enhancement occurring in the resonant channel near an electronic transition with energy @xmath128 .
( _ c _ ) schematics of a conventional diffractometer , with kinematics for the scattering / diffraction process outlined in ( _ d _ ) . ]
central in the theory of elastic scattering is the calculation of the probability of transition from a state @xmath129 to a state @xmath130 , where the photon states @xmath131 and @xmath132 are as given in the previous paragraph , and where we further assume that the electronic part of the initial and final state is in the ground state @xmath133 .
the transition probability @xmath134 between the initial and final quantum states can be calculated using the generalized fermi s golden rule @xcite : @xmath135 where the @xmath136-matrix is defined as follows : @xmath137 here the first operator on the right - hand side represents the first - order perturbation term , the second operator represents the second - order perturbation term , and so on . in eq.[eq : t_matrix ] , @xmath138 is the interaction operator and @xmath139 is the unperturbed hamiltonian in our case @xmath140 and @xmath141 , respectively . in scattering , we require operator combinations of the kind @xmath142 inside the @xmath136-matrix , which originate from interaction terms that are quadratic in the vector potential @xmath143 , since the latter can be expressed in second - quantized notation as @xmath144 $ ] ( @xmath145 denoting the polarization vector of the polarization state @xmath96 ) .
these combinations are generated by using the quadratic interaction operator @xmath109 in the first - order term of eq.[eq : t_matrix ] , and by the linear interaction operator @xmath108 in the second - order term of eq.[eq : t_matrix ] .
the corresponding first [ @xmath146 and second [ @xmath147 order perturbative transition probabilities can then be obtained from eq.[eq : fgr ] : @xmath148 where for convenience we have dropped the @xmath149-function ( enforcing conservation of the total energy in the scattering process ) originally present in eq.[eq : fgr ] . in eq.[eq : wfi_res ] , @xmath150 represents a generic ( excited ) quantum state of the light - matter system , with corresponding energy @xmath151 and lifetime @xmath152 .
the squared vector potential can then be expanded as : @xmath153 \times { \sum}_{{\mathbf{q}}^{\prime } , { \nu}^{\prime } } { \boldsymbol \varepsilon}_{{\nu}^{\prime } } \left [ { e}^ { i ( { \mathbf{q}}^{\prime } \cdot \mathbf{r } - { \omega}^{\prime } t ) } \ : { a}_{{\nu}^{\prime}}^{\dagger } \left ( { \mathbf{q}}^{\prime } \right ) + h.c .
\right ] \nonumber \\ & \propto & \left ( { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{in } } } \cdot { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{out } } } \right ) \left [ { e}^ { i ( { \mathbf{q}}_{\mathrm{out } } \cdot \mathbf{r } - { \omega}_{\mathrm{out } } t ) } \cdot { a}_{{\nu}_{\mathrm{out}}}^{\dagger } \left ( { \mathbf{q}}_{\mathrm{out } } \right ) \cdot { e}^ { -i ( { \mathbf{q}}_{\mathrm{in } } \cdot \mathbf{r } - { \omega}_{\mathrm{in } } t ) } \cdot { a}_{{\nu}_{\mathrm{in } } } \left ( { \mathbf{q}}_{\mathrm{in } } \right ) \right ] \nonumber\\ & \propto & \left ( { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{in } } } \cdot { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{out } } } \right ) \times { e}^{\left [ i \left ( { \mathbf{q}}_{\mathrm{out } } - { \mathbf{q}}_{\mathrm{in } } \right ) \cdot \mathbf{r } \right ] } \cdot { a}_{{\nu}_{\mathrm{out}}}^{\dagger } \left ( { \mathbf{q}}_{\mathrm{out } } \right ) { a}_{{\nu}_{\mathrm{in } } } \left ( { \mathbf{q}}_{\mathrm{in } } \right ) \label{eq : a_squared}\end{aligned}\ ] ] where @xmath154 and @xmath155 represent the polarization vector of the incoming and scattered photons , and in the last step of eq.[eq : a_squared ] we assumed the scattering process to be elastic , i.e. @xmath156 . at this point , by using the previous definitions for the initial and final states @xmath157 and @xmath158 , together with the fact that @xmath159 $ ] and @xmath160 $ ] , and considering that @xmath161 , we can rewrite eqs.[eq : wfi_nonres ] and [ eq : wfi_res ] as follows : @xmath162 where @xmath163 is the momentum transferred by the photon field to the sample , and @xmath164 is the fourier transform of the electron density operator @xmath165 . in the x - ray regime ,
the transition channel represented by eq.[eq : wfi_res_el ] involves the excitation of a high - energy many - body state with a core hole ( @xmath166 ) . in a more intuitive perspective , this is equivalent to the excitation of a core electron into an intermediate state through absorption of the first photon , followed by re - emission of a ( scattered ) photon once the core hole is filled back . on the other hand , in the first - order perturbative term expressed by eq.[eq : wfi_nonres ] , the scattering process is instantaneous as it does not involve the excitation of an intermediate state . as a consequence ,
the first - order mechanism ( known as thomson scattering ) , proportional to the squared amplitude of the total electronic density in the ground state , is _ non - resonant _ and controls the signal in conventional x - ray diffraction ( xrd ) .
the second - order process is instead _ resonant _ , and is therefore associated to rxs .
equation[eq : wfi_res_el ] is a particular version of the kramers - heisenberg formula , which represents the general solution to the problem of a photon scattering from an electron .
it can be further simplified under the assumption that the x - ray excitation process is local , which implies that : ( i ) the local orbitals ( at the lattice site @xmath167 ) can be used for the electronic basis set : @xmath168 ; ( ii ) that all matrix elements @xmath169 , i.e. that they vanish for orbitals belonging to different lattice sites ; and ( iii ) that , due to the localization of the initial ( core ) electron around the position @xmath170 of its parent atom , the phase due to the photon field can be approximated as @xmath171 . at this stage , it is convenient to introduce a new quantity , the _ form factor _ @xmath172 , which is a photon energy- and site - dependent complex tensor defined as follows:@xmath173 here @xmath174 and @xmath175 represent the initial and intermediate single - particle electronic states at site @xmath170 ( with energies @xmath176 and @xmath177 , respectively ) involved in the light - induced transition @xmath178 .
@xmath179 is the inverse lifetime ( @xmath180 ) of the intermediate state with an electron in @xmath175 and a hole in @xmath174 . the resonant scattering cross - section , through the form factor @xmath181 ,
can be shown to bear a close connection to the x - ray absorption ( xas ) , which is a first order process in the radiation - matter interaction hamiltonian : @xmath182 \label{xas_def } \\ { i}^{\mathrm{rxs } }
\left ( \mathbf{q } , \hbar \omega \right ) \ ! & \propto & \ ! { \left\vert { \sum}_{p q } { \left ( { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{in } } } \right)}_{p } \cdot \left [ { \sum}_{n } \ : { f}_{pq}^{(n ) } \left ( \hbar \omega \right ) { e}^{i \mathbf{q } { \textstyle\cdot } \ : { \mathbf{r}}_{n } } \right ] \cdot { \left ( { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{out } } } \right)}_{q } \right\vert}^{2}\nonumber\\ \ ! & = & \ ! { \left\vert { \sum}_{p q } { \left ( { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{in } } } \right)}_{p } \cdot { f}_{pq } \left ( \hbar \omega \right ) \cdot { \left ( { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{out } } } \right)}_{q } \right\vert}^{2 } \ : . \label{rsxs_def}\end{aligned}\ ] ] in eq.[rsxs_def ] , @xmath183 represents the scattering tensor , which _ is not _ a local quantity ( does not depend on the lattice position @xmath170 ) and is more directly related to the physical observable in rxs experiments ( @xmath184 ) .
moreover , from the above equations it follows that xas only depends on the incoming light polarization @xmath185 , whereas the rxs signal depends on the outgoing light polarization @xmath186 , as well .
the difference between resonant and non - resonant scattering is schematized in fig.[mechanism_rsxs](a ) .
the mechanism corresponding to xrd involves a single step , in virtue of its first - order nature ; conversely rxs , being a second - order transition , proceeds in two stages involving an intermediate state .
this clearly reflects in the very different photon energy ( @xmath187 ) dependence of the two channels [ see fig.[mechanism_rsxs](b ) ] : whereas xrd is nearly energy - independent ( red dashed curve ) , the cross - section for rxs is strongly peaked around the energy of the electronic transition ( blue curve ) , where the experimental signal undergoes a strong enhancement ( while decaying to zero away from the resonance ) .
this usually occurs in correspondence of an absorption edge , i.e. when electronic transition from a deeply bound core state into the valence band ( and beyond into the continuum ) take place . as a consequence , rxs gains a strong sensitivity [ as large as 82-fold in the cuprates @xcite ] to partial modulations of the charge density involving a single electronic band , whereas the xrd signal reflects the total electronic density ( see eq.[eq : wfi_nonres_el ] ) and therefore suffers from a weak sensitivity to spatial variations of the latter , unless they are accompanied by a distortion of the lattice , which would involve all the electrons ( core and valence ) . in an actual scattering or diffraction experiment ,
a monochromatic x - ray beam with wavevector @xmath188 , photon energy @xmath189 and polarization @xmath190 impinges on a sample and a scattered photon will be detected along the direction of the wavevector @xmath191 using an energy - integrating photon detector ( pd ) or an energy - resolving spectrometer ( see fig.[mechanism_rsxs]c ) . at the end of the process , a net momentum and energy have been transferred to the sample , which can be derived from the corresponding conservation laws : @xmath192 in the case of elastic scattering @xmath193 and there is no energy transfer with the sample ( @xmath194 ) , whereas the case @xmath195 defines inelastic scattering events . strictly speaking , elastic scattering probes the static component of the charge and magnetization density occurring in the system under study , whereas inelastic scattering is sensitive to dynamical processes and low - energy excitations .
however , due to the spectrometer - characteristic finite energy resolution @xmath196 ( which in the soft x - ray regime ranges between 30mev and 1ev , while hard x - ray spectrometers reach down to 1mev ) , purely elastic scattering can not be experimentally accessed , and it is more appropriate to use the term quasi - elastic scattering , which probes a regime which is static up to a timescale @xmath197 .
from a more practical standpoint , the energy - integrated measurement in most cases yields a reliable representation of the momentum structure of the ordered state , due to fact that the inelastic part of the spectra usually evolves very smoothly and can be discarded as background in rxs , especially if it exhibits a different temperature dependence with respect to the zero energy - loss feature ( see also discussion of fig.[rsxs_ybco_ghiringhelli_fig ] ) .
hereafter , we will focus on the momentum structure of rxs measurements and , unless otherwise specified , will assume the use of energy - integrated mode . from eq.[comin
: eq : totmomentum_rsxs ] , and using @xmath193 , the magnitude of the exchanged momentum can be expressed as @xmath198 [ @xmath199 is the scattering angle , see fig.[mechanism_rsxs](c ) ] .
projecting @xmath200 into the plane defining the sample surface then yields the in - plane ( @xmath201 ) and out - of - plane ( @xmath202 ) components of the transferred momentum , which will be often referenced in the rest of this work [ see fig.[mechanism_rsxs](d ) ] .
the realization of this geometry is illustrated in fig.[mechanism_rsxs](c ) , and described in greater detail in ref .. in general the photon detector moves on a single - circle , that is , a single angular goniometer ( the corresponding variable is often denominated @xmath203 and corresponds to the scattering angle @xmath199 ) .
the sample stage usually involves translational motion ( xyz ) and various rotations , whose number defines the type of diffractometer .
two - circle diffractometers represent the most common choice for soft x - ray experiments , featuring a single angular motion ( with angular variable denominated @xmath204 ) for the sample ( @xmath205 circle ) , whose axis is perpendicular to the scattering plane ( the one spanned by the vectors @xmath188 and @xmath191 ) , in addition to the detector rotation ( @xmath203 ) , which represents the @xmath206 circle .
the first generation of soft x - ray diffractometers also includes two additional rotational degrees of freedom ( denominated @xmath207 and @xmath208 , see again fig.[mechanism_rsxs]c ) , allowing a fine alignment of the sample axes with respect to the scattering plane ( however , @xmath207 and @xmath208 typically cover a limited angular range of @xmath209 to @xmath210 ) .
new designs are being developed using different geometries to extend the angular range and control for the sample orientation .
diffractometers with more circles ( up to 6 3 for the sample , 3 for the detector ) covering an ample angular range are routinely used at higher photon energies ( hard x - rays , @xmath211kev ) , where the added complication of the all - vacuum environment required for soft x - rays is lifted .
consequently , hard x - ray diffractometers can typically access a wider portion of reciprocal space .
typically , the experimental signal is comprised of both resonant and non - resonant contributions , and in the two possible regimes @xmath212 or @xmath213 one ends up probing different phenomena ( see again fig.[mechanism_rsxs]b ) . in the first case , where non - resonant processes are dominant ,
all electrons contribute equally to the measured signal , which is therefore simply proportional to the atomic number @xmath214 . the diffraction signal will be dominated by the core electrons , which usually outnumber the valence ones ( @xmath215 ) , with the exception of lighter elements which as a result are not probed very effectively in xrd .
in addition , since core states are very tightly bound to their parent nucleus , in a conventional diffraction experiment one mainly probes the ionic lattice in reciprocal space , which is why xrd is used primarily for structural studies . in the second case
, the scattering process has a strong enhancement in correspondence of a very specific electronic transition . as a result
the signal bears the signature of the electronic distribution ( in reciprocal space ) of the final state of such transition .
this characteristic of resonant scattering allows it to be not only element - specific ( whenever the absorption edges of different chemical species are spaced sufficiently apart in photon energy ) , but also orbital - selective .
this unique capability of rxs has been established and employed in many different systems .
charge - ordering in cuprates @xcite and cobaltates @xcite , and orbital - ordering in the manganites @xcite , are among the most spectacular case studies .
fifteen years after the original discovery of stripe order by tranquada
_ et al . _
, charge order had been observed in the doping region around 12% hole doping directly in the 3 families of la- , bi- and ca - based cuprates : lnsco @xcite , lbco @xcite , and lesco @xcite using neutron and x - ray scattering ( in the case of refs resonant x - ray scattering was used , in particular ) ; bi2212 @xcite , na - ccoc @xcite , and bi2201 @xcite using stm .
however , no clear evidence of charge order had been found in the ybco compounds , where the possibility of introducing doped carriers via the fractional filling of the cu - o chain layer helps reducing the electronic inhomogeneity in the cuo@xmath9 planes @xcite .
cu@xmath17o@xmath18 . *
( _ a _ ) resonant inelastic x - ray scattering ( rixs ) measurements of underdoped nd@xmath216ba@xmath217cu@xmath17o@xmath218 ( @xmath219k ) , as a function of energy loss and momentum [ along @xmath220 , showing the emergence of a quasi - elastic peak around @xmath221r.l.u . ; inset : three - dimensional view of the lattice structure of yba@xmath9cu@xmath17o@xmath18 .
( _ b , c _ ) momentum dependence of the quasi - elastic ( _ b _ ) and of the energy - integrated rxs intensity ( _ c _ ) for a series of temperatures across the superconducting transition @xmath222 . ( _ d , e _ ) temperature evolution of the charge order peak intensity ( _ d _ ) and full - width - at - half - maximum ( _ e _ ) , showing a cusp at @xmath222 , thereby providing evidence of competition between charge order and superconductivity . readapted from ref
the first indication of electronic order in ybco came in 2007 thanks to a series of pioneering measurements at high magnetic fields ( up to 62 t ) and low temperatures ( down to 1.5k ) revealing the presence of quantum oscillations in the hall resistance @xmath223 in the normal state of underdoped yba@xmath9cu@xmath17o@xmath224 ( @xmath225 , @xmath226k ) samples @xcite .
the frequency of these oscillations vs. the inverse magnetic field provided evidence for the emergence of small fermi pockets at high fields @xcite , arguably of electron - like nature in light of the negative sign for the hall coefficient @xmath227 @xcite .
these findings indicated a change in the fermi surface topology from the large hole - like fermi ` barrels ' in the overdoped regime @xcite to smaller pockets in the underdoped regime , thereby suggesting _ a reconstruction of the fermi surface caused by the onset of a density - wave phase , as is thought to occur in the electron - doped copper oxides near the onset of antiferromagnetic order " _
the striking similarity between the hall and seebeck coefficients of ybco and those of lesco @xcite made a strong case for a form of order in ybco akin to the stripe order in lesco .
the first direct evidence that the fermi surface reconstruction is due to charge order was provided by high - field nuclear magnetic resonance ( nmr ) measurements on yba@xmath9cu@xmath17o@xmath228 ( @xmath229 , @xmath226k ) , where a splitting in the @xmath230cu(2 ) lines was observed at 28.5 t and below 50k , signaling a change in the quadrupole frequency which was ascribed to a periodic variation in the charge density at planar cu sites or at the oxygen sites bridging them @xcite .
the simplest scenario seemingly compatible with the nmr results was a unidirectional density modulation with @xmath231 ( four unit cells ) periodicity .
furthermore , the observation of the charge - order - induced nmr splitting for strong fields perpendicular but not parallel to the conducting cuo@xmath9 planes provided evidence of a competition between superconductivity and charge order
. the quantum oscillations and nmr results seemed to point to the necessity of completely suppressing superconductivity to observe the emergence of a seemingly competing charge ordered state .
following early rxs explorations by hawthorn _ _ et al.__@xcite , the first direct observation of charge density wave in reciprocal space , and at zero magnetic field , was obtained in 2012 by ghiringhelli _
_ @xcite using energy - resolved and energy - integrated rxs on a series of ybco doping levels around 12% . the central experimental data uncovering the momentum location , and therefore the periodicity , of charge order in ybco are shown in fig.[rsxs_ybco_ghiringhelli_fig]a , and consist of a series of scans of the x - ray energy loss ( horizontal axis ) for different values of the planar projection ( @xmath232 ) of the momentum @xmath233 along the @xmath234 direction in reciprocal space .
the energy structure of the resonant inelastic x - ray scattering spectra bears three main contributions : ( i ) a quasi - elastic line at zero energy loss ; ( ii ) a low - energy peak / shoulder representing spin excitations ; and ( iii ) intra - band ( dd ) particle - hole excitations .
while the spectral weight of the spin and dd excitations is nearly momentum - independent , a clear enhancement of the quasi - elastic line can be seen around a planar momentum @xmath235 r.l.u . , which reveals the presence of a periodic modulation of the electronic density representing the momentum structure of a zero - field ordered state possibly connected to the one previously identified with nmr [ recent high - field diffraction measurements , which will be discussed later , helped clarify the nature of this connection @xcite ] .
the rxs data further revealed charge modulations to be present along both planar crystallographic axes , albeit with different amplitudes ( as discussed more extensively in ref . ) .
the same momentum - space peak could be equally well identified following the quasi - elastic component in energy resolved spectra ( fig.[rsxs_ybco_ghiringhelli_fig]b ) and the energy - integrated rxs ( fig.[rsxs_ybco_ghiringhelli_fig]c ) , which confirms : ( i ) the importance of the resonant enhancement at the cu-@xmath74 edge ; ( ii ) only the quasi - elastic scattering contributes to a pronounced momentum - resolved structure , as previously discussed . the detailed temperature dependence of energy - resolved and energy - integrated rxs momentum scans in figs.[rsxs_ybco_ghiringhelli_fig]b and c points to an onset temperature @xmath236k and , most importantly , to a partial suppression of the charge order peak below the superconducting transition temperature @xmath0 .
this finding provides direct evidence of a competition between superconductivity and charge order , and is further substantiated by the temperature evolution of the charge order peak intensity and full - width - at - half - maximum ( fwhm , inversely proportional to the correlation lengths ) reported in fig.[rsxs_ybco_ghiringhelli_fig]d and e for the case of energy - resolved and energy - integrated rxs , respectively .
the intensity and fwhm data indicate a clear weakening of the charge ordered state both in its amplitude and spatial correlations at the emergence of the superconducting state . around the same time , the wavevector of the charge - density - wave in ybco , and its competition with the superconducting state were measured by non - resonant hard x - ray diffraction experiments , which revealed an enhancement of the charge order amplitude below @xmath0 when superconductivity was actively weakened by applying magnetic fields up to 17 t @xcite .
field - induced enhancement of charge order was similarly reported in lbco , using hard x - ray diffraction @xcite .
the discovery and identification of the charge order state in ybco represented a breakthrough in the field of copper - oxide superconductors , as it suggested that the charge ordered state could be a defining instability of the cuo@xmath9 planes , and strongly contributed to revitalizing this research direction .
stimulated by the possibility to provide a unifying view of charge order in the cuprates , intense efforts at the experimental and theoretical level were put forth in order to identify a common thread across the different manifestations of charge order that emerged over the years . ) and vertical ( @xmath237 ) incoming polarization ( top ) and comparison between experiment and the prediction of the energy - shift model for polarization along the * a * axis .
readapted from ref .. ] a series of studies analyzed the photon energy dependence of the charge order signal from lesco @xcite , lnsco @xcite , and ybco @xcite , which is inherited from the photon energy dependence of the complex form factor @xmath238 ( see eqs.[form_factor ] and [ rsxs_def ] ) .
the imaginary part of the latter is directly related to the x - ray absorption signal , and the real part can be extracted by a kramers - kronig transformation . for a transition into a single cu-@xmath239 hole ( which is the case at the cu-@xmath74 edge in the cuprates ) there is no multiplet structure , and the ( site - dependent ) form factor can be approximated by a simple lorentzian lineshape as @xmath240 , so that the rxs intensity can be expressed as : @xmath241 where the polarization dependence has been neglected for simplicity .
eq.[rsxs_def_achkar ] shows that there are three variables controlling the rxs signal which can vary spatially : ( i ) the transition amplitude @xmath242 ; ( ii ) the lattice position @xmath170 ; and ( iii ) the transition energy @xmath243 .
correspondingly , achkar _ et al .
_ evaluated the impact on the rxs signal of assuming a periodic modulation for ( see fig.[rsxs_e_shifts_fig]a ) : ( i ) the valence modulation ; ( ii ) the lattice displacements ; ( iii ) the transition energy shifts ( this term was introduced in refs ) .
the comparison between the experimental rxs peak amplitude and the model calculations vs. photon energy is shown in fig.[rsxs_e_shifts_fig]b for lnsco and in fig.[rsxs_e_shifts_fig]c for ybco . in both cases
, it was determined that the rxs signal is predominantly controlled by the spatial variation of the x - ray transition energy shifts , while in the earlier work on lesco the rxs lineshape was explained with the nonlinear increase of the charge - carrier peak as a function of doping @xmath244 , due to a reduction of correlation effects ( @xmath245 ) at higher doping concentrations .
more recently , an alternative theoretical framework has been devised that explains the photon energy - dependent rxs signal by accounting for the delocalized character of intermediate states @xcite . a closely following series of studies provided a complete mapping of the detailed characteristics of this phenomenology the relative amplitude of the order parameter between ybco and lbco @xcite , the connection between charge order and magnetic instabilities @xcite , the origin of finite correlations and the role of disorder in the chain layer @xcite , the feedback on lattice dynamics @xcite as a function of doping , temperature , and magnetic field .
this tremendous amount of progress made it possible to shed new light on the role of charge order within the phase diagram and its connection to coexisting and neighbouring electronic orders and phases .
the complete doping dependence of the salient properties of charge order in ybco , and its comparison to la - based cuprates , were reported by blanco - canosa _ _ et al.__@xcite and hucker _ _ et al.__@xcite .
one of the open questions is the connection between the charge order seen in high magnetic fields via nmr @xcite and the charge modulations seen in zero field via x - ray diffraction .
a field - induced thermodynamic phase transition was detected in the sound velocity of underdoped ybco @xcite , suggesting that the high - field and low - field states are distinct , and in particular that the high - field charge order is two - dimensional in nature .
nmr measurements , owing to their ability to probe both the high - field and the zero - field regimes , also established that charge order in these two regimes manifests itself as distinct phases , which furthermore coexist at low temperatures and high fields @xcite . a recent x - ray diffraction study in pulsed magnetic fields up to 28
t revealed that there is indeed a field - induced crossover from a short - ranged order to a long - ranged charge modulations with different period along the * c * axis @xcite .
the discovery of cdws in ybco reinforced the idea that charge order might be a genuinely universal instability of the cuo@xmath9 planes , and the combination of real - space ( stm ) and reciprocal - space ( neutron scattering , xrd , rxs ) techniques had provided support for one and the same underlying general phenomenology in underdoped cuprates around 12% hole doping . despite such mounting evidence and the several unifying traits linking the various cuprate families , the real - space phenomenology of charge order in bi - based compounds appeared very granular , at variance with the well - defined structures in reciprocal space as observed by scattering probes in la- and y - based cuprates .
the very different probing depth
few angstroms for stm vs. hundreds of nanometer ( and more ) for scattering techniques and a possible dichotomy between surface and bulk [
as observed for charge - density waves along the cu - o bond direction in lsco @xcite and along the zone diagonal in bi2201 @xcite ] , also contributed to a perceived disconnect between the domains ( real space / surface vs. momentum space / bulk ) and materials ( bi- vs. la- and y - based cuprates ) investigated by these two classes of experimental methods .
sr@xmath9cuo@xmath54 . *
( _ a _ ) rxs scans along @xmath246 for an underdoped ( @xmath247k ) bi2201 at low ( blue ) and high ( red ) temperature , showing the emergence of a broad charge order peak around @xmath248 r.l.u .. ( _ b _ ) @xmath246 linecut of the fourier - transformed differential tunnelling conductance on freshly - cleaved surfaces of the same bi2201 samples , indicating the equivalence of the charge modulations detected using stm .
( _ c _ ) temperature dependence of the charge order peak intensity ( from rxs ) for three doping levels ( @xmath247k , @xmath249k , and @xmath250k ) in the under - to - optimal doping range , and correlation of the charge order onset with the pseudogap temperature @xmath55 ( grey boxes ) .
( _ d _ ) experimental ( bottom ) and calculated ( top ) fermi surface for ud15k bi2201 , highlighting the fermi arcs characterizing the pseudogap regime , and the wavevector ( yellow connector ) linking the arc - tips , or ` hot - spots ' , in agreement with the rxs experimental results .
readapted from ref .. ] two recent works @xcite addressed this aspect by investigating charge order on the same materials using stm and rxs . due to the complications in exposing atomically flat surfaces in any cuprate material other than bi - based compounds ,
these two studies were performed on la - doped bi2201 ( in the doping range @xmath251 ) @xcite and bi2212 ( in the doping range @xmath252 ) @xcite . in bi2201 , charge order was found with a wavevector between 0.243 and 0.265 r.l.u . for decreasing doping , whereas in bi2212
the ordering wavevector spans across a more extended range , from 0.25 to 0.31 r.l.u .
, a finding which was later corroborated by the detection of charge order also in optimally - doped bi2212 ( @xmath56 , @xmath253k ) @xcite .
most importantly , both studies demonstrated that the ldos modulations imaged by stm and the reflections measured by rxs in reciprocal space originate from the very same microscopic entity .
this is shown in fig.[rsxs_bi2201_fig]a and b by the comparison between the rxs momentum scans ( panel a ) in underdoped bi2201 ( @xmath254k , @xmath255 ) , revealing a charge order peak emerging at low temperature at @xmath256 , and the fourier - transformed differential conductance data ( panel b ) taken on the very same samples , which exhibit a peak at the same momentum value .
a similar correspondence was found in bi2212 .
the temperature dependence of the rxs charge order peak for three different bi2201 doping levels is reported in fig.[rsxs_bi2201_fig]c , showing a very gradual onset of the density modulations which occurs on a temperature scale which is proximate to the pseudogap temperature @xmath55 as determined from knight shift measurements @xcite . in bi2212 ,
rxs data also reveal a rather slow temperature evolution of the peak intensity , which is accompanied by a drop below @xmath0 .
although this drop is smaller than its analogue in ybco @xcite , the competition with superconductivity is clearly demonstrated by temperature - dependent stm measurements that also reveal that the charge order is more pronounced for the unoccupied states @xcite .
furthermore , in bi2201 comparison between the rxs results and the fermi surface measured using angle - resolved photoemission spectroscopy ( arpes ) revealed a quantitative link between the observed charge order wavevector and the momentum vector connecting the tips of the fermi arcs , which in this case coincide with the so - called ` hot - spots ' ( see fig.[rsxs_bi2201_fig]d ) .
this correspondence suggests a possible link between the density modulations and the low - energy electronic structure , an element which is consistent with the doping evolution of the wavevector and which had been hinted by previous studies @xcite .
recent theoretical works have discussed the possible special role played by the hot - spots in the context of a magnetically - driven charge order instability @xcite , as well as the possibility that charge order might arise from a @xmath257 instability of the antinodal points within a pair - density - wave framework @xcite . additionally , we note that the rxs detection of charge order around @xmath258r.l.u . in bi2201
confirms previously unpublished findings of a phonon anomaly at the same momentum location using inelastic x - ray scattering @xcite , which might be suggestive of a related anomaly in the electronic susceptibility or otherwise of a very strongly momentum - dependent electron - phonon coupling mechanism at work . .
] in the wake of the results on y- and bi - based cuprates , evidence had been mounting that charge order could be a universal phenomenon in the cuprates . in 2012 ,
wu _ et al . _
detected charge order in the prototypical superconductor lsco , around a doping of 12% @xcite , a discovery which revealed that the low - temperature tetragonal structure in la - based cuprates is not essential for the appearance of stripe order ( but perhaps relevant for stabilizing it ) . in this study , the comparison between resonant and non - resonant scattering data was interpreted in terms of a surface enhancement of the stripe order in lsco .
more recently , stripe order was found in the bulk of lsco using hard x - ray diffraction @xcite and resonant x - ray scattering @xcite .
having exhausted all other families , hg - based cuprates remained the last hole - doped family to be investigated .
in particular , hgba@xmath9cuo@xmath30 ( hg1201 ) represents the only series of compounds with a pristine , tetragonal unit cell , therefore possessing the highest structural symmetry among all cuprates .
the first indirect evidence of broken translational symmetry in hg1201 came from the discovery of fermi - surface reconstruction via high - field measurements of the hall and seebeck coefficients in underdoped hg1201 @xcite , followed by the detection of quantum oscillations @xcite . in 2014 , tabis _ et al .
_ reported the very first evidence of charge order below 200k in underdoped ( @xmath259 , @xmath260k ) hg1201 from rxs and rixs measurements @xcite .
the charge order peak was found at @xmath261r.l.u .
comparable to the values found in bi2212 and intermediate between ybco and la - based compounds and required using resonant excitation , indicating that charge order is a rather subtle phenomenon in the hg1201 .
this instance , combined with the fact that hg1201 exhibits record - high @xmath0 among single - layered cuprates , suggests a possible anticorrelation between charge order and superconductivity . furthermore , the study revealed a direct connection between the onset of charge correlations and @xmath262 , _ the temperature at which the seebeck coefficient reaches its maximum value , and below which conventional fermi - liquid planar charge transport is observed"_. it also suggested a possible common origin of the main instabilities in the cuo@xmath9 planes , namely _ the possibility that the sequence of ordering tendencies ( @xmath263 order precedes charge order , which in turn precedes superconducting order ) and the phase diagram as a whole are driven by short - range antiferromagnetic correlations"_.
lastly , tabis _ et al . _
successfully established the link between the charge order - induced reconstruction of the fermi surface and the size of the electron pockets observed by quantum oscillations in ybco @xcite and hg1201 @xcite , and were able to correctly predict the qo frequencies over an extended doping range .
figure [ cdw_all ] provides a graphical compendium of all charge order observations in the hole - doped cuprates for what concerns the doping dependence of the onset temperature ( fig.[cdw_all]a ) and of the wavevector ( fig.[cdw_all]b ) as observed using both spatial- ( open symbols ) and momentum - resolved ( full symbols ) probes .
the shaded phase diagram in fig.[cdw_all]a is representative of ybco and includes the antiferromagnetic ( af ) phase near zero doping and the superconducting ( sc ) dome at higher hole doping levels .
all coloured lines are guides - to - the - eye , except for the @xmath264 line which interpolates the doping dependence of the ordering wavevector in lsco below 12% doping according to the picture of perfect stripes with 1/2 hole per unit cell along the charged rivers .
note that the points from refs .
are from neutron scattering measurements of the antiferromagnetic peaks in lsco , from which the charge order wavevector @xmath265 was derived using the phenomenological relation @xmath266 .
it is clear from the phase diagram of fig.[cdw_all]a how in all cases charge order emerges at the highest temperature scales around 12% hole doping and spans a doping range tentatively bound by two endpoints ( @xmath267k ) , along the @xmath73 reciprocal axis only , with wavevector @xmath268r.l.u . and onset temperature @xmath269k @xcite . ] @xcite , thereby hinting at a possible intimate connection between this phenomenology and quantum criticality @xcite .
interestingly , the onset temperatures appear to scale inversely with the charge order amplitudes and correspondingly correlation lengths ( which are maximum for la - based cuprates , intermediate for ybco , and weak in the case of bi - based compounds ) . at the same time , the extent of the suppression of @xmath0 near 12% doping is larger for the families exhibiting stronger charge order , consistent with the experimentally - observed competition between these two instabilities
. finally , we note that two main phenomenological differences exist between the stripy la - based cuprates and all other families : ( i ) in la - based compounds charge order and incommensurate antiferromagnetic spin order are simultaneously present , whereas they are mutually exclusive in the other compounds ; ( ii ) the doping evolution of the charge order wavevector exhibits the opposite sign for la - cuprates with respect to the other families , which are putatively compatible with a nesting scenario ( see also guides - to - the - eye in fig.[cdw_all]b ) .
there are therefore strong indications for a common charge instability in all families of doped cuprates , but at the same time there are also stark differences between the manifestations of charge order in the conventional stripy compounds and in all other families . with charge order consistently detected across all flavours of hole - doped cuprates around 12% doping , the next frontier to be crossed was represented by the exploration of the electron - doped side of the phase diagram .
early insights came from inelastic x - ray scattering work @xcite reporting anomalies in the dispersion of optical phonons in nd@xmath19ce@xmath8cuo@xmath30 ( ncco ) , followed by quantum oscillation studies providing evidence for a reconstructed small - pocket fermi surface in the same compound @xcite and , more recently , by evidence from time - resolved studies of a broken - symmetry phase @xcite .
recently , the first hints of charge order were reported by lee _
_ @xcite and ishii _ et al . _
@xcite , who used rixs to map out the momentum - dependent structure of low - energy bosonic excitations in ncco , for electron - doping values spanning the phase diagram from the antiferromagnetic to the superconducting phase . in both studies
, rixs measurements revealed the presence of spin waves associated to the antiferromagnetic order at low doping , which evolved into paramagnon excitations in the superconducting state ( @xmath270 ) . around this doping , an additional branch of a rapidly - dispersing
, gapped excitation was found , which was interpreted as a particle - hole excitations in ref . and
as a charge amplitude mode in ref .
, therefore suggesting the possible signature of a symmetry - broken quantum state over an extended temperature range beyond the superconducting phase . direct evidence for charge order in ncco
was obtained by da silva neto _
_ @xcite by means of rxs measurements in reciprocal space , which revealed a broad reflection at a wavevector @xmath258r.l.u . ,
i.e. very proximate to the findings in hole - doped cuprates , once again pointing to a unified phenomenology and a possible universal instability of the cuo@xmath9 planes .
the rxs scans for a superconducting ncco sample ( @xmath271 ) are shown in fig.[rsxs_ncco_fig]a and b as a function of photon energy and temperature , respectively .
the charge order peaks appear to be rather broad , with a correlation length @xmath272 , and a very smooth evolution as a function of temperature and a finite peak amplitude seemingly persisting up to rather high temperatures ( 350 - 400k ) , as reported in fig.[rsxs_ncco_fig]c .
such a high onset temperature rules out a direct connection of charge order with the pseudogap phenomenology in electron - doped cuprates ; instead , the charge order signal appears to partly correlate with the temperature evolution of the antiferromagnetic fluctuations in the same material @xcite , uncovering a possible connection between the charge and magnetic instabilities , as previously suggested for the hole - doped cuprates @xcite .
most importantly , and independently of the detailed temperature dependence , the detection of charge order in ncco conclusively demonstrates that this phenomenon is truly universal in the cuprates , establishing a powerful and robust paradigm for the physics of the lightly - doped cuo@xmath9 planes , which transcends the asymmetry inherent to the different orbital character of doped holes ( largely on o-@xmath3 states ) vs. doped electrons ( occupying the cu - derived upper hubbard band ) . ] of the charge order peak around @xmath248 r.l.u . in superconducting nd@xmath19ce@xmath8cuo@xmath30 with @xmath271 .
( _ c _ ) gradual temperature evolution of the charge order intensity , with a tentative onset temperature around 350 - 400k .
readapted from ref .. ] as the case for the universality of charge order was receiving increasingly supporting evidence , the theoretical understanding of the origin of this phenomenon and of its interplay with coexisting instabilities , as well as the detailed exploration of the microscopic structure of the ordered state , all regained a central role in the context of the physics of copper - based high - temperature superconductors .
@xcite in a single - band model charge order would be a scalar field , which can be expressed as the periodic modulation of the occupation of a given electronic orbital as a function of the spatial coordinate .
however , the low - energy electronic structure of the cuo@xmath9 involves three different orbitals : cu-@xmath273 ( at the copper sites ) , o-@xmath274 ( at the horizontally - bridging oxygen sites ) , and o-@xmath275 ( at the vertically - bridging oxygen sites ) .
consequently , charge order can be expressed as a vector field with three components @xcite : ( i ) a site - centered modulation , corresponding to an extra charge residing on the cu-@xmath4 orbital ( fig.[rsxs_symmetry_ybco_fig]c , top ) ; ( ii ) an extended _ s_-wave bond - order , where the spatially - modulated density is on the o-@xmath3 states , and the maxima along the x and y directions coincide ( fig.[rsxs_symmetry_ybco_fig]c , middle ) ; ( iii ) a _ d_-wave bond - order , where the charge modulation changes sign between x- and y - coordinated oxygen atoms , and the maxima are shifted by a half wavelength ( fig.[rsxs_symmetry_ybco_fig]c , bottom ) . the notation , as well as the denomination of ` symmetry terms ' for the components introduced above , follow from the resulting angular distribution of the phases within the cuo@xmath10 plaquette ; this is encoded via an internal momentum variable @xmath276 ( restricted to the first brillouin zone ) in addition to the lattice momentum @xmath200 , in the definition of the charge order parameter : @xmath277 .
this definition enables a full decomposition of the orbital - dependent modulation of the electronic density as a linear combination of the symmetry components introduced above , which therefore serve as a basis set for the charge order parameter , so that the latter can be expressed as @xcite : around the transferred momentum @xmath200 , which is consequently preserved in the frame of reference of the sample .
( _ c _ ) different ( but not orthogonal ) symmetry representations for the charge ordered state in a 3-orbital system ( cu-@xmath4 , o-@xmath274 , and o-@xmath275 ) , such as the cuprates .
( _ d _ ) the modulation of the rxs signal vs. azimuthal angle @xmath278 is evident from the raw rxs scans across the charge order peak in underdoped yba@xmath9cu@xmath17o@xmath279 , for both @xmath237 and @xmath280 polarizations .
( _ e _ ) azimuthal dependence of the ratio of the @xmath237 and @xmath280 charge order peak intensities , and comparison to the numerical prediction for different binary combinations of the symmetry terms introduced in ( _ c _ ) , suggesting a prominent @xmath1-wave component .
readapted from ref .. ] and @xmath280 , with theoretical fit profiles achieving best agreement in the case of predominant @xmath281 symmetry .
( _ b _ ) ratio @xmath282 of the intensities for the two polarizations , and corresponding best fit profile .
( _ c _ ) diagram of the relative weight of @xmath281- and @xmath1-wave components as a function of the sign and magnitude of the in - plane ( @xmath283 ) to the out - of - plane ( @xmath284 ) x - ray transition amplitudes ; the red marker and bar indicates the parameter range yielding the best agreement with the data .
readapted from ref .. ] @xmath285 where @xmath286 , @xmath287 , and @xmath288 represent the magnitude of the _ s_- , _
s_- , and _ d_-wave terms .
theoretical predictions in the context of the _
t_-_j _ model @xcite and of the spin - fermion model @xcite concluded that a _ d_-wave pattern of electronic charges is energetically favoured over the other terms . in order to test these theoretical scenarios ,
an alternative rxs scheme has been recently devised @xcite and applied to the study of the local symmetry of charge order in bi2201 and ybco @xcite , and ybco and lbco @xcite .
this approach relies on the definition of a local form factor @xmath289 ( which is a tensorial quantity , see eq.[form_factor ] ) capable of discriminating between the different symmetry components of charge order .
once the magnitudes of the different components are built into the local form factor [ @xmath290 , the scattering yield can be written as : @xmath291 \cdot { \left ( { \boldsymbol \varepsilon}_{{\nu}_{\mathrm{out } } } \right)}_{q } \right\vert}^{2 } \label{rsxs_azimuthal}\ ] ] where @xmath154 and @xmath155 represent the polarization vectors for incoming and outgoing photons , respectively , while @xmath292 is the ordering wavevector . at this point
, a single measurement of the rxs intensity will not suffice to resolve the @xmath293 terms .
this issue is overcome by performing the rxs measurement as a function of the sample rotation around the azimuthal axis ( see schematic of experimental geometry in fig.[rsxs_symmetry_ybco_fig]a and b ) , which is collinear with the ordering wavevector .
this procedure allows modulating the projection of the form factor onto the light polarization as a function of the azimuthal angle @xmath278 . following this approach ,
the rxs scans can be measured for a range of azimuthal values ( see fig.[rsxs_symmetry_ybco_fig]d for the case of ybco , from ref . ) .
the resulting rxs intensities , shown in fig.[rsxs_symmetry_ybco_fig]e for the case of ybco at the cu-@xmath74 edge @xcite and in fig.[rsxs_symmetry_lbco_fig]a and b for the case of lbco at the o-@xmath73 edge @xcite , can be fitted to the theoretical model of eq.[rsxs_azimuthal ] to evaluate the magnitudes of the symmetry components ( which are treated as variational parameters ) . from these studies , it was concluded that ybco possesses a prominent _ d_-wave symmetry ( fig.[rsxs_symmetry_ybco_fig]e and ref . ) , while the stripe order in lbco is best described by a _
s_-wave pattern ( fig.[rsxs_symmetry_lbco_fig]c and ref . ) .
the determination of the charge order symmetry in bi2201 was instead not conclusive using rxs @xcite . while these early rxs results are poised to stimulate further work to identify and classify these types of symmetries in momentum space , in a recent stm study of bi2212 and na - ccoc , fujita _
_ @xcite have successfully implemented the decomposition of eq.[cdw_op_definition ] by analyzing the reciprocal space representation of the different local symmetries in real space . in particular , these authors first noted how a _
d_-wave [ _ s_-wave ] symmetry suppresses the fourier amplitudes of the @xmath294 and @xmath295 [ @xmath296 , @xmath297 , @xmath298 , and @xmath299 peaks , and subsequently were able to resolve the extinction of the inner charge order peaks in reciprocal space starting from the spatially - resolved , partial orbital occupation of the @xmath300 , and @xmath301 sites .
the detailed analysis of the stm conductance maps finally revealed that charge order in bi2212 and na - ccoc also possess a predominant @xmath1-wave form factor .
another key aspect regarding the microscopic description of charge order , and one that has been debated for long time @xcite , is whether charge order has a checkerboard ( bidirectional ) or stripe ( unidirectional ) character .
this problem found an early answer in the la - based cuprates , thanks to the coexistence of spin and charge order with a precise wavevector relation that rules out a checkerboard state .
however , magnetic and charge order tend to avoid each other in the phase diagram of all other cuprates , a circumstance which , combined with the typical observation of orthogonal [ @xmath302 and @xmath303 charge order reflections , has hindered a conclusive resolution of the checkerboard / stripe dichotomy .
furthermore , evidence in real space using local probes ( stm ) has been traditionally hampered by the disorder characterizing bi - based compounds , known to blur the distinction between native uni- and bi - directional ordering instabilities @xcite ; however , recent advancements in the analysis of stm datasets have enabled the visualization of a predominant unidirectional character of electronic modulations in bi2212 @xcite .
( red ellipse ) and @xmath304 ( blue ellipse ) for a yba@xmath9cu@xmath17o@xmath224 sample , as determined from fitting the azimuthal - dependent peak widths ( red and blue markers ) .
bottom - right inset : schematic representation of the original peak shapes ( full ellipses ) and their @xmath305 rotated versions ( hollow ellipses ) .
( _ b _ ) temperature - dependent longitudinal ( green ) and transverse ( yellow ) correlation lengths at @xmath306 ( left ) and @xmath304 ( right ) , showing a clear anisotropy in the evolution across @xmath0 .
( _ c _ ) illustration of the anisotropy reported in ( _ b _ ) , showing how the onset of the superconducting phase reduces the density - density correlations preferentially across the stripes .
readapted from ref .. ] the first indications of unidirectional charge order in underdoped ybco came from blackburn _ _ et al.__@xcite and blanco - canosa _ et al . _
@xcite , showing experimental evidence for very unequal amplitudes between the charge order peak along the * b * axis ( strong ) peak and along the * a * axis ( weak ) in ybco ortho - ii , a direct signature of a tendency to unidirectional behavior in the charge order parameter . a recent attempt to assess
the character of density modulations was put forth by resolving the full reciprocal - space structure of the charge order peaks in ybco using rxs @xcite .
the same geometry was used as outlined in fig.[rsxs_symmetry_ybco_fig]a and b , which effectively allows to slice through the ordering peak in the @xmath307 plane , along different directions in reciprocal space . in this study ,
the linewidth from the rxs scans ( see again fig.[rsxs_symmetry_ybco_fig]d ) was extracted in order to reconstruct the two - dimensional peak shape as shown in fig.[rsxs_stripes_ybco_fig]a for a yba@xmath9cu@xmath17o@xmath224 sample .
the elongation of the charge order peaks located along the @xmath308 direction ( @xmath309 ) and the @xmath12 direction ( @xmath310 ) suggests a locking between the direction of long correlation ( narrow peak linewidth ) and the wavevector , occurring for the yba@xmath9cu@xmath17o@xmath224 and yba@xmath9cu@xmath17o@xmath311 samples whereas the charge order peaks in yba@xmath9cu@xmath17o@xmath279 exhibits the same elongation .
this locking mechanism suggests a breaking of four - fold rotational ( @xmath312 ) symmetry which is independent for the two ordering components , and is therefore incompatible with a checkerboard state
the latter being an equal superposition of density modulations along @xmath308 and @xmath12 , imposes the same peak structure for the charge order peaks @xmath309 and @xmath310 .
the momentum structure of the charge order peaks already rules out a checkerboard state in favour of a unidirectional ( stripy ) instability ( where stripes can be segregated or overlapping and still generate the same structure in reciprocal space ) .
further support is provided by the temperature dependence of the longitudinal ( @xmath313 ) and transverse ( @xmath314 ) correlation lengths , which can be derived as the inverse of the peak linewidth along the direction parallel and perpendicular to the wavevector , respectively , and are shown in fig.[rsxs_stripes_ybco_fig]b .
the temperature evolution of the correlation lengths suggests a larger drop , below the superconducting transition temperature @xmath0 , for the longitudinal ( i.e. , across the stripes ) correlations ( see cartoon in fig.[rsxs_stripes_ybco_fig]c ) , again leading to a preferential , q - dependent locking of the density - density correlations and to a breaking of fourfold rotational symmetry . a similar tendency to
a unidirectional character of the charge modulations has also been identified in the different orbital symmetry of the @xmath309 and @xmath310 peaks , measured with rxs @xcite .
this in - plane anisotropy of the short - range charge modulations in ybco , also detected in nmr measurements @xcite , have been argued to cause the large in - plane anisotropy of the nernst signal seen in ybco near @xmath315 @xcite , since the nernst anisotropy grows upon cooling in tandem with the growth in modulation amplitude @xcite .
despite the recent flurry of experimental findings and theoretical insights on charge order in the cuprates , there is still a lot to learn and to understand .
first and foremost , the mechanism driving the doped holes into breaking translational symmetry has not been conclusively pinned down . in particular , it is unclear whether the cuprates exhibit peierls physics similar to other low - dimensional compounds @xcite or whether a new and unconventional mechanism is at play .
several studies have been performed since the discovery of stripe order , aimed at elucidating the electron - lattice interplay and its relevance for charge order instabilities @xcite .
recently , the work by le tacon _ et al . _ @xcite revealed a strong and sharp ( in momentum ) softening of the low - energy acoustic and optical phonons at the charge order wavevector in underdoped ybco [ see fig.[rsxs_new_developments]a similar observations were also made for ncco @xcite and bi2201 @xcite ] .
this result _ per se _ already provides direct evidence of a pronounced electron - lattice coupling mechanism , while the partial yet incomplete softening ( with the frequency remaining nonzero ) further clarifies that the charge order in cuprates is not due to a phonon mode freezing into a static lattice distortion , consistent with the short - ranged nature of charge correlations at zero magnetic field .
the charge order peak remains confined to the quasi - elastic line , which bears the typical signatures of a ` central peak ' , which is characteristic of the presence of ordered nanodomains which gradually coalesce into a state with longer correlations .
however , the most puzzling finding emerges from the temperature dependence of the softening effect , which is found to be enhanced in the superconducting phase , where charge order is paradoxically weakened .
this behavior reflects a strong superconductivity - induced phonon anomaly , occurring at the wavevector of the charge instability , thus exposing a very complex intertwining between these competing phases and might deserve further explorations for a recent theoretical discussion of this phenomenon see ref . ) .
k ) , showing a sharp softening occurring at the charge order wavevector @xmath316 in the superconducting phase ( @xmath317 ) . readapted from ref .. ( _ b _ ) charge order peak in underdoped ybco ( @xmath318k ) , acquired using ultrafast soft x - ray pulses from the linear coherent light source .
( _ c _ ) corresponding time - resolved dynamics of the charge order peak amplitude following impulsive photoexcitation ( @xmath319 ) of the apical oxygen mode ( @xmath320 m ) using 400fs pulses .
readapted from ref .. ] time - resolved experimental methods , which probe the recovery dynamics of the system following an ultrafast perturbation , have rapidly emerged as an alternative tool to gain further insights on correlated materials .
depending on the nature of the excitation , pump - probe techniques can probe both the microscopic coupling between different degrees of freedom near equilibrium ( i.e. , in the linear response regime attained at low excitation intensities ) @xcite or explore completely new physics in the strongly perturbative regime ( at high excitation intensities ) where photoinduced phases can be accessed that are often inaccessible at equilibrium @xcite .
the field of time - resolved scattering and diffraction has seen rapid advancements since its inception at the end of the 1990 s @xcite .
new avenues in the x - ray study of ordering phenomena out - of - equilibrium have been enabled by the recent development of free - electron - laser ( fel ) sources , which generate bright ultrashort light pulses over a broad range of energies from the extreme ultraviolet to the hard x - rays , thereby creating new opportunities for the investigation of structural and electron dynamics with high time resolution .
the spectral control of the photoexcitation process is an additional experimental parameter , enabling the selective perturbation of specific degrees of freedom via the nonlinear coupling between light and lattice @xcite .
when this framework is combined with photon energy tunability , resonant scattering can be used to probe the recovery of electronic orders that are brought out - of - equilibrium by an ultrafast pump @xcite . in the context of charge order in the cuprates
, these measurements were performed for the first time by frst _
@xcite , using ultrafast thz light pulses to photoexcite a particular lattice vibration ( the apical oxygen mode ) in underdoped ybco and soft x - ray fel pulses ( at the cu-@xmath74 ) to track the evolution of the charge order peak ( fig.[rsxs_new_developments]b ) as a function of time .
the photoexcitation process strengthens the superconducting state by promoting coherent interlayer transport and the charge order amplitude is correspondingly weakened by a factor two ( see fig.[rsxs_new_developments]c ) , thus providing new insights on the competition between charge order and superconductivity in a regime where phase coherence is artificially enhanced with light @xcite [ similar results were also reported in lbco @xcite ] .
more in general , the new fel capabilities are poised to considerably extend the domain of application of resonant x - ray methods toward the study of nanoscale ordering phenomena at unprecedented length- and time - scales .
another new frontier for the use of high - brightness x - ray sources is in the context of high - magnetic field studies .
the use of high fields within different experimental techniques , such as quantum oscillations or nmr , has been transformative for our understanding of high - temperature superconductors , enabling unprecedented insights into the nature and interplay of competing orders in these complex materials .
it was then natural to envisage the extension of high - field capabilities to scattering and diffraction experiments and , notwithstanding the complications caused by the requirement for optical access for incoming and outgoing photon beams , several efforts have been successfully carried forward in this direction in recent years .
@xcite used a superconducting cryomagnet specifically designed for low - angle forward - scattering measurements in transmission geometry and using very penetrating high energy photons ( @xmath321kev ) , which enabled the first observation of magnetic field - induced enhancement of the charge order signal .
very recently , an innovative experimental scheme was introduced by gerber _
_ @xcite that is based on the use of a high - field ( 28 t ) split - pair pulsed magnet synchronized with the ultra - bright and ultrafast x - ray pulses ( with photon energy of 8.8kev ) of the linear coherent light source .
this approach , leveraging the high single - pulse photon flux of the fel beam , enabled the acquisition of two - dimensional single - shot diffraction patterns with sufficient photon counts in spite of the very low sampling frequency ( with one spectrum every 2 to 25 minutes , due to the recovery time of the pulsed magnetic field apparatus ) .
the momentum - space maps of charge order in ybco , from 0 to 25 t , reveal an ostensible evolution in the momentum structure of the charge order peak as a function of both the in - plane ( @xmath73 ) and out - of - plane ( @xmath90 ) wavevectors .
in particular , the out - of - plane character of the charge order peak changes from a very broad , elongated structure at zero field to a well - defined peak ( with @xmath322 correlation length ) located around @xmath323 .
this is accompanied by a sharpening of the in - plane peak linewidth , with a corresponding increase of charge correlations from @xmath324 to @xmath325 , and a concomitant enhancement of the diffracted intensity .
these results disclose the complexity of charge order and its rich and unconventional phenomenology as a function of doping as well as magnetic field . in particular
, this study not only uncovers the momentum - resolved crossover between the low- and high - field regimes , previously investigated using nmr @xcite , but also clarifies that a full reconstruction of the fermi surface as seen by quantum oscillations only happens in high field when long range order sets in , thereby elucidating the absence of folding and small pockets in the arpes measurements ( which can only be performed in zero field ) on ybco @xcite .
to conclude , we would like to emphasize that the further development of these novel approaches and methodologies will enable the exploration of completely new dimensions .
this will bring us many more surprises and much deeper insights in the study of symmetry breaking instabilities in cuprates , and will lead to a fuller understanding of these phenomena and of their intimate interplay with high - temperature superconductivity . 1 .
* technical advancements * : the challenge posed by the detection and characterization of charge - density waves in the context of the rich phenomenology of the doped cuo@xmath9 planes has propelled tremendous advancements in the field of soft x - ray scattering methods .
presently , rxs beamlines are operational or under construction at several facilities worldwide : als , aps , bessy , cls , desy , diamond , esrf , nsls - ii , nsrrc , sls , soleil , spring-8 , ssrl .
2 . * universality * : to - date , evidence of charge order has been found for all hole - doped cuprate families , as well as in nd - based electron - doped compounds , using a variety of complementary experimental probes .
* resolution * : resonant x - ray scattering has been successfully used to detect charge density modulations with a spatial coherence as short as 15 - 20 , thus extending to momentum space capabilities that were previously accessible only to spatially - resolved techniques ( stm ) .
* symmetry * : the richness of the rxs information , and the multiple control parameters available polarization , photon energy , sample orientation have opened up the possibility to explore new microscopic aspects of the ordered state such as its dimensionality ( 1d vs. 2d ) and its local symmetry ( _ s_- , _
s_- , and _
d_-wave ) . 1
. unveiling the driving force behind charge order in cuprates .
2 . find new approaches to modulate and , in general , control charge order .
3 . elucidate the relative role of mott physics and the proximity to quantum - critical behavior for charge order , as well as superconductivity , in cuprates .
4 . probe the evolution of charge order and its interplay with superconductivity under extreme conditions high magnetic fields , high pressures , ultrafast optical pumping .
the authors are not aware of any affiliations , memberships , funding , or financial holdings that might be perceived as affecting the objectivity of this review .
the authors are indebted to p. abbamonte , l. braicovich , a. cavalleri , e. h. da silva neto , j. c. davis , j. fink , j. geck , g. ghiringhelli , c. giannetti , r. l. greene , m. greven , d. g. hawthorn , f. he , j. e. hoffman , m .- h .
julien , b. keimer , m. le tacon , p. a. lee , w. s. lee , c. mazzoli , m. r. norman , s. sachdev , g. a. sawatzky , r. sutarto , l. taillefer , and j. m. tranquada for valuable discussions and for the critical reading of this review .
this work was supported by the max planck - ubc centre for quantum materials , the killam , a. p. sloan , a. von humboldt , and nserc s steacie memorial fellowships ( a.d . ) , the canada research chairs program ( a.d . ) , nserc , cfi and cifar quantum materials .
fujita k , hamidian m , firmo i , mukhopadhyay s , kim ck , et al . 2015 .
strongly correlated systems , chap .
spectroscopic imaging stm : atomic - scale visualization of electronic structure and symmetry in underdoped cuprates .
springer series in solid - state sciences roper m , van der laan g , drr h , dudzik e , collins s , et al . 2001 . _ nuclear instruments and methods in physics research section a : accelerators , spectrometers , detectors and associated equipment _
467468 , part 2:1101 1104 .
proceedings of the 7th int . conf .
on synchrotron radiation instru mentation | x - ray techniques have been used for more than a century to study the atomic and electronic structure in virtually any type of material .
the advent of correlated electron systems , in particular complex oxides , brought about new scientific challenges and opportunities for the advancement of conventional x - ray methods . in this context , the need for new approaches capable of selectively sensing new forms of orders involving all degrees of freedom charge , orbital , spin , and lattice paved the way for the emergence and success of resonant x - ray scattering , which has become an increasingly popular and powerful tool for the study of electronic ordering phenomena in solids . here
we review the recent resonant x - ray scattering breakthroughs in the copper oxide high - temperature superconductors , in particular regarding the phenomenon of charge order a broken - symmetry state occurring when valence electrons self - organize into periodic structures . after a brief historical perspective on charge order , we outline the milestones in the development of resonant x - ray scattering , as well as the basic theoretical formalism underlying its unique capabilities .
the rest of the review will focus on the recent contributions of resonant scattering to the tremendous advancements in our description and understanding of charge order . to conclude
, we propose a series of present and upcoming challenges , and discuss the future outlook for this technique .
charge - density - wave ; charge order ; resonant soft x - ray scattering ; high-@xmath0 cuprates ; superconductivity . |
for many years , the tully - fisher relation ( tully & fisher , 1977 ) has been used to estimate distances to the virgo cluster .
these studies were based on optical or near infrared photometry ( gavazzi 1999 ) , together with gas kinematics from hi linewidths ( e.g. , yasuda 1997 ) or h@xmath0 rotation curves ( rubin 1999 ) .
these techniques have been useful in the determination of distances to individual galaxies within the cluster , allowing attempts to determine its 3-d structure ( i.e. yasuda 1997 , gavazzi 1999 , and more recently solanes 2002 ) .
recent results show that there is a population of galaxies in virgo with hi line
widths narrower than those expected from the mean tully - fisher relation derived for the cluster .
we refer to these as tully - fisher deviant " galaxies .
either these galaxies are foreground to the virgo cluster , or they have peculiarities which make their observed gas line widths poor tracers of galaxy mass , and hence luminosity .
the rubin ( 1999 ) study of h@xmath0 rotation curves of 89 virgo disk galaxies identified a population ( @xmath1 6% of total sample ) which appears to deviate significantly from the tully - fisher relation , showing anomalously low gas rotation " velocities at apparently every radius , which they called `` low rotators galaxies '' .
solanes ( 2002 ) studied the 3-d structure of the cluster using the tully fisher relation based on hi linewidths and , most recently , 13 hi deficient spiral galaxies with apparently large 3-d barycentric distances located far away in front and behind the core of virgo ( sanchis et al .
seven galaxies were identified as being in the foreground of the cluster core ( @xmath2 13 mpc ) , supposedly in areas where the icm is too tenuous to be responsible for the stripping of the ism gas from these galaxies . of these galaxies
only three had large projected distances ( @xmath3 7@xmath4 ) , whereas the rest had small projected distance ( @xmath5 4@xmath4 ) .
some of the solanes s small hi line
width galaxies are also h@xmath0 rubin low rotator " galaxies . in subsequent studies by the sanchis group ( sanchis 2002 , mamon 2004 , and sanchis 2004 ) ,
the possibility was analyzed that these 13 hi - deficient objects had crossed the core of the cluster , losing their gas content as a result of icm - ism stripping .
they concluded these galaxies lie farther away than the maximum radius to which these galaxies can bounce out , making core - crossing improbable , and thus difficult for icm - ism stripping to be the cause of the observed hi - deficiency .
these authors suggested alternative scenarios including the possibility of erroneous distance estimations , hi - deficiency caused by tidal - interactions , heating of the gas by mergers , and even questioned the claims for hi - deficiency .
the application of the techniques discussed above to hi deficient galaxies suffers from the possibility that environmental effects can reduce the spatial extent of the atomic gas , cause non - circular motions or other disturbances , thus reducing the measured line - widths and therefore causing derived distances to be underestimated ( guhathakurta 1988 and teerikorpi 1992 ) .
indeed these small line - widths could be caused either by truncation of the gas disks by icm - ism stripping , or because the gas velocities are intrinsically small at every radius ( e.g. rubin 1999 ) , which can happen if the gas motion is non - equilibrium or non - circular , or the gas geometry is not understood .
these conclusions raise obvious questions about these objects , such as : could these `` tully - fisher deviant '' galaxies be the result of an underestimation of the distances produced by environmental effects , leading to an artificial distortion of the true cluster shape ?
we have carried out an extensive study of the kinematics and morphology of 13 peculiar virgo cluster galaxies ( corts 2006 ; corts 2008 in prep ) .
six galaxies within our sample , which includes four tentative foreground galaxies , are objects with small hi line - widths and therefore `` tully - fisher deviant '' . in order to answer these questions , in this paper distances to these galaxies
are estimated through an approach that relies on stellar kinematics and is therefore less affected by environmental effects .
this paper is structured as follows .
we describe in 2 the galaxy sample . in 3 , the method for estimating distances via stellar kinematics is introduced .
the detailed application of this method to our data , the results on the distances to individual galaxies of the sample , and the nature of the `` tully - fisher deviant '' galaxies is studied in 4 .
a discussion about the existence of the putative foreground group of galaxies in virgo as well as of the causes for these galaxies hi - deficiency is presented in 5 . finally , suggested improvements to the technique , summary and concluding remarks are found in 8 .
the sample consists of eleven peculiar virgo cluster spiral galaxies , spanning a variety of optical morphologies ( table @xmath6 ) .
our sample includes more early type than late type galaxies , since most of the strongly disturbed cluster galaxies are early types .
morphological selection was made using the r and h@xmath0 atlas of virgo cluster galaxies of koopmann ( 2001 ) , whereas the kinematical selection made use of the published h@xmath0 kinematics on 89 virgo cluster spirals by rubin ( 1999 ) .
the virgo cluster has about 110 spiral and lenticular galaxies brighter than 0.1 @xmath7 ( bingelli 1985 , bst ) , so we have observed about 12% of the brightest virgo cluster spiral galaxies .
the sample galaxies are spread around the cluster core , spanning projected distances ranging from a few hundred kiloparsecs to 2.5 mpc ( fig .
@xmath8 ) .
while the sample selection is not uniform , it is designed to include bright virgo spirals whose peculiarities are most poorly understood , and to include representatives of the different h@xmath0 types identified by koopmann & kenney ( 2004 ) . in choosing sample galaxies within a given h@xmath0 type , we gave preference to those with kinematical peculiarities .
following the use of the star formation classes introduced by koopmann & kenney ( 2004 ) , we can group our sample galaxies as follow * _ normal galaxies : _ ngc 4651 . * _ truncated / normal galaxies : _ ngc 4351 , ngc 4569 , ngc 4580 , and ngc 4694 . * _ truncated / compact galaxies : _ ngc 4064 , ngc 4424 , and ngc 4606 .
* _ anemic or truncated / anemic galaxies : _ ngc 4293 , ngc 4429 , and ngc 4450 .
although our main goal is to study galaxy morphological evolution in the cluster , we have at least six galaxies in the sample with small hi linewidths ( ngc 4064 , ngc 4351 , ngc 4424 , ngc 4580 , ngc 4606 , and ngc 4694 ) , including four galaxies that sanchis ( 2004 ) suggest may belong to a foreground group .
the determination of distances to hi - deficient galaxies using hi line - widths is vulnerable to the reduction in the spatial extension of the hi disks resulting from environmental effects , and by the possibility that gas velocities be intrinsically smaller at every radius . on the other hand ,
the use of stellar kinematics for distance determination ( stellar kinematics based distances " , hereafter skb distances ) has the advantage that it is in principle unaffected by physical processes other than gravitational processes , whereas as we have seen , the kinematics of the gaseous content can be affected by a number of non - gravitational physical processes resulting from environmental interactions ( e.g. icm - ism stripping ) .
however , stellar kinematics are far more difficult to measure and analyze than gas velocities , on account of the faintness of the stellar component as well as the collisionless nature of the stars which makes the pressure supported component important .
the latter implies that rotation velocities are no longer fully representative of the gradient of the potential .
star velocities however , are as good tracers of the mass of the galaxy as gas velocities are , provided one is able to properly handle their physical complexities
. we must therefore remember that ; a ) it is important to take into account the contribution of the stellar velocity dispersion @xmath9 to the real circular velocity , and b ) we have to restrict the tully - fisher analysis to a region within the galaxy where the surface brightness is high enough to allow reliable measurement of stellar velocities .
we have devised a distance estimation method based on stellar kinematics rather than gas velocities .
this method consists of the evaluation of the circular velocity from the stellar kinematics through the use of two - integral dynamical models ( see binney , davies & illingworth 1990 ; van der marel 1990 ; cinzano & van der marel 1994 ; magorrian 1998 ; cretton 1999 ) which are then used in a tully - fisher relation .
the method can be summarized as follow : \1 ) the stellar kinematics are modeled via two - integral dynamical models , to yield the true circular velocity @xmath10 .
\2 ) because this circular velocity is valid only within the region that we have measured reliably it can not be used in a traditional tully - fisher relation which requires sampling to large galactocentric distances . to resolve this
we construct synthetic ionized gas rotation curves for galaxies spanning a range of absolute magnitudes , using near infrared tully - fisher diagrams at each radius , as in rubin ( 1985 ) .
these synthetic rotation curves represent the typical rotation curves as a function of absolute magnitude for `` normal '' galaxies .
\3 ) finally , we compare these synthetic rotation curves with the circular velocities derived in 1 ) .
the best match gives the absolute magnitude , and therefore the distance modulus . in the following sections we provide a more detailed explanation of the estimation of the circular velocity and the construction of the synthetic rotation curves .
the multi - gaussian expansion mge formalism ( emsellem 1994 ; cappellari 2002 ) provides one of the simplest ways to build reliable two - integral dynamical models .
the mge formalism allows us to parametrize the surface brightness of a galaxy as a sum of gaussians with major axis widths @xmath11 , axial ratios @xmath12 , and luminosities @xmath13 , from which we can find a unique luminosity density @xmath14 ( cappellari 2002 ) , assuming axisymmetry .
this approach offers advantages over the standard bulge + disk decomposition as a description of the light distribution .
it provides an analytical and efficient way to do the luminosity deprojection , and it allows kinematical quantities such as the gravitational potential , velocity dispersion , and circular velocity to be estimated with a simple integration , thus simplifying the calculations .
these models can succesfully reproduce the observed stellar kinematics in ( presumably ) axisymmetric galaxies .
however , we must keep in mind that issues such as dust obscuration , triaxiality , kinematically distinct features in the stellar disks , and the assumed lack of dark matter can easily make these models discordant with the observed kinematics .
the bias introduced by these issues are discussed in 3.3 . to estimate the circular velocity , two - integral self - consistent models under the mge formalism
were built for eleven galaxies of the sample .
these models assume axisymmetry and @xmath15 .
they probably do not reproduce properly the kinematics , since the galaxies can exhibit important anisotropies in their velocity ellipsoid , but for the purposes of this work these models are acceptable since our aim is to estimate the circular velocity rather than fully reproduce the internal kinematics .
it is important to notice , that these models are purely stellar and do not include the contribution from dark matter . in
most of our galaxies our measurements are restricted to the inner parts ( @xmath1 20% of the optical radius ) making it impossible to constraint properly the dark matter halo parameters .
thus , these skb distances must in all the cases represent a lower limit to the distances . however , precisely because we are sampling the central regions where dark matter is the least important its effect on our distances is probably minor .
the detailed modeling proceeded as follows : \1 ) the mge formalism was used to find the luminosity density @xmath16 from r - band flux
calibrated images ( koopmann 2001 ) , assuming an oblate system .
the dust lanes were carefully masked to minimize the effects of dust obscuration .
for a given surface brightness @xmath17 , the formalism finds the best mge model ( fig @xmath18 ) , which can be de - projected to yield a unique luminosity density .
the inclination was assumed to be the minimum required for an oblate distribution . in most cases this inclination was consistent with the inclination obtained by koopmann ( 2001 ) .
\2 ) if we consider the axisymmetric case , following cappellari ( 2002 ) , the mge mass density can be written as @xmath19},\ ] ] where , @xmath20 is the stellar mass - to - light ratio , @xmath21 is the number of adopted gaussian components , each with major axis width @xmath11 , intrisic axial ratio @xmath22 , and luminosity @xmath13 . also , the mge gravitational potential can be written as @xmath23 where , @xmath24 is the gravitional constant , and @xmath25 is an dimensionless function given by @xmath26}}{\sqrt{1 - \epsilon_{i } t^{2}}},\ ] ] with @xmath27 .
the circular velocity at each radius in the plane of the galaxies was calculated directly from the potential .
the models have as input parameters the mass - to - light ratio @xmath20 and the distance @xmath28 , but the luminosity scales with the distance as @xmath29 .
this lead us to conclude that the circular velocity obeys @xmath30 here @xmath31 , and @xmath32 is @xmath33 which represents the radial variation of the circular velocity , and does nt depend on the distance .
thus , the models can be rescaled to any desired value of @xmath0 .
\3 ) the @xmath34 moment was calculated , for any point in the sky , following equations 6163 from emsellem ( 1994 ) , after taking into account the typo error noticed by cappellari ( 2006 ) .
to include the effects of seeing in the computation of @xmath35 , we weighted the data when appropriate by the seeing convolved surface brightness .
this moment has the advantage that along the major axis it does not depend on the shape of the velocity ellipsoid . ] .
it can be shown that @xmath36 scales with @xmath0 , which is the quantity we wish to determine .
thus , we can choose a model with an initial guess for @xmath0 ( called @xmath37 , see table @xmath38 for the value of @xmath37 used in each case ) , based on reasonable guesses for the galaxy distance and @xmath20 , and rescale it to any mass - to - light ratio @xmath20 and distance @xmath28 as @xmath39 \4 ) finally , the observed @xmath35 was calculated using the stellar kinematics data from corts 2008 ( fig .
@xmath40 ) .
the model - derived @xmath41 moment and the observed @xmath35 were compared , and @xmath0 was calculated along the major axis of the galaxy as @xmath42 yielding the mean @xmath0 factor ( e.g. fig @xmath43 ) .
errors were estimated as the standard deviation in @xmath0 .
therefore , the sought - after circular velocity in equation ( 4 ) corresponds to that of the model which best matches the observed @xmath35 giving the best @xmath0 .
it is important to notice that here we have derived @xmath0 , but not @xmath20 and @xmath28 independently .
these should be disentangled from @xmath0 .
so far we have derived the best circular velocity using two - integral dynamical models . in what follows
we focus on making use of these velocities to obtain the distances to our galaxies . in placing our circular velocities on a tully - fisher plot to derive the distance and additionally disentangle @xmath20 from the @xmath0 factor , we must remember that we have measured velocities only until @xmath45 0.4 @xmath46 ( where @xmath46 is the radius at which the galaxy has a surface brightness in h - band of 20.5 mag arcsec@xmath47 ) .
this makes it impossible , as discussed , to place our velocities on the usual tully - fisher relation , usually determined at @xmath1@xmath46 , in order to obtain reliable distances .
to overcome this problem , we use the approach of building synthetic rotation curves based on h@xmath0 gas velocities as in rubin ( 1985 ) .
these synthetic rotation curves represent the mean rotation curves for galaxies with different absolute magnitudes .
the procedure to derive the synthetic rotation curves was the following : \1 ) we constructed @xmath48-band infrared pseudo tully - fisher ( ptf ) relations for different normalized radii ( 0.05 , 0.1 , 0.2 , 0.3 , 0.6 , 0.7 , 0.8 , 0.9 r@xmath49 ) , using the h@xmath0 rotation curves for virgo cluster galaxies obtained by rubin 1999 .
@xmath50 ) .
the ptfs were determined using the h - band magnitudes ( table @xmath51 ) derived by gavazzi ( 1999 ) , and jarrett ( 2003 ) for the galaxies of the rubin s sample .
the advantages of using infrared magnitudes rather than @xmath52 magnitudes lie in the reduction of the dependence of the tully - fisher relation on hubble type , in the reduction of the spread due to dust absorption and stellar population variations , and in the improved ability to trace the luminous matter ( e.g. pierce & tully 1998 , and gavazzi 1996 ) .
we excluded from the t - f sample those galaxies that were identified as belonging to possible background groups ( @xmath53 mpc ) by solanes ( 2002 ) , whose could have introduced undesirable systematic errors .
we finally ended up with 34 galaxies , all with h@xmath0 kinematics and infrared photometry , which we used to build the ptfs .
we supposed that all these galaxies were located at a mean distance of 16.8 mpc , which corresponds to the distance to m 87 ( gavazzi 1999 ; nielsen & tsvetanov 2000 ; tonry 2001 ) . the usual standard tully - fisher relation corresponds to a ptf measured at @xmath1@xmath46 .
\2 ) we fitted a straight line to each of the above ptf relations corresponding to a given radius .
then , for every magnitude , we derived an interpolated gas velocity foe each radius from the family of fitted straight lines ( fig @xmath54 , top panel ) .
\3 ) the final curves for each magnitude ( and thus mass ) were constructed using the gas velocities obtained for each radius ( fig @xmath55 bottom panel ) .
these synthetic curves represent the expected shapes of the rotation curves for galaxies with different luminosities ( mass ) belonging to the virgo cluster , and obeying the tully - fisher relation .
as we see in the graph , the synthetic rotation curves do not overlap , depending only on the absolute magnitude .
the absolute magnitude corresponding to each synthetic curve was calculated by assuming that each curve correspond to a galaxy located at a distance @xmath56 mpc .
finally , the absolute magnitudes @xmath57 of our sample galaxies could be found by locating the synthetic rotation curve that best matches the stellar circular velocity curve for each sample galaxy .
the latter can be determined by minimizing the @xmath58 ( figure @xmath59 ) between the stellar circular velocity and the synthetic rotation curves .
the distance modulus follows from the observed apparent magnitude @xmath60 .
errors were estimated using the variation @xmath61 , as 3@xmath9 errors ( @xmath61=9 ) .
having derived the absolute magnitude by matching the best model circular velocity and the synthetic rotation curves , the distance @xmath28 is simply derived as @xmath62 moreover , after deriving @xmath28 we can obtain the mass - to - light ratio as @xmath63 .
the comparison between gas - based synthetic rotation curves and stellar circular velocities is justified , if we exclude the galaxies with peculiar rotation curves such as the low rotators , since gas - based synthetic rotation curves represent the typical rotation curve , and thus the gradient of the potential .
it is important to notice however that discrepancies between gas - based synthetic rotation curves , and stellar circular velocities could be found in the inner parts of the galaxies ( @xmath64 0.1 @xmath46 ) , where the gas velocity dispersion or non - circular gas motions could be important enough to affect the determination of the gas - based synthetic rotation curves , and peculiar stellar structures as circumnuclear disk can be found .
thus we restrict our comparision between stellar circular velocity and gas - based synthetic rotation curves to radii greater than 0.1 @xmath46 .
our skb distances are not free of systematic effects .
we recall that these effects are : \1 ) the presence of dust in our galaxies can reduce the amount of luminous mass that contributes to the potential , so we could underestimate the expected circular velocity .
this problem arises preferently in galaxies with prominent dust lanes .
\2 ) triaxiality and non - axisymmetric structures are ignored in the pursuit of oblate models .
this in turn introduces either an overestimate or underestimate of the mass by altering the amplitude and shape of the rotation curve .
galaxies with strong bars could be seriously affected . moreover , triaxial features are very difficult to model and their inclusion is beyond the scope of this work .
\3 ) kinematically distinct components are also important .
some galaxies have circumnuclear stellar disks ( e.g. ngc 4429 ) or counter - rotating disks which can be easily detected in the velocity fields but are hard to detect photometrically .
the latter makes the inclusion of such components in the models difficult making our models discordant with the observations in regions dominated by such features .
thus , we must exclude in our comparison the regions where these features are important .
\4 ) the dark matter halo is an important component of the potential of the galaxy especially in the outer parts .
the dark matter contribution can be represented by a logarithmic potential , @xmath65 which is the simplest potential producing a flat circular velocity @xmath66 at larger radii ( @xmath67 ) . in figure @xmath68
, we show the effect of the dark matter halo in two sample galaxies ; ngc 4450 and ngc 4569 .
the parameters @xmath69 and @xmath66 were chosen in order to match the observed h@xmath0 rotation curve from rubin ( 1999 ) .
this figure shows that the dark matter halo is indeed important in the outer parts , but our observations are limited to the inner 30 " , so we are unable to constrain the dark matter halo parameters properly , implying that we are underestimating the circular velocity .
most of these problems imply an underestimation of the circular velocity . for example , dust and dark matter have the similar effect of hiding the amount of mass that contributes to the potential .
this translates to an underestimate of the luminosity and therefore the model masses .
thus , we should consider these skb distances as lower limits to the real distance ; galaxies will tend to be somewhat further away than estimated here .
now that we have discussed in detail the methodology , we focus on the actual fittings and their errors .
the mge fits to the light distribution are presented in fig .
we see that most galaxies are reasonably well fit by the mge algorithm .
differences in surface brightness between the galaxy images and the mge model are less than 10% , which translates to velocity uncertainties of the order of 5 - 10% .
galaxies such as ngc 4064 and ngc 4450 present bigger uncertainties resulting from the existence of a central bar that obviously can not be accounted by an axisymmetric decomposition .
the @xmath70 ratio salong the major axis are displayed in fig @xmath43 . in most galaxies ,
the values are roughly constant along the major axis outside the central 5 `` , except in ngc 4293 , ngc 4424 , ngc 4429 , and ngc 4450 . in ngc 4429
there is a prominent circumnuclear disk which can not be reproduced with these simple models . the best @xmath70 ratios , therefore @xmath0 , were determined ( table @xmath38 ) as the average along the major axis outside the central 5 '' , except for ngc 4429 where we use distances greater than 12 " in order to exclude the circumnuclear disk .
the average @xmath70 ratio is displayed as a solid line in figure @xmath43 ( left panels ) , while the errors appear displayed as dashed lines .
the comparisons between the observed and best model @xmath35 moments along the major axis for the galaxies are displayed in fig @xmath43 ( right panels ) on these plots , the solid line represents the best model defined by the mean @xmath0 factor , the shaded region represents the range of uncertainty given by the standard deviation in @xmath0 . along the major axis ,
the model and the observed @xmath35 in most cases agree reasonably . to compare the synthetic rotation curves with the circular velocity curves , the latter were rescaled in radius in terms of @xmath46 in the h - band ( gavazzi 1999 ) .
the exceptions were ngc 4064 and ngc 4424 whose values were determined from h - band surface brightness profiles derived by jarrett ( 2003 ) , since gavazzi ( 1999 ) do not have measurements for these galaxies .
this is shown in fig . @xmath59 . the model - based circular velocities extend further than the stellar kinematic data .
this is justified since the circular velocities are derived from self - consistent models based on optical images which extend into the outer galaxies .
the comparison between the stellar circular velocities and the synthetic rotation curves were performed by monitoring the @xmath58 between the synthetic rotation curves and the stellar circular velocities , within the range of 0.1 0.8 @xmath46 .
this range was chosen in order to avoid large variations in velocity existing within the inner 0.05 @xmath46 ( e.g. ngc 4429 , ngc 4450 , and ngc 4569 ) , and to minimize the effect of not including a dark matter halo , which can be significant if we compare only the outer parts ( i.e. , using a range of radii of 0.6 0.8 @xmath46 ) .
comparison between the present circular velocities with those derived by rubin in h@xmath0 show that in most cases rubin s values are of the order of or lower than our circular velocities .
the exceptions are ngc 4450 , and ngc 4569 . in both cases
, this implies that the galaxies are more massive than predicted by our models . in the case of ngc 4450 , the mge fit ( fig .
@xmath71 exhibits a big discrepancy in p.a on one of the isophotes , indicating the presence of an important non - axisymmetric structure which might account for the missing mass . in ngc 4569 ,
dust lanes are prominent until at least 5 kpc through the disk , which prevented us from including all the luminous mass .
once the absolute magnitudes m@xmath72 for each galaxy is determined as is described in section 3.2 , the line - of - sight distance can be found from the distance modulus .
@xmath20 can also be derived from @xmath28 and @xmath73(table @xmath74 ) . to test whether the results on @xmath20 are physically meaningful , we plot in fig @xmath75 the relation between the r - band luminosity , and the r - band mass - to - light ratio . the best fit straight line in @xmath76
, @xmath77 space is @xmath78 this fit is roughly consistent within the errors with the fundamental plane correlation @xmath79 predicted from the virial theorem ( e.g. , faber 1987 ; bender , burstein , & faber 1992 ) .
this consistency lends credibility to the modelling presented here .
the observed variations in @xmath20 are mostly due to the variation in @xmath80 . however , there are galaxies that present very low @xmath20 .
for example ,
ngc 4694 has a value of @xmath20 of 1.1 @xmath81 0.3 @xmath82 , which is a factor @xmath1 2 lower than the expected .
this low value and its @xmath83 color ( @xmath1 0.62 ) , suggest a recent starburst .
figure @xmath84 which displays the galaxy distance distribution shows that our sample galaxies are located between 10 to 27 mpc from us , with the majority of them between 13 and 20 mpc .
two galaxies , ngc 4450 and ngc 4569 , have model locations in front of the virgo main core at about 1012 mpc from us , although the models underestimate the mass for these galaxies , so it is likely that these galaxies are not foreground ( see notes in table @xmath74 ) .
on the other hand , ngc 4651 is located at 27 mpc and is clearly a background galaxy .
it is interesting to compare our stellar kinematics - based distance estimates with solanes s hi - based distance estimates . in figure @xmath85 , we show the correlation between solanes s distances and the ratio between the skb distances and solanes s , with 3@xmath9 error bars .
> from inspection of this figure , we conclude that ngc 4293 , ngc 4450 , ngc 4569 , ngc 4580 , and ngc 4651 have skb distances that agree within the errors with those of solanes . but others such as ngc 4064 , ngc 4351 , ngc 4424 , and ngc 4606 have significant discrepancies worthy of further analysis .
ngc 4064 and ngc 4424 are particularly interesting because both galaxies have been regarded as possible foreground galaxies .
moreover , both galaxies , together with ngc 4694 , belong to the class of rubin s low rotator " galaxies , although for the case of ngc 4064 rubin s h@xmath0 velocities are much smaller than our ionized gas velocity measurements ( corts 2006 ) . solanes s distance estimations indicate a distance of 9.8 @xmath81 0.4 mpc for ngc 4064 , and 4.1 @xmath81 0.2 mpc for ngc 4424 errors ] in stark contrast with our estimations of 18.0 @xmath86 mpc , and 15.2@xmath87 mpc respectively .
the discrepancies are of a factor 1.8 for ngc 4064 and 3.7 for ngc 4424 , which are statistically significant . in the case of the other low rotator " , ngc 4694
, we determine a distance of 13.4 @xmath88 mpc . clearly
, ngc 4064 and ngc 4424 belong to the virgo cluster .
ngc 4694 might be affected by the dark matter bias in our models , so its distance should be considered a lower limit .
all these galaxies present truncated h@xmath0 disks ( koopmann & kenney 2004 ) , and their gas kinematics are peculiar ( see corts 2006 ) , indicating that they are heavily disturbed by the environment .
they all show evidence for strong gravitational interactions or mergers , which may account for their low gas velocities .
this is cautionary of the risk of using gas velocities as distance estimators for hi deficient and/or disturbed galaxies .
we conclude that the low rotators " found by rubin do not exhibit a low gas rotation velocity because they are intrinsically faint and thus closer to us , but rather because they have been heavily affected by gravitational interactions perhaps in addition to ram pressure stripping .
consider now the cases of ngc 4351 , and ngc 4606 .
both galaxies belong to rubin s sample but they were not classified as low rotators " . solanes s estimation gives distances of 11.2 @xmath81 1.3 mpc , and 12.7 @xmath81 0.3 mpc respectively which again differ from our results of 20.3 @xmath89 mpc , and 19.9 @xmath90 mpc . in ngc 4351 the discrepancy , could be due to truncation of the hi disk via icm - ism stripping , a very likely effect on this galaxy considering its high relative velocity with respect to the cluster mean ( @xmath1 1200 @xmath91 ) .
ngc 4606 has a star formation class truncated / compact " , same as ngc 4064 , and ngc 4424 , and as these galaxies it has features that suggest gravitational interactions ( corts 2006 ) , so again its small line - width could be due to the action of environmental effects .
this galaxy has a skb estimated distance of 9.9 @xmath92 mpc , still in front of virgo .
however , the skb distance for ngc 4569 may be an underestimate , because of its strong dust lanes and possibly large influence of its dark matter halo ( fig @xmath68 ) . if we include the dark matter halo in order to match the rubin s h@xmath0 rotation curve we find a barycentric distance of 3.8 mpc , locating this galaxy inside the maximum rebound radius .
the skb distance is slightly bigger than solanes s hi - based distances , although there are different hi linewidths reported in the literature for ngc 4569 , and one gets a different distance depending on which linewidths one uses .
the hi linewidth in figure @xmath59 comes from the arecibo general catalogue , a private database mantained by riccardo giovanelli and martha p. haynes , partially published by solanes ( 2002 ) .
this is not necessarly consistent with the homogenized hi - distance of solanes ( 2002 ) , which is an _ average _ between different tully - fisher ( tf ) catalogues .
if fact , the hi - distance in the tf catalogues used by solanes can be as low as 7.9 mpc ( kraan - korteweg 1988 ) and as high as 16.1 mpc ( ekholm 2000 ) . as it was shown in 4.2 , in the cases of ngc 4293 , ngc 4450 , ngc 4569 , ngc 4580 , and ngc 4651
, hi - based distances agree with our skb distances .
these galaxies have star formation classes ( koopmann & kenney 2004 ) ranging from truncated / anemic " for ngc 4293 , and ngc 4450 , to truncated / normal " for ngc 4569 and ngc 4580 , and to normal " in the case of ngc 4651 .
in contrast for the truncated / compact " galaxies the skb distances do nt agree with hi - based distances .
this suggests that truncated / compact " galaxies are not only very truncated in their gas content , but that they also have non - circular gas kinematics or gas motions which are not in the plane of the stellar disk . here
we comment on the possible existence of several galaxies in the foreground of virgo at @xmath110 mpc ( solanes 2002 ) , which were tentatively identified as a group by sanchis ( 2002 ) .
four of our sample galaxies are considered foreground by solanes .
all of them except for ngc 4569 are truncated / compact " ( i.e. , ngc 4064 , ngc 4424 , and ngc 4606 ) with skb distances in the range 1620 mpc .
ngc 4569 has a stellar - kinematics estimated distance of 9.9 @xmath92 mpc , still in front of virgo .
however , this galaxy shows strong evidence for ongoing icm - ism stripping ( vollmer 2004 ; kenney 2008 in prep ) , and its angular distance ( 1.7@xmath4 ) suggests that it lies within a region where the icm medium is dense enough for efficient stripping of its gas .
it is important to recall that our skb distances should be considered as a lower limit due to the biases .
we should expect that the real distances are larger than our estimates , so the discrepancies between solanes and our determinations should increase if we correct for the biases .
we conclude , on the basis of our overlapping sample ( we have analyzed 60% ( 4/7 ) of their foreground candidates ) , that there is no compelling evidence for the existence of multiple foreground galaxies at d@xmath9310 mpc , as suggested by solanes ( 2002 ) .
in order to determine how far these galaxies are with respect to the core of the virgo cluster , we calculated their barycentric distances @xmath94 to m87 , assumed to be at a distance of @xmath95=16.8 mpc ( gavazzi 1999 ; nielsen & tsvetanov 2000 ; tonry 2001 ) .
these distances can be derived from @xmath96 where @xmath28 is the line - of - sight distance of the galaxies , and @xmath97 the angular distance to m87 .
the results can be found in table @xmath74 and are shown in fig @xmath84 .
all galaxies are plotted on a p - v diagram .
the solid line represents the `` virial radius '' ( @xmath98 1.65 @xmath99 mpc , where @xmath100/(66.7 km s@xmath101 mpc@xmath101 ) ) derived by mamon ( 2004 ) , and the dashed line represents the largest``maximum rebound radius '' ( @xmath102 12.5 @xmath103 ) of all the sample galaxies .
the latter corresponds to the maximum radius to which particles that cross the core can bounce to . in their calculations
mamon ( 2004 ) considered only one single halo , the m 87 subcluster , which is the most massive substructure in virgo .
we see in fig @xmath84 that most of our galaxies are within the largest maximum rebound radius , and such galaxies could have crossed the core of the cluster and suffered icm - ism stripping , which reducing the size of their gas disks .
while many of the hi - deficient spirals in the outer cluster probably lost most of their gas during a core passage , others have apparently experienced icm - ism stripping in the cluster outskirts .
the virgo spiral ngc 4522 is a clear example of outer cluster icm - ism stripping ( kenney 2004 , crowl & kenney 2006 ) , which is thought to be caused by ram pressure which is enhanced due to an icm which is dynamic and lumpy rather than static and smooth .
others likely owe their hi - deficiency to gravitational interactions , such as ngc 4064 ( corts 2006 ) and ngc 4293 , which exhibit kinematical peculiarities and disturbed dust distributions , both signatures of gravitational interactions .
ngc 4450 , ngc 4569 , ngc 4580 , and ngc 4651 are galaxies further out from the maximum rebound radius , according the skb distances .
the first three galaxies are hi deficient .
ngc 4450 is an anemic galaxy with no evidence of recent icm stripping .
its maximum circular velocity as determined by us is lower than the maximum h@xmath0 velocity , indicating that our model contains less mass than needed for reproducing the observed h@xmath0 velocities , probably due to either non - axisymmetric features which were not included or a significant dark matter halo .
if we include a dark matter halo in order to match the h@xmath0 rotation curve ( fig @xmath68 ) , it is likely located inside the maximum rebound radius ( @xmath104 3.3 mpc ) .
ngc 4580 has a truncated / normal " star formation class , so it is likely to have been affected by icm
ism stripping .
however it could nt have crossed the core according to the simple cluster model , and the age of stellar population in outer disk reveals that it must have been stripped outside the cluster core ( crowl & kenney 2008 ) .
this galaxy is closer ( in the plane of the sky ) to m49 than m87 , so perhaps it is bound to the m49 subcluster rather than the m87 subcluster , and affected by the m49 environment .
the m49 subcluster is 4 times less massive than the m87 subcluster , and has a less dense icm , so it is unlikely to strip the gas from a galaxy if the icm is static and smooth , but could if the icm were dynamic and lumpy
. in general substructures in the cluster can be important in several ways : by stirring up the icm through sub - cluster mergers and locally increasing the strength of ram pressure , by changing the value of the maximum rebound radius , or by the pre - processing of galaxies in infalling groups .
the case of ngc 4651 is clearer : this galaxy is very far from the core of virgo ( + 10.3 @xmath105 mpc ) , and its star formation class is normal " , indicating that it has not been affected by the icm medium , so it probably correspond to a background galaxy rather than a virgo member . finally , the case of ngc 4569 is the most complicated .
the skb distance places this galaxy outside of the maximum rebound radius ( @xmath106 -7.0 @xmath1070.2 mpc ) , and in front of the cluster .
its high hi deficiency , truncated hi and h@xmath0 distribution , and extraplanar gas arm ( vollmer etal 2004 ; kenney etal 2008 ) are clear evidence for recent icm
ism stripping . simulations ( vollmer 2004 ) suggest that peak pressure occurred 300 myrs ago .
its extreme line - of - sight velocity ( -1200 @xmath91 relative to the cluster mean ) is consistent with a galaxy after peak pressure located on the near side of the core , but it is hard to understand how it could be as much as 6 mpc closer than the core .
the prominent dust lanes are an issue here , biasing the distance determination , so it might simply happen that this galaxy actually lies inside the maximum rebound radius . also ,
if we test the inclusion of a dark matter halo in order to match the h@xmath0 rotation curve ( fig @xmath68 ) , the corrected distance is about 13 mpc , putting this galaxy inside the maximum rebound radius .
in this paper , we have shown that stellar kinematics can be reliably used for determining distances to spiral and peculiar galaxies .
this technique presents a considerable advantage with respect to gas kinematics based techniques since , unlike the latter , it is in principle unaffected by physical processes other than gravity . * _ overcoming dust obscuration _ by using near - infrared imaging . * _ extending observations of galaxies further out_. this has the advantage of going beyond bars , and bulges , and getting a better constraint on the dark matter halo . * _ improving the reliability of synthetic rotation curves _ by increasing the number of sample galaxies used to build these curves , and extend the sample to non - cluster undisturbed galaxies . using stellar kinematics techniques , we have obtained skb distances to 11 peculiar virgo cluster galaxies .
our results can be summarized as follows : + + 1.- of the four galaxies in our sample with hi line - of - sight distances estimated by solanes ( 2002 ) to be significantly foreground to virgo , three galaxies have skb distances within 4 mpc of the cluster core , only one ( ngc 4569 ) has an skb distance ( within 3@xmath9 ) consistent with the hi los distance .
+ + 2.- galaxies with star - formation classes other than truncated / compact " have skb distances that agree with the hi - based distances within 3@xmath9 errors . on the other hand , the truncated / compact " galaxies ngc 4064 and ngc 4424 exhibit large discrepancies between hi - based and skb distances .
skb predicts distances 2 to 4 times farther away than hi - based distances .
thus , environmental effects can indeed affect both their gas content and kinematics , rendering hi - based distances unreliable .
+ + 3.- skb distance determinations show that low rotator " galaxies belong to the cluster , strongly suggesting that environmental effects are the cause of the their low gas velocity .
galaxies with highly truncated gas disks and other peculiarities such as disturbed stellar disks and non - circular or non - planar gas motions give gas - based distances smaller than skb distances .
these galaxies were perhaps subject to gravitational interactions plus icm
ism stripping . on the other hand , galaxies with normal or truncated gas disks , and normal stellar disks , presumably affected only by icm
ism stripping , have skb distances consistent within the errors.with gas based distances .
+ + 4.- most of the galaxies in our sample are within the bounce
back zone ( e.g. distances within 4 mpc of m87 ) , indicating they could have crossed the core of the cluster .
if this is the case , icm - ism stripping could explain the existence of hi - deficient galaxies in the outskirts of the cluster , although there are some galaxies that were clearly stripped outside of the core .
this is probably due to the existence of a dynamic lumpy icm stirred up by sub - cluster mergers , or gravitational interactions plus associated gas stripping .
+ we thank roeland van der marel for invaluable comments , crucial to this work and an anonymous referee for useful suggestions .
funding for this research has been provided by fundacin andes chile , fondap project grant 15010003 , chile , and nsf grant ast-0071251 .
bertola , f. , corsini , e. m. , vega beltrn , j. c. , pizella , a. , sarzi , m. , cappellari , m. , & funes , j. g. 1999 , apj , 519 , l127 bender , r. , burstein , d. , & faber , s. m. 1992 , apj , 399 , 462 bingelli , b. , tammann , g. a. , & sandage , a. ( bst ) 1987 , aj , 94 , 241 binney , j. , j. , davies , r. l. , & illingworth , g. d. 1990 , apj , 361 , 78 bohringer , h. , briel , u. g. , schwarz , r. a. , voges , w. , hartner g. , & trumper j. 1994 , nature , 368 , 828 bottema , r. 1993 , a&a , 275 , 16 cappellari , m. 2002 , mnras , 333 , 400 cappellari , m. , 2006 , mnras , 366 , 1126 cinzano , p. , & van der marel , r. p. 1994 , mnras , 270 , 325 corts , j. r. , kenney j. d. p. , & hardy , e. 2006 , aj , 131 , 747 cretton , n. , van den bosch , f. c. 1999 , apj , 514 , 704 crowl , h. h. , & kenney , j. d. p. 2006 ,
apj , 649 , l75 crowl , h. h. , & kenney , j. d. p. 2008
, apj , in press ekholm , lanoix , p. , teerikorpi , p. , fouqu , p. , & paturel , g. 2000 , a&a , 335 , 835 emsellem , e. , monnet , g. , & bacon , r. 1994 , a&a , 285 , 723 devaucouleurs , g. , devaucouleurs , a. , corwin , h. g. , buta , r. j. , paturel , g. , fouqu , p. 1991
, _ third reference catalog of bright galaxies _ , ( new york : springer - verlag ) faber , s. m. , dressler , a. , davies , r. l. , burstein , d. , lyndell - bell , d. , terlevich , r. , & wegner , g. 1987 , in nearly normal galaxies , ed . s. m. faber ( new york : springer ) , 317 falcn - barroso , j. 2006 , mnras , 369 , 529 gavazzi , g. , pierini , d. , & boselli , a. 1996 , a&a , 312 , 397 gavazzi , g. , boselli , a. , scodeggio , m. , pierini , d. , & belsole , e. 1999 , mnras , 304 , 595 guhathakurta , p. , van gorkom , j. h. , kotanyi , c. g. , & balkowski , c. 1988 , aj , 96 , 851 jarrett , t.h . ,
chester , t. , cutri , r. , schneider , s. e. , & huchra , j. p. 2003
, aj , 125 , 525 kenney , j. d. p. , van gorkom , j. h. , & vollmer , b. 2004 , aj , 127 , 3361 koopmann , r. a. , kenney , j. d. p. , & young , j. 2001 , apjs , 135 , 125 koopmann , r. a. , & kenney , j. d. p. 2004 , apj , 613 , 866 kraan - korteweg , r. c , cameron , l. m. , & tammann , g. a. 1998 , apj , 331 , 620 magorrian , j. , tremaine , s. , richstone , d. , bender , r. , bower , g. , dressler , a. , faber , s. m. , gebhardt , k. , green , r. , grillmair , c. , kormendy , j. , & lauer , t. 1998 , aj , 115 , 2285 mamon , g. a. , sanchis , t. , salvador - sol , e. , & solanes , j. m. 2004 , a&a , 414 , 445 van der marel , r. , binney , j. , & davies r. l 1990 , mnras , 245 , 582 neilsen , e. h. , & tsvetanov , z. i. 2000 , apj , 536 , 255 pierce , m. j. , & tully , r. b. 1988 , apj , 330 , 579 rubin , v. c. , burstein , d. , ford , w. k. , jr . , & thonnard , n. 1985 , apj , 289 , 81 rubin , v. c. , waterman , a. h. , & kenney , j. d. p. 1999
, aj , 118 , 236 sanchis , t. , solanes , j. m. , salvador - sol , e. , fouqu , p. , & manrique , a. 2002 , apj , 580 , 164 sanchis , t. , mamon , g. a. , salvador - sol , e. & solanes j. m. 2004 , a&a , 418 , 393 sandage , a. , & becdke , j. 1994 , _ the carnegie atlas of galaxies _ , ( washington : carnegie ) sandage , a. , & tammann , g. 1987 , _ a revised shapley - ames catalog of bright galaxies _ , ( washington : carnegie ) solanes , j. m. , sanchis , t. , salvador - sol , r. , giovanelli r. , & haynes , m. p. 2002 , aj , 124 , 2440 teerikorpi , p. , bottinelli , l. , gouguenheim , l. , & paturel , g. 1992 , a&a , 260,17 tonry , j. l. , dressler , a. , blakeslee , j. p. , 2001
, apj , 546 , 681 tully , r. b. , & fisher , j. r. 1977 , aa , 54 , 661 vollmer , b. , balkowski , c. , cayatte , v. , van driel , w. , & huchtmeier , w. 2004 , a&a , 419 , 35 yasuda , n. , ohamura , s. , fukugita , m. , & ohamura , s. 1997 , apjs , 108 , 417 gas rotation curves are represented by open symbols .
hi maximum velocities ( solanes 2002 ) are represented by the dashed line .
the numbers to the right of the synthetic rotation curves represent the absolute @xmath108 magnitudes for a galaxy having such a rotation curve at the distance of the virgo cluster . _
right panels ; _ @xmath61 between the mge circular velocity and synthetic rotation curves , dashed lines represent the 3@xmath9 confidence level . ,
scaledwidth=70.0% ] ngc 4064 & 12 04 11.2 & 18 26 36 & sbc(s ) : & sb(s)a : pec & t / c & 12.30 & 70 & 150 & 931 @xmath81 12 & 8.8 + ngc 4293 & 12 21 12.8 & 18 22 57 & sa pec & ( r)sb(s)0/a & t / a & 11.20 & 67 & 66 & 930 @xmath81 14 & 6.4 + ngc 4351 & 12 24 01.6 & 12 12 18 & sc(s ) ii.3 & sb(rs)ab : pec & t / n[s ] & 13.0 4 & 47 & 80 & 2317 @xmath81 16 & 1.7 + ngc 4424 & 12 27 11.5 & 09 25 15 & sa pec & sb(s)a : & t / c & 12.32 & 60 & 90 & 440 @xmath81 6 & 3.1 + ngc 4429 & 12 27 26.4 & 11 06 29 & s0@xmath109(6)/sa pec & sa(r)0@xmath110 & & 10.9 & 62 & 90 & 1127 @xmath81 31 & 1.5
+ ngc 4450 & 12 28 29.3 & 17 05 07 & sab pec & sa(s)ab & t / a & 10.93 & 46 & 175 & 1958 @xmath81 6 & 4.7 + ngc 4569 & 12 36 49.8 & 13 09 46 & sab(s ) i - ii & sab(rs)ab & t / n[s ] & 10.2 5 & 64 & 23 & -232 @xmath81 22 & 1.7 + ngc 4580 & 12 37 48.6 & 05 22 06 & sc / sa & sab(rs)a pec & t / n[s ] & 12.49 & 45 & 158 & 1036 @xmath81 7 & 7.2 + ngc 4606 & 12 40 57.6 & 11 54 44 & sa pec & sb(s)a : & t / c & 12.69 & 67 & 40 & 1655 @xmath81 16 & 2.5 + ngc 4651 & 12 43 42.6 & 16 23 36 & sc(r ) i - ii & sa(rs)c & n & 11.36 & 51 & 71 & 804 @xmath81 10 & 5.1 + ngc 4694 & 12 48 15.1 & 10 59 00 & amorph & sb0 pec & t / n & 12.19 & 42 & 140 & 1177 @xmath81 11 & 4.5 + ccc ngc 4064 & 8.78 & 64.0 + ngc 4293 & 7.35 & 156.30 + ngc 4351 & 10.44 & 35.50 + ngc
4424 & 9.17 & 75.69 + ngc 4429 & 6.76 & 166.50 + ngc 4450 & 6.90 & 147.4 + ngc 4569 & 6.77 & 195.00 + ngc 4580 & 8.77 & 58.79 + ngc 4606 & 9.30 & 60.93 + ngc 4651 & 8.25 & 69.06 + ngc 4694 & 9.03 & 61.30 + cccccc ngc 4064 & 70 & 1.82 @xmath81 0.26 & 32 & 58 @xmath81 8 & major axis , mean @xmath111 5 `` + ngc 4293 & 70 & 2.68 @xmath81 0.72 & 32 & 86 @xmath81 23 & major axis , mean @xmath111 5 '' + ngc 4351 & 47 & 1.12 @xmath81 0.13 & 32 & 36 @xmath81 4 & major axis , mean @xmath111 5 `` + ngc 4424 & 72 & 1.17 @xmath81 0.33 & 32 & 37 @xmath81 11 & major axis , mean @xmath111 5 '' + ngc 4429 & 66 & 3.0 @xmath81 0.5 & 32 & 96 @xmath81 16 & major axis , mean @xmath111 12 `` + ngc 4450 & 50 & 2.02 @xmath81 0.34 & 32 & 65 @xmath81 1 & major axis , mean @xmath111 5 '' + ngc 4569 & 66 & 1.14 @xmath81 0.09 & 40 & 46 @xmath81 4 & major axis , mean @xmath111 5 `` + ngc 4580 & 46 & 2.19 @xmath81 0.09 & 32 & 70 @xmath81 3 & major axis , mean @xmath111 5 '' + ngc 4606 & 68 & 1.35 @xmath81 0.15 & 32 & 43 @xmath81 5 & major axis , mean @xmath111 5 `` + ngc 4651 & 51 & 1.53 @xmath81 0.15 & 55 & 84 @xmath81 8 & major axis , mean @xmath111 5 '' + ngc 4694 & 59 & 0.43 @xmath81 0.09 & 32 & 14 @xmath81 3 & major axis , mean @xmath111 5 " + ngc 4064 & -22.5 @xmath112 & 31.3@xmath113 & 18.0 @xmath86 & 3.2 @xmath114 & 8.8 & 2.9 @xmath115 & 9.8 @xmath81 0.4 & 929 @xmath81 3 & 1.84 @xmath116 + ngc 4293 & -23.4 @xmath117 & 30.8
@xmath118 & 14.1 @xmath119 & 6.1 @xmath120 & 6.4 & 3.2 @xmath121 & 14.2 @xmath122 & 926 @xmath81 4 & 0.99 @xmath123 + ngc 4351 & -21.1 @xmath81 0.1 & 31.5 @xmath81 0.1 & 20.3 @xmath89 & 1.8 @xmath81 0.2 & 1.7 & 3.6 @xmath124 & 11.2 @xmath125 & 2310 @xmath81 2 & 1.8 @xmath126 + ngc 4424 & -21.7 @xmath81 0.3 & 30.9 @xmath81 0.3 & 15.2 @xmath81 1.9 & 2.4 @xmath81 0.8 & 3.1 & 1.8 @xmath81 1.6 & 4.1 @xmath81 0.2 & 442 @xmath81 4 & 3.7 @xmath81 0.7 + ngc 4429 & -24.3 @xmath81 0.1 & 31.1 @xmath81 0.1 & 16.3 @xmath81 0.9 & 7.0 @xmath81 1.9 & 1.5 & 0.7 @xmath81 0.7 & & 1094 @xmath81 6 & + ngc 4450 & -23.6 @xmath112 & 30.5 @xmath113 & 12.6 @xmath127 & 5.9 @xmath81 1.0 & 4.7 & 4.4
@xmath128 & 14.1 @xmath129 & 1953 @xmath81 2 & 0.89 @xmath130 + ngc 4569 & -23.20 @xmath131 & 29.97 @xmath132 & 9.9 @xmath81 0.2 & 4.7 @xmath81 0.4 & 1.7 & 7.0 @xmath81 0.2 & 9.4 @xmath133 & -222 @xmath81 6 & 1.1 @xmath126 + ngc 4580 & -22.95 @xmath81 0.05 & 31.72 @xmath81 0.05 & 22.1 @xmath81 0.5 & 3.2 @xmath81 0.2 & 7.2 & 5.8 @xmath81 0.5 & 19.3 @xmath81 1.1 & 1031 @xmath81 4 & 1.1 @xmath81 0.1 + ngc 4606 & -22.20 @xmath134 & 31.50 @xmath135 & 19.9 @xmath90 & 2.16 @xmath81 0.3 & 2.5 & 3.25 @xmath136 & 12.7 @xmath118 & 1636 @xmath81 5 & 1.6 @xmath81 0.1 + ngc 4651 & -23.90 @xmath137 & 32.15 @xmath138 & 26.9 @xmath105 & 3.1 @xmath81 0.3 & 5.1 & 10.3 @xmath105 & 22.3 @xmath139 & 784 @xmath81 2 & 1.2 @xmath81 0.4 + ngc 4694 & -21.6 @xmath81 0.2 & 30.6 @xmath81 0.2 & 13.4 @xmath88 & 1.1 @xmath81 0.3 & 4.5 & 3.6 @xmath140 & & 1174 @xmath81 22 & + | we present distance estimates for eleven peculiar virgo cluster spiral galaxies based on measurements of the stellar kinematics of their central 2 kpc .
stellar circular velocities were obtained using two - integral dynamical models .
distances were obtained by comparing , at each radius , the stellar circular velocities with synthetic h@xmath0 rotation curves derived from nir tully - fisher relations .
the results show that most of our galaxies are located within 4 mpc of the core of the cluster .
three of these galaxies , previously classified as low rotator galaxies " or with truncated / compact " h@xmath0 radial distributions , have stellar kinematics - based distances that are discrepant with hi - based distances by at least 60% , and are likely to be located within the virial radius of the cluster .
these discrepancies appear due to very truncated gas distributions plus non - circular gas motions or gas motions not in the plane of the stellar disk , perhaps as the result of gravitational interactions .
our results show that environmental effects can significantly reduce the measured hi linewidths for some disturbed cluster galaxies , thus affecting the accurate determination of distances based on gas kinematics methods . |
strassen s algorithm @xcite was the first sub - cubic algorithm for matrix multiplication .
its improvement by winograd @xcite led to a highly practical algorithm .
the best asymptotic complexity for this computation has been successively improved since then , down to @xmath0 in @xcite ( see @xcite for a review ) , but strassen - winograd s still remains one of the most practicable .
former studies on how to turn this algorithm into practice can be found in @xcite and references therein for numerical computation and in @xcite for computations over a finite field .
+ in this paper , we propose new schedules of the algorithm , that reduce the extra memory allocation , by three different means : by introducing a few pre - additions , by overwriting the input matrices , or by using a first recursive level of classical multiplication .
these schedules can prove useful for instance for memory efficient computations of the rank , determinant , nullspace basis , system resolution , matrix inversion ... indeed , the matrix multiplication based lqup factorization of @xcite can be computed with no other temporary allocations than the ones involved in its block matrix multiplications @xcite . therefore the improvements on the memory requirements of the matrix multiplication , used together for instance with cache optimization strategies @xcite , will directly improve these higher level computations .
we only consider here the computational complexity and space complexity , counting the number of arithmetic operations and memory allocations .
the focus here is neither on stability issues , nor really on speed improvements .
we rather study potential memory space savings .
further studies have thus to be made to assess for some gains for in - core computations or to use these schedules for numerical computations .
they are nonetheless already useful for exact computations , for instance on integer / rational or finite field applications @xcite .
the remainder of this paper is organized as follows : we review strassen - winograd s algorithm and existing memory schedules in sections [ sec : algo ] and [ ssec : constinput ] .
we then present in section [ sec : pebble ] the dynamic program we used to search for schedules .
this allows us to give several schedules overwriting their inputs in section [ sec : overwrite ] , and then a new schedule for @xmath1 using only two extra temporaries in section [ sec : hyb ] , all of them preserving the leading term of the arithmetic complexity .
finally , in section [ sec : mix ] , we present a generic way of transforming non in - place matrix multiplication algorithms into in - place ones ( without any extra temporary space ) , with a small constant factor overhead
. then we recapitulate in table [ tab : resume ] the different available schedules and give their respective features .
we first review strassen - winograd s algorithm , and setup the notations that will be used throughout the paper .
+ let @xmath2 and @xmath3 be powers of @xmath4 .
let @xmath5 and @xmath6 be two matrices of dimension @xmath7 and @xmath8 and let @xmath9 . consider the natural block decomposition : @xmath10 where @xmath11 and @xmath12 respectively have dimensions @xmath13 and @xmath14 .
winograd s algorithm computes the @xmath15 matrix @xmath9 with the following 22 block operations : + 8 additions : @xmath16 7 recursive multiplications : @xmath17 7 final additions : @xmath18 the result is the matrix : @xmath19 $ ] .
figure [ fig : winotask ] illustrates the dependencies between these tasks .
unlike the classic multiplication algorithm , winograd s algorithm requires some extra temporary memory allocations to perform its 22 block operations .
we first consider the basic operation @xmath20 .
the best known schedule for this case was given by @xcite .
we reproduce a similar schedule in table [ tab : schedule : ab ] . 1 & @xmath21 & @xmath22 & 12 & @xmath23 & @xmath22 + 2 & @xmath24 & @xmath25 & 13 & @xmath26 & @xmath27 + 3 & @xmath28 & @xmath29 & 14 & @xmath30 & @xmath29 + 4 & @xmath31 & @xmath22 & 15 & @xmath32 & @xmath27 + 5 & @xmath33 & @xmath25 & 16 & @xmath34 & @xmath35 + 6 & @xmath36 & @xmath35 & 17 & @xmath37 & @xmath27 + 7 & @xmath38 & @xmath22 & 18 & @xmath39 & @xmath25 + 8 & @xmath40 & @xmath25 & 19 & @xmath41 & @xmath42 + 9 & @xmath43 & @xmath27 & 20 & @xmath44 & @xmath29 + 10 & @xmath45 & @xmath22 & 21 & @xmath46 & @xmath42 + 11 & @xmath47 & @xmath42 & 22 & @xmath48 & @xmath42 + it requires two temporary blocks @xmath22 and @xmath25 whose dimensions are respectively equal to @xmath49 and @xmath50 .
thus the extra memory used is : @xmath51 summing these temporary allocations over every recursive levels leads to a total amount of memory , where for brevity @xmath52 : @xmath53 we can prove in the same manner the following lemma : [ lem : sum ] let @xmath54 , @xmath3 and @xmath55 be powers of two , @xmath56 be homogeneous , @xmath57 and @xmath58 be a function such that @xmath59 then @xmath60 . in the remainder of the paper
, we use @xmath61 to denote the amount of extra memory used in table number @xmath62 .
the amount of extra memory we consider is always the sum up to the last recursion level .
finally , assuming @xmath63 gives a total extra memory requirement of @xmath64 for the more general operation @xmath65 , a first nave method would compute the product @xmath66 using the scheduling of table [ tab : schedule : ab ] , into a temporary matrix @xmath67 and finally compute @xmath68 .
it would require @xmath69 extra memory allocations in the square case .
+ now the schedule of table [ tab : schedule : abc ] due to @xcite only requires 3 temporary blocks for the same number of operations ( @xmath70 multiplications and @xmath71 additions ) . 1 & @xmath31 & @xmath22 & 12 & @xmath45 & @xmath22 + 2 & @xmath33 & @xmath25 & 13 & @xmath39 & @xmath25 + 3 & @xmath72 & @xmath73 & 14 & @xmath74 & @xmath27 + 4 & @xmath75 & @xmath35 & 15 & @xmath76 & @xmath27 + 5 & @xmath77 & @xmath27 & 16 & @xmath78 & @xmath29 + 6 & @xmath38 & @xmath22 & 17 & @xmath21 & @xmath22 + 7 & @xmath79 & @xmath25 & 18 & @xmath24 & @xmath25 + 8 & @xmath80 & @xmath73 & 19 & @xmath81 & @xmath73 + 9 & @xmath82 & @xmath42 & 20 & @xmath83 & @xmath35 + 10 & @xmath84 & @xmath73 & 21 & @xmath85 & @xmath29 + 11&@xmath86 & @xmath42 & 22 & & + the required three temporary blocks @xmath87 have dimensions @xmath88 , @xmath50 and @xmath89 . since the two temporary blocks in schedule [ tab : schedule : ab ] are smaller than the three ones here , we have @xmath90 .
hence , using lemma [ lem : sum ] , we get @xmath91 with @xmath63 , this gives @xmath92 + we propose in table [ tab : abc:2tmp ] a new schedule for the same operation @xmath93 only requiring two temporary blocks . +
our new schedule is more efficient if some inner calls overwrite their temporary input matrices .
we now present some overwriting schedules and the dynamic program we used to find them .
we used a brute force search algorithm to get some of the new schedules that will be presented in the following sections .
it is very similar to the pebble game of huss - lederman et al .
@xcite . + a sequence of computations is represented as a directed graph , just like figure [ fig : winotask ] is built from winograd s algorithm . +
a node represents a program variable .
the nodes can be classified as initials ( when they correspond to inputs ) , temporaries ( for intermediate computations ) or finals ( results or nodes that we want to keep , such as ready - only inputs ) .
+ the edges represent the operations ; they point from the operands to the result . + a pebble represents an allocated memory .
we can put pebbles on any nodes , move or remove them according to a set of simple rules shown below .
+ when a pebble arrives to a node , the computation at the associated variable starts , and can be `` partially '' or `` fully '' executed .
if not specified , it is assumed that the computation is fully executed .
+ edges can be removed , when the corresponding operation has been computed .
+ the last two points are especially useful for accumulation operations : for example , it is possible to try schedule the multiplication separately from the addition in an otherwise recursive @xmath94 call ; the edges involved in the multiplication operation would then be removed first and the accumulated part later .
they are also useful if we do not want to fix the way some additions are performed : if @xmath95 the associativity allows different ways of computing the sum and we let the program explore these possibilities . at the beginning of the exploration , each initial node has a pebble and we may have a few extra available pebbles .
the program then tries to apply the following rules , in order , on each node .
the program stops when every final node has a pebble or when no further moves of pebbles are possible : _ computing a result / removing edges .
_ if a node has a pebble and parents with pebbles , then the operation can be performed and the corresponding edges removed .
the node is then at least partially computed . _
freeing some memory / removing a pebble . _
if a node is isolated and not final , its pebble is freed .
this means that we can reclaim the memory here because this node has been fully computed ( no edge pointing to it ) and is no longer in use as an operand ( no edge initiating from it ) .
_ computing in place / moving a pebble . _
if a node @xmath96 has a full pebble and a single empty child node @xmath97 and if other parents of @xmath97 have pebbles on them , then the pebble on @xmath96 may be transferred to @xmath97 ( corresponding edges are removed ) .
this means an operation has been made in place in the parent @xmath96 s pebble . _ using more memory / adding a pebble . _
if parents of an empty node @xmath98 have pebbles and a free pebble is available , then this pebble can be assigned to @xmath98 and the corresponding edges are removed .
this means that the operation is computed in a new memory location . _ copying some memory / duplicating a pebble .
_ a computed node having a pebble can be duplicated .
the edges pointed to or from the original node are then rearranged between them .
this means that a temporary result has been copied into some free place to allow more flexibility .
we now relax some constraints on the previous problem : the input matrices @xmath5 and @xmath6 can be overwritten , as proposed by @xcite . for the sake of simplicity , we first give schedules only working for square matrices ( i.e. @xmath63 and any memory location
is supposed to be able to receive any result of any size ) .
we nevertheless give the memory requirements of each schedule as a function of @xmath54 ; @xmath3 and @xmath55 .
therefore it is easier in the last part of this section to adapt the proposed schedules partially for the general case . in the tables , the notation @xmath99 ( resp .
@xmath100 denotes the use of the algorithm from table [ tab : schedule : ab ] ( resp .
table [ tab : schedule : abc ] ) as a subroutine .
otherwise we use the notation @xmath101 to denote a recursive call or the use of one of our new schedules as a subroutine .
we propose in table [ tab : ab : inplace ] a new schedule that computes the product @xmath20 without any temporary memory allocation .
the idea here is to find an ordering where the recursive calls can be made also in place such that the operands of a multiplication are no longer in use after the multiplication has completed because they are overwritten .
an exhaustive search showed that no schedule exists overwriting less than four sub - blocks . 1 & @xmath21 & @xmath42 & 12 & @xmath45 & @xmath102 + 2 & @xmath31 & @xmath103 & 13 & @xmath104 & @xmath35 + 3 & @xmath33 & @xmath35 & 14 & @xmath26 & @xmath35 + 4 & @xmath24 & @xmath105 & 15 & @xmath106 & @xmath27 + 5 & @xmath107 ) & @xmath29 & 16 & @xmath48 & * @xmath42 + 6 & @xmath38 & @xmath27 & 17 & @xmath32 & @xmath27 + 7 & @xmath108 & @xmath42 & 18 & @xmath30 & @xmath35 + 8 & @xmath40 & @xmath12 & 19 & @xmath44 & * @xmath29 + 9 & @xmath109 & @xmath11 & 20 & @xmath34 & * @xmath35 + 10 & @xmath39 & @xmath35 & 21 & @xmath110 & @xmath111 + 11 & @xmath112 & @xmath103 & 22 & @xmath37 & * @xmath27 + * * * * note that this schedule uses only two blocks of @xmath6 and the whole of @xmath5 but overwrites all of @xmath5 and @xmath6 . for instance
the recursive computation of @xmath113 requires overwriting parts of @xmath111 and @xmath114 too . using another schedule as well as back - ups of overwritten parts into some available memory in the following
, we will denote by ` ip`for ` i`nplace , either one of these two schedules .
+ we present in tables [ tab : ab : ipleft ] and [ tab : ab : ipright ] two new schedules overwriting only one of the two input matrices , but requiring an extra temporary space .
these two schedules are denoted ` ovl`and ` ovr ` .
the exhaustive search also showed that no schedule exists overwriting only one of @xmath5 and @xmath6 and using no extra temporary . 1 & @xmath21 & @xmath35 & 12 & @xmath115 & @xmath29 + 2 & @xmath31 & @xmath103 & 13 & @xmath39 & @xmath11 + 3 & @xmath38 & @xmath27 & 14 & @xmath26 & @xmath29 + 4 & @xmath33 & @xmath29 & 15 & @xmath32 & @xmath27 + 5 & @xmath116 & @xmath42 & 16 & @xmath30 & @xmath29 + 6 & @xmath24 & @xmath11 & 17 & @xmath34 & * @xmath35 + 7 & @xmath117 & @xmath22 & 18 & @xmath37 & * @xmath27 + 8 & @xmath40 & @xmath11 & 19 & @xmath118 & @xmath22 + 9 & @xmath109 & @xmath35 & 20 & @xmath48 & * @xmath42 + 10 & @xmath45 & @xmath29 & 21 & @xmath112 & @xmath103 + 11 & @xmath119 & @xmath103 & 22 & @xmath44 & * @xmath29 + * * * * 1 & @xmath21 & @xmath35 & 12 & @xmath120 & @xmath105 + 2 & @xmath31 & @xmath29 & 13 & @xmath45 & @xmath12 + 3 & @xmath33 & @xmath27 & 14 & @xmath26 & @xmath29 + 4 & @xmath121 & @xmath42 & 15 & @xmath32 & @xmath27 + 5 & @xmath38 & @xmath12 & 16 & @xmath30 & @xmath29 + 6 & @xmath24 & @xmath105 & 17 & @xmath34 & * @xmath35 + 7 & @xmath117 & @xmath22 & 18 & @xmath44 & * @xmath29 + 8 & @xmath40 & @xmath105 & 19 & @xmath110 & @xmath105 + 9 & @xmath109 & @xmath35 & 20 & @xmath37 & * @xmath27 + 10 & @xmath39 & @xmath27 & 21 & @xmath122 & @xmath105 + 11 & @xmath123 & @xmath29 & 22 & @xmath48 & * @xmath42 + * * * * we note that we can overwrite only two blocks of @xmath5 in ` ovl`when the schedule is modified as follows : 18bis & @xmath124 & @xmath103 + 19bis & @xmath125 & @xmath111 + 21 & @xmath120 & @xmath103 + similarly , for ` ovr ` , we can overwrite only two blocks of @xmath6 using copies on lines 20 and 21 and ` ovl`on line 19 . +
we now compute the extra memory needed for the schedule of table [ tab : ab : ipright ] .
the size of the temporary block @xmath22 is @xmath126 , the extra memory required for table [ tab : ab : ipright ] hence satisfies : @xmath127 .
we now consider the operation @xmath128 , where the input matrices @xmath5 and @xmath6 can be overwritten .
we propose in table [ tab : abc : overwrite ] a schedule that only requires @xmath4 temporary block matrices , instead of the @xmath129 in table [ tab : schedule : abc ] .
this is achieved by overwriting the inputs and by using two additional pre - additions ( @xmath130 and @xmath131 ) on the matrix @xmath132 . 1 & @xmath133 & @xmath35 & 13 & @xmath134 & @xmath29 + 2 & @xmath31 & @xmath22 & 14 & @xmath45 & @xmath102 + 3 & @xmath33 & @xmath25 & 15 & @xmath135 & @xmath22 + 4 & @xmath136 & @xmath29 & 16 & @xmath137 & @xmath42 + 5 & @xmath24 & @xmath105 & 17 & @xmath48 & @xmath42 + 6 & @xmath21 & @xmath103 & 18 & @xmath26 & @xmath22 + 5 & @xmath138 & @xmath35 & 17 & @xmath30 & @xmath35 + 8 & @xmath38 & @xmath103 & 20 & @xmath32 & @xmath22 + 9 & @xmath40 & @xmath105 & 21 & @xmath44 & @xmath29 + 10 & @xmath139 & @xmath27 & 22 & @xmath34 & @xmath35 + 11 & @xmath140 & @xmath25 & 23 & @xmath141 & @xmath27 + 12 & @xmath39 & @xmath22 & 24 & @xmath37 & @xmath27 + we also propose in table [ tab : abc : overright ] a schedule similar to table [ tab : abc : overwrite ] overwriting only for instance the right input matrix .
it also uses only two temporaries , but has to call the ` ovr`schedule . the extra memory required by @xmath22 and @xmath25 in table [ tab : abc : overwrite ]
is @xmath142 .
hence , using lemma [ lem : sum ] : @xmath143 1 & @xmath144 & @xmath35 & 13 & @xmath145 & @xmath42 + 2 & @xmath33 & @xmath22 & 14 & @xmath38 & @xmath25 + 3 & @xmath136 & @xmath29 & 15 & @xmath146 & @xmath114 + 4 & @xmath24 & @xmath105 & 16 & @xmath45 & @xmath25 + 5 & @xmath147 & @xmath25 & 17 & @xmath26 & @xmath114 + 6 & @xmath148 & @xmath35 & 18 & @xmath30 & @xmath35 + 7 & @xmath31 & @xmath25 & 19 & @xmath32 & @xmath114 + 8 & @xmath40 & @xmath105 & 20 & @xmath44 & @xmath29 + 9 & @xmath149 & @xmath27 & 21 & @xmath48 & @xmath42 + 10 & @xmath39 & @xmath22 & 22 & @xmath34 & @xmath35 + 11 & @xmath150 & @xmath29 & 23 & @xmath141 & @xmath27 + 12 & @xmath151 & @xmath22 & 24 & @xmath37 & @xmath27 + the extra memory @xmath152 required for table [ tab : abc : overright ] in the top level of recursion is : @xmath153 we clearly have @xmath154 and : @xmath155 compared with the schedule of table [ tab : schedule : abc ] , the possibility to overwrite the input matrices makes it possible to have further in place calls and replace recursive calls with accumulation by calls without accumulation .
we show in theorem [ thm : cost ] that this enables us to almost compensate for the extra additions performed .
we now examine the sizes of the temporary locations used , when the matrices involved do not have identical sizes .
we want to make use of table [ tab : ab : inplace ] for the general case .
+ firstly , the sizes of @xmath5 and @xmath6 must not be bigger than that of @xmath132 ( we need @xmath156 ) .
indeed , let s play a pebble game that we start with pebbles on the inputs and @xmath157 extra pebbles that are the size of a @xmath158 .
no initial pebble can be moved since at least two edges initiate from the initial nodes .
if the size of @xmath159 is larger that the size of the free pebbles , then we can not put a free pebble on the @xmath160 nodes ( they are too large ) .
we can not put either a pebble on @xmath161 or @xmath113 since their operands would be overwritten .
so the size of @xmath159 is smaller or equal than that of @xmath158 .
the same reasoning applies for @xmath162 .
+ then , if we consider a pebble game that was successful , we can prove in the same fashion that either the size of @xmath5 or the size of @xmath6 can not be smaller that of @xmath132 ( so one of them has the same size as @xmath132 ) .
+ finally , table [ tab : ab : inplace ] shows that this is indeed possible , with @xmath163 .
it is also possible to switch the roles of @xmath54 and @xmath55 .
+ now in tables 4 to 7 , we need that @xmath5 , @xmath6 and @xmath132 have the same size .
generalizing table [ tab : ab : inplace ] whenever we do not have a dedicated in - place schedule can then done by cutting the larger matrices in squares of dimension @xmath164 and doing the multiplications / product with accumulations on these smaller matrices using algorithm 1 to 7 and free space from @xmath5 , @xmath6 or @xmath132.since algorithms 1 to 7 require less than @xmath165 extra memory , we can use them as soon as one small matrix is free .
+ we now propose an example in algorithm [ alg : ipmm0 ] for the case @xmath166 : let @xmath167 and @xmath168 .
split @xmath169 $ ] , @xmath170 $ ] and @xmath171 $ ] @xmath172 now we use @xmath173 as temporary space .
@xmath174 @xmath175 algorithm [ alg : ipmm0 ] computes the product @xmath176 in place , overwriting @xmath5 and @xmath6 .
finally , we generalize the accumulation operation from table [ tab : abc : overright ] to the rectangular case
. we can no longer use dedicated square algorithms .
this is done in table [ tab : abc : overright : gen ] , overwriting only one of the inputs and using only two temporaries , but with 5 recursive accumulation calls : 1 & @xmath144 & @xmath35 & 13 & @xmath177 & @xmath42 + 2 & @xmath33 & @xmath22 & 14 & @xmath48 & @xmath42 + 3 & @xmath136 & @xmath29 & 15 & @xmath38 & @xmath25 + 4 & @xmath24 & @xmath105 & 16 & @xmath178 & @xmath22 + 5 & @xmath21 & @xmath25 & 17 & @xmath30 & @xmath35 + 6 & @xmath179 & @xmath35 & 18 & @xmath44 & @xmath29 + 7 & @xmath31 & @xmath25 & 19 & @xmath34 & @xmath35 + 8 & @xmath40 & @xmath105 & 20 & @xmath32 & @xmath22 + 9 & @xmath180 & @xmath27 & 21 & @xmath45 & @xmath25 + 10 & @xmath39 & @xmath22 & 22 & @xmath181 & @xmath27 + 11 & @xmath182 & @xmath29 & 23 & @xmath37 & @xmath27 + 12 & @xmath80 & @xmath22 & 24 & & + for instance , in table [ tab : abc : overright : gen ] , the last multiplication ( line 22 , @xmath181 ) could have been made by a call to the in place algorithm , would @xmath27 be large enough .
this is not always the case in a rectangular setting .
now , the size of the extra temporaries required in table [ tab : abc : overright : gen ] is @xmath183 and @xmath184 is equal to : @xmath185 if @xmath186 or @xmath187 , then @xmath188 : @xmath189 otherwise @xmath190 and : @xmath191 in the square case , this simplifies into @xmath192 + in addition , if the size of @xmath6 is bigger than that of @xmath5 , then one can store @xmath193 , for instance within @xmath105 , and separate the recursive call @xmath194 into a multiplication and an addition , which reduces the arithmetic complexity .
otherwise , a scheduling with only 4 recursive calls exists too , but we need for instance to recompute @xmath195 at step @xmath196 .
by combining techniques from sections [ ssec : constinput ] and [ sec : overwrite ] , we now propose in table [ tab : abc:2tmp ] a hybrid algorithm that performs the computation @xmath128 with constant input matrices @xmath5 and @xmath6 , with a lower extra memory requirement than the scheduling of @xcite ( table [ tab : schedule : abc ] ) . we have to pay a price of order @xmath197 extra operations , as we need to compute the temporary variable @xmath198 twice . 1 & @xmath133 & @xmath35 & 14 & @xmath199 & @xmath42 + 2 & @xmath200 & @xmath27 & 15 & @xmath48 & @xmath42 + 3 & @xmath31 & @xmath22 & 16 & @xmath201 & @xmath27 + 4 & @xmath33 & @xmath25 & 17 & @xmath21 & @xmath22 + 5 & @xmath202 & @xmath27 & 18 & @xmath24 & @xmath25 + 6 & @xmath38 & @xmath22 & 19 & @xmath203 & @xmath29 + 7 & @xmath79 & @xmath25 & & @xmath204 & + 8 & @xmath205 & @xmath29 & 20 & @xmath206 & @xmath35 + 9 & @xmath45 & @xmath22 & 21 & @xmath207 & @xmath25 + 10 & @xmath208 & @xmath35 & 22 & @xmath209 & @xmath25 + 11 & @xmath210 & @xmath27 & 23 & @xmath211 & @xmath25 + 12 & @xmath80 & @xmath22 & 24 & @xmath212 & @xmath29 + 13 & @xmath213 & @xmath29 & & @xmath214 & + again , the two temporary blocks @xmath22 and @xmath25 have dimensions @xmath215 so that : @xmath216 in all cases , @xmath217 but @xmath218 is not as large as the size of the two temporaries in table [ tab : abc : overwrite ] . we therefore get : @xmath219 assuming @xmath63 , one gets @xmath220 which is smaller than the extra memory requirement of table [ tab : schedule : abc ] . +
following the improvements of the previous section , the question was raised whether extra memory allocation was intrinsic to sub - cubic matrix multiplication algorithms .
more precisely , is there a matrix multiplication algorithm computing @xmath20 in @xmath221 arithmetic operations without extra memory allocation and without overwriting its input arguments ?
we show in this section that a combination of winograd s algorithm and a classic block algorithm provides a positive answer .
furthermore this algorithm also improves the extra memory requirement for the product with accumulation @xmath128 .
the key idea is to split the result matrix @xmath132 into four quadrants of dimension @xmath222 .
the first three quadrants @xmath223 and @xmath29 are computed using fast rectangular matrix multiplication , which accounts for @xmath224 standard winograd multiplications on blocks of dimension @xmath222 .
the temporary memory for these computations is stored in @xmath35 .
lastly , the block @xmath35 is computed recursively up to a base case , as shown on algorithm [ alg : ipmm ] .
this base case , when the matrix is too small to benefit from the fast routine , is then computed with the classical matrix multiplication .
split @xmath225 , @xmath226 $ ] and @xmath227 $ ] where each @xmath228 and @xmath229 have dimension @xmath222 .
@xmath230 @xmath231 @xmath232 @xmath233 @xmath234 @xmath235 @xmath236 the complexity of algorithm [ alg : ipmm ] is : @xmath237 when @xmath238 .
recall that the cost of winograd s algorithm for square matrices is @xmath239 for the operation @xmath20 and @xmath240 for the operation @xmath241 .
the cost @xmath242 of algorithm [ alg : ipmm ] is given by the relation @xmath243 the base case being a classical dot product : @xmath244 .
thus , @xmath245 . for any @xmath54 ,
@xmath55 and @xmath3 , algorithm [ alg : ipmm ] is in place .
w.l.o.g , we assume that @xmath246 ( otherwise we could use the transpose ) . the exact amount of extra memory from algorithms in table [ tab : schedule : ab ] and [ tab : schedule : abc ] is respectively given by eq .
( [ eq : schedule : ab ] ) and ( [ eq : schedule : abc ] ) .
+ if we cut @xmath6 into @xmath247 stripes at recursion level @xmath62 , then the sizes for the involved submatrices of @xmath5 ( resp .
@xmath6 ) are @xmath248 ( reps .
@xmath249 ) . the lower right corner submatrix of @xmath132 that we would like to use as temporary space has a size @xmath250 .
thus we need to ensure that the following inequality holds : @xmath251 it is clear that @xmath252 which simplifies the previous inequality .
let us now write @xmath253 , @xmath254 and @xmath255 .
we need to find , for every @xmath62 an integer @xmath256 so that eq .
( [ eq : aim ] ) holds . in other words , let us show that there exists some @xmath257 such that , for any @xmath258 , the inequality @xmath259 holds . then the fact that @xmath260 provides at least one such @xmath261 .
+ as the requirements in algorithm [ alg : ipmm ] ensure that @xmath262 and @xmath263 , there just remains to prove that @xmath264 . since @xmath265 and again @xmath266 , algorithm [ alg : ipmm ] is indeed in place .
hence a fully in - place @xmath221 algorithm is obtained for matrix multiplication .
the overhead of this approach appears in the multiplicative constant of the leading term of the complexity , growing from @xmath267 to @xmath268 . +
this approach extends to the case of matrices with general dimensions , using for instance peeling or padding techniques . + it is also useful if any sub - cubic algorithm is used instead of winograd s .
for instance , in the square case , one can use the product with accumulation in table [ tab : abc:2tmp ] instead of table [ tab : schedule : abc ] . in the case of computing the product with accumulation
, the matrix @xmath132 can no longer be used as temporary storage , and extra memory allocation can not be avoided .
again we can use the idea of the classical block matrix multiplication at the higher level and call winograd algorithm for the block multiplications . as in the previous subsection
, @xmath132 can be divided into four blocks and then the product can be made with 8 calls to winograd algorithm for the smaller blocks , with only one extra temporary block of dimension @xmath269 . + more generally , for square @xmath270 matrices , @xmath132 can be divided in @xmath271 blocks of dimension @xmath272 .
then one can compute each block with winograd algorithm using only one extra memory chunk of size @xmath273 .
the complexity is changed to @xmath274 which is @xmath275 for an accumulation product with winograd s algorithm . using the parameter @xmath276 , one can then balance the memory usage and the extra arithmetic operations .
for example , with @xmath277 , @xmath278 and with @xmath279 , @xmath280 note that one can use the algorithm of table [ tab : abc:2tmp ] instead of the classical winograd accumulation as the base case algorithm .
then the memory overhead drops down to @xmath281 and the arithmetic complexity increases to @xmath282 .
with constant input matrices , we reduced the number of extra memory allocations for the operation @xmath283 from @xmath165 to @xmath284 , by introducing two extra pre - additions .
as shown below , the overhead induced by these supplementary additions is amortized by the gains in number of memory allocations . if the input matrices can be overwritten , we proposed a fully _ in - place _
schedule for the operation @xmath20 without any extra operations .
we also proposed variants for the operation @xmath285 , where only one of the input matrices is being overwritten and one temporary is required .
these subroutines allow us to reduce the extra memory allocations required for the @xmath286 operation without overwrite : the extra required temporary space drops from @xmath165 to only @xmath284 , at a negligible cost . some algorithms with an even more reduced memory usage , but with some increase in arithmetic complexity , are also shown .
table [ tab : resume ] gives a summary of the features of each schedule that has been presented .
the complexities are given only for @xmath287 being a power of @xmath4 .
& total extra memory & total # of extra allocations & arithmetic complexity + & & & & & & + & table [ tab : schedule : ab ] @xcite & constant & @xmath4 & @xmath288 & @xmath289 & @xmath290 + & table [ tab : ab : inplace ] & both overwritten & @xmath291 & @xmath291 & @xmath291 & @xmath290 + & table [ tab : ab : ipleft ] or [ tab : ab : ipright ] & @xmath5 or @xmath6 overwritten & @xmath292 & @xmath293 & @xmath294 & @xmath290 + & [ sec : fully ] & constant & @xmath291 & @xmath291 & @xmath291 & @xmath295 + & & & & & & + & table [ tab : schedule : abc ] @xcite & constant & @xmath129 & @xmath165 & @xmath296 & @xmath297 + & table [ tab : abc : overwrite ] & both overwritten & @xmath4 & @xmath288 & @xmath298 & @xmath299 + & table [ tab : abc : overright ] & @xmath6 overwritten & @xmath4 & @xmath288 & @xmath300 & @xmath299 + & table [ tab : abc:2tmp ] & constant & @xmath4 & @xmath288 & @xmath301 & @xmath302 + & [ sec : red ] & constant & n / a & @xmath303 & @xmath303 & @xmath304 + & [ sec : red ] & constant & n / a & @xmath305 & @xmath305 & @xmath306 + for @xmath318 , the arithmetic complexity of @xcitesatisfies @xmath319 hence @xmath320 ; its memory overhead satisfies @xmath321 which is @xmath322 ; its total number of allocations satisfies @xmath323 which is @xmath324 the arithmetic complexity of the schedule of table [ tab : abc : overwrite ] satisfies @xmath325
so that @xmath326 ; its number of extra memory satisfies @xmath327 which is @xmath328 ; its total number of allocations satisfies @xmath329 which is @xmath330 .
the arithmetic complexity of table [ tab : abc : overright ] schedule satisfies @xmath331 so that @xmath332 ; its number of extra memory satisfies @xmath333 which is @xmath334 ; its total number of allocations satisfies @xmath335 which is @xmath336 .
the arithmetic complexity of the schedule of table [ tab : abc:2tmp ] satisfies @xmath337 so that @xmath338 ; its number of extra memory satisfies @xmath339 which is @xmath340 ; its total number of allocations satisfies @xmath341 which is @xmath342 .
for instance , by adding up allocations and arithmetic operations in table [ tab : resume ] , one sees that the overhead in arithmetic operations of the schedule of table [ tab : abc:2tmp ] is somehow amortized by the decrease of memory allocations .
thus it makes it theoretically competitive with the algorithm of @xcite as soon as @xmath343 .
also , problems with dimensions that are not powers of two can be handled by combining the cuttings of algorithms [ alg : ipmm0 ] and [ alg : ipmm ] with peeling or padding techniques .
moreover , some cut - off can be set in order to stop the recursion and switch to the classical algorithm .
the use of these cut - offs will in general decrease both the extra memory requirements and the arithmetic complexity overhead . for instance we show on table [
tab : ipmmperf ] the relative speed of different multiplication procedures for some double floating point rectangular matrices .
we use atlas-3.9.4 for the blas and a cut - off of 1024 .
we see that pour new schedules perform quite competitively with the previous ones and that the savings in memory enable larger computations ( mt for memory thrashing ) .
s. huss - lederman , e. m. jacobson , j. r. johnson , a. tsao , and t. turnbull .
implementation of strassen s algorithm for matrix multiplication . in acm ,
editor , _ supercomputing 96 conference proceedings : november 1722 , pittsburgh , pa_. acm press and ieee computer society press , 1996 .
www.supercomp.org / sc96/proceedings / sc96proc / jacobson/. s. huss - lederman , e. m. jacobson , j. r. johnson , a. tsao , and t. turnbull .
s algorithm for matrix multiplication : modeling analysis , and implementation .
technical report , center for computing sciences , nov .
ccs - tr-96 - 17 .
a. kreczmar . on memory requirements of strassen s algorithms . in a.
mazurkiewicz , editor , _ proceedings of the 5th symposium on mathematical foundations of computer science _ , volume 45 of _ lncs _ , pages 404407 , gdask , poland , sept . | we propose several new schedules for strassen - winograd s matrix multiplication algorithm , they reduce the extra memory allocation requirements by three different means : by introducing a few pre - additions , by overwriting the input matrices , or by using a first recursive level of classical multiplication . in particular , we show two fully in - place schedules : one having the same number of operations , if the input matrices can be overwritten ; the other one , slightly increasing the constant of the leading term of the complexity , if the input matrices are read - only . many of these schedules have been found by an implementation of an exhaustive search algorithm based on a pebble game .
* keywords : * matrix multiplication , strassen - winograd s algorithm , memory placement . |
cosmic history from the decoupling ( @xmath4 ) to the epoch of lighting of the first luminous sources ( at redshifts @xmath5 to 5 ) is one of the biggest unknowns of present day observational cosmology .
high redshifts and dust extinction during early active phases both degrade the energetic optical uv photons emitted by massive stars , decaying particles , or more exotic energy sources , to the infrared wavelengths . a fundamental information on the total energy budget associated with astrophysical processes occurring at high redshifts is then provided by observations of the cosmic background at infrared wavelengths ( cirb ) .
unfortunately , the infrared domain presents various levels of difficulty to the observational astronomer , because of the huge backgrounds from the earth s atmosphere , the interplanetary dust ( ipd ) , and diffuse dust in the milky way , in addition to the background produced by the telescope itself . also the sensitivity and stability of infrared detectors are far poorer than those used in the optical .
because of all this , the detection and characterization of the diffuse background flux of low - energy photons coming from primeval structures has been exceedingly difficult so far .
even dedicated experiments exploiting cooled platforms outside the atmosphere , among which the most important is dirbe on cobe ( @xcite ) , have failed so far to detect significant signals from the cirb above the intense foregrounds .
the upper limits allowed by the foreground emission deconvolution of dirbe maps are significantly higher than expectations over a substantial and crucial @xmath6range from a few to @xmath7 @xmath8
the situation , at these wavelengths in particular , is not likely to improve in the future , until a mission flying to the outer solar system will get rid of the fundamental limitation set by the ipd ( both the scattered light and dust re radiation ) . under these circumstances ,
a very interesting alternative to the direct detection of cirb photons has been suggested by stecker and de jager ( 1993 ) soon after the discovery of high - energy photon fluxes coming from distant blazars ( with gro , @xcite and with the whipple observatory , @xcite ) .
the idea is to infer the cirb spectral intensity from combined gev and tev observations of a set of active galactic nuclei ( agn ) , by exploiting the @xmath9 interactions and pair production between the agn high energy photons and low
energy background photons in the line of sight to the source .
the interaction is expected to produce an absorption feature , testable in principle , in the source tev spectrum .
interesting limits have been discussed by stecker and de jager ( 1993 ) , de jager , stecker and salamon ( 1994 ) , and dwek and slavin ( 1994 ) , _ all based on tev observations of the blazar mkn 421_. to summarize , the best current upper limits on the cirb in the spectral range from 10 to 40 micron are those reported by de jager , stecker & salamon ( 1994 ) , with a @xmath10 upper limit of @xmath11 . in the same large waveband interval ,
marginal detections , at levels of @xmath12 were reported by de jager , stecker & salamon ( 1994 ) and dwek & slavin ( 1994 ) . at shorter wavelengths ,
1 to 10 @xmath8 m , upper limits have been obtained by stecker & de jager ( 1993 ) , stecker ( 1996 ) , dwek & slavin ( 1994 ) and biller et al .
the most conservative bounds , accounting for the precise spectral shape of the cirb , given by dwek & slavin ( 1994 , @xmath13 ) , still keep a substantial factor ( @xmath1410 ) higher than the expected contribution of known sources . while all previous analyses relied uniquely on tev observations of the blazar mkn 421 ,
we exploit here a new high quality dataset of tev gamma
ray observations by hegra of mkn 501 during a state of high activity ( aharonian et al 1997a ) to further constrain the intensity of the diffuse ir background .
two sets of constraints on the cirb spectral intensity are derived in section 2 .
one is based on the assumption that the background spectrum is dominated by the contribution of distant galaxies , and thus reflects the average galactic ir spectrum .
the other constraint is free of model
dependent assumptions and treats the cirb as a combination of twelve bins wherein the background spectrum is flat ( in @xmath15 ) with independent arbitrary normalizations .
extremely tight constraints on the cirb ensue from this analysis , reflecting the missing evidence of any absorption in the tev spectrum of mkn 501 .
a discussion is given in section 3 in terms of a low average past emissivity both of galaxies and of primeval energy sources .
indeed the limits are so severe that they begin to conflict with the integrated ir flux of distant galaxies , recently detected in large numbers by ground - based and space observatories .
we finally emphasize the alternative possibility that the interaction of high - energy gamma - rays with low - energy photons is more complex than previously supposed .
@xmath16 is assumed throughout the paper .
we use the data set taken in march and april 1997 with the hegra stereoscopic system of four imaging cherenkov telescopes ( @xcite ) . at that time mkn 501 was in an extremely high state , the far brightest gamma - ray source in the sky .
the flare was observed by all tev @xmath0ray observatories in the northern hemisphere ( @xcite ) .
the high gamma ray flux allowed the hegra observatory to measure the gamma
ray energy spectrum in short time intervals .
the spectrum measured between march 15 to 20 covers 1.2 orders of magnitude in photon energy , from 0.8 to 12.6 tev , divided into twelve logarithmically spaced energy bins .
the photon number spectrum at tev @xmath0ray energies is consistent with a power law of differential spectral index @xmath17 = [email protected] .
this spectrum is shown in figure 1 ( data points ) together with the allowed range of best fitting power
laws derived by the observational team ( shaded region ) .
similar power - law spectra for mkn 501 have been observed by various other groups ( see e.g. protheroe et al .
1997 ) . the main difficulty with the derivation of the cirb spectral intensity from tev @xmath0ray absorption is the lack of knowledge of the production spectrum , which may deviate from a pure power - law and/or show absorption at the source .
spectra measured within short time intervals , during which the source s activity state is not likely to vary , are very valuable in this respect , although there is not an evidence for strong spectral variability of the mkn501 emission(@xcite ) . for this reason we confined our analysis to a spectrum based on data collected between march 15 to 20 , rather than the one using the whole march - april database in the updated version of aharonian et al ( 1997a ) . as a first step in understanding the observed spectrum
, we attempted to derive the allowed range of differential spectral indices @xmath17 at the source and the cirb absorption by fitting the observed @xmath0ray spectrum with a variety of assumptions for @xmath17 and for the intensity @xmath19 of the extragalactic ir background .
we first assumed that the source spectrum in the observed range is a pure power - law and that the low - energy background is dominated by the integrated contribution of distant galaxies .
in such a case , the shape of the cirb is well constrained , and should reflect the average galactic spectrum , which has a minimum at @xmath20 corresponding to the intersection of the stellar with the dust emission component .
we refer here to the detailed spectral shape estimated by franceschini et al ( 1991 ) .
then the fit was performed in a two parameter space consisting of the source gamma - ray spectral index @xmath17 and the cirb intensity @xmath19 normalized to the model spectrum @xmath21 ) by franceschini et al ( 1991 ) . for a given spectral index @xmath17 the absorption due to the cirb
was first calculated and the resulting spectrum was then normalized to the detected @xmath0ray flux above 1 tev .
this procedure enabled us to avoid introducing a third parameter , the source spectrum normalization .
table 1 shows the optical depth assuming @xmath22 = 1 , for all twelve experimental energy bins .
figure 2 summarizes the results of fitting the observed mkn 501 spectrum .
it plots contours of the @xmath23 2d distribution in the parameter plane , including the unphysical region corresponding to negative @xmath22 values , where gamma
rays are ` created ' in collisions with cirb photons .
the darker an area is the better the fit is ( except for fits with @xmath24 1.45 , corresponding to the blank inner region ) .
the formal best fit ( @xmath25 = 1.41 p.d.f . for 10 degrees of freedom , confidence level 0.16 ) is in the `` unphysical '' region : @xmath17 = 2.52 , @xmath22 = -0.3 .
the best fit in the `` physical '' region occurs at @xmath26 = 0 ( no - absorption , @xmath25 = 1.42 p.d.f . ) and yields @xmath17 = 2.48 , in agreement with the result of the experimental group .
flatter tev @xmath0ray source spectra allow for higher intergalactic absorption : a value for the spectral index of @xmath27 , corresponding to @xmath22 = 1.70 , is the hardest spectrum with @xmath28 1.80 ( corresponding to the 95% confidence limit ) .
the best - fit spectra corresponding to @xmath22=0 and the flattest one assuming @xmath22 = 1.7 are also shown in fig . 1 .
it is important to understand that the fit quality , and correspondingly the limits on the cirb , depend crucially on the manner in which the ` theoretical ' power - law fluxes are normalized in the fitting procedure .
the normalization to the observed @xmath0ray flux above 1 tev causes very different spectral indices to fit the observations equally well and generates a broad @xmath23 valley in fig.2 .
this would change significantly if the observed @xmath0ray energy range were extended on either side .
an observation at @xmath29 300 gev ( the threshold energy of the whipple telescope ) would improve the overall normalization and allow us to distinguish better between different spectral indices ` at production ' .
an extension to higher @xmath0ray energy , hence larger optical depths ( see table 1 ) would make easier the detection of any absorption by the cirb .
we finally attempted to obtain model
independent limits on the cirb , with no a - priori guess at the background spectrum . to compute them we conservatively assumed that the high - energy source spectrum is the flattest allowed by the fits of fig . 1 , i.e. @xmath17 = 2.16 .
we normalize the spectrum at the source to be @xmath30 in the first experimental bin .
this normalization factor is a source of uncertainty .
it is partially justified by the fact that mkn 501 is close enough ( 136 mpc ) compared to the optical depth ( 600 mpc for @xmath31 ) that the content of this bin would be absorbed only by 20% .
the absorption of @xmath0rays of energy @xmath32 tev is due to ir photons within a certain wavelength interval around the maximum absorption at @xmath33 = @xmath34 ( @xmath35 m ) .
we have then assumed that the absorption of any given @xmath0ray energy bin is caused by cirb with flat @xmath36 spectrum .
the envelope of all upper limits to the cirb intensity obtained in this way is shown as a histogram in fig .
the limits are assigned within wavelength intervals that contribute 90% of the optical depth for that bin .
as we see , the limits become less stringent at longer wavelengths .
the limit in the 3.4 @xmath8 m to 24 @xmath8 m range comes from one single energy bin ( 5.0 to 6.3 tev ) where a particularly large @xmath0ray flux was measured .
_ note that the obtained upper limits fall already very close to recent direct evaluations of the ir background based on deep ir surveys ( @xcite)_. as shown in fig .
3 , the upper limits heavily rely on the error in the tev flux measurement and are less stringent for @xmath37 10 @xmath8 m , where the tev @xmath0ray statistics is not as good .
in either case , both assuming the cirb shape and relaxing it , the constraints on the cirb intensity appear dramatic . essentially the mkn 501 gamma - ray spectrum does not display the expected effect of absorption , it rather shows a simple @xmath38 power - law spectrum .
how reliable are these limits in view of the possible systematic errors of @xmath3925% in the energy estimates of the hegra telescopes ( @xcite ) ?
this is very easy to estimate in the case of a flat binned @xmath15 cirb spectrum .
the limits in figure 3 would move upward and towards shorter wavelengths with the fractional amount of energy overestimate .
similar relaxation would occur also in the case of a more specialized model cirb spectrum .
one could use the optical depths from table 1 to estimate the amount of relaxation .
similarly , a higher normalization of the source spectrum would relax the model
independent limits by the ratio of the two normalizations .
is this featureless spectrum of mkn 501 inconsistent with that observed for mkn 421 , which is almost at the same distance ?
the latter has been interpreted by some authors ( e.g. @xcite ) as showing a turnover at @xmath40 , which was attributed to @xmath1 absorption with the cirb .
in fact , new observations of this source during a high activity state do not appear to confirm the presence of absorption ( krennrich et al .
1997 ) , and show significant counting rate above 5 tev .
so , at least during this high state , mkn 421 seems to show a power law spectrum similar to the spectrum discussed here for mkn 501 .
the constraints on the cirb intensity from tev observations of mkn 501 start to approach the `` measured '' lower limits at 2.2 , 6.7 and 15 @xmath8 m given by the integrated emission of galaxies already resolved in deep integrations at those wavelengths .
deep surveys have been performed from ground in the k - band and from space by the mid ir camera ( isocam , see @xcite ) on the iso satellite in the two latter bands .
a summary of these `` direct '' determinations of the galaxy contribution to the cirb , and a discussion of the related uncertainties , are given by franceschini et al ( 1997 ) and oliver et al .
( 1997 ) . in any case
, the cirb can not be lower than reported at these three wavelengths .
few possibilities are left .
the first one is that the cirb is very close to the limits allowed by the gamma ray spectrum of mkn 501 observed by aharonian et al .
( 1997a ) .
this would imply a very strong constraint on any signals unrelated to the emission of distant galaxies ( see e.g. @xcite , for a review ) .
but , in view of the fact that the same power - law spectral shape as show in fig . 1 for mkn 501 has been confirmed by later integrations on this source ( aharonian et al .
1997a ; protheroe et al .
1997 ) , that apparently a similar shape is also suggested for mkn 421 , and because an appreciable cirb flux has already been detected and resolved into discrete sources , we find more likely that _ we have to revise our concepts about the propagation of tev gamma - rays into the intergalactic space , and that something complicates the process_. the question is : why the propagation of tev gamma - rays in intergalactic space should not produce the expected absorption in high energy spectra of distant sources ?
a possible solution could be that part of the source spectrum is regenerated in @xmath0ray cascading ( pair production + inverse compton ) . in such a cascading process , the @xmath0ray spectrum of the source is depleted around the region of maximum absorption .
if the @xmath0ray emission of mkn 501 at the source extends above @xmath41 ev , the spectrum would be depleted in collisions with microwave background photons .
the @xmath42 pairs generated on the microwave background would inverse compton scatter on the microwave background to regenerate photons of lower ( tev ) energy , thus generating ` bumps ' on a power - law production spectrum ( @xcite ) .
the resulting @xmath0ray spectrum may then appear unattenuated at observation .
this would however require not only a @xmath0ray spectrum extending to very high energy , but also a very low ( @xmath43 gauss ) value for the extragalactic magnetic field in the direction of mkn501 .
otherwise the @xmath42 pairs would deflect in the magnetic field and form a halo around the source , well outside of the angular resolution of the hegra detector . a good deal of constraints useful to disentangle between these two possibilities are soon expected by improved observations of the mkn501 outburst ( which has been observed by the whipple and cat cherenkov telescopes , @xcite ) with different energy thresholds and wavelength bands and by refined forthcoming data on mkn 421 . +
* acknowledgements . *
the authors appreciate the contribution of an anonymous referee to the improvement of the paper .
ts thanks j. buckley and t.k .
gaisser for the careful reading of the manuscript and e. dwek for comments .
the research of ts is funded in part by nasa grant nag55106 .
franceschini , a. , aussel , h. , bressan , a. , cesarsky , c.j . ,
danese , l. , de zotti , g. , elbaz , d. , granato , g.l . ,
mazzei , p. , silva , l. : 1997 , _ source counts and background radiation _
, in proc .
esa symposium `` the far infrared and submillimeter universe '' , esa sp-401 , in press : astro - ph/9707080 .
crrcrrc @xmath44 & & + ( tev ) & @xmath45 & @xmath46 & @xmath47 & @xmath45 & @xmath46 & @xmath47 + & & ( @xmath8 m ) & ( @xmath8 m ) & & ( @xmath8 m ) & ( @xmath8 m ) + 0.79 1.00 & 0.25 & 1.7 & 0.53 3.8 & 0.23 & 2.1 & 0.53 3.8 + 1.00 1.26 & 0.29 & 2.0 & 0.60 4.8 & 0.28 & 2.7 & 0.67 4.8 + 1.26 1.59 & 0.32 & 2.1 & 0.67 6.0 & 0.36 & 3.4 & 0.84 6.0 + 1.58 1.99 & 0.33 & 2.2 & 0.75 7.5 & 0.45 & 4.2 & 1.1 7.5 + 2.00 2.51 & 0.34 & 2.4 & 0.84 9.5 & 0.57 & 5.3 & 1.3 9.5 + 2.51 3.16 & 0.35 & 2.5 & 0.95 12 . &
0.71 & 6.7 & 1.7 12 .
+ 3.16 3.98 & 0.37 & 3.0 & 1.1 15 . & 0.86 & 8.4 & 2.1 15 .
+ 3.98 5.01 & 0.41 & 13.4 & 1.2 19 . & 1.13 & 10.9 & 2.7 19 .
+ 5.01 6.31 & 0.48 & 15.0 & 1.5 24 . & 1.43 & 13.4 & 3.4 24 .
+ 6.31 7.94 & 0.59 & 16.8 & 1.9 30 . & 1.80 & 16.8 & 4.2 30 .
+ 7.94 10.0 & 0.76 & 21.2 & 2.4 38 . & 2.27 & 21.2 & 5.3 38 .
+ 10.00 12.6 & 1.00 & 30.0 & 3.4 48 . & 2.85 & 26.7 & 6.7 48 . | we use the new results of the hegra detector on the tev @xmath0ray emission from mkn 501 to set upper limits on the energy density of the cosmic infrared background ( cirb ) .
contrary to previous interpretations of the @xmath0ray spectrum of mkn 421 as showing an intergalactic absorption cutoff at 5 tev , the observed spectrum of mkn 501 extends beyond 10 tev and appears to be unattenuated by @xmath1 collisions with the low - energy cirb photons .
the upper limits on the cirb intensity derived both assuming an _ a priori _ shape for the cirb spectrum and without model - dependent assumptions
are thus quite strong and come almost in conflict with the observational evaluations based on deep surveys of extragalactic sources in the near- and mid - ir .
if spectra at tev energies for extragalactic @xmath2-ray sources like this for mkn 501 will be confirmed with improved statistics , we may be forced to conclude that the process of @xmath1 interaction in the intergalactic space is more complex than expected and the average intergalactic magnetic field extremely weak ( @xmath3 ) . |
electromagnetic fields are the main tool in the control and manipulation of the motion of polarizable dipoles , such as atoms and molecules , which are key objects in physics and chemistry .
optical traps @xcite and laser cooling @xcite are widely used to induce external mechanical forces on individual dipoles , whereas interactions between atoms are often tuned with the help of external magnetic or electric fields @xcite . in this work ,
we address laser - induced dipole - dipole interactions ( liddi ) by establishing their relation to resonant inter - dipolar interaction ( excitation exchange ) in any confined geometry .
polarizable dipoles subject to static electric fields interact via the electrostatic dipole - dipole interaction that scales in free space as @xmath0 , @xmath1 being the inter - dipolar distance .
a dynamic analog is obtained when a laser field , far - detuned from the dipolar resonant frequency , illuminates the ( induced ) dipoles . in the retarded regime ,
where @xmath2 with @xmath3 the laser wavenumber , the resulting liddi scales like @xmath4 , whereas in the non - retarded quasistatic case , where @xmath5 , the electrostatic scaling @xmath0 is restored @xcite .
the retarded long - range liddi was shown to be associated with peculiar many - body effects , such as self trapping @xcite , density modulations @xcite and the existence of a roton - like collective excitations @xcite , in an atomic bose - einstein condensate .
recently , the possibility to control and shape the space - dependence of the inter - dipolar liddi potential was considered .
one option is to tune the laser parameters @xcite or the radiation spectrum @xcite .
another , more recent approach , is to consider dipoles coupled to structures that support confined photon modes @xcite . once illuminated by off - resonant light and virtually excited , these dipoles interact via confined virtual photons whose spatial - mode structure determines the resulting liddi space - dependence .
for example , in the case of many atoms that are trapped in the vicinity of an optical fiber @xcite and free to move along its axis as in @xcite , the fiber - mediated liddi can effectively become one - dimensional ( 1d ) , such that it extends to any range and the atoms may self - organize @xcite .
when the fiber incorporates a bragg grating ( 1d photonic crystal ) , the relaxation dynamics of the resulting many - atom system was shown to reveal the inequivalence of statistical ensembles typical of non - additive systems @xcite .
self organization and dynamics of laser - illuminated atoms inside a cavity were also studied theoretically and experimentally @xcite .
a similar situation arises when considering the spatial dependence of the dispersive interaction between the internal degrees of freedom of the dipoles , the so - called resonant dipole - dipole interaction ( rddi ) , where a dipole , e.g. an atom initially in the excited state , periodically exchanges its excitation with another dipole , e.g. an atom initially in the ground state @xcite .
rddi is mediated by the exchange of virtual photons between the atoms , hence its space - dependence is determined by the spatial structure / propagation of the mediating photon modes . in free space and for @xmath6 , @xmath7 being the typical dipolar transition wavelength , the rddi exchange frequency scales as @xmath0 whereas in the retarded regime @xmath8 it takes the form @xmath9 .
such rddi retardation effects are of interest for quantum information processing @xcite
. studies of geometries where rddi is mediated by confined photon modes include , for example , rddi via surface - plasmon - polariton and coplanar waveguide modes in one dimension @xcite or in geometries that create cutoffs or bandgaps in the photonic mode spectrum @xcite , where spontaneous emission is absent and enhanced long - range rddi may become dominant @xcite . the extensive experimental progress in the ability to couple atoms / dipoles to various confined photonic structures @xcite , makes the discussion of liddi in such structures ever more relevant . in this study
we derive a general formalism for liddi in _ any geometry _ by relating it to the underlying rddi process , thus providing a simple and intuitive recipe for calculating liddi through the less complex rddi .
we also show that the same holds for the relation between scattering of laser photons and spontaneous emission , respectively .
our formalism has several appealing features : ( 1 ) it is fully quantum mechanical , with the laser radiation taken to be in a coherent state ( in contrast to the less realistic number state @xcite ) .
( 2 ) it interprets _ liddi as rddi between dressed dipoles _
, thus allowing to account for liddi in strong radiation fields .
this considerably simplifies the calculation of liddi in a general geometry to merely calculating the rddi in that geometry , thereby avoiding the need for the rather cumbersome fourth - order perturbative qed calculation which contains 24 diagrams @xcite .
( 3 ) it enables the analysis of _ nonlinear effects _ in liddi , resulting from the nonlinear response of the dipoles to the laser , unaccounted for previously .
( 4 ) for the case of far - detuned laser field , it reveals a simple relation between the pairwise rddi energy @xmath10 and the liddi potential @xmath11 , where @xmath12 is the laser rabi frequency and @xmath13 the dipole - laser detuning .
this provides an intuitive picture for liddi : the dipole / atom is virtually excited by the off - resonant laser with probability @xmath14 and , once excited , interacts with the other dipole / atom via rddi .
the paper is organized as the following .
section ii is dedicated to the derivation of the general formalism that relates liddi and scattering to their underlying rddi and spontaneous emission , respectively , arriving at our central , general result for the liddi potential , eq .
( [ liddi1 ] ) .
the large - detuning limit is discussed in section iii , whereas the application of our formalism to the analysis of liddi in a cavity is illustrated in section iv . in section
v we put forward intuitive and simple arguments that explain the origin of liddi effects caused by nonlinear dipole response to the laser .
our conclusions are presented in section vi .
we adopt a two - level atom model for the polarizable dipole .
thus , while still capturing the essence of nonlinear response of the dipoles , the discussion is kept simple and intuitive .
the two - level atom approximation is justified for the typical situation where a laser frequency is close to a specific dipole - allowed transition of a system with discrete energy levels , such as atoms , molecules , quantum dots or even superconducting qubits .
the atoms , with excited state @xmath15 and ground state @xmath16 interact with a laser ( @xmath17 below ) . considering a laser mode in a coherent state , an _ exact _
transformation due to mollow @xcite allows to treat the interacting system of atoms + single field mode in a coherent state ( laser ) + vacuum in the rest of the modes , as a system of atoms + single - mode external field ( laser ) + vacuum in all field modes @xcite .
the atoms are coupled to the vacuum modes @xmath18 in the considered geometry ( @xmath19 below ) via dipole couplings ( @xmath20 below ) which ultimately lead to the inter - atomic interaction .
the hamiltonian @xmath21 reads @xmath22 , \nonumber\\ h_v&=&\sum_k\hbar\omega_k\hat{a}^{\dag}_k\hat{a}_k , \nonumber \\
h_{av}&=&\sum_{\nu=1}^2\sum_k\hbar\left[i g_{k\nu } \hat{a}_k - i g^{\ast}_{k\nu } \hat{a}^{\dag}_k \right]\left[\hat{\sigma}_{\nu}^{-}+\hat{\sigma}_{\nu}^{+}\right ] .
\nonumber \\ \label{h}\end{aligned}\ ] ] here @xmath23 is the electromagnetic vacuum mode @xmath24 lowering operator , whereas @xmath25 and @xmath26 are the laser frequency and wavevector with rabi frequency @xmath27 , @xmath28 and @xmath29 being its electric field amplitude and polarization vector , respectively , and @xmath30 denoting the atomic dipole matrix element . in order to find the pairwise - liddi potential it is enough to focus on a single pair of atoms with indices @xmath31 and corresponding atomic operators @xmath32 , @xmath33 .
the atom - vacuum dipolar couplings are @xmath34 , @xmath35 being the location of atom @xmath36 , and @xmath37 the spatial function of mode @xmath24 . moving to the interaction picture with respect to ( w.r.t ) @xmath38 the hamiltonian becomes @xmath39 .
\label{hi1}\ ] ] here the rotating - wave approximation @xmath40 is taken , and the interaction - picture atomic operators are @xmath41 with @xmath42 , where @xmath43 denotes time ordering .
we are now interested in finding the interaction - picture operator @xmath44 of the coupled atom - laser system .
this is done for each atom @xmath31 separately , hence the index @xmath36 is suppressed here . the operator @xmath44 can be viewed as a heisenberg - picture operator w.r.t the hamiltonian @xmath17 from eq .
( [ h ] ) .
then , dividing @xmath17 into @xmath45 and @xmath46 , we move to the so - called laser - rotated frame , which is an interaction picture w.r.t @xmath47 , where the hamiltonian @xmath17 becomes @xmath48 with @xmath49 and where @xmath50 is the atom - laser detuning .
the eigenstates @xmath51 and eigenvalues @xmath52 of @xmath53 are found to be ( see also @xcite ) @xmath54 where @xmath55 is the effective rabi frequency .
considering the relation between operators in the heisenberg and interaction pictures we find @xmath56 where @xmath57 and @xmath58 , such that @xmath59 .
then , recalling @xmath60 , @xmath61 , we can use the transformation ( [ tran ] ) to express the operators @xmath62 in terms of the @xmath51-dressed - basis projection operators and easily apply @xmath63 on these operators , obtaining @xmath64 where @xmath65 and @xmath66 are the spin operators in the dressed - basis .
this result can be written in a more compact form by absorbing the time - independent coefficients into the operators , as @xmath67 eqs .
( [ sig1 ] ) and ( [ sig2 ] ) describe the dynamics of the atom operator @xmath68 , dressed by the laser , as composed of a linear combination of the three spin operators in the dressed - basis @xmath69 , each oscillating in time with a distinct frequency @xmath70 , hence allowing all possible transitions in this basis . inserting @xmath71 from eq .
( [ sig2 ] ) into eq .
( [ hi1 ] ) , we obtain the interaction picture hamiltonian in the form @xmath72 . \label{hi2}\ ] ] this hamiltonian describes the interaction of atoms , dressed by the laser ( operators @xmath73 ) , with the electromagnetic vacuum ( operators @xmath23 ) .
we are interested to use lowest order perturbation theory in order to find the effective interaction between the dressed atoms , mediated by the vacuum , similar to the case of rddi @xcite .
equivalently , this can be performed by the derivation of the master equation for the density operator of the two - atom system @xmath74 when it interacts with the vacuum reservoir in a stationary vacuum state @xmath75 .
to lowest order ( born approximation ) the master equation is given by @xcite @xmath76\right]\right\}. \label{me1}\ ] ] inserting the hamiltonian from eq .
( [ hi2 ] ) into eq .
( [ me1 ] ) and using standard methods ( see appendix a ) , we obtain the markovian master equation @xmath77+\sum_{\nu,\nu'=1}^2\sum_{i=\pm , z}\left[\gamma_{\nu \nu'}^i\tilde{s}_{i\nu}^{\dag}\rho\tilde{s}_{i\nu ' } \right .
\nonumber \\ & & \left .
-\frac{1}{2}\gamma_{\nu \nu'}^i\left(\tilde{s}_{i\nu}\tilde{s}_{i\nu'}^{\dag}\rho+\rho\tilde{s}_{i\nu}\tilde{s}_{i\nu'}^{\dag}\right)\right ] , \label{me2}\end{aligned}\ ] ] with the effective dipole - dipole hamiltonian @xmath78 the dispersive , virtual - photon - mediated , rddi energy @xmath79 and the corresponding cooperative emission rate @xmath80 , are related by kramers - kronig relation and given by @xmath81 where @xmath82 is the vacuum - reservoir two - point ( autocorrelation ) spectrum defined in the continuum limit @xmath83 .
the geometry dependence of the liddi and related effects discussed below is thus encoded in the function @xmath82 , since this function depends on the dipole couplings @xmath84 , which in turn depend on the spatial photon modes of the confined geometry @xmath85 .
two important assumptions regarding the temporal resolution of interest ( coarse - graining time ) @xmath86 were made in the derivation of eqs .
( [ me2 ] ) , ( [ hdd ] ) and ( [ d ] ) ( appendix a ) : ( 1 ) @xmath87 , i.e. @xmath88 .
this allows to take terms oscillating as @xmath89 to be kronecker deltas @xmath90 .
( 2 ) @xmath91 , where @xmath92 is the correlation time of the reservoir , namely @xmath93 is the width of the functions @xmath94 around @xmath70 .
this allows to take the markov approximation and obtain a master equation which is local in time ( see appendix a and refs .
@xcite ) . in the absence of the external laser light ,
i.e. taking the limit @xmath95 in eqs .
( [ sig2 ] ) and ( [ tran ] ) , we obtain @xmath96 , @xmath97 , @xmath98 and @xmath99 , such that the effective dipole - dipole hamiltonian , eq .
( [ hdd ] ) , reduces to that of rddi , @xmath100 then , once one of the atoms happens to be excited , it coherently and reversibly exchanges its excitation with the other , initially unexcited , atom . the complementary incoherent and radiative cooperative process , namely that of irreversible excitation exchange between the atoms via the emission of a real photon at the atomic frequency @xmath101 @xcite , is described in the non - hamiltonian part of the master equation ( [ me2 ] ) by terms proportional to @xmath102 , e.g. @xmath103 .
these bare - atom processes are depicted by the diagram in fig .
1(a ) , where the atoms exchange their excitation via a photon at frequency @xmath104 . in the radiative process , the photon is real and hence has the frequency of the atomic transition @xmath105 , whereas the dispersive effect involves the summation over all diagrams with virtual photons @xmath104 around @xmath101 , as can be seen by the sampling of the functions @xmath106 and @xmath107 in eq .
( [ d ] ) , at @xmath99 for @xmath108 and @xmath109 , respectively .
both rddi and its irreversible counterpart rely on the fact that one of the atoms is excited , and are absent when both atoms are in their ground states . as we shall see below
, the essence of liddi revealed by our formalism is the possibility to excite the atoms by the laser , thus allowing them to interact via rddi . in the more general case , where the laser is turned on ,
we write the effective dipole - dipole hamiltonian ( [ hdd ] ) explicitly using eq .
( [ sig2 ] ) , as @xmath110 , \nonumber \\ & -&\hbar\delta^-_{12}\frac{(\delta+\bar{\omega})^2}{4\bar{\omega}^2}\left[e^{-i\mathbf{k}_l\cdot\mathbf{r}_{12}}\hat{s}^-_1\hat{s}^+_2+e^{i\mathbf{k}_l\cdot\mathbf{r}_{12}}\hat{s}^+_1\hat{s}^-_2\right ] , \nonumber \\ \label{hdd_d}\end{aligned}\ ] ] where @xmath111 .
we also note that here the single - atom @xmath112 ( @xmath113 ) terms are not considered , since they are not important for the interaction between the dipoles , and that we have assumed the typical situation where @xmath114 is real .
the complementary dissipative effects include terms of similar form , e.g. with @xmath115 and @xmath116 replacing the corresponding @xmath117 terms in ( [ hdd_d ] ) . in analogy to the bare - atom case , the above hamiltonian and
its dissipative counterparts describe the rddi and cooperative emission , respectively , between two dressed atoms , whose level scheme , consisting of two levels separated by an energy @xmath118 [ eq .
( [ tran ] ) ] , is presented in fig .
however , as opposed to the bare - atom case , where only one exchange process exists [ see eq .
( [ rddi ] ) and fig .
1(a ) ] , here several processes can occur : + ( i ) the first line in eq .
( [ hdd_d ] ) describes a process where the internal states of the dressed atoms do not change , giving rise to a cooperative energy shift to the dressed states .
+ ( ii ) the process described in the second line , shown in fig .
1(c ) , is analogous to the bare - atom rddi in eq .
( [ rddi ] ) and fig .
1(a ) , where the dressed states @xmath119 and @xmath120 take the role of the excited and ground states @xmath15 and @xmath16 , respectively .
the relevant atomic transition is then @xmath121 , which in fig .
1(b ) is shown to be resonant with @xmath118 . however , since the dressed - states ( [ tran ] ) are written in a frame rotating with frequency @xmath25 , the actual transition frequency is @xmath122 .
then , a real - photon , dissipative effect involves a photon exchange with frequency @xmath123 , whereas the dispersive rddi results from the equivalent virtual - photon effect , just as seen by the sampling of @xmath124 and @xmath125 at @xmath123 in eq .
( [ d ] ) for @xmath126 and @xmath127 , respectively .
+ ( iii ) the last process , described by the third line of ( [ hdd_d ] ) is that of the atomic transition @xmath128 , shown in fig .
namely , atom 1 , initially in state @xmath120 , transfers a photon to atom 2 which is initially in state @xmath119 , and as a result both of them flip their states .
the emission of a photon from the state @xmath120 and its absorption by the state @xmath119 underlying this process , is enabled by the fact that the dressed state @xmath120 contains a component of the excited state and likewise the dressed state @xmath119 contains a ground - state component .
the transition frequency in the dressed - state basis is then @xmath129 , so that the actual frequency is @xmath130 .
the three processes surveyed above , describe the interaction between the atoms mediated by photon modes at frequencies @xmath131 and @xmath132 , respectively , in analogy with the three spectral components of the well - known mollow - triplet of resonance fluorescence @xcite .
the distinction between these three processes and their physical meaning are further discussed in sec .
v c below . we now present a central insight of this paper , namely , that liddi ( and scattering ) originate from rddi ( and emission ) of dressed atoms .
the average interaction energy between the dressed atoms , interpreted as the liddi potential , is obtained by averaging quantum mechanically over the hamiltonian @xmath133 , eq .
( [ hdd_d ] ) , @xmath134 , \label{liddi}\ ] ] where @xmath135 is the density matrix of the two - atom state .
we note that the hamiltonian @xmath133 is assumed to be time - independent in this formalism .
in fact , @xmath133 depends on the external degrees of freedom of the atoms , namely their positions @xmath136 , and hence becomes time - dependent due to their dynamics which is driven by the liddi potential @xmath137 .
however , assuming this dynamics is slow enough , namely , @xmath138 is much smaller than any relevant energy scale such as @xmath118 , a description using the potential @xmath137 from eq .
( [ liddi ] ) can be viewed as a valid adiabatic approximation .
( [ hdd_d ] ) and ( [ liddi ] ) allow us to write the liddi potential for a general two - atom density operator @xmath135 .
using the two - atom basis , @xmath139 and denoting the corresponding density matrix elements @xmath140 , we find @xmath141=1 - 2(\rho_{22}+\rho_{33})$ ] and @xmath142=\rho_{23}$ ] .
the liddi potential between a pair of atoms is then , @xmath143 , \nonumber \\
u^{\pm}(\mathbf{r}_{12})&=&-\hbar\delta^{\pm}_{12}\frac{(\delta\mp\bar{\omega})^2}{4\bar{\omega}^2}\left[e^{\pm i\mathbf{k}_l\cdot\mathbf{r}_{12}}\rho_{23}+e^{\mp i\mathbf{k}_l\cdot\mathbf{r}_{12}}\rho^{\ast}_{23}\right ] .
\nonumber \\ \label{liddi1}\end{aligned}\ ] ] here we have separated the potential @xmath144 into its components @xmath145 according to its contributions from processes mediated by different photon mode frequencies @xmath70 ( @xmath69 ) .
for example , if @xmath135 does not allow for a superposition of @xmath51 states in each atom , then @xmath146 vanishes and the only contribution to the liddi originates from the @xmath132 mediated process , later interpreted here as the linear component of liddi .
we recall that @xmath12 and @xmath147 contain the atomic - dipole matrix element @xmath30 which , in spherically - symmetric cases , is determined by the laser polarization @xmath29 , the only preferred direction ( see appendix b ) . the two - atom density matrix @xmath135 , to be used in eq . ( [ liddi1 ] ) for the liddi potential ,
is determined by the solution of the master equation ( [ me2 ] ) . in general , this yields a time - dependent potential @xmath137 obtained by averaging in eq .
( [ liddi ] ) with the time - dependent solution @xmath74 .
if however the time - resolution of interest is longer than the typical decay times @xmath148^{-1}$ ] and @xmath149^{-1}$ ] , then the steady - state solution of the master equation , found by setting @xmath150 in ( [ me2 ] ) , can be used to obtain the stationary liddi potential . on the contrary , if the dynamics of interest occurs at times much shorter than the decay times @xmath151 ( @xmath69 ) , then @xmath74 and hence @xmath137 , may be obtained by simply solving eq .
( [ me2 ] ) without the dissipative terms .
these two limiting cases are illustrated in sec .
iv below for atoms in a cavity .
turning to the dissipative counterpart of the liddi potential , the dressed atoms can scatter laser photons via spontaneous emission , resulting in a random diffusive motion for the atoms , on top of that driven by the liddi potential . upon ignoring ( for the time being ) the cooperative scattering terms , the scattering rate from a single atom ( e.g. atom 1 without loss of generality )
can be obtained by averaging the dissipative , imaginary - hamiltonian - like , terms of the master equation for @xmath152 . in analogy to eq .
( [ liddi1 ] ) , we obtain the single atom scattering rate @xmath153 as @xmath154 where @xmath155 $ ] is the density operator of atom 1 , obtained by tracing over atom 2 , and @xmath51 are the states of atom 1 . here
we used @xmath156=1 $ ] and @xmath157=\langle \pm|\rho^{(1 ) } |\pm\rangle$ ] .
the scattering rate given above is composed of scattering rates of photons at three distinct frequencies : @xmath132 , the frequency of the incoming laser light , and two sidebands at frequencies @xmath158 , which reproduce the mollow - triplet fluorescence spectrum .
the amplitudes of the sidebands are determined by their corresponding scattering processes : an atom that scatters a @xmath159 photon is initially in state @xmath51 and finally in state @xmath160 .
hence , the amplitude is related to the probability of the atom to be in the initial state , i.e. @xmath161 that appears in the expressions for @xmath162 .
likewise , scattering at @xmath25 does not change the internal state , which explains its independence of the atomic state @xmath135 .
our liddi result of eq .
( [ liddi1 ] ) reveals that liddi is mediated by virtual photons centered around three frequencies ; namely , that of the incident laser , @xmath25 , and those of the sidebands @xmath159 .
the former , linear , process gives rise to the liddi term proportional to @xmath163 , whereas the latter give rise to a nonlinear liddi process manifest by @xmath164 . here , the term `` linear '' is more easily understood when we consider the scattering rate @xmath153 [ see eq .
( [ r ] ) ] : if the scattered real photons are at the same frequency as the incident light , as in the term @xmath165 , then the process is linear , as in elastic scattering . however , if the scattered photons are at a different frequency , here @xmath131 , then the scattering is inelastic and we call it a nonlinear process . likewise ,
viewing liddi as the virtual - photon ( and cooperative ) counterpart of scattering , we call the process mediated by virtual photons around @xmath132 `` linear '' and that mediated by @xmath159 `` nonlinear '' . as discussed below , previous treatments of liddi @xcite obtained only the linear contribution , whereas our formalism accounts also for the nonlinear response of the atoms to the laser light .
most of the previous results of liddi , e.g. those of refs .
@xcite , are obtained in the limit of weak laser amplitude @xmath12 compared to atom - laser detuning @xmath13 , assuming also that the atoms are in the ground state .
navely , this result can be reproduced by our formalism in this large - detuning limit @xmath166 as follows : we first note that @xmath167 and further approximate @xmath168 for each atom , such that @xmath169 .
we then have @xmath170 , such that we find that the nonlinear liddi vanishes , @xmath171 , and only the linear part remains , @xmath172 by recalling the dynamical polarizability of atoms in their ground - state and in the limit of far detuned laser , @xmath173 , and considering the known solution for the rddi @xmath163 in free space @xcite , we recover the free - space liddi result from @xcite .
this result fails to capture the nonlinear part of liddi which is clearly important beyond the weak laser ( i.e. large detuning ) approximation .
interestingly enough however , the above result , eq .
( [ ul ] ) , appears to be inconsistent even within the @xmath166 approximation , from two main reasons : ( 1 ) the assumption that the atoms always stay in the ground state , which although seems reasonable for large detuning , is not always true , especially for long times ( see sec .
iv b below ) , and generally the state of the atoms @xmath135 has to be found by solving the master equation ( [ me2 ] ) .
( 2 ) even if we assume that the atoms are in the ground state , i.e. @xmath174 , and take the approximate @xmath16 from ( [ g ] ) instead of just taking @xmath168 , we find @xmath175
. then , to second order in @xmath176 , there exists also a nonlinear term , describing liddi mediated at ac - stark shifted atomic frequency ( for e.g. @xmath177 ) @xmath178 , @xmath179 which is totally missed by previous treatments and eq .
( [ ul ] ) .
this is because in previous treatments linearity , namely , that the mediating virtual photons are at the same frequency as the incident laser photons , was imposed by the method of calculation . in appendix c ,
we briefly review two such previous approaches and show how linearity is imposed by them , thus clarifying the reason for the absence of the nonlinear term therein . in fact , as shown in sec .
iv below , the above results , ( [ ul ] ) and ( [ unl ] ) , are obtained for liddi in the large - detuning and transient regime , for atoms initially in their ground states , whereas in steady - state only the linear part ( [ ul ] ) survives . moreover , eqs .
( [ ul ] ) and ( [ unl ] ) provide an interesting insight into the nature of liddi in the large - detuning limit : looking at the coefficient @xmath180 in front of the rddi rate @xmath147 in both equations , and recalling that it is proportional to the probability to excite an initially ground - state atom in this regime , we can interpret liddi as the excitation of an atom by the laser at probability @xmath180 thus allowing it to interact with another atom via rddi at rate @xmath147 .
let us illustrate our formalism by considering the liddi potential for atoms inside a cavity geometry .
such an analysis is very relevant for current experimental and theoretical research in many - atom systems inside cavities @xcite .
we treat this example in two cases , the first case where dissipation ( emission / scattering ) is significant such that the relevant liddi is that in the steady state , and the second case in the transient regime where dissipation is negligible . for both cases
however , we first have to find the underlying effects of rddi and emission . for simplicity , we assume a perfect cavity of length @xmath181 and effective area @xmath182 , and consider a single transverse mode , such that the relevant photon modes are those in the longitudinal direction @xmath183 , @xmath184 with frequencies @xmath185 and where @xmath186 is the polarization index , e.g. @xmath187 .
assuming all three @xmath70 s ( @xmath69 ) are between and do not coincide with the cavity modes @xmath188 , there exists no atom dissipation to the cavity modes , i.e. they do not contribute to @xmath80 , whereas their contribution to the rddi , given by eqs .
( [ d ] ) , ( [ un ] ) , is @xmath189 with @xmath190 ( for a more complete treatment of rddi in a cavity see @xcite ) . since the transverse ( @xmath191 ) area of the cavity mirrors is finite , the atoms are also coupled to non - confined photon modes .
we then expect that their spatial scaling is similar to that of free - space modes , such that their contribution to the cooperative effects @xmath192 and @xmath117 falls off initially like @xmath193 and is negligible for typical atomic distances larger than the atomic wavelength . then these modes only contribute to the single - atom spontaneous emission rate , which is approximated as that of free - space . finally , we have @xmath194 and @xmath195 from eq .
( [ dc ] ) . assuming the dissipation rate is larger or comparable to the rddi rate , i.e. @xmath196 , there are no interesting dynamics at time - scales much shorter than @xmath197^{-1}$ ]
, such that dissipation is important .
then , the relevant liddi is that in the steady - state , which is calculated by solving the master equation ( [ me2 ] ) for @xmath150 with the dissipation and rddi parameters from eqs .
( [ dc ] ) and ( [ gfs ] ) . for the density matrix elements of interest
, we find the steady - state equation , @xmath198 \\
\mathrm{im}[\rho_{23 } ] \\
\end{array } \right ) = \left ( \begin{array}{c } 0 \\ -\tilde{\gamma}^- \\ -\tilde{\gamma}^- \\ 0
\\ 0\\ \end{array } \right ) , \label{mest}\ ] ] with @xmath199 , @xmath200 and @xmath201 .
the steady - state solution is then @xmath202 and @xmath203 .
inserting this solution into the state - dependent liddi potential in eq .
( [ liddi1 ] ) we find only linear liddi in steady - state , @xmath204 ^ 2}\right ] \nonumber \\ & & \times \hbar\delta^z_{12}\cos(\mathbf{k}_l\cdot\mathbf{r}_{12 } ) .
\label{ust}\end{aligned}\ ] ] for the large detuning limit , @xmath166 , we expand the above result to lowest order and find , @xmath205 reproducing the linear - analysis result shown in sec .
( [ ul ] ) .
our analysis shows that while rddi is mediated by confined cavity photon modes , the dissipation is driven by the non - confined free - space - like modes . for @xmath25 ( and hence @xmath70 s )
close to a cavity mode @xmath188 , this leads to the possibility of rddi much stronger than dissipation , as revealed by eqs .
( [ dc ] ) and ( [ gfs ] ) .
therefore , in such a regime , it is relevant to consider the dynamics , and hence liddi , at time - scales much shorter than the dissipation time @xmath206^{-1}$ ] .
this allows to set @xmath207 in the master equation ( [ me2 ] ) for @xmath135 , obtaining , @xmath208 \\
\end{array } \right ) = \left ( \begin{array}{ccc } 0 & 0 & 2\delta \\ 0 & 0 & -2\delta \\
-\delta & \delta & 0 \\ \end{array } \right ) \left ( \begin{array}{c } \rho_{22 } \\
\rho_{33 } \\
\mathrm{im}[\rho_{23 } ] \\ \end{array } \right ) , \label{met}\ ] ] and @xmath209=0 $ ] , where @xmath210 . taking the initial state where both atoms are in the ground state , @xmath211 , we find @xmath212 .
the liddi potential in the transient regime @xmath213^{-1}$ ] then becomes @xmath214 .
\nonumber\\ \label{ut}\end{aligned}\ ] ] here we divided the liddi into its linear and nonlinear contributions , @xmath215 and @xmath216 , respectively . in the large - detuning approximation @xmath166 we find @xmath217 while the linear part of this result is the same as that of the steady - state result , here there is an additional non - linear contribution , which in the large - detuning regime is equivalent to that discussed in sec .
( [ unl ] ) .
since previous treatments @xcite have only yielded linear liddi processes , it is interesting to focus on the mechanism that leads to nonlinear effects in our more general approach .
these effects are manifest in the probing of @xmath125 at frequencies different than that of the incident laser @xmath25 in the expression for liddi , eq .
( [ liddi1 ] ) , which is relevant at time resolution @xmath218 , with @xmath55 , as revealed by our dynamic master - equation - based formulation ( sec . ii b ) . in
what follows , we discuss , from different perspectives , the origin of the nonlinear effects and when they are expected to be significant .
first , let us take an intuitive dynamic approach towards the fluorescence of an illuminated atom described by hamiltonian ( [ ha ] ) .
the solution for the dynamics of the atomic excitation - probability @xmath219 is known to be @xcite @xmath220.\ ] ] hence , the excitation probability @xmath219 , and consequently the scattering rate from a single atom , is modulated . since
this scattered radiation is centered around @xmath25 [ recalling that hamiltonian ( [ ha ] ) is written in a frame rotated by @xmath25 ] , this expression implies an amplitude - modulated scattered signal , whose spectrum contains peaks around @xmath132 and @xmath131 , as found above .
the distance between the peaks is @xmath118 , hence they can be resolved only at times larger than @xmath222 .
this provides an intuitive interpretation for the scattering rate in eq .
( [ r ] ) and the liddi related to it , eq . ( [ liddi1 ] ) : at times @xmath218 it is possible to distinguish between scattered or liddi - mediating photons at three frequencies @xmath70 ( @xmath69 ) ; hence , @xmath223 or @xmath125 , respectively , are probed at these frequencies for such a time - resolution .
let us recall the master - equation approach used to derive the liddi and scattering rate expressions above : in sec .
ii a we first solved for the relevant operator of the combined atom - laser , or dressed - atom , system that couples it to the reservoir , i.e. @xmath224 .
in analogy to the bare - atom case , where @xmath225 , we then proceeded to calculate the rddi between atoms , however this time they are dressed , with the time - dependent @xmath226 replacing the bare time - independent @xmath227 .
then , from the point of view of the reservoir , it interacts with a dipole that oscillates at all frequencies contained in @xmath228 ( around @xmath101 ) , and hence responds resonantly with all these frequencies , which are the same @xmath70 ( @xmath69 ) as before .
this explains why @xmath223 or @xmath125 are probed at these ( resonant ) frequencies .
moreover , it provides a frequency - domain picture for the time - scales at which the nonlinear effect , namely , the probing of @xmath223 and @xmath125 at frequencies other than @xmath132 , becomes significant : the overlap integral in eq .
( [ a1f ] ) of appendix a gives distinct results for different @xmath70 ( @xmath69 ) when the sinc function @xmath229 is narrower than the differences between different @xmath70 , @xmath218 and as long as the widths of @xmath223 and @xmath125 around @xmath70 , @xmath93 , are not larger than their difference , @xmath230 .
this approach to our results is related to known theories for the interaction between modulated systems and a reservoir @xcite : there , modulation of system parameters , such as atomic energy levels , by external fields , yields the replacement of the bare @xmath231 operator by @xmath232 in the system - reservoir coupling , @xmath233 being a time - dependent function determined by the modulation , in analogy to @xmath234 in our case . by writing @xmath235 ,
one then obtains that the reservoir response , e.g. @xmath223 , becomes @xmath236 instead of the bare - system @xmath237 @xcite , in analogy with our liddi and bare - atom - rddi results , respectively . in section ii c
we have treated the effects that underlie liddi and scattering as dipolar interactions driven by three distinct processes , each involving a different transition between the dressed states of fig .
1(b ) . equations ( [ hdd_d ] ) , ( [ liddi1 ] ) show that each of these processes has a different amplitude : @xmath238 for the process mediated by @xmath132 and @xmath239 for the process mediated by @xmath131 . in the following
we show that these amplitudes are directly related to those of corresponding transitions between laser - quantized dressed states @xcite . in the spirit of chapter vi in ref .
@xcite , we consider the coupling of a quantized laser mode @xmath240 to an atom via @xmath241 and assume the laser to be in a coherent state with average photon occupation @xmath242 . for large enough @xmath242 , we can ignore the laser photon fluctuations around the average @xmath243 compared to the average @xmath242 , such that in the @xmath242-manifold of the atom - laser space , @xmath244 , the relevant rabi frequency is @xmath245 and the corresponding dressed states are @xmath246 with eigenenergies @xmath247 , respectively .
the energy level diagram of two adjacent manifolds is plotted in fig .
2 . radiative transitions , i.e. emission of a photon from the atom to the vacuum , do not change the number of laser photons but take the atom from the excited to the ground state @xmath248 , i.e , involve only a @xmath249 operation . hence
, such transitions can only occur between adjacent manifolds .
there are four possible transitions , marked by dashed lines in fig .
2 : the two at the center represent the linear process which involves an emission of a photon at the laser frequency @xmath132 and the other two transitions at frequencies @xmath131 represent the nonlinear processes . hence , the strength of any process that involves an emission or exchange of a photon at frequency @xmath70 ( @xmath69 ) as in the scattering and liddi of eqs . ( [ r ] ) and ( [ liddi1 ] ) should be proportional to the probability of the transition between dressed states with energy difference @xmath250 in the emitting atom .
figure 2 shows that a nonlinear process with a @xmath122 photon involves the transition from @xmath251 to @xmath252 with transition amplitude @xmath253 [ inferred from eq .
( [ drs ] ) ] .
the associated probability is then @xmath254 , just as in the prefactor that appears in the @xmath255 scattering and liddi of eqs .
( [ r ] ) and ( [ liddi1 ] ) .
likewise , the relevant nonlinear @xmath256 process seen in fig .
2 is that of the @xmath257 transition , with amplitude @xmath258 and probability @xmath259 , as in the prefactor of the @xmath260 terms of eqs .
( [ liddi1]),([r ] ) . finally , there are two different processes that involve a @xmath25 photon , i.e. @xmath261 and @xmath262 with identical transition probabilities @xmath263 . summing the probabilities of both processes
we obtain @xmath238 as in the prefactor of the @xmath264 terms of eqs .
( [ liddi1]),([r ] ) .
this study has addressed laser - induced interactions between polarizable dipoles ( liddi ) , e.g. atoms , molecules or nanoparticles , in a general geometry .
it is expected to prove useful to the understanding and design of liddi between polarizable dipoles in confined electromagnetic environments , encountered in various experimental systems @xcite .
we have analyzed in detail liddi in a general geometry without assuming that the dipoles respond linearly to the applied laser field , and have found that it can be described as the resonant dipole - dipole interaction ( rddi ) between laser - dressed atoms in the same geometry .
this description has allowed us to obtain general formulae for both liddi and scattering in terms of the corresponding rddi and spontaneous emission , respectively [ eqs .
( [ liddi1 ] ) and ( [ r ] ) ] .
the liddi and scattering revealed by our formulae contain contributions due to the _ nonlinear _ response of the dipoles to the laser light , which are missing in previous treatments , e.g. those of ref .
@xcite . by reviewing the principles of these previous treatments ,
we have demonstrated their imposed linearity as the source of the absence of nonlinear liddi / scattering effects in these treatments .
we have then explained , using several simple and intuitive approaches , the origin of the nonlinear liddi and scattering terms in our expressions .
considerable progress has been reported in the ability to couple atoms to cavities @xcite and nano - fibers @xcite .
this promising direction has already lead to exciting predictions @xcite and experiments @xcite concerning the many - body physics of illuminated atoms in confined geometries . in this context , the generality of our formalism and its illustration for the analysis of liddi in a cavity , suggest its applicability for further exploration of this timely line of research .
we appreciate useful discussions with ilya averbukh . the support of isf and bsf is acknowledged .
here we present the derivation of the master equation ( me ) , eq .
( [ me2 ] ) . beginning with the double commutator in the me ( [ me1 ] ) ,
@xmath265 $ ] , it includes four terms , @xmath266 .
\label{com}\end{aligned}\ ] ] the rest of the derivation is similar for these four terms , hence in the following we present in detail only that of the first term . inserting @xmath267 from eq .
( [ hi2 ] ) into the first term in ( [ com ] ) we obtain @xmath268 . \nonumber\\ \label{a1a}\end{aligned}\ ] ] taking the trace @xmath269 in eq .
( [ me1 ] ) over the reservoir in the vacuum state , we have @xmath270=\langle \hat{a}_k \hat{a}_{k'}^{\dag } \rangle=\delta_{kk ' } , \label{tr}\ ] ] and similarly @xmath271=\mathrm{tr}_v[\rho_v\hat{a}_k^{\dag } \hat{a}_{k'}]=\mathrm{tr}_v[\rho_v\hat{a}_k^{\dag } \hat{a}_{k'}^{\dag}]=0 $ ] , such that only the third term in ( [ a1a ] ) contributes , and the trace over ( [ a1a ] ) becomes @xmath272 assuming that the relevant time - scale for the effects we are interested to resolve ( coarse graining resolution ) , @xmath86 , is much larger than @xmath273 for any @xmath274 , or simply that @xmath275 , we can view the exponent on the right hand side of ( [ a1b ] ) @xmath276 as a fast oscillation that averages out for @xmath274 and hence replace it with a kronecker delta @xmath90 , yielding for ( [ a1b ] ) @xmath277 taking the continuum limit @xmath278 , where @xmath279 is the density of photon modes @xmath18 in the considered geometry [ e.g. @xmath280 in free - space or @xmath281 in a waveguide with on - axis wavenumber @xmath282 , we identify the vacuum reservoir two - point ( autocorrelation ) spectrum as @xmath283 then , performing the integration @xmath284 in the me ( [ me1 ] ) and using ( [ g ] ) , the term in ( [ a1c ] ) becomes @xmath285 such that the me has @xmath286 in its left - hand side ( lhs ) and the above term ( [ a1d ] ) , along with 4 similar terms , in the right - hand side ( rhs ) .
_ markov approximation_. assume now that we are interested to solve the me for very short times @xmath287 , much shorter than the typical time - scale of variation for @xmath74 .
namely , we assume that @xmath288 hardly changes within the integral from @xmath289 to @xmath290 , such that it can be taken out of the integral and approximated as @xmath291 , which is the so - called markov approximation . at first
, this seems like a strange assumption , considering the fact that eventually we are interested to find an equation for the dynamics of @xmath74 on time - scales @xmath290 where @xmath74 does change appreciably , but this will become clearer in the following , where @xmath86 is interpreted as a coarse - graining time .
next , we need to solve the resulting integral over @xmath292 , @xmath293 where the coordinate transformation @xmath294 was used and @xmath295 is the heaviside step function . using also the relations @xmath296 where @xmath297 denotes the principal value upon integration , the me term ( [ a1d ] ) becomes @xmath298 .
\label{a1e}\end{aligned}\ ] ] denoting the sinc function @xmath299 , ( [ a1e ] ) becomes @xmath300 , \nonumber\\ \label{a1f}\end{aligned}\ ] ] with @xmath106 and @xmath107 from eq .
( [ d ] ) in the main text .
this expression describes an overlap integral between the complex linear response ( spectrum ) of the vacuum reservoir to the two - atom system @xmath301 and the sinc function @xmath229 of width @xmath302 around @xmath303 . denoting the typical width of the vacuum spectral response functions @xmath106 and @xmath107 around @xmath70 by @xmath93 , and assuming @xmath91 , we can approximate @xmath304 and finally get @xmath305\tilde{s}_{i\nu}\tilde{s}_{i\nu'}^{\dag}\rho(t ) .
\label{a1g}\ ] ] the rest of the three terms that appear on the rhs of the me can be treated similarly , thus obtaining the me in eq .
( [ me2 ] ) .
however , here it is written for short times @xmath287 w.r.t the typical time - variation of @xmath74 .
nevertheless , since the coefficients of the me are constant , it has the form @xmath306 with @xmath307 a time - independent superoperator .
then , its formal solution for @xmath287 is @xmath308 , and similarly @xmath309 for @xmath310 .
this procedure can be repeated successively for any time duration @xmath290 , so that the me is in fact valid for arbitrary long time @xmath290 , including time durations where the system state @xmath74 changes considerably .
_ separation of time - scales_. the coefficients of the me terms become time - independent in the markov approximation , namely by assuming that the time - resolution of interest , @xmath86 , is much larger than the so - called correlation time of the reservoir @xmath92 . on the other hand , in order to have sufficient resolution in probing the system dynamics , recall that we assumed @xmath86 to be much shorter than the typical time of variation for the system , that can be directly read from the me to be @xmath311 , i.e. the inverse of the rates @xmath312 .
this implies , @xmath313 namely , that the time - scales for the reservoir memory / correlation , @xmath92 , and the system dynamics , @xmath314 , are well separated .
let us consider the dependence of the liddi potential , eq .
( [ liddi1 ] ) , on the orientation of the dipolar matrix element @xmath30 and laser polarization .
we note that both @xmath12 and @xmath147 ( via @xmath315 ) depend on @xmath30 , which we recall to be @xmath316 , where @xmath317 is the dipole operator ( e.g. the electron charge times electron position operator in an atom ) . for the case of polarizable dipoles with a fixed orientation ,
@xmath30 is fixed , and so are the products @xmath318 and @xmath319 , such that @xmath12 , @xmath147 and hence the liddi @xmath144 can be calculated .
however , for cases where the atomic polarizability is effectively isotropic , the only preferred direction is that of the laser polarization , such that it effectively determines the orientation of @xmath30 . in the following we illustrate this idea for two typical scenarios of atomic and molecular dipoles
_ isotropic atoms_. consider the ground state @xmath16 to have a spherical symmetric electronic wavefunction , i.e. with spherical harmonic quantum numbers @xmath320 : @xmath321 . then
, the quantization axis @xmath183 that fixes the quantum number @xmath322 is arbitrary and , assuming the laser is linearly polarized , we can choose it to be that of the laser polarization , i.e. @xmath323 .
consider now the possible dipole - allowed excited states in a spherically symmetric potential for the electron , @xmath324 with @xmath325 .
then , by dipole selection rules , the dipole operator is generally written as @xmath326 where @xmath327 is a right / left circular polarization unit vector , respectively .
then , the product @xmath328 in @xmath12 implies that the only state to be excited by the laser is that of @xmath329 , namely , @xmath330 , such that the relevant dipolar matrix element that appears in the liddi potential @xmath144 is @xmath331 . considering a circular laser polarization , @xmath332 , the arbitrary quantization axis @xmath183 is now chosen to be perpendicular to both circular polarizations , and accordingly the product @xmath319 in @xmath12 imposes @xmath333 and @xmath334 , again parallel to the laser polarization . _
randomly oriented anisotropic molecules_. consider now molecules with an anisotropic polarizability , e.g. where for a given quantization axis not all possible @xmath325 excited states exist , and the dipole matrix element is some fixed @xmath30 .
however , by allowing for random molecule orientation , the quantization axis of each molecule ( with index @xmath36 ) becomes random and so does the orientation of its corresponding dipole matrix element @xmath335 .
we recall , @xmath336 where @xmath337 is the unit vector of @xmath338 and where we considered for simplicity @xmath339 in the large - detuning approximation as in eqs .
( [ ul ] ) and ( [ unl ] ) .
the , choosing the spherical coordinates of the vector @xmath335 with angles @xmath340 and @xmath341 , setting @xmath323 , we have @xmath342 \nonumber \\ \mathbf{d}_{\nu}\cdot \mathbf{e}_l&=&|\mathbf{d}|\cos(\theta_{\nu}),\end{aligned}\ ] ] with @xmath343 ( @xmath344 ) . inserting the above products into @xmath144 in ( [ u12 ] ) , and assuming no correlation between the random orientation of different molecules , we obtain @xmath345,\end{aligned}\ ] ] where @xmath346 denotes averaging over orientation with an isotropic distribution .
performing the averaging we finally get @xmath347 this result is equivalent to taking @xmath348 in @xmath147 and @xmath12 .
therefore , again , yet for a different spherically symmetric system , we effectively obtained @xmath349 .
in section iii we have mentioned that the apparent discrepancy between our liddi results and those obtained by previous treatments ( e.g. @xcite ) is due to the linearity imposed in those treatments . here
we would like to briefly review the essence of two typical approaches used previously , and highlight how linearity is imposed in them , thus preventing them to capture the nonlinear terms of liddi revealed by our formalism .
consider the following description of liddi depicted in fig .
3(a ) : a strong laser beam illuminates two atoms , such that atom 1 is subject to the laser electric field @xmath350 and atom 2 to the field @xmath351 . focusing on atom 1 , the laser induces its polarization , which to first order in @xmath28 is @xmath352 , @xmath353 being the atomic linear polarizability at the laser frequency @xmath25 . then , since the polarized atom 1 is an oscillating dipole , it scatters a field @xmath354 , which then arrives at atom 2 where it is given to lowest order ( born approximation ) by @xmath355 . here
@xmath356 is proportional to the green s function of the electromagnetic - field propagation at frequency @xmath25 from @xmath357 to @xmath358 in the considered geometry .
since the electromagnetic energy of atom 2 ( like any dipole ) is @xmath359 , @xmath360 being the total electric field at @xmath358 , the lowest - order interaction energy @xmath144 , interpreted as liddi , becomes , @xmath361 where the last step ( right - hand side ) is valid for atoms in their ground state and for large detuning .
the result and description above render the essence of treatments used in refs .
. its analogy to the linear part of our result ( [ ul ] ) , @xmath362 , becomes transparent upon recalling that the rddi term @xmath125 is directly related to the green s function @xmath363 @xcite .
the origin of linearity in this treatment is clear : the atoms are modeled as polarizable dipoles by using their linear response to electric fields , @xmath364 , at the same frequency as the incident radiation .
this leads to the sampling of the green s function at frequency @xmath25 , thereby obtaining @xmath365 . clearly , by imposing a description where the highly nonlinear two - level atoms linearly respond to the strong laser , effects due to real or virtual photons at frequencies other than @xmath25 can not be revealed .
a different approach , applied in e.g. refs.@xcite , uses a qed time - independent perturbative treatment similar to that used to calculate the van der waals and casimir - polder forces . taking the state of the laser as a number state for the time - being ,
the laser and the vacuum are treated as a single system ( the photons ) , where there are @xmath242 photons in the mode with @xmath25 and vacuum in the rest of the modes .
then , the liddi potential is interpreted as the lowest ( fourth ) order correction ( in the atom - photon coupling ) of the state @xmath366 , namely , both atoms in the ground state , @xmath242 photons in the laser mode @xmath181 and vacuum ( @xmath289 ) in the rest of the modes .
this energy correction is then given by @xcite @xmath367 with @xmath20 from eq .
( [ h ] ) .
this is a sum over intermediate / virtual states , @xmath368 , @xmath369 and @xmath370 , where @xmath371 is the energy of state @xmath372 without the interaction @xmath20 , e.g. @xmath373 ( taking the atomic ground - state energy to zero here ) .
the possible sets of virtual states can be represented by diagrams , such as that in fig .
3(b ) , with a total of 24 diagrams @xcite . in order to illustrate the essence of this approach ,
it is enough to focus on the diagram in fig .
3(b ) : it describes the intermediate states @xmath374 , @xmath375 and @xmath376 with energies @xmath377 , @xmath378 and @xmath379 , respectively , where @xmath104 denotes the mode of the virtual photon that mediates the interaction between vertices ii and iii , with @xmath380 . by identifying the laser rabi frequency as @xmath381 , @xmath382 being the dipole coupling to the laser mode , the resulting energy correction for this diagram becomes @xmath383 yielding again the linear liddi .
the crucial step that leads to linearity here is the treatment of atom - laser interaction in vertices i and iv : these vertices give rise to the factor @xmath384 which is indeed small for large detuning .
however , such a treatment , where the interaction with the strong laser is taken only in two bare atom - laser vertices , does not allow to account for the nonlinear effect .
c. j. pethick and h. smith , _ bose - einstein condensation in dilute gases _ ( cabridge university press , 2002 ) .
c. cohen - tannoudji , _ atomic motion in laser light _ , in _ fundamental systems in quantum optics _
, les houches , session liii , 1990 , pp . 1 - 164 ( elsevier science publisher b.v . , 1992 )
. g. kurizki , i. e. mazets , d. h. odell and w. p. schleich , int .
b * 18 * , 961 ( 2004 ) .
d. p. craig and t. thirunamachandran , _ molecular quantum electrodynamics _
( academic , london , 1984 ) .
t. thirunamachandran , mol . phys . * 40 * , 393 ( 1980 ) . a. salam , advances in quantum chemistry * 62 * , 1 ( 2011 ) .
d. odell , s. giovanazzi , g. kurizki and v. m. akulin , phys .
* 84 * , 5687 ( 2000 ) . s. giovanazzi , d. odell and g. kurizki , phys . rev
. lett . * 88 * , 130402 ( 2002 ) .
d. h. j. odell , s. giovanazzi and g. kurizki , phys .
lett . * 90 * , 110402 ( 2002 ) .
m. lemeshko , phys .
a * 83 * , 051402(r ) ( 2011 ) .
j. rodrguez and d. l. andrews , opt . com .
* 282 * , 2267 ( 2009 ) .
d. e. chang , j. i. cirac and h. j. kimble , phys .
* 110 * , 113606 ( 2013 ) .
e. shahmoon , i. mazets and g. kurizki , arxiv:1309.0555 ( 2013 ) .
a. goban , k. s. choi , d. j. alton , d. ding , c. lacroute , m. pototschnig , t. thiele , n. p. stern and h. j. kimble , phys .
lett . * 109 * , 033603 ( 2012 ) .
e. vetsch , d. reitz , g. sagu , r. schmidt , s. t. dawkins , and a. rauschenbeutel , phys .
lett . * 104 * , 203603 ( 2010 ) .
h. ritsch , p. domokos , f. brennecke and t. esslinger , rev .
phys . * 85 * , 553 ( 2013 ) .
r. h. lehmberg , phys . rev . a * 2 * , 883 ( 1970 ) .
d. petrosyan and g. kurizki , phys .
lett . * 89 * , 207902 ( 2002 ) ; t. opatrny , b. deb and g. kurizki _ ibid . _ * 90 * , 250404 ( 2003 ) ; g. kurizki , a. g. kofman , and v. yudson , phys . rev .
a * 53 * , r35 ( 1996 ) .
a. gonzalez - tudela , d. martin - cano , e. moreno , l. martin - moreno , c. tejedor and f. j. garcia - vidal , phys . rev .
lett * 106 * , 020501 ( 2011 ) .
kevin lalumire , barry c. sanders , arjan f. van loo , arkady fedorov , andreas wallraff and alexandre blais , arxiv:1305.7135 ( 2013 ) g. kurizki , phys .
a * 42 * , 2915 ( 1990 )
; g. kurizki and j. w. haus , j. mod . opt . * 41 * , 171 ( 1994 ) .
t. kobayashi , q. zheng and t. sekiguchi , phys .
a * 52 * , 2835 ( 1995 ) .
g. i. kweon and n. m. lawandy , j. mod
. opt . * 41 * , 311 ( 1994 ) .
e. shahmoon and g. kurizki , phys .
a * 87 * , 033831 ( 2013 ) .
v. venkataraman , k. saha and a. l. gaeta , nature photon .
* 7 * , 138 - 141 ( 2012 ) .
j. m. raimond , m. brune , and s. haroche , rev .
phys . * 73 * , 565 ( 2001 ) .
j. d. thompson , t. g. tiecke , n. p. de leon , j. feist , a. v. akimov , m. gullans , a. s. zibrov , v. vuleti and m. d. lukin , science * 340 * , 1202 ( 2013 ) .
j. koch , t. m. yu , j. gambetta , a. a. houck , d. i. schuster , j. majer , a. blais , m. h. devoret , s. m. girvin , and r. j. schoelkopf , phys . rev .
a * 76 , * 042319 ( 2007 ) .
b. r. mollow , phys .
a * 12 * , 1919 ( 1975 ) . c. cohen - tannoudji , j. dupont - roc , and g. grynberg , _
atom - photon interactions : basic processes and applications _ , ( wiley - vch , 2004 ) . s. barnett and p. m. radmore , _ methods in theoretical quantum optics _ ( oxford univeristy press , 2002 )
. h. j. carmichael , _ statistical methods in quantum optics 1 _ ( springer , 1998 ) .
r. h. dicke , phys . rev . * 93 * , 99 ( 1954 ) .
b. r. mollow , phys . rev .
* 188 * , 1969 ( 1969 ) .
j. rodrguez and d. l. andrews , phys .
a * 79 * , 022106 ( 2009 ) .
d. dzsotjan , a. s. srensen and m. fleischhauer , phys .
b * 82 * , 075427 ( 2010 ) . a. g. kofman and g. kurizki , phys . rev
. lett . * 93 * , 130406 ( 2004 ) .
e. shahmoon and g. kurizki , phys .
a * 87 * , 013841 ( 2013 ) .
w. niedenzu , t. grieer and h. ritsch , eur .
. lett . * 96 * , 43001 ( 2011 ) .
r. mottl , f. brennecke , k. baumann , r. landig , t. donner and t. esslinger , science * 336 * , 1570 ( 2012 ) . | polarizable dipoles , such as atoms , molecules or nanoparticles , subject to laser radiation , may attract or repel each other . we derive a general formalism in which such laser - induced dipole - dipole interactions ( liddi ) in any geometry and for any laser strength are described in terms of the resonant dipole - dipole interaction ( rddi ) between dipoles dressed by the laser .
this approach provides a simple route towards the analysis of liddi in a general geometry .
our general results reveal liddi effects due to nonlinear dipolar response to the laser , previously unaccounted for .
the origin of these nonlinear effects is discussed .
our general formalism is illustrated for liddi between atoms in a cavity . |
of the most predominant protocols of the internet is the transmission control protocol ( tcp ) , which provides a high level of reliability on end - to - end connections .
it regulates the transmission rate between the two ends of a connection based on the changes of the underlying network . in other words , it estimates the condition of the connection and adjusts its congestion window ( _ cwnd _ ) accordingly . in order to enhance the total performance of tcp over high - speed networks , many congestion control algorithms ( ccas )
have been proposed in the literature such as scalable tcp @xcite , hs - tcp @xcite , h - tcp @xcite , bic @xcite , tcp africa @xcite , tcp compound @xcite , fusion @xcite , yeah @xcite , tcp illinois @xcite , cubic , @xcite and hcc @xcite .
the main three approaches , which are employed by these ccas , are either loss - based , delay - based or loss - delay - based approach @xcite .
the delay - based approach relies on the variation of delay resulted by big buffers and/or long rtts , which are presented in high - bdp networks .
however , the low - bdp networks employ small buffer sizes and generate very short rtts resulting in a very trivial delay variation , which makes it worthless to use a delay - based approach . these characteristics of such networks allow tcp to rely only on packet losses since it is the only indicator to congestion . in turn
, this behavior makes tcp very sensitive to packet losses , which negatively affects its total performance .
one of the latest versions of loss - based tcp cca is agile - sd @xcite , which has been proposed with a view to mitigating the sensitivity to packet losses .
it proposes agility factor ( @xmath2 ) , which is used to shorten the epoch time ( the epoch is the time needed by a cca to increase its @xmath3 from the time of reduction to the time of attaining the maximum utilization of the link ) , as shown in fig.[fig : agile ] .
the unique agility factor mechanism ( afm ) of agile - sd has been evaluated by extensive simulation experiments , which confirmed that agile - sd has the ability to reduce the sensitivity to packet losses and to improve the performance of tcp to a reasonable extent , especially over low - bdp networks . in the literature , many mathematical models have been proposed in order to study and evaluate the performance of tcp .
the authors of @xcite proposed a clear model to calculate the steady - state throughput of tcp as a function of rtt and loss rate .
also , the authors of @xcite evaluated the throughput of tcp using periodic loss based model .
the authors of @xcite used the markov chains and the stationary distribution to predict the behavior of tcp dealing with red - based routers .
the authors of @xcite used the random matrix model to evaluate the performance of multiple aimd tcp flows via drop - tail queue management system .
the authors of @xcite derived the stationary distribution of markov chains to calculate the steady - state throughput of tcp cubic .
as aforementioned , the models in @xcite calculate the steady - state throughput of tcp as a function of rtt and loss rate , while they did not take the buffer size into account . in order to calculate the throughput of tcp as a function of buffer , rtt and loss rate , we propose a novel mathematical model to calculate the average throughput of both newreno and agile - sd over high - speed networks .
the main contributions of this model are : first , to validate the simulation results of agile - sd by comparing it to the numerical results of this model and to the results of newreno as a benchmark .
second , to study the impact of @xmath0 parameter on the throughput and epoch time , where the epoch time is a period of time confined between two consecutive losses .
third , to formulate an equation for automating the configuration of optimal @xmath0 based on the given system parameter @xmath4 in order to increase the scalability of agile - sd .
consider a link between two computers and suppose that the link speed is _ c _ ( in kbps ) and the source has a large file to send to the destination .
assume that the packet size is @xmath5 ( in kbits ) .
also , assume that the rtt ( in seconds ) is constant , which is a common assumption in loss - based tcp mathematical models @xcite .
since the bandwidth - delay product ( bdp ) of a link is equal to the multiplication of _ c _ by rtt , consequently , the maximum congestion window size @xmath6 is calculated as equation @xcite , @xmath7 as for the buffer size , we necessarily need to understand its impact on the behavior of tcp to know how it could affect this model . based on equation
, suppose that we have a scenario in which the maximum window size is 100 packets , which allows 100 packets , maximum , to be in - flight . then , let us say that the buffer size used in this link is 20 packets , which allows the maximum packets in - flight to be 120 packets .
accordingly , we can conclude that the buffer size is playing a role in extending the capacity of the link , thus , equation can be reformulated as below , @xmath8 where @xmath9 is a constant indicates to the buffer size in packets .
at the congestion avoidance stage , agile - sd increases its congestion window @xmath10 by a small fraction after every reception of acknowledgment , similarly like the standard newreno .
however , newreno calculates this increase as 1 over @xmath10 , which gives : @xmath11 while in agile - sd this increase is calculated as @xmath2 over @xmath10 , which gives : @xmath12 where the cycle index @xmath13 @xmath14 , @xmath15 is the amount of degradation in congestion window after loss , @xmath16 , @xmath1 is the multiplicative decrease factor of tcp , the epoch index @xmath17 @xmath18 and @xmath19 is the number of epochs needed for tcp connection to finish its data transmission , where @xmath20 and tt is the total transmission time , as shown in fig.[fig : agile ] . as for @xmath21 , it represents the agility factor of agile - sd at the cycle @xmath13 in the epoch @xmath17 , which is calculated as in equation below , @xmath22 where @xmath23 and @xmath0 are system parameters , @xmath24 , @xmath25 , , @xmath26 is the congestion window on the time of reduction of the previous epoch , where , @xmath27 is the congestion window just before the time of reduction of the previous epoch and @xmath28 is a system parameter representing the minimum allowed congestion window . as for max
( ) function , it is used to always guarantee that @xmath29 . from fig.[fig : agile ] , it is very clear that the epoch is a period of time confined between two consecutive losses @xcite .
each epoch @xmath17 contains a number of sequent cycles and every cycle @xmath13 is a sub - period of time confined between two consecutive increases or between a consecutive degradation and increase . as known ,
the standard tcp variants are rtt - dependent ; in which tcp needs to receive all acks of the previous _ cwnd _ to increase their _ cwnd _ by one .
consequently , these tcp variants consume a complete rtt per cycle in order to achieve that increase .
since the operating systems allow their _ cwnd _ to be only an integer number , we use the flooring function during the assignment of a new value to the _ cwnd _ of the operating system .
thus , the transmission rate ( tr ) of data at cycle @xmath13 in epoch @xmath17 is , @xmath30 differently , agile - sd is an rtt - independent , which consumes only @xmath31 to increase its _ cwnd _ by one .
thus , agile - sd consumes shorter time than the time needed by the rtt - dependent tcp variants to reach the maximum _
cwnd_. in other words , agile - sd shows shorter epoch time than rtt - dependent tcps , as shown in fig.[fig : epochsa ] and fig.[fig : epochsb ] .
[ h ! ] in fact , agile - sd is still subject to the same transmission rate of standard tcp as equation , however , agile - sd sends only @xmath32 packets in every cycle , where the cycle duration time is only @xmath33 seconds .
this behavior does not increase the transmission rate per cycle , but it controls the period of transmission time given to every cycle .
thus , it shortens the time given for the cycles with low rates and lengthens the time given to the cycles with high rates , which results in shorter epochs ( as in fig.[fig : epochsb ] ) than the rtt - dependent tcp variants ( as in fig.[fig : epochsa ] ) .
let us now calculate the transmission rate per epoch , which is equal to the total data sent , in an epoch , over its duration time .
since the data sent at a cycle @xmath13 is equal to @xmath34 and the duration time of a cycle @xmath13 is equal to @xmath35 , thus , the epoch average transmission rate ( eatr@xmath36 ) is equal to the summation of data sent over all cycles in epoch @xmath17 divided by the duration time of the epoch cycles , as in equation below , @xmath37 where @xmath20 , tt is the total transmission time , @xmath15 is the number of cycles in epoch @xmath17 , , @xmath5 is the packet size , @xmath26 is the congestion window at the time of reduction of the previous epoch where , @xmath38 , and @xmath1 is the multiplicative decrease factor of tcp .
however , calculating the transmission rate per epoch is not a final target , it is only to be used for calculating the total average transmission rate ( tatr ) of a connection .
fundamentally , any connection passes through an @xmath19 epochs , where every epoch @xmath17 sends an amount of data equal to @xmath39 over a duration time equal to @xmath40 .
thus , the tatr of a connection can be calculated as the summation of total data sent through the connection over the total duration of its transmission time , as in equation below , @xmath41 assume that packet losses are either _ congestion loss _ or _ random packet loss_. congestion loss happens when the transmission rate attains the maximum capacity @xmath42 of the bottleneck link or the maximum window size @xmath6 , and it also happens when the buffers are overflowed . as for random packet loss , it can be caused by collision , interference and/or fading , where the random packet loss is subject to the poisson distribution with rate @xmath43 @xcite .
thus , the probability density function @xmath44 for a given congestion window @xmath10 is : @xmath45 where @xmath46 is the probability of occurrence of @xmath47 packet losses in congestion window @xmath10 .
let @xmath48 denote the range of congestion window size , which represents a system with an @xmath49 states , where @xmath50 and @xmath51 , which results in , @xmath52 where @xmath6 and @xmath28 are the maximum and minimum allowed congestion window , respectively .
let @xmath53 denote the transition probability matrix of the markov chains for the system of @xmath49 states , given as , @xmath54 } & _ { v[1,2 ] } & \dots & _ { v[1,n-1 ] } & _ { v[1,n ] } \\ _ { v[2,1 ] } & _ { v[2,2 ] } & \dots & _ { v[2,n-1 ] } & _ { v[2,n ] } \\
\vdots & \vdots & \ddots & \vdots & \vdots \\ _ { v[n-1,1 ] } & _ { v[n-1,2 ] } & \dots & _ { v[n-1,n-1 ] } & _ { v[n-1,n ] } \\ _ { v[n,1 ] } & _ { v[n,2 ] } & \dots & _ { v[n , n-1 ] } & _ { v[n , n ] } \\ \end{bmatrix},\end{aligned}\ ] ] where @xmath55 $ ] represents the transition probability of the system to move from the @xmath56 state to the @xmath57 state , where .
if the _ cwnd _ size is in the @xmath56 state , it is denoted by @xmath58 .
this finite set of _ cwnd _ sizes corresponds to the markov chain with @xmath49 states .
let us consider an example of a system with @xmath59 and @xmath60 .
based on equation , the system has only 5 states ( @xmath61 ) , where the finite set of the system states is @xmath62 where @xmath63 .
also , let us suppose that @xmath64 .
thus , the transition diagram of the system can be drawn as shown in fig.[fig : tdiagram ] , where the transition probabilities are calculated based on the probability distribution shown in equation @xcite below : @xmath65 = \left\{\begin{array}{ll } p_x(w ) & ,
j = \lfloor\beta \times i\rfloor\\ & , i = [ 1 , n-1]\\ \\
1-p_x(w ) & , j = i+1\\ & , i = [ 1 , n-1]\\ \\ 1 & , j = \lfloor\beta \times
i\rfloor\\ & , i = n\\ \\ 0 & , \text{otherwise } , \end{array}\right . \label{eq11}\end{aligned}\ ] ] where @xmath10 is the congestion window , @xmath55 $ ] is the transition probability of the system to move from the @xmath56 state to the @xmath57 state , and .
, @xmath66 , @xmath62 and @xmath63 . ]
let us now put the probabilities of all represented transitions shown in fig.[fig : tdiagram ] into their relevant places in the transition matrix @xmath53 .
then , let us fill the unrepresented transitions by zeros .
thus , the resulted transition matrix @xmath53 will be as shown in fig.[fig : tmatrix ] .
let @xmath67 be the @xmath56 row vector in @xmath53 , which represents the probabilities for the system to move from the @xmath56 state to all possible _ n _ states of the system , @xmath68 , v[i,2 ] , \ldots , v[i , n]\big],\end{aligned}\ ] ] where every element represents the probability to move from the @xmath56 state to one @xmath57 state . [ !
, @xmath66 , @xmath62 and @xmath63.,title="fig : " ] from fig.[fig : tmatrix ] , we can see that the @xmath69 is a special case , in which the system will transit from @xmath70 to @xmath71 due to a random loss with @xmath72 probability or a congestion loss with .
thus , the total probability will be . in order to validate the distribution of these probabilities in the matrix @xmath53 ,
the summation of the probabilities in every row vector must always be equal to one .
@xmath73 } = 1 , \forall i \in \{1 , 2 , \ldots , n\}.\end{aligned}\ ] ] let @xmath74 be the probability distribution of the system states at the transition @xmath75 , where @xmath74 is equal to the product of the previous probability distribution of the system states at the transition @xmath76 and the matrix @xmath53 , as follows , @xmath77 where @xmath78 is the initial state , which is the transition probability distribution of system states at the time zero . since this model is for the average throughput , we set the initial _ cwnd _ of the system to @xmath79 , which represents a system in the state of loss degradation at the congestion avoidance stage .
thus , the initial state row vector of the markov chain is denoted by @xmath78 , in which the probabilities are distributed as below :
@xmath80 = \left\{\begin{array}{ll } 1 & , j = \lceil\beta n\rceil\\ \\ 0 & , \text{otherwise } \end{array}\right . , \forall j
\in \{1,2,\ldots , n\ } , \label{eq111}\end{aligned}\ ] ] where @xmath17 is the item index in @xmath78 . in order to calculate the average throughput ( ath ) of the system
while taking the congestion and random losses into account , we employ equation into equation to count all data packets received at the destination over the given connection .
since the source sends @xmath81 data at transition @xmath75 and the destination receives only @xmath82 data at the same transition , therefor , the ath can be calculated as in equation below , @xmath83 where @xmath84 is the number of iterations needed for the system to reach the steady - state , @xmath85 is the transpose of the sample space vector @xmath86 , where @xmath87 $ ] , and @xmath88 is the agility factor at @xmath89 , as , @xmath90 where @xmath23 and @xmath0 are system parameters , @xmath24 , @xmath25 , . indeed , another markovian model to account for the changing agility factor ( @xmath2 ) is needed to be integrated with equation to find the stationary distribution of the system states , by calculating the eigenvector of @xmath53 with eigenvalue 1 .
however , the integration of two markovian models will increase the complexity of the main model and may hinder the process of understanding the model . in order for us to avoid this complexity and
since @xmath88 is easy to be determined at every transition @xmath75 in the main markov chain of this model , we account for @xmath88 using equation at every cycle in equation .
based on observation , we found that the system starts entering to the steady state after about 4 thousand iterations , however , 10 thousand iterations are used in order to increase the precision of results .
in this section , we compare the numerical results for both newreno and agile - sd to the simulation results in order to confirm the validity of the proposed model .
indeed , our mathematical model is able to represent both newreno and agile - sd , where the numerical results of newreno can be obtained using this model with @xmath91 . using simulation
, we confirmed that agile - sd would exactly perform like newreno if @xmath0 was set to 1 , as shown in fig.[fig : lambda ] . and
newreno , under different pers . ] in order to obtain the numerical results , we use octave version 4.0 , which is a major free alternative software to matlab . more specifically , we run our model with both @xmath23 and @xmath0 equal to 1 to obtain the results of newreno .
thereafter , we run it again with @xmath23 and @xmath0 equal to 1 and 5 , respectively , to obtain the results of agile - sd . for better understanding , table [ params1 ]
shows the setting of the model .
p4.3cm|ll & + + buffer size ( @xmath9 ) & + loss rate ( @xmath92 ) & + packetsize ( @xmath5 ) & + link capacity ( @xmath42 ) & + 2-way link delay ( rtt ) & + minimum allowed window ( @xmath28 ) & + maximum allowed window ( @xmath6 ) & + system states ( @xmath49 ) & + sample space of the system ( @xmath86 ) & + iterations ( @xmath84 ) & + congestion control algorithms & & + & + multiplicative decrease factor @xmath4 & & + minimum @xmath2 @xmath93 & & + maximum @xmath2 @xmath94 & & + + further , we present the impact of some parameters , such as buffer size and @xmath0 , on the epoch time and average throughput of agile - sd . at the end
, an automated algorithm configuration and parameter tuning ( aacpt ) @xcite process is exploited to automate the calculation of the optimal @xmath0 for a given multiplicative decrease factor @xmath1 , instead of configuring it manually by the system administrators as a preset parameter . for simplicity , a node - to - node topology , as shown in fig.[fig : topology ] , is used to validate the results of this mathematical model using ns2 simulator version 2.35 .
the sender - node ( s1 ) sends an ftp data to the receiver - node ( d1 ) for 100 seconds over a full - duplex tcp connection , in which the bandwidth is 1 gbps , the two - way propagation delay is 10 milliseconds , and the buffer size in each node is varied from 4 to 128 packets subject to drop - tail , while the per is varied from @xmath95 to @xmath96 .
furthermore , @xmath1 and @xmath0 are set to 0.5 and 5 , respectively .
in fact , these setting ( as shown in table [ params ] ) are used to mimic the worst short - distance network configurations , where the buffer sizes are very small and the pers are very high .
p4.3cm|l & + + ccas & agile - sd , newreno + @xmath1 & 0.5 + @xmath97 & 1 , 5 only for agile - sd + buffer size & from 4 to 128 packets + packet size & 1000 bytes + link capacity & 1gbps + 2-way link delay & 10ms + loss rate ( per ) & from @xmath95 to @xmath96 + queuing algo & drop - tail + traffic type & ftp + sack , fack & disabled + simulation time & 100 seconds + simulator version & ns-2 ver2.35 + + [ params ] in order to obtain the simulation results shown in fig.[fig : lr ] , where the buffer size is fixed to @xmath98 packets , we run the simulation for 10 times to calculate the average throughput for every per . as for the simulation results shown in fig.[fig : comparison ] , where the per is fixed to @xmath95 , we also run the simulation for 10 time to calculate the average throughput for every buffer size . additionally , the numerical results are collected using the proposed model under the same conditions used in the simulations in order to present a fair comparison . as it is clear in fig.[fig : lr ] , the simulation and mathematical results are very close to each other
. however , the mathematical curve slightly diverges from the simulation curve , due to neglecting the slow start phase in our model and also due to the additional processing overhead occur during the simulation .
nevertheless , the results in fig.[fig : lr ] confirms the validity of this model supported by the results in fig.[fig : comparison ] and fig.[fig : rtt ] , which compare between the throughput of agile - sd and the standard newreno under different conditions of buffer size and rtt , respectively . as shown in fig.[fig : comparison ] , agile - sd overcomes the standard newreno in all cases .
nevertheless , both algorithms show the same pattern for a proportional relationship between throughput and buffer size , which supports our assumption in equation . as well as , agile - sd can perform better than the standard newreno in most rtt cases , especially when the rtt and the used buffer size are very small , as shown in fig.[fig : rtt ] .
since these network characteristics are seen in short - distance and near - zero buffer networks ( such as fiber optic networks ) , thus , the behavior of agile - sd would be promising for improving tcp performance if such networks are targeted . per where @xmath99 and @xmath100 .
] [ h ! ] per where @xmath99 , @xmath100 and buffer size is only 4 packets.,title="fig : " ] due to its unique agility factor mechanism ( afm ) , agile - sd shows shorter epoch time than the standard newreno , as shown in fig.[fig : cwnd ] .
this behavior reduces the underutilized area of bandwidth and , in turn , it improves the average throughput of tcp , as shown in equation , where the epoch time is represented by @xmath33 at the denominator . as shown in fig.[fig
: lambda - impact ] , it is very clear that the epoch time is inversely correlated to the value of @xmath0 parameter , while the average throughput is positively correlated to @xmath0 .
thus , the greater @xmath0 , the shorter epoch time and the higher average throughput , and vice versa .
hence , it can be deduced that the main player , which directly affects the performance of agile - sd , is the parameter @xmath0 .
thus , this parameter @xmath94 must be carefully configured based on the value of @xmath1 . [
t ! ] on the average throughput and epoch time.,title="fig : " ]
the designers and system users of parameterized algorithms routinely encounter a problem of how to find the optimum configurations or parameter settings to obtain the best possible results . in this section ,
an aacpt @xcite technique is used to find the optimal parameter settings of @xmath0 based on a set of problem instances of @xmath1 .
the optimum value of @xmath0 should reflect the minimum setting to obtain the maximum average throughput for a given @xmath1 .
henceforth , @xmath0 will be denoted by @xmath101 for the sake of simplifying the presentation .
let @xmath102 donate the set of possible configurations of @xmath101 , and let @xmath103 denote the set of problem instances of @xmath1 , where @xmath104 $ ] and @xmath105 $ ] .
let @xmath106 be the parameter configuration combination matrix with the size @xmath107 $ ] , which contains all possible @xmath108 combinations , where @xmath109 , @xmath110 , @xmath111 , @xmath112 , and @xmath113 .
thus , each row @xmath114 in @xmath106 merely denotes the possible @xmath102 configurations for a given problem instance @xmath115 , as shown in equation , @xmath116 \label{eq - row } , \forall i \in \{1 , 2 , \ldots , m\}.\end{aligned}\ ] ] let us now calculate the average throughput ( at ) of agile - sd for each configuration combination @xmath108 using equation , where the result is stored into the relevant at@xmath117 element in at , which is an @xmath107 $ ] matrix .
let us also track @xmath118 , which is the optimal @xmath101 configuration in @xmath102 for a given @xmath115 .
then , let us save every @xmath118 into the relevant position in the vector of the optimum setting ( @xmath119 ) , which is an @xmath120 $ ] matrix .
for more understanding , algorithm [ algo01 ] explains the process of aacpt in detail .
afterward , the at results matrix of agile - sd under all possible configurations is drawn as a graph , as shown in fig.[fig:3d ] , to show the complete perspective behind this aacpt .
the circled points in fig.[fig:3d ] reflect the vector of the optimal configurations @xmath121 for the given @xmath103 .
for each @xmath115 in problem instance space @xmath103 , the optimal @xmath101 configuration is @xmath118 from @xmath119 .
the relationship between the problem instance space @xmath103 and its optimal configuration space @xmath118 , as shown in fig.[fig : apc ] is directly extracted from fig.[fig:3d ] . in order to facilitate the process of formulating an equation to fit this relationship , the free on - line tool , namely @xcite ,
is used to find the trend line of the optimal points . eventually , the employment of tool produces a simple linear equation , which is able to directly calculate @xmath118 based on a given @xmath115 , as shown in equation , @xmath122 where @xmath123 is the length of @xmath119 , @xmath124 .
fortunately , the result of this equation is identical to the result of the aforementioned aacpt process , as shown in fig.[fig : apc ] .
[ h ! ]
* initialization : * + @xmath125 $ ] ; + @xmath126 $ ] ; + define @xmath127 at as @xmath128 $ ] matrices ; + define @xmath119 as @xmath129 $ ] matrix ; + set @xmath130 ; + [ h ! ] and @xmath1.,title="fig : " ] it is very clear that equation has the ability to improve the average throughput of agile - sd compared to newreno without the need for manually configuring the @xmath101 , as shown in fig.[fig : optimumlambda ] .
thus , this equation is highly recommended to be used in agile - sd since it can automate the calculation of the optimum value of @xmath101 parameter based on the preset value of @xmath1 , where @xmath1 is one of the main parameters in the standard tcp interface .
this automation of @xmath101 increases the scalability of agile - sd and maximizes its throughput .
more importantly , it helps for keeping the standard interface of tcp as it is , which facilitates the implementation of agile - sd into the real operating systems .
[ h ! ] , per@xmath131 , and buffer size is only 4 packets.,title="fig : " ]
in this paper , we propose a mathematical model to calculate the average throughput of agile - sd and newreno .
this model is designed based on markov chains , in which the congestion loss and packet error rate are considered . in order to validate the correctness of this model , a number of simulation experiments
are carried out , in which the results are compared to the outputs of this model to validate the results presented in @xcite . besides , this model evaluates the average throughput of agile - sd under different pers , rtts , and buffer sizes , in which agile - sd overcomes newreno even in the cases of small buffers and short rtts .
based on this results , agile - sd could be a promising congestion control algorithm for the short - distance networks , especially that with near - zero buffers such as fiber optic networks .
also , the proposed model shows that the average throughput is positively correlated to the values of @xmath1 and @xmath101 .
thus , the greater the @xmath1 and @xmath101 , the higher the throughput and vice versa . from the other side
, there must be an inverse correlation between @xmath1 and @xmath101 in order to keep the aggressiveness of agile - sd balanced .
in other words , whenever @xmath1 is increased , @xmath101 must be decreased and vice versa . for more robustness , an aacpt process is exploited to formulate an equation , which calculates the optimal configuration of @xmath101 based on a given @xmath1 , as in equation ( [ eq - opt ] ) . for future work
, we plan to implement this equation in agile - sd to make its interface compatible with the standard tcp interface .
this work has been partially supported by the malaysian ministry of education under the fundamental research grant frgs/1/2014/ict03/upm/01/1 for financial support . l. xu , k. harfoush , and i. rhee , `` binary increase congestion control ( bic ) for fast long - distance networks , '' in _
infocom 2004 .
twenty - third annual joint conference of the ieee computer and communications societies _ , vol . 4 , 2004 ,
. 25142524 . r. king , r. baraniuk , and r. riedi , `` tcp - africa : an adaptive and fair rapid increase rule for scalable tcp , '' in _
infocom 2005 .
24th annual joint conference of the ieee computer and communications societies .
proceedings ieee .
_ , 2005 , pp . 111
. k. tan and j. song , `` compound tcp : a scalable and tcp - friendly congestion control for high - speed networks , '' in _ 4th international workshop on protocols for fast long - distance networks ( pfldnet ) _ , 2006 , pp .
k. kaneko , t. fujikawa , z. su , and j. katto , `` tcp - fusion : a hybrid congestion control algorithm for high - speed networks , '' in _ proc .
pfldnet , isi , marina del rey ( los angeles ) , california .
_ , 2007 , pp .
3136 . ,
mohamed othman , borhanuddin ali , and zurina mohd hanapi , `` comparative study of high - speed linux tcp variants over high - bdp networks , '' _ journal of network and computer applications _ , vol .
43 , pp . 6675 , 2014 .
j. padhye , v. firoiu , d. towsley , and j. kurose , `` modeling tcp throughput : a simple model and its empirical validation , '' _ acm sigcomm computer communication review _ , vol .
28 , no . 4 , pp . 303314 , 1998 .
j. padhye , v. firoiu , d. f. towsley , and j. f. kurose , `` modeling tcp reno performance : a simple model and its empirical validation , '' _ ieee / acm transactions on networking ( ton ) _ , vol . 8 , no . 2 , pp .
133145 , 2000 .
a. misra and t. j. ott , `` the window distribution of idealized tcp congestion avoidance with variable packet loss , '' in _ proceedings .
ieee infocom99 , eighteenth annual joint conference of the ieee computer and communications societies _ , vol .
3.1em plus 0.5em minus 0.4emieee , mar 1999 , pp .
15641572 .
i. huh , j. y. lee , and b .- c .
kim , `` decision of maximum congestion window size for tcp performance improvement by bandwidth and rtt measurement in wireless multi - hop networks . '' _ international journal of information processing systems _ , vol . 2 , no . 1 ,
3438 , 2006 .
a. jansang , a. phonphoem , and b. paillassa , `` analytical model for expected packet delay evaluation in ieee 802.11 e , '' in _ communications and mobile computing , 2009 . cmc09 .
wri international conference on _ , vol .
2.1em plus 0.5em minus 0.4emieee , 2009 , pp . 344348 .
a. jansang and a. phonphoem , `` a simple analytical model for expected frame waiting time evaluation in ieee 802.11e hcca mode , '' _ wireless personal communications _ ,
69 , no . 4 , pp . 18991924 , 2013 . mohamed a. alrshah received his bsc degree in computer science from naser university - libya , in 2000 , and his msc degree in computer networks in may 2009 from universiti putra malaysia .
and now he is a ph.d candidate in the faculty of computer science and information technology , universiti putra malaysia .
he has published a number of articles in high impact factor scientific journals .
his research interests are in the field of high speed tcp protocols , high speed network , parallel and distributed algorithms , software defined networking , network design and management , wireless networks . mohamed othman received his phd from the universiti kebangsaan malaysia with distinction ( best phd thesis in 2000 awarded by sime darby malaysia and malaysian mathematical science society ) .
now , he is a professor in the faculty of computer science and information technology , universiti putra malaysia ( upm ) .
he is also an associate researcher at the lab of computational science and mathematical physics , institute of mathematical research ( inspem ) , upm .
he published more than 160 international journals and 230 proceeding papers .
his main research interests are in the fields of high speed network , parallel and distributed algorithms , wireless network ( mpdu- and msdu - frame aggregation , tcp performance , mac layer , resource management , and traffic monitoring ) and scientific telegraph equation and modeling .
borhanuddin mohd .
ali received his b.sc ( hons . ) electrical and electronics engineering from loughborough university of technology in 1979 , his m.sc and phd in electromagnetic engineering , from the university of wales , in 1981 - 1985 , respectively .
borhanuddin is now a full time professor in department of computer and communication engineering , faculty of engineering upm .
he was on secondment to mimos bhd , a government research lab on ict , heading the centre of wireless communications .
previously , he served as the director of the institute of multimedia and software , a centre of excellence within the same university , 2002 - 2006 , spent 1 year at celcom r&d in 1995 as visiting scientist , and 2 years at might in 1996 - 97 as a senior manager , charged with coming up with research and policy direction for malaysian telecommunication industry . in 1996
he helped to realize the formation of teman project , and later was made the chairman of the myren research community , a national research testbed .
he is a chartered engineer and a member of the iet , and senior member of ieee .
he was the chair of ieee malaysia section 2002 - 2004 , and previously the chair of comsoc chapter , 1999 - 2002 .
his research interest spans wireless and broadband communications , and network engineering .
binti mohd hanapi received her b.sc . in computer and electronic system engineering , strathclyde university in 1999 , her m.sc . in computer and communication systems engineering , universiti putra malaysia in 2004 and her ph.d.in electrical , electronic , and system engineering , universiti kebangsaan malaysia in 2011
now , she is working as a lecturer at universiti putra malaysia .
she has received an excellence teaching awards in 2005 , 2006 and 2012 . and she has received silver medal in 2004 and bronze medal in 2012 . she is a leader of some research projects and she has published many conference and journal papers .
her research interests in routing , wireless sensor network , wireless communication , distributed computing , network security , crpytography , and intelligent systems . | agile - sd is one of the latest versions of loss - based congestion control algorithm ( cca ) , which has been proposed to improve the total performance of tcp over high - speed and short - distance networks .
it has introduced a new mechanism , called agility factor mechanism ( afm ) , which shortens the epoch time to reduce the sensitivity to packet losses and in turn to increase the average throughput .
agile - sd has only been tested via simulation , however , it has not been mathematically proven or evaluated .
the contribution of this paper is twofold : first , a new mathematical model for the throughput of newreno and agile - sd is proposed .
this model is designed using the well - known markov chains to validate the correctness of agile - sd and to show the impact of buffer size , multiplicative decrease factor and maximum limit of agility factor ( @xmath0 ) on the total performance .
second , an automated algorithm configuration and parameter tuning ( aacpt ) technique is employed to optimize and automate the configuration of @xmath0 .
further , the numerical results for both newreno and agile - sd are compared to the simulation results in which the validity of the proposed model is confirmed .
moreover , the output of aacpt is exploited to formulate a new equation which calculates the optimal @xmath0 from a given @xmath1 in order to conserve the standard interface of tcp .
this equation increases the scalability of agile - sd and improves its total performance .
addtoresetfootnotepage agile - sd , transmission control protocol , congestion control , markov chains , average throughput . |
coronal mass ejections ( cmes ) are large scale transient eruptions from the sun that travel into the heliosphere .
the impact of these transients onto earth are the dominant source of the most severe space weather incidents @xcite .
the velocity and magnetic field vectors within a cme are often considered the primary properties of interest for estimating their space weather effects .
cmes also play a significant role in removing magnetic helicity which is generated as the sun evolves ( e.g. * ? ? ?
it has also been shown to provide a significant contribution in modulating the open solar flux over a solar cycle ( e.g. * ? ? ? * ) .
this open solar flux has been correlated to the long term variations of the heliospheric magnetic field ( e.g. * ? ? ?
* ; * ? ? ?
therefore understanding the distribution and amount of toroidal flux contained within cmes are important for estimating the severity of a space weather incident as well as for long term solar influence on earth s climate .
the morphology of a cme containing a magnetic flux rope ( fr ) during its propagation towards earth is often described as ` pancaking ' @xcite .
historically , this is supported by remote observations from coronagraphs which display cmes propagating radially away from the sun , and more recently with heliospheric imagers on board stereo @xcite .
also , these cameras often display a flux rope structure as a black void within the images such that the observed white light marks the outer boundary , such as sheath regions ( e.g. * ? ? ?
* ; * ? ? ?
synthetic white light images from simulations with a magnetic flux rope have also been shown to produce a dark cavity @xcite .
these observations are also further supported by simulations that are often used in space weather forecasting ( e.g. enlil model ) .
the enlil model describes a cme as a high density packet without a coherent magnetic field structure @xcite .
both the stereo observations and the enlil model display a cme uniformly stretching perpendicular to the radial flow , while contradictory evidence is often presented with in situ measurements .
flux rope models that satisfy the lundquist solution ( e.g. * ? ? ?
* ; * ? ? ?
* ) are often successfully shown to reproduce similar results to the magnetic vectors from in situ measurements within cmes .
the morphological structure of this lundquist - solution flux rope model is cylindrical with a circular cross section ( i.e. not stretched into a ` pancaking ' structure ) . in this paper
we present evidence , through the use of simulations , that a cme can display both the ` pancaking ' properties seen remotely and circular cross section measured in situ .
flux ropes are often characterised by magnetic fields that are stronger than the ambient solar wind and are topologically defined to have a coherent rotation of their magnetic field lines around a central axis .
these flux rope cmes have often been observed to possess a relatively low proton temperature when detected in situ , and were labelled magnetic clouds ( m.c .
the fraction of cmes that contain a fr structure is under debate , with statistical arguments suggesting that there is a solar cycle dependence as well there being a variation between @xmath3 @xcite and @xmath4 @xcite .
recently , further evidence suggests that all cmes may in fact be associated with a magnetic flux rope topology by hypothesising certain events without these signatures are due to the spacecraft trajectory or a misidentification @xcite .
a successfully simple technique to model the magnetic field structure of flux ropes from solar wind in situ measurements has been a force - free model .
some of the earliest attempts assumed a cylindrically symmetric condition @xcite and due to their success , continue to be used to this day ( e.g. * ? ? ?
* ; * ? ? ?
they used a constant alpha value in the bessel function solutions to @xmath5 @xcite .
further developments to these models have included ad hoc expansion to the field vectors @xcite , incorporating an elliptical cross section @xcite and curved flux rope axis @xcite .
an alternative technique to the modelling has been developed by @xcite . by using magnetic field and plasma measurements and integrating the nonlinear , planar grad - shafranov ( gs ) equation ,
the technique is able to estimate some global cme properties similar to that of flux rope models mentioned earlier . unlike the models
, the gs technique does not enforce any fixed global morphology for the 2d cross - section of the flux rope .
this leads to estimates of magnetic flux and flux rope size that are not constrained to a predetermined model morphology .
both of these parameters are pertinent in this study .
the popularity and success of the cylindrical flux rope has often led research to progress by implementing an idealised morphology as a first approximation , e.g. in the case of a 3d reconstruction technique for coronagraphs @xcite , or for the growth rate of the cross - sectional width further in the heliosphere @xcite .
extensions to these studies have been carried out to better understand the distortions towards a more elliptical shape as the cme propagates into the heliosphere @xcite .
however the variations in the ambient solar wind can further develop the morphology into a more complex structure ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
cmes are often considered to propagate radially away from the sun .
this behaviour forces an expansion onto the object that is perpendicular to its radial propagation .
this effective velocity would usually be different from the internal radial expansion which is commonly attributed to a pressure force .
a kinematic study by @xcite inferred that a cme would drastically deviate away from the force - free cylindrical configuration and labelled the expansion as ` pancaking ' . however a more detailed investigation into the aspect ratio of the cme cross section by @xcite revealed that the global morphology of an initially circular structure would approach a steady state .
if the radial expansion is considered to equal a fixed fraction of the bulk flow then the aspect ratio would tend towards a constant value of @xmath6 , depending on the original size of the cme .
however , the success of the circular fr fits may be due to the cme structure being much more circular on a local scale .
this suggests that at certain locations , the magnetic tension or external forces due to structured solar wind conditions could restore the magnetic vectors towards the minimum energy configuration of the bessel functions @xcite .
a study by @xcite supports this scenario as the authors investigated the shock stand - off distance of cmes to infer the aspect ratio .
the results from 45 cmes gave an average aspect ratio of @xmath7 .
also , a visual inspection of the several studies with reconstructed cme morphology using the gs technique have shown ratios consistently @xmath8 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
estimates of the total flux content within cmes have important consequences for estimating the long term trends of the total open solar flux @xcite .
the study by @xcite assumed the heliosphere contained a constant background of open flux with a time - varying component due to interplanetary cmes .
the contribution of cme flux was estimated from 132 magnetic clouds with the analysis assuming a constant-@xmath9 force free fr with a circular cross - section @xcite .
this paper tests the validity of assuming a circular cross - section for estimating the flux content as it is often the assumption used ( e.g. * ? ? ?
we have implemented an axisymmetric mhd simulation of a flux rope cme propagating out to 300 solar radii ( rs ) .
the ambient solar wind is initialised with an inner boundary at 4rs and propagated at a uniform supersonic speed of 300km / s .
the magnetic flux rope is super - imposed at 10 rs as a semi - analytically calculated toriodal solution to the azimuthally symmetric grad - shafranov equation ( where @xmath10 is defined as the toroidal direction and parallel to the fr axis ) .
initially the fr is given a radial width of 8rs with a speed of 700 km / s . in figure 1
, we summarize the evolution of the flux rope at three intervals .
the green line indicates the boundary of the fr and the black contours display magnetic field lines which follow the flux function , @xmath11 , defined by @xmath12 where @xmath13 is the in - the - plane magnetic field .
detailed mathematical formulation of the simulation is presented elsewhere @xcite and we restrict our discussion to a brief description of the initial conditions and the evolution of the fr .
the simulated fr is placed into the heliospheric region with an instantaneous velocity equal to 700 km / s . while this does not capture the initiation and low coronal phases of the cme evolution
, it is a reasonable approximation for fast cmes that often originate in active regions .
the flux rope was initiated with a toroidal ( axial , or out - of - the - plane ) flux of @xmath14mx which is of a typical order of magnitude observed .
figure 1 displays the plasma @xmath0 on the colour scale to highlight the regions that are more magnetically dominated than others .
the plasma @xmath0 is defined as the ratio of the plasma pressure @xmath15 to the magnetic pressure @xmath16 .
the temperature and density of the background solar wind are not replicated accurately relative to observations , as such , the quantitative values in the plasma @xmath0 is therefore not realistic ( i.e. at 1 au the simulation density and temperature are larger than observed ) .
however , the focus of this investigation is the relative changes in the plasma @xmath0 from one region to the next , which is expected to be well reproduced by the simulation .
the entire fr begins with low plasma @xmath0 and its fast release quickly produces a leading shock front ahead .
figure [ beta ] shows that within the sheath region ahead of the fr leading edge , the plasma @xmath0 increases significantly above both the ambient solar wind and fr values . from the beginning of the fr propagation
, the plasma dominated sheath region can be seen to drag the open field lines locally around fr .
after some time , these regions also interact with the outer edges of the fr such that the plasma @xmath0 in these areas also increases .
these outer ` layers ' of the fr are then also dragged away from the ecliptic plane with the ambient solar wind ( see also figure 3 in @xcite ) . by the time the fr approaches 215rs ( @xmath17 )
, the outer boundary of the flux rope ( figure [ beta ] , green curve ) more closely resembles an evenly pancaked structure ( as shown by @xcite ) than earlier ( @xmath18 ) in its propagation .
the fr edge is defined by a boundary between closed and open field lines .
upstream of the fr is a sheath region containing ` shocked ' solar wind material which is downstream of the shock wave . in the early evolution of the flux rope , the outer boundary appears closer to a circular substructure with extended ` wing ' segments .
the formation of a wake is quickly established downstream of the ` wings ' . in our case ,
similar to the case shown by , the outer layers of the flux rope with weaker magnetic field are unable to resist the incipient external flow , so that they are bodily convected to the rear " .
the plasma @xmath0 within the outer boundary is predominately below ten in the initial conditions as well as in the early evolution , whereas the sheath region is significantly higher .
however , as the flux rope propagates the higher plasma @xmath0 of the sheath region smoothly changes to a lower value in the cme core .
we see that the concentrated magnetic flux seen in the earlier evolution of the fr still persists out to 1 au . in order to investigate the distribution of flux , we manually selected an arc between the outer most edge of the fr ( green curve ) to the center as defined by the maximum flux function ( green cross ) .
we choose to define the points to approximately occur at the positions where the magnetic field line is a maximum distance away from the center .
a circular arc was then created to best fit the selected data points ( black dashed line in figure [ beta ] and [ fluxcont ] ) .
as this paper is interested in understanding how the flux is ` stretched ' in the out - of - ecliptic plane ( z direction ) , eleven positions along the arc were chosen to be uniformly distributed along the z direction ( figure [ fluxcont ] , red squares ) .
the toroidal flux contained within the contours of these constant flux function values were then calculated ( curves in figure [ fluxcont ] ) .
the distance was then calculated as the position on the contour furthest away from the fr center .
this method enables the flux content to be sampled at an approximately uniform rate with regards to the distance away from flux rope center .
other methods were investigated , but they often concentrated the points within the inner core of the flux rope with the outer edge of the flux rope being under - sampled .
also , this procedure meant that the exact positions from the initial manual selections does not matter as the maximum distance of the field line and the out - of - plane flux contained within the closed loop are estimated independently .
figure [ fracflux ] displays the distribution of toroidal flux as a function of the distance to the fr center . in order to better compare the results from different time frames
, the toroidal flux values have been normalised by the maximum value at the outer edge ( @xmath19 ) ; and the distance away from the fr center was normalised by the maximum diameter of the fr ( @xmath20 ) .
the top panel of figure [ fracflux ] shows the initial distribution ( t=0 hours ) of the fr which is a solution of toriodal grad - shafranov equation . for comparison
, the flux distribution from the cylindrical bessel function is also displayed as a thick green curve .
the flux distribution is clearly not uniform at @xmath21 rs , but interestingly the distribution remains non - uniform even when the morphology may appear otherwise .
the results show that beyond 100rs the majority of the flux ( @xmath1 ) remains within an inner core that is about@xmath22 of the overall size .
figure [ betaprof ] displays the plasma @xmath0 values for the eleven positions within each of the frames shown in figure [ beta ] .
a transition region between the more magnetically and plasma dominated regimes ( low to high plasma @xmath0 ) is found at @xmath23 of the total toroidal flux content .
this boundary is shown as dashed blue lines in figure [ fracflux ] and [ betaprof ] .
the formation of these stretched high @xmath0 regions leads to a morphology of ` wings ' and the disconnected wake ( vortical eddys)to the rear . to understand the causality of the quasi - cylindrical magnetic core we investigate the source of the resistance to deformation which leads to the observed transition from low to high plasma @xmath0 .
the two competing forces acting on the periphery of the flux rope are : 1 . the act of deformation which can be considered as the kinetic energy density of the relative flow onto the flux rope ; 2 .
the act of restoration which is represented by the energy density of the poloidal magnetic field ( i.e. the tension force of a twisted field ) .
the kinetic energy density is defined by , @xmath24 . where @xmath25 is the relative velocity of the solar wind onto the flux rope .
the poloidal - field energy density is defined by @xmath26 .
where @xmath27 is the ( in - the - plane ) poloidal component of the magnetic field .
figure 5a displays a circular ring of high intensity @xmath27 that is part of the core fr region . from the ring to the center of the fr , the @xmath27 field strength decreases rapidly as the central region of the fr becomes more dominated by an axial field direction ( toroidal direction ) .
the contours of the flux function are displayed as white curves on figure [ 2by2]a , with a thicker curve marking the outer boundary of the fr as determined by a outermost closed field line .
other regions ( out of the equatorial plane and to the rear of the fr ) also have a high @xmath27 field strength .
these regions surround the wake that forms behind the fr ` wings ' .
we define the equipartition line by the same method as em98 ; i.e. when the two energy densities are equal .
this definition is also equivalent to identifying the locus of all points where the magnetic weber number equals one .
the magnetic weber number is defined as : @xmath28 the weber number can therefore also be expressed in relation to the poloidal component of the alfvn velocity , @xmath29 : @xmath30 in the hydrodynamical literature , the weber number is a dimensionless unit that measures the relative importance of a fluid s inertia compared to its surface tension . in our magnetohydrodynamic case ,
the role of the surface tension is replaced by the magnetic tension of the poloidal field component .
em98 studied the physics of a rising twisted magnetic flux tube in two dimensions , under conditions akin to the solar convection zone .
in this situation , the flux tube rose slowly without the formation of a shock front and through a non - magnetic background medium . under this scenario
, it was appropriate to define the relative velocity as @xmath31 .
where @xmath32 is the velocity at a single location - the apex of the flux tube . by choosing the relative velocity with respect to the location of the flux rope center we display the weber equipartition line ( white curve ) in figure [ 2by2]b
the colour scale displayed in figure [ 2by2]b is the magnitude of the relative velocity . for the front and rear boundary of the flux rope , near to the equatorial plane , the weber boundary accurately follows the drastic change in the plasma @xmath0 .
the weber boundary reflects a quasi - circular morphology surrounding the core region of the fr .
however , at the rear of the fr the weber boundary bridges to a another region out of the equatorial plane . for regions significantly outside the equatorial plane ,
the weber boundary as defined by em98 , inappropriately extends further into the ` wing ' regions of the flux rope .
this region also extends significantly into the magnetically disconnected wake .
to better understand the complex nature of the weber line , the contour of we=1 is over - plotted onto an intensity map of the @xmath33 ( see figure [ 2by2]c ) .
figure [ 2by2]c displays the weber contour ( grey curve ) with an arbitrarily chosen contour of @xmath34 ( white curve ) . with regards to the core of the fr , both boundaries display very similar quasi - circular morphology that remain within the closed magnetic field lines of the fr .
the extended regions of the weber line that stretch into the wake also occurs for the @xmath33 line .
the bridge that connects to the extended regions out of the equatorial plane appears to be caused by a high poloidal magnetic field strength . whereas , the extended regions themselves seem to be caused by a reduction in the relative flow velocity ( @xmath35 ) .
a significant difference between the study performed by em98 and the one presented here is the direction of the solar wind flow falling onto the leading edge of the flux rope .
the presence of a leading shock front deflects the solar wind within the sheath region prior to arriving at the flux rope leading edge .
the presence of a shock front clearly adds further complication to our scenario .
therefore to correct for the weber equipartition line , we re - define the relative velocity to equal the component parallel to the in - plane magnetic field and relative to flux rope center .
i.e we define it as @xmath36 this method now correctly expresses the relative shear velocity of the background solar wind flowing tangentially to the flux rope magnetic field .
the velocity vectors of this newly - defined tangential flow is displayed in figure [ 2by2]a .
the intensity of the @xmath37 is displayed in figure [ 2by2]d along with 2 contour lines : @xmath34 ( white curve as shown in figure [ 2by2]c ) , and @xmath38 - i.e. using the velocity from equation [ eqve ] ( grey curve ) . the newer weber equipartition line ( @xmath39 ) surrounds a larger part of the fr than the original weber line .
@xmath39 line displays an extremely extended region outside , but connected to , the fr leading edge and is out of the equatorial plane .
this thin region is within the sheath region and approximately follows the curvature of the shock front .
the reason this region is included within the equipartition line is similar to that of the extended regions of the we=1 boundary - namely the @xmath37 is smaller than its surroundings . to understand how the sharp transition in plasma @xmath0 is related to the different equipartition lines we display the intensity of the plasma @xmath0 in figure [ webernew ] using the same time frame as figure [ 2by2 ]
the morphology of fr core in the figure , with a low plasma @xmath0 , can be seen to be quasi - circular .
figure [ webernew]b displays the same plasma @xmath0 intensity , but with 2 layers of semi - transparent grey shaded regions over - plotted .
the boundary demarcating the light and dark shading of each layer are the same as those displayed in figure [ 2by2]d .
one of the semi - transparent layers displays a white region for @xmath40 and black otherwise .
the other layer shows a white region for @xmath41 .
therefore figure [ webernew]b highlights the regions defined by the mathematical equipartition lines and darkens the surroundings .
we see that the @xmath42 equipartition line , for the most part , tightly surrounds the low plasma @xmath0 core .
the only exception is that of the thin extending section outside and to the rear of the fr .
this region is the lightest region in the figure as it is also within the boundary of the second transparent layer .
the region defined as @xmath40 but with @xmath41 is the light grey region . concentrating on this light grey region that remains within the magnetic flux rope , we notice that the area extends a little further in the ` wing ' sections of the fr . the boundary of the @xmath43 demarcates the region where the plasma @xmath0 has just transitioned into the higher values .
therefore the region inside the fr and between the two equipartition lines ( light grey region ) identifies the section where the sharp plasma beta transition occurs . inside the transition region
( i.e. closer to the fr center ) the topology of the fr displays a quasi - cylindrical structure .
recent developments in heliospheric imagery from the stereo mission have often shown that cmes continue to propagate radially away from the sun and deform into oblate shapes .
this is further supported by heliospheric simulations such as the one often used in space weather forecasting , enlil . however strong conflicting evidence from in situ data along one dimension inside
cmes suggest that they often have a magnetically circular cross - section . in this paper
we hypothesise that a fr may form a magnetic substructure core as an explanation as to why remote observations sometimes may appear to contradict the topology suggested by in situ data .
the weber equipartition line defines the portion of the obstacle that will stay together .
but there is a boundary layer surrounding the line where the plasma properties between the two regions smoothly but drastically change .
this boundary layer is where the sw is able to drag the outer flux rope field lines away .
therefore the drastic plasma @xmath0 change is a consequence of the tension force not restraining the field at the outer edges .
although our simulation would benefit from more advanced and realistic background solar wind conditions , we investigate the plasma @xmath0 within the cme and how this affects the distribution of flux . as the cme propagates the plasma @xmath0 transitions between low to high , which at first starts as a sharp boundary at the front edge of the fr structure ( i.e. inner boundary of the sheath ) .
this boundary then diffuses into a layer that surrounds the weber equipartition line and moves within the fr field itself .
this leads to a single cme behaving with two different properties : an inner core that is dominated by a stronger twist magnetic field which will favour a circular shape due to magnetic tension and the peripheral edge which has a stronger interaction with the ambient solar wind and is therefore prone to deformation . in a sense , the weber equipartition line acts as a belt or ` rubber band ' around the core region , restraining flux from stretching out . but with the presence of a boundary layer surrounding the equipartition line , magnetic flux ( field lines ) are able to slowly leak through the weber line to form part of the extend wings and then eventually disconnect from the flux rope all together . in our simulation we show that the flux rope core contains approximately @xmath1 of the original toroidal flux while remaining about @xmath2 of the overall size .
em98 performed a similar simulation of a twisted flux tube within a high - plasma @xmath0 regime , akin to a tube rising in the solar convection zone , to show that the conversion of the tube into two vortex rolls can be suppressed by increasing the magnetic field twist .
this work showed that the fr maintained its structure when conditions regarding the inflow velocity and poloidal alfvn velocity were met .
em98 performed a parametric study by varying the amount of magnetic field twist within the flux rope .
such a parametric study would be a natural extension to the work presented here . therefore further detailed studies which involve varying the tension force or changing the plasma interaction ( by changing the background plasma density or plasma @xmath0 ) should be performed to better understand the extent to which the substructure occurs .
a useful focus would be to understand whether this plasma @xmath0 transition scenario is a common occurrence or if it exists for only extreme ( geoeffective ) events .
another useful extension would be to perform a parametric study of cme speeds to better understand the cause of the complexity in the weber boundary for conditions with / without a propagating shock front .
understanding these thresholds may also provide clues into the requirements and limitations of magnetic field twist during the initial configuration above the solar surface . developing more advanced simulations that realistically replicate the background solar wind in three dimensions
would aid further understanding of this phenomena . by injecting a cme instantaneously with a large magnetic flux
, we emulate an extreme acceleration profile for a fast cme entering the inner heliosphere which is akin to that expected for a carrington event or other extreme cases . in this simulation ,
the propagation speed of the flux rope through the heliosphere is more comparable to a typical fast cme . under these conditions
it is clear that a head - on impact onto the earth will have a drastically different effect to one that , at first sight , would appear to be only moderately away from the sun - earth line .
however further work could benefit from investigating a more physical acceleration process similar to @xcite . the initiation process of launching a fr by @xcite shows that a propagating cme might display the characteristic of two regimes separated by surface of @xmath44 ( see figure 13 within the authors work ) .
the enlil model is often used for space weather forecasting .
however this model is usually implemented by assuming the cme can be initially modelled as a high density blob with no internal magnetic field .
the work in this paper suggests that without a restraining force from a twisted field the cme would indeed deform and spread out into a pancake structure .
but interestingly , the magnetic field from the background solar wind often drapes around the cme during its propagation .
this sheath region field could potentially serve as a restorative force preventing the cme from largely disintegrating by the time it arrives at earth .
further work into understanding the extent to which draping field ahead of a cme can act as a restraining force should also be pursued . at terrestrial distances ,
remote observations of cmes using the hi-2 camera on board the stereo mission have often identified flattened ` pancaked ' structures .
however , this is commonly performed by locating a dark cavity or a bright ` sheath ' region that surrounds the probable flux rope ( e.g. * ? ? ?
* ; * ? ? ?
while the sheath region has an elongated structure due to the overall morphology of a magnetic flux rope , the imagery does not express the topological distribution of flux within the observed cavity .
recent articles from in situ analysis have alluded to the possibility that different parts of a single cme may have differing in situ properties . by using in situ data and a cylindrical model ,
@xcite investigated the flux content of 20 cmes and showed the inner core ( @xmath45 of its total size ) remained more symmetric .
later @xcite created a flux rope model that relaxed the cylindrical symmetry explicitly to better estimate the cross sectional shape from in situ data .
dmoulin et al . investigated elliptical shapes with a variety of eccentricities .
the authors observed that the sharper ( more deformed ) parts of the boundary impose a stronger curvature , therefore the stronger magnetic tension reduces the field line bending inside the flux rope . for larger bending of the outer boundary , the magnetic tension increases such that the core field lines slightly shrink towards the center , forming a more circular shape .
in addition , for elliptically shaped flux ropes with significant deformations , the authors showed that the magnetic pressure is lower at the flux rope sides than the center . assuming the plasma pressure remains uniform , they argue that even for very low plasma @xmath0 , the force - free approximation is no longer valid .
they therefore expect these limb ( ` wing ' ) regions to be swept away by the solar wind .
they further hypothesis that reconnection with the solar wind magnetic field will contribute to removing these parts , while the core field maintains its identity .
@xcite also highlighted this conflict by hypothesising the magnetic tension may favour a circular shape and interpreting the external part as an interaction with the ambient solar wind .
our simulation support such a scenario of a core flux rope cme which is significantly smaller than the full extent of the original flux rope .
therefore if a spacecraft were to travel through the center of a flux rope , the in situ magnetic field measurements made would suggest a quasi - cylindrical obstacle similar to the core substructure shown in our simulations .
several authors have previously shown schematics suggesting that a quasi - cylindrical fr may be part of a larger magnetic obstacle ( e.g. * ? ? ?
* ; * ? ? ?
thus the trajectory of a spacecraft will dictate whether a cylindrical fr topology is measured or not .
this then brings into question whether all cmes are in fact frs or not @xcite - or more accurately speaking , if a fr topology always exists somewhere within the entire cme structure .
@xcite investigated the plasma @xmath0 properties of 181 cmes using a list of events provided by @xcite .
@xcite showed that the coherent structure of in situ measured frs often have a significantly lower plasma @xmath0 value than events without a fr topology ( see figure 3 of their paper ) .
these authors also find that the plasma @xmath0 is one of the most significant predictor variables for the presence of a fr topology .
the weber equipartition analysis in this paper provides a scientific explanation to their result .
several other interpretations for creating substructures within a cme which ultimately leads to a quasi - cylindrical flux rope at terrestrial distances have been proposed .
for example , @xcite proposed large current sheets can lead to magnetic reconnection which , in an extreme case , could lead towards forming multiple small flux ropes from a single larger one .
a review of other potential scenarios for creating substructures can be found in @xcite .
if a cme contains a magnetic substructure which is significantly more cylindrical than the entire morphology , then in situ measurements should show cme - like properties outside the magnetic flux rope as indicated by a multi - spacecraft analysis by @xcite .
a study by @xcite investigated cmes over solar cycle 23 and defined boundaries by the ionisation charge states of certain heavy ions .
figure 2 and 3 from their work shows that this may be a frequent occurrence .
however the charge states do not always overlap with the magnetic signatures ( e.g. * ? ? ?
* ; * ? ? ?
* ) and so can not be used as a definitive explanation .
therefore it is important for future work to investigate simulated time - series at different locations within the cme ( similar to * ? ? ?
was partially supported by the nasa living with a star jack eddy postdoctoral fellowship program , administered by the ucar visiting scientist programs and hosted by the naval research laboratory .
a.v . and m.g.l
were supported by nasa and the office of naval research .
the numerical calculations were performed using the supercomputing cluster at the solar - terrestrial environment laboratory , nagoya university . , b. t. , & gonzalez , w. d. 1997 , in washington dc american geophysical union geophysical monograph series , vol .
98 , washington dc american geophysical union geophysical monograph series , ed .
b. t. tsurutani , w. d. gonzalez , y. kamide , & j. k. arballo , 7789 ( colour scale ) and magnetic flux function(black contours ) are displayed for three times during an mhd calculation .
the edge of the closed field lines for the magnetic flux rope are displayed as a thick line(green ) .
the black crosses denote manually selected positions through the flux rope , and the dashed line is an arc of a circle that is optimally chosen from the crosses .
the middle panel identifies the uniformly distributed positions along the arc that is later used in our analysis ( red squares ) .
the figure shows that as the flux rope propagates the outer layers increase in plasma @xmath0 . ] against the distance away from the center of the flux rope .
the flux rope is normalised to the values of the outer edge .
the three panels represent the same time intervals displayed in figure [ beta ] .
the black crosses are the @xmath0 values for the same positions shown in figure [ fluxcont ] and [ fracflux].this figure shows that the flux rope has a sharp transition in plasma @xmath0 which coincides with approximately 80 @xmath46 of the toroidal magnetic flux . ] solar radii .
( a ) displays the poloidal component of the magnetic field @xmath47 ( colour scale ) and magnetic flux function ( white contours ) with the velocity vectors of the solar wind ( @xmath48 ) .
the @xmath37 is the parallel component to the magnetic field and measured relative to the fr center .
( b ) displays the magnitude of the velocity relative to the fr center ( colour scale ) with the weber equipartition line , we , which was determined using @xmath35 similar to @xcite ( white curve ) ( c ) displays the poloidal component of the alfvn speed ( colour scale ) with the we line ( grey curve ) and the contour of @xmath49 ( white curve ) .
( d ) displays the @xmath37 ( colour scale ) with the @xmath39 equipartition line which was determined using @xmath37 ( grey curve ) and the contour of @xmath49 ( white curve ) . ]
( colour scale ) with position of the shock front for the same heliocentric distance as shown in figure [ 2by2 ] .
the magnetic flux function are shown as black contours with the thicker black line demarcating the outer boundary of the fr .
( right ) the same as the left figure but with two additional semi - transparent layers overplotted .
one of the semi - transparent layers displays a white region for @xmath40 and black otherwise .
the other layer shows a white region for @xmath41 .
therefore the figure highlights the regions defined by the mathematical equipartition lines and darkens the surroundings . ] | we present a 2.5d mhd simulation of a magnetic flux rope ( fr ) propagating in the heliosphere and investigate the cause of the observed sharp plasma @xmath0 transition .
specifically , we consider a strong internal magnetic field and an explosive fast start , such that the plasma @xmath0 is significantly lower in the fr than the sheath region that is formed ahead .
this leads to an unusual fr morphology in the first stage of propagation , while the more traditional view ( e.g. from space weather simulations like enlil ) of a ` pancake ' shaped fr is observed as it approaches 1au .
we investigate how an equipartition line , defined by a magnetic weber number , surrounding a core region of a propagating fr can demarcate a boundary layer where there is a sharp transition in the plasma @xmath0 .
the substructure affects the distribution of toroidal flux , with the majority of the flux remaining in a small core region which maintains a quasi - cylindrical structure .
quantitatively , we investigate a locus of points where the kinetic energy density of the relative inflow field is equal to the energy density of the transverse magnetic field ( i.e. effective tension force ) .
the simulation provides compelling evidence that at all heliocentric distances the distribution of toroidal magnetic flux away from the fr axis is not linear ; with @xmath1 of the toroidal flux occurring within @xmath2 of the distance from the fr axis .
thus our simulation displays evidence that the competing ideas of a pancaking structure observed remotely can coexist with a quasi - cylindrical magnetic structure seen in situ . |
the study of soliton and black hole solutions of the einstein - yang - mills ( eym ) equations has been an active subject for some twenty - five years , triggered by the discovery of regular , static , spherically symmetric , solitons @xcite and `` coloured '' black holes @xcite in @xmath2 eym in four - dimensional asymptotically flat space - time . in these configurations ,
the nonabelian gauge field is purely magnetic @xcite and described by a single function @xmath4 of the radial coordinate @xmath5 .
numerical investigations @xcite show that @xmath4 must have at least one zero , and this is has also been proven analytically @xcite .
both the particle - like and black hole solutions arise at discrete points in the phase space of parameters , and can be indexed by @xmath6 , the number of zeros of @xmath4 .
the @xmath2 solitons and black holes have no magnetic charge at infinity .
the black holes , in particular , are indistinguishable from a schwarzschild black hole at infinity , and therefore present counter - examples to the `` no - hair '' conjecture @xcite . however , both the black hole and soliton solutions are unstable under linear , spherically symmetric , perturbations of the gauge field and metric @xcite . as a consequence of this ,
bizon postulated a modified `` no - hair '' conjecture @xcite , heuristically stated as `` a stable black hole is uniquely characterized by global charges '' .
the `` coloured '' black holes do not contradict this conjecture due to their instability .
notwithstanding the instability of the @xmath2 eym solitons and black holes in asymptotically flat space - time , their discovery sparked extensive research into soliton and black hole solutions of more general matter models involving nonabelian gauge fields ( see @xcite for a review ) . in four - dimensional asymptotically flat space - time , enlarging the gauge group to @xmath0 , a richer phase space arises @xcite , but solutions still occur at discrete points in the phase space .
furthermore , there is a very general result that all purely magnetic , asymptotically flat , soliton or black hole solutions of the @xmath0 eym equations in four - dimensional space - time are unstable @xcite .
the situation is very different if one considers eym with a negative cosmological constant @xmath7 , so that the space - time is asymptotically anti - de sitter ( ads ) rather than asymptotically flat .
four - dimensional , purely magnetic , black hole @xcite and soliton @xcite solutions of the @xmath2 eym equations in ads are found in continuous regions of parameter space rather than at discrete points as in the asymptotically flat case .
furthermore , if @xmath8 is sufficiently large , there are solutions for which the single magnetic gauge field function @xmath4 has no zeros .
these nodeless solutions have no analogue in asymptotically flat space - time and at least some of them are stable under linear , spherically symmetric perturbations @xcite . the proof of stability
can be extended to general linear perturbations of the metric and gauge field providing @xmath9 is sufficiently large @xcite . for the larger @xmath0 gauge group ,
purely magnetic soliton and black hole solutions of eym in ads exist in continuous regions of the parameter space @xcite . in this case
the purely magnetic gauge field is described by @xmath1 functions @xmath10 , @xmath11 , of the radial coordinate @xmath5 . there
exist both soliton and black hole solutions for which all the @xmath10 have no zeros provided @xmath9 is sufficiently large @xcite .
these solutions are of particular interest because it can be proven that at least some of them are stable under linear , spherically symmetric , perturbations of the metric and gauge field @xcite .
although these stable black holes are coupled to potentially unlimited amounts of gauge field hair , they can nonetheless be characterized by global charges defined far from the black hole @xcite , in accordance with the modified `` no - hair '' conjecture @xcite ( see also the reviews @xcite ) .
returning to @xmath2 eym in ads , the space of solutions revealed another surprise @xcite . in asymptotically flat space - time , all nontrivial ( that is , not corresponding to embedded schwarzschild or abelian reissner - nordstrm configurations ) @xmath2 solitons and black holes must have gauge fields with vanishing electric parts @xcite .
this is not the case in asymptotically ads space - time , with dyonic solitons and black holes existing , for which the gauge field has nontrivial electric and magnetic components @xcite . for @xmath2 gauge group ,
the electric part is described by a single function @xmath12 and the magnetic part by a single function @xmath4 , where @xmath5 is the radial coordinate .
the electric gauge field function @xmath12 is monotonic and has no zeros ; there exist solitons and black holes for which the magnetic gauge field function @xmath4 also has no zeros @xcite .
although these solutions were found numerically fifteen years ago @xcite , it was only recently that it was shown that at least some of the solutions for which @xmath4 is nodeless are stable under linear , spherically symmetric perturbations @xcite . in this paper
we explore the phase space of dyonic eym solitons and black holes in ads when the gauge group is enlarged to @xmath0 . in the purely magnetic case , with the larger gauge group
the phase space has a very rich structure @xcite and we anticipate that the same will be true for dyonic solutions .
dyonic solutions of @xmath0 eym will be described by @xmath1 electric gauge field functions @xmath13 , @xmath14 , and @xmath1 magnetic gauge field functions @xmath10 , @xmath14 .
recently the existence of soliton and black hole solutions ( when @xmath9 is sufficiently large ) for which all the @xmath15 functions have no zeros was proven @xcite .
these nodeless solutions are of particular interest because , in the light of results in the purely magnetic sector @xcite and for @xmath2 dyonic solutions @xcite , it is anticipated that at least some of them will be stable under linear , spherically symmetric perturbations . in this paper
we present the first numerical dyonic solutions of the @xmath0 eym equations in ads , focusing not just on the nodeless solutions whose existence is proven in @xcite , but also more generally on the rich features of the phase space of solutions .
the outline of this paper is as follows . in sec .
[ sec : equations ] we introduce the @xmath0 eym model and field equations , and describe our static , spherically symmetric ansatz for the metric and gauge field potential .
we also discuss the boundary conditions at the origin ( for regular solitons ) , event horizon ( for black holes ) and at infinity .
solutions of the field equations are discussed in detail in sec . [
sec : solutions ] .
we describe our numerical method , and explore the phase space of dyonic soliton and black hole solutions with @xmath2 and @xmath3 gauge group .
our conclusions are presented in sec .
[ sec : conc ] .
in this section we describe our gauge field and metric ansatz , the field equations and the boundary conditions for static , spherically symmetric dyon and dyonic black hole solutions of @xmath0 eym in ads .
we consider static , spherically symmetric , four - dimensional solitons and black holes with metric @xmath16 where the metric functions @xmath17 and @xmath18 depend only on the radial coordinate @xmath5 .
we write the metric function @xmath19 in the form @xmath20 where @xmath7 is the cosmological constant . in ( [ eq : metric ] ) and throughout this paper , the metric has signature @xmath21 and we use units in which @xmath22 . for the static , spherically symmetric @xmath0 gauge field potential @xmath23
, we use the ansatz @xcite @xmath24 d\phi , \label{eq : gaugepot}\end{aligned}\ ] ] where we have set the gauge coupling @xmath25 . in ( [ eq : gaugepot ] ) , @xmath26 , @xmath27 , @xmath28 and @xmath29 are @xmath30 matrices which depend only on the radial coordinate @xmath5 .
the matrices @xmath26 and @xmath27 are purely imaginary , diagonal and traceless .
the matrix @xmath28 ( with hermitian conjugate @xmath31 ) is upper triangular , with the only nonzero entries immediately above the diagonal , which are written as : @xmath32 where @xmath10 and @xmath33 are real functions .
the matrix @xmath29 is constant , diagonal and traceless , and given by @xmath34 the ansatz ( [ eq : gaugepot ] ) has some residual gauge freedom which can be used to set @xmath35 @xcite .
it should be emphasized that the choice of ansatz ( [ eq : gaugepot ] ) is not unique @xcite . for purely magnetic solutions @xcite one
makes the choice @xmath36 .
however , in this paper we are interested in solutions for which the gauge field has nontrivial electric and magnetic parts
. therefore @xmath26 will not vanish .
we write @xmath26 in terms of matrices @xmath37 , which are the diagonal generators of the cartan subalgebra of @xmath0 .
we define the @xmath37 in a similar way to ref .
@xcite , but using a slightly different normalization .
the nonzero entries of the matrix @xmath37 are : @xmath38 _ { j , k } = - \frac { i}{{\sqrt { 2\ell \left ( \ell + 1 \right ) } } } \left [ \sum _ { p=1}^{\ell } \left ( \delta _ { j , p } \delta _ { k , p } \right ) - \ell \delta _ { j , \ell+1 } \delta _ { k,\ell+1 } \right ] , \label{eq : hdef}\ ] ] where @xmath39 is the usual kronecker delta . for @xmath2
eym , there is a single generator of the cartan subalgebra : @xmath40 while for @xmath3 eym , there are two @xmath41 matrices : @xmath42 for general @xmath43 , there are @xmath1 matrices @xmath37 ( [ eq : hdef ] ) , since @xmath1 is the rank of the @xmath0 lie algebra . in terms of the @xmath37 matrices ,
the electric part of the gauge potential @xmath23 ( [ eq : gaugepot ] ) takes the form @xmath44 in terms of @xmath1 real scalar functions @xmath45 , depending on the radial coordinate @xmath5 only .
our ansatz ( [ eq : adef ] ) for the electric part of the gauge potential at first sight looks rather different from that conventionally used in the literature @xcite , although it is equivalent .
usually the @xmath43 diagonal entries of the traceless matrix @xmath26 are denoted by @xmath46 , @xmath47 ( where the @xmath48 are real functions of @xmath5 ) , and written in terms of @xmath1 real quantities @xmath49 as follows @xcite : @xmath50 so that @xmath51 from ( [ eq : adef ] ) , the @xmath48 can be written in terms of the @xmath52 as : @xmath53 and consequently @xmath54 we write the relation ( [ eq : caledef ] ) between the @xmath1 functions @xmath52 and the @xmath1 quantities @xmath55 as @xmath56 where @xmath57 , @xmath58 and @xmath59 is the lower - triangular @xmath60 matrix with entries @xmath61 by inverting @xmath59 ( [ eq : calfdef ] ) , we can write the @xmath52 in terms of the @xmath55 : @xmath62 for the magnetic part of the gauge field potential ( [ eq : gaugepot ] ) , we now assume that all the functions @xmath10 ( [ eq : cj ] ) in the matrix @xmath28 are nonzero . in this case one of the yang - mills equations reduces to @xmath63 for all @xmath64 and @xmath5 @xcite .
our ansatz for the gauge potential ( [ eq : gaugepot ] ) then reduces to : @xmath65 d\phi , \label{eq : gaugepotsimp}\end{aligned}\ ] ] and the nonzero entries in the matrix @xmath28 are now simply @xmath66 the above ansatze for the matrices @xmath26 and @xmath28 appearing in the electric and magnetic parts of the gauge field respectively , are related to the expansion of the @xmath0 gauge field in terms of simple roots ( such an expansion is well - known in the construction of the @xmath0 monopole , see , for example , @xcite for a review ) .
the matrix @xmath26 ( [ eq : adef ] ) is a linear combination of the generators @xmath67 , where the @xmath68 are the positive roots of @xmath0 , while the matrix @xmath28 with entries given by ( [ eq : cjsimp ] ) is a linear combination of the generators @xmath69 ( the raising operators ) corresponding to the simple roots of @xmath0 . in summary ,
the gauge field is described by @xmath70 functions of @xmath5 : the @xmath1 electric gauge functions @xmath45 and the @xmath1 magnetic gauge functions @xmath10 .
the components of the yang - mills gauge field are given in terms of the gauge potential components ( [ eq : gaugepotsimp ] ) as follows : @xmath71 , \label{eq : gaugefield}\ ] ] when the gauge coupling constant @xmath25 .
the yang - mills equations take the form @xmath72 = 0 .
\label{eq : yme}\ ] ] the stress - energy tensor of the yang - mills field is @xmath73 , \label{eq : stress}\ ] ] where @xmath74 denotes a lie algebra trace .
the stress - energy tensor @xmath75 acts as the source in the einstein equations : @xmath76 since we are using units in which @xmath77 . the yang - mills equations ( [ eq : yme ] ) for the gauge field with potential ( [ eq : gaugepotsimp ] ) take the form ( a prime @xmath78 denotes differentiation with respect to
the radial coordinate @xmath5 ) : [ eq : ymefinal ] @xmath79 for @xmath80 .
the einstein equations ( [ eq : ee ] ) give two equations for the metric functions @xmath81 and @xmath82 : [ eq : eefinal ] @xmath83 + \sum_{k=1}^{n-1 } \left [ \mu \omega _ { k}'^{2 } + \frac { k(k+1)}{4r^{2 } } \left ( 1 - \frac { \omega_{k}^{2}}{k } + \frac { \omega _ { k+1}^{2}}{k+1 } \right ) ^{2 } \right ] , \label{eq : mprime } \\
\sigma ' & = & \sum_{k=1}^{n-1 } \left [ \frac { 2\sigma \omega _ { k}'^{2}}{r } + \frac { 2\omega _ { k}^{2}}{\sigma \mu ^{2 } r } \left ( h_{k } { \sqrt { \frac { k+1}{2k } } } - h_{k-1 } { \sqrt { \frac { k-1}{2k } } } \right ) ^{2 } \right ] .
\label{eq : sigmaprime}\end{aligned}\ ] ] setting @xmath84 for all @xmath85 , the field equations ( [ eq : ymefinal][eq : eefinal ] ) reduce to the purely magnetic field equations studied in @xcite .
the field equations ( [ eq : ymefinal][eq : eefinal ] ) are singular at the origin @xmath86 ( relevant for soliton solutions ) , the black hole event horizon @xmath87 ( if there is one ) and infinity @xmath88 .
we therefore need to specify appropriate boundary conditions at each of these points . as with the purely magnetic solutions @xcite the boundary conditions at the origin are the most complex
. local existence of solutions of the field equations near the singular points @xmath86 , @xmath87 and @xmath88 is proven in @xcite , using a different representation of the electric part of the gauge field potential . in this section
we cast the results of @xcite into our notation for completeness .
we assume that the space - time is asymptotically ads and that the field variables have regular taylor series expansions about @xmath88 : @xmath89 substituting the expressions ( [ eq : infinity ] ) into the field equations ( [ eq : ymefinal][eq : eefinal ] ) , we find that @xmath90 , @xmath91 , @xmath92 and @xmath93 are free parameters , @xmath94 and that @xmath95 and @xmath96 are given by ( cf .
@xcite ) @xmath97 , \nonumber \\
\sigma _ { 4 } & = & - \frac { 1}{2 } \sum_{k=1}^{n-1 } \left [ l^{4 } \omega _ { k,\infty } ^{2 } \left ( h_{k,\infty } { \sqrt { \frac { k+1}{2k } } } - h_{k-1,\infty } { \sqrt { \frac { k-1}{2k } } } \right ) ^{2 } + c_{k,1}^{2 } \right ] , \label{eq : m1sigma4infinity}\end{aligned}\ ] ] where @xmath98 is the ads radius of curvature .
we assume that there is a regular , nonextremal black hole event horizon at @xmath87 . at @xmath87 , the metric function @xmath19 ( [ eq : mu ] )
has a single zero , so that @xmath99 to avoid a singularity at the event horizon , it must be the case that the electric gauge functions @xmath100 vanish at @xmath87 .
we therefore assume the following taylor series expansions near the event horizon : @xmath101 from the field equations , we find that @xmath102 , @xmath103 and @xmath104 are free parameters and that @xmath105 , @xmath106 and @xmath107 are given in terms of them as follows ( cf .
@xcite ) : @xmath108 , \nonumber \\
\sigma ' ( r_{h } ) & = & 2\sum _ { k=1}^{n-1 } \left [ \frac { \omega _ { k}(r_{h})^{2}}{\sigma ( r_{h } ) \mu ' ( r_{h})^{2 } r_{h } } \left ( h_{k}'(r_{h } ) { \sqrt { \frac { k+1}{2k } } } - h_{k-1}'(r_{h } ) { \sqrt { \frac { k-1}{2k } } } \right ) ^{2 } \right ] + 2\sum_{k=1}^{n-1 } \frac { \sigma ( r_{h } ) \omega _ { k}'(r_{h})^{2}}{r_{h } } , \nonumber \\ \omega _ { k}'(r_{h } ) & = & \frac { \omega _ { k}(r_{h})}{\mu ' ( r_{h } ) r_{h}^{2 } } \left ( \omega _ { k}(r_{h})^{2 } - 1 - \frac{1}{2 } \left [ \omega _ { k-1}(r_{h})^{2 } + \omega _ { k+1}(r_{h})^{2 } \right ] \right ) .
\label{eq : horexp}\end{aligned}\ ] ] the value of @xmath102 is fixed in practice by the requirement that the metric function @xmath82 approaches unity as @xmath88 .
this leaves the @xmath109 free parameters @xmath103 and @xmath104 for @xmath80 , whose values are restricted by the requirement @xmath110 which is needed for a regular nonextremal horizon at @xmath87 . in the purely magnetic case @xcite ,
the boundary conditions for the @xmath0 gauge potential near the origin are rather complicated , with a power series expansion up to @xmath111 necessary in order to completely specify the gauge field functions .
it is no surprise that the addition of a nontrivial electric part to the gauge field potential only complicates matters further @xcite .
we begin by assuming regular taylor series expansions for all field variables in a neighbourhood of the origin @xmath86 : @xmath112 the constant @xmath113 must be nonzero in order for the metric ( [ eq : metric ] ) to be regular at the origin .
it is otherwise arbitrary as far as the expansions near the origin are concerned , and will be fixed in practice by the requirement that @xmath114 as @xmath88 .
regularity of the field equations ( [ eq : ymefinal][eq : eefinal ] ) , metric ( [ eq : metric ] ) and curvature at @xmath86 gives @xcite @xmath115 from the equation for @xmath116 ( [ eq : mprime ] ) , we also have @xmath117 ^{2 } = 0 , \label{eq : originconds2}\ ] ] which is solved by @xmath118 the field equations ( [ eq : ymefinal][eq : eefinal ] ) are unchanged by the transformation @xmath119 , for each @xmath85 separately .
therefore we take the positive sign in ( [ eq : omegaorigin ] ) without loss of generality .
we consider next the magnetic gauge field functions @xmath120 .
the coupling between these and the electric gauge field functions @xmath100 in the yang - mills equation for @xmath120 ( [ eq : omegakeqn ] ) does not affect the first two terms in the expansion of this equation near @xmath86 , because @xmath121 as @xmath122 .
following @xcite , we define two vectors @xmath123 and @xmath124 .
the first two terms in the expansion about @xmath86 of the yang - mills equation ( [ eq : omegakeqn ] ) then give the following equations for @xmath125 and @xmath126 : @xmath127 where @xmath128 is the @xmath129 matrix @xcite @xmath130 as discussed in @xcite , the eigenvalues of the matrix @xmath128 are @xmath131 and the eigenvectors have a complicated form involving hahn polynomials @xcite . writing a basis of normalized eigenvectors of @xmath128 as @xmath132 ( where @xmath133 is the eigenvector having eigenvalue @xmath134 ( [ eq : eigenvalues ] ) ) , we have , from ( [ eq : omega23 ] ) , @xmath135 where @xmath136 and @xmath137 are arbitrary constants .
we proceed in a similar way for the electric gauge field functions @xmath100 , defining two vectors @xmath138 and @xmath139 . as with the magnetic gauge field functions , the coupling between the @xmath100 and @xmath120 in the yang - mills equation ( [ eq : hkeqn ] )
does not affect the first two terms in the expansion of this equation about @xmath86 , which then give the following two equations : @xmath140 where @xmath141 is the @xmath129 matrix @xmath142 in @xcite , an expansion similar to ( [ eq : origin ] ) is performed for the alternative electric gauge field functions @xmath143 , defined in terms of the @xmath144 functions that we use here by ( [ eq : caledef ] , [ eq : caleh ] ) : @xmath145 again we have @xmath146 for all @xmath85 , and the vectors @xmath147 and @xmath148 satisfy the equations @xcite @xmath149 using the relationship between @xmath150 and @xmath151 in ( [ eq : caleh ] ) , it is clear that the matrices @xmath128 ( [ eq : calm ] ) and @xmath141 ( [ eq : caln ] ) are related by @xmath152 and therefore the matrices @xmath128 and @xmath141 have the same eigenvalues @xmath134 ( [ eq : eigenvalues ] ) .
in particular , we have , from ( [ eq : h23 ] ) , @xmath153 where @xmath154 and @xmath155 are arbitrary constants and the vectors @xmath156 , @xmath157 , are eigenvectors of @xmath141 , given by @xmath158 where an overall multiplicative constant is required so that the @xmath156 are normalized .
once we have found the eigenvectors @xmath125 , @xmath126 , @xmath159 and @xmath160 , the first two terms in the einstein equations ( [ eq : mprime ] , [ eq : sigmaprime ] ) give the values of @xmath161 and @xmath162 ( which also depend on the cosmological constant @xmath163 ) @xcite . from the above analysis
, the expansion of the @xmath100 to order @xmath5 and @xmath120 functions to order @xmath164 depends on just two arbitrary constants , @xmath136 ( [ eq : omega23sol ] ) and @xmath154 ( [ eq : h23sol ] ) , while the expansion to next order in @xmath5 ( that is , to order @xmath164 for the electric gauge field functions @xmath100 and order @xmath165 for the magnetic gauge field functions @xmath120 ) adds a further two arbitrary constants , @xmath137 and @xmath155 .
it is shown in proposition 8 in @xcite , in analogy with the purely magnetic case @xcite , that a total of @xmath166 arbitrary constants are required to completely specify the gauge field functions in a neighbourhood of the origin .
each additional power of @xmath5 in the expansion of the @xmath100 and @xmath120 depends on just two further arbitrary constants , one for the @xmath100 and one for the @xmath120 . to see this ,
define the vectors @xmath167 examination of the appropriate term in the expansion of the relevant yang - mills equation ( [ eq : ymefinal ] ) shows that the vectors @xmath168 , @xmath169 satisfy equations of the form @xmath170 { { \mbox{{\boldmath{{$\omega $ } } } } } } _ { j+1 } & = & { { \mbox{{\boldmath{{$p$}}}}}}_{j+1 } , \nonumber \\ \left
[ { \mathcal { n}}_{n-1 } - j \left ( j + 1 \right ) \right ] { { \mbox{{\boldmath{{$h$}}}}}}_{j } & = & { { \mbox{{\boldmath{{$q$}}}}}}_{j+1 } , \label{eq : higherorigin}\end{aligned}\ ] ] where @xmath171 and @xmath172 are complicated vectors depending on @xmath173 , @xmath174 , @xmath175 , @xmath176 , whose detailed form can be found in @xcite .
the solutions of ( [ eq : higherorigin ] ) are @xcite : @xmath177 where @xmath178 and @xmath179 are arbitrary constants . in ( [ eq : highersol ] ) ,
the vectors @xmath180 and @xmath181 are particular solutions of ( [ eq : higherorigin ] ) .
it is shown in @xcite that these particular solutions can be chosen such that @xmath180 and @xmath181 are linear combinations of @xmath182 .
we can therefore write the electric and magnetic gauge field functions in the following vectorial form , where @xmath183 and @xmath184 : @xmath185 where @xmath186 and the @xmath187 and @xmath188 functions have the following behaviour near the origin : @xmath189 as is explained in more detail in the next section , in our numerical procedure we integrate the field equations for the @xmath187 and @xmath188 functions rather than @xmath100 and @xmath120 , to improve numerical accuracy .
in this paper we present dyonic solutions for the @xmath190 and @xmath191 cases only , so it is sufficient for our purposes to find @xmath160 , @xmath192 , @xmath125 and @xmath126 .
we do not need to consider the complicated vectors @xmath171 , @xmath172 ( [ eq : higherorigin ] ) or @xmath193 , @xmath194 ( [ eq : highersol ] ) .
we now present our new soliton and black hole solutions of the field equations ( [ eq : ymefinal][eq : eefinal ] ) .
these field equations have a number of trivial solutions , which we describe first in sec .
[ sec : trivial ] , before discussing our numerical method in sec .
[ sec : numerics ] and the nontrivial solutions for @xmath2 and @xmath3 gauge groups in secs .
[ sec : su2 ] and [ sec : su3 ] respectively .
the first trivial solution arises on setting the electric gauge functions @xmath195 for @xmath80 , and the magnetic gauge functions @xmath120 to be the following constants : @xmath196 the metric functions @xmath81 and @xmath82 are then constants .
setting @xmath197 without loss of generality gives the schwarzschild - ads metric with @xmath198 .
the second trivial solution is reissner - nordstrm - ads .
the metric function @xmath197 and @xmath19 takes the form @xmath199 where the mass @xmath200 and charge @xmath201 are constants .
this solution of the field equations arises on setting the magnetic gauge functions @xmath202 for all @xmath80 , in which case the electric gauge functions are exactly @xmath203 where the @xmath90 and @xmath91 are constants . from the einstein equation ( [ eq : mprime ] )
, the charge @xmath201 is given by @xmath204 the charge @xmath201 ( [ eq : rncharge ] ) is an effective charge , with @xmath205 having two components .
the first , @xmath206 , is a magnetic charge , and the second , @xmath207 , is an electric charge .
the reissner - nordstrm - ads solution is therefore dyonic in this case . setting all the @xmath208 yields the purely magnetically charged reissner - nordstrm - ads solution .
note that purely electrically charged reissner - nordstrm - ads is _ not _ a solution of the field equations ( [ eq : ymefinal][eq : eefinal ] ) due to the coupling between the electric gauge field functions @xmath100 and the magnetic gauge field functions @xmath120 .
the third class of trivial solutions is @xmath2 embedded solutions , obtained in @xcite with an alternative parametrization of the electric part of the gauge field potential .
we start by writing the @xmath1 magnetic gauge functions @xmath120 in terms of a single function @xmath209 , and the @xmath1 electric gauge functions @xmath100 in terms of a single function @xmath210 , as follows : @xmath211 where @xmath212 and @xmath213 are constants . the yang - mills equations ( [ eq : ymefinal ] ) reduce to those for the @xmath2 case with gauge functions @xmath210 and @xmath209 if the following conditions hold : @xmath214 substituting ( [ eq : su2gauge ] ) into the einstein equations ( [ eq : eefinal ] )
, we obtain the @xmath2 equations if @xmath215 and @xmath216 the conditions ( [ eq : ymsu2conds][eq : esu2conds2 ] ) are solved by taking @xmath217 the values of @xmath212 are the same as those used to embed purely magnetic @xmath2 solutions into @xmath0 eym @xcite . substituting ( [ eq : su2gauge ] , [ eq : absu2 ] ) into the field equations ( [ eq : ymefinal][eq : eefinal ] ) and defining new rescaled variables as follows @xcite : @xmath218 where @xmath219 gives the @xmath2 field equations @xmath220 \right ) , \end{aligned}\ ] ]
where the magnetic gauge field function @xmath209 and the metric functions @xmath19 and @xmath82 are not scaled . setting @xmath221 gives , as expected , the purely magnetic embedded @xmath2 equations .
the field equations ( [ eq : ymefinal][eq : eefinal ] ) form a set of @xmath222 ordinary differential equations .
note that , unlike the purely magnetic case @xcite , here the equation for the metric function @xmath18 ( [ eq : sigmaprime ] ) does not decouple from the other equations .
this does not complicate the numerical method significantly .
we employ standard `` shooting '' techniques @xcite , using a bulirsch - stoer algorithm in c++ to integrate the ordinary differential equations .
the field equations are singular at the origin or a black hole event horizon . for black hole solutions ,
we start our integration just outside the event horizon , at @xmath223 , using the expansions ( [ eq : horizon ] ) as initial conditions .
we then integrate outwards with @xmath5 increasing until the solution either becomes singular or the field variables @xmath224 and @xmath82 have converged to constant asymptotic values to within a suitable numerical tolerance . for soliton solutions ,
we start our integration close to the origin . the need to include higher order terms in the expansions of the gauge field functions @xmath100 and
@xmath120 means that these functions are not suitable for numerical integration . with limited numerical precision
, we can not keep adding powers of @xmath5 in our initial conditions ( the expansions ( [ eq : origin ] ) ) without losing accuracy . for each @xmath43
we therefore first make a change of variables , writing the electric gauge functions @xmath100 in terms of new variables @xmath33 , @xmath14 , and the magnetic gauge functions @xmath120 in terms of new variables @xmath225 , @xmath14 ( [ eq : betagamma ] ) , where the @xmath225 and @xmath33 are chosen so that their expansions near the origin have the form ( [ eq : betagammaorigin ] ) . in secs .
[ sec : su2dyons ] and [ sec : su3dyons ] the details of this change of variables will be presented for the @xmath190 and @xmath191 cases respectively . in the following sections , we present examples of numerical solutions of the field equations ( [ eq : ymefinal][eq : eefinal ] ) representing both dyons and dyonic black holes , for @xmath190 and @xmath191 . following @xcite , we also study the structure of the space of solutions by examining the phase space of parameters characterizing the solutions near the event horizon or origin , as applicable . in the @xmath2 case ,
it is straightforward to show that the single electric gauge field function @xmath12 has no zeros , as follows .
the equation for @xmath12 takes the form ( [ eq : hkeqn ] ) @xmath226 if the function @xmath12 has a turning point at @xmath227 , then @xmath228 and ( [ eq : h1su2 ] ) gives @xmath229 since the metric function @xmath19 is strictly positive for all @xmath230 for soliton solutions and for all @xmath231 for black hole solutions , we conclude from ( [ eq : h1ppsu2r0 ] ) that the turning point is a minimum if @xmath232 and a maximum if @xmath233 .
from the expansions ( [ eq : origin ] ) , noting that @xmath234 , we see that @xmath235 has the same sign as @xmath12 very close to the origin .
similarly , near a black hole event horizon , @xmath235 and @xmath12 have the same sign from ( [ eq : horizon ] ) .
we deduce that it is not possible for @xmath12 to have a turning point , and therefore it is monotonic and has no zeros . for @xmath0 ,
it is proven in @xcite that the electric gauge field functions @xmath55 , defined in terms of the @xmath100 by ( [ eq : caledef ] ) , are monotonic and have no zeros . while it is not necessarily the case that our alternative electric gauge field functions @xmath100 have no zeros , since all the @xmath55 are nodeless any zeros of @xmath100 are a quirk of our parametrization of the electric part of the gauge potential , rather than revealing any underlying structure of the space of solutions .
we therefore divide our numerical solutions into classes depending on the numbers of zeros @xmath236 of the magnetic gauge field functions @xmath120 respectively .
we use coloured plots to show this phase space structure , but hope that the key features will still be apparent to readers using black and white .
dyons and dyonic black holes in @xmath2 eym in ads were first found by bjoraker and hosotani @xcite . in this section
we study the phase space of @xmath2 solutions , checking that we reproduce the results of @xcite and exploring in more detail those key features which will extend to the larger @xmath43 case .
four parameters are required to describe the @xmath2 black hole solutions : @xmath237 , @xmath163 , @xmath238 and @xmath239 ( [ eq : horizon ] , [ eq : horexp ] ) .
we fix @xmath240 .
eym in ads .
the parameters are : @xmath240 , @xmath241 , @xmath242 and @xmath243 .
the electric gauge field function @xmath12 is monotonic and nodeless ; the magnetic gauge field function @xmath4 has a single zero.,width=321 ] a typical black hole solution with @xmath241 is shown in fig .
as anticipated , the electric gauge field function @xmath12 is monotonic and has no zeros . in fig .
[ fig1 ] , we have chosen the initial values @xmath242 and @xmath244 . for these initial values
we see that the magnetic gauge field function @xmath4 has a single zero .
we now study the phase space by fixing @xmath163 and varying @xmath238 and @xmath239 .
setting @xmath245 gives purely magnetic solutions .
the field equations ( [ eq : ymefinal][eq : eefinal ] ) are invariant under the separate transformations @xmath246 and @xmath247 .
therefore it suffices to consider @xmath248 and @xmath249 .
the values of @xmath238 and @xmath239 are not completely free : the constraint ( [ eq : murh ] ) must be satisfied . as in the purely magnetic case @xcite , for each value of @xmath163 studied we find a region in the @xmath250-plane for which ( [ eq : murh ] ) is satisfied , but we are unable to find regular black hole solutions which converge as @xmath88 .
eym , with @xmath240 and @xmath241 .
all shaded points in the plot correspond to black hole solutions .
the solutions are indexed by @xmath251 , the number of zeros of the magnetic gauge field function @xmath4 .
the different values of @xmath251 are indicated by colour - coding the regions - in black and white the different colours are different shades of grey .
solutions with the largest values of @xmath251 are found towards the right - hand - side of the coloured region.,width=321 ] solutions from fig .
all shaded points in the plot correspond to black hole solutions .
the @xmath252 region ( the small red or dark grey region near the point ( 0,1 ) ) is in agreement with ref .
@xcite.,width=321 ] in fig .
[ fig2 ] we show the phase space for black holes with @xmath240 and @xmath241 , part of which has previously been shown in @xcite .
all points in the plot in fig .
[ fig2 ] represent black hole solutions with particular values of @xmath238 and @xmath239 . in fig .
[ fig2 ] we find a richly structured parameter space , with solutions for which @xmath4 has up to 17 nodes . the number of zeros of @xmath4 increases as @xmath238 increases for each fixed value of @xmath239 .
the corresponding plot in @xcite focused on the small region near @xmath245 , @xmath253 for which there are nodeless @xmath252 solutions , so for comparison with @xcite , in fig .
[ fig3 ] we show a close - up of the parameter space near the @xmath252 region , which is in agreement with @xcite .
eym , with @xmath240 and @xmath254 .
all black hole solutions that we find have @xmath252 .
the region labeled `` no solution '' ( the red / darker grey region containing larger values of @xmath238 and @xmath239 ) is the region where the constraint ( [ eq : murh ] ) is satisfied , giving a regular event horizon , but we have not found any regular black hole solutions.,width=321 ] for purely magnetic solutions , increasing the magnitude of the negative cosmological constant @xmath9 increased the size of the region of phase space where solutions were found @xcite .
the @xmath252 region of nodeless solutions also expanded as a proportion of the total solution space @xcite .
we find the same effects for dyonic black hole solutions . to illustrate this , in fig .
[ fig4 ] we show the phase space of solutions for @xmath254 and @xmath240 . in this case
the only solutions we find are such that the magnetic gauge field function @xmath4 has no zeros . in fig .
[ fig4 ] , we have also shown the region of the @xmath250-plane where the constraint ( [ eq : murh ] ) is satisfied but we have not been able to find regular solutions .
this region is marked `` no solution '' in fig .
[ fig4 ] .
eym with @xmath240 and @xmath255 .
the solutions are indexed by @xmath251 , the number of zeros of the magnetic gauge function @xmath4 .
regions with different colours ( or shades of grey ) correspond to different values of @xmath251 .
we find solutions with very large values of @xmath251 as @xmath9 decreases .
the blue / darkest region ( second from the right ) corresponds to @xmath256 black hole solutions , and extends to @xmath257 , with @xmath258 in this limit.,width=321 ] eym with @xmath240 and @xmath259 .
the solutions are indexed by @xmath251 , the number of zeros of the magnetic gauge function @xmath4 .
regions with different colours ( or shades of grey ) correspond to different values of @xmath251 .
we find solutions with very large values of @xmath251 as @xmath9 decreases .
however , the phase space of solutions does not extend to @xmath257 as there is no solution in this limit for this value of @xmath239.,width=321 ] bjoraker and hosotani @xcite found a rich , `` fractal''-like , structure in the phase space of solutions as @xmath257 ( see also @xcite for similar behaviour for solitons in the purely magnetic case ) . in asymptotically flat space ,
there are no dyonic black hole solutions of @xmath2 eym @xcite
. however , there are purely magnetic solutions found by setting the electric part of the gauge potential equal to zero @xcite . in figs .
[ fig5 ] and [ fig6 ] we investigate the structure of the phase space as @xmath257 .
we fix @xmath239 and vary @xmath238 and @xmath163 . in fig .
[ fig5 ] we set @xmath260 , which is the value for the first `` coloured '' black hole solution which exists in the limit @xmath257 @xcite . in fig .
[ fig6 ] we set @xmath259 , which does not correspond to a regular black hole solution in the limit @xmath257 . in both figs .
[ fig5 ] and [ fig6 ] , we have marked `` no solution '' the region where the constraint ( [ eq : murh ] ) for a regular event horizon is satisfied , but our numerical solution becomes singular before @xmath88 . for @xmath9 sufficiently large , for both values of @xmath239 the solutions are such that @xmath4 is nodeless . as @xmath9 decreases , the number of zeros of @xmath4 increases rapidly .
we find solutions for which @xmath4 has more than fifty zeros .
the phase spaces shown in figs . [ fig5 ] and [ fig6 ] are subtly different when @xmath9 is very small . when @xmath260 ( fig .
[ fig5 ] ) , the @xmath256 part of the phase space ( the second region from the right ) extends to @xmath257 as we have the first `` coloured '' black hole solution in this limit @xcite .
however , for @xmath259 , there is no solution in the limit @xmath257 and so the phase space ends at a small but nonzero value of @xmath9 .
we are unable to find solutions for @xmath259 and @xmath9 smaller than about @xmath261 . as well as the cosmological constant @xmath163 ,
dyonic solitons in @xmath2 eym are parameterized by the quantities @xmath262 and @xmath263 , so that the expansions ( [ eq : origin ] ) of the gauge field functions take the form @xmath264 unlike the black hole case , where the constraint ( [ eq : murh ] ) for a regular nonextremal event horizon restricts the values of the parameters describing the solutions , for solitons there are no _ a priori _ constraints on the values that @xmath262 and @xmath263 can take . in particular
, @xmath262 can take either positive or negative values .
however , since the field equations are invariant under the transformation @xmath265 , we can restrict attention to @xmath266 without loss of generality . to avoid numerical errors in terms in the field equations ( [ eq : ymefinal][eq : eefinal ] ) ,
we define a new variable @xmath267 by @xmath268 which satisfies the first order differential equation @xmath269 and add this differential equation to those ( [ eq : ymefinal][eq : eefinal ] ) to be integrated numerically . near the origin ,
@xmath270 in the @xmath2 case , the relation ( [ eq : betagamma ] ) for the gauge field functions takes the form @xmath271 therefore our new variable @xmath267 ( [ eq : psidef ] ) is related to @xmath272 by @xmath273 and , using ( [ eq : betagammaorigin ] ) , @xmath274 eym in ads .
the parameters are : @xmath241 , @xmath275 , @xmath276 .
the electric gauge field function @xmath12 is monotonic and nodeless ; the magnetic gauge field function @xmath4 has a single zero.,width=321 ] a typical soliton solution with @xmath241 is shown in fig .
[ fig7 ] , where the parameters are @xmath277 and @xmath276 .
for these initial values we see that the magnetic gauge field function @xmath4 has a single zero . as expected , the electric gauge field function @xmath12 is monotonic and has no zeros .
eym , with @xmath241 .
all shaded points in the plot correspond to soliton solutions .
the solutions are indexed by @xmath251 , the number of zeros of the magnetic gauge field function @xmath4 .
the different values of @xmath251 are indicated by colour - coding the regions - in black and white the different colours are different shades of grey .
solutions with the largest values of @xmath251 are found towards the right - hand - side of the coloured region.,width=321 ] solutions from fig .
the @xmath252 region ( the green or light grey region containing the point ( 0,0 ) ) is in agreement with ref .
the region labelled `` no solution '' ( the red / darker grey region containing larger values of @xmath262 for smaller values of @xmath263 ) is the region where we have not found any regular soliton solutions.,width=321 ] the entire phase space of soliton solutions for @xmath241 is shown in fig .
all points in the plot in fig .
[ fig8 ] represent dyonic soliton solutions with particular values of @xmath263 and @xmath262 . in fig .
[ fig8 ] , as in the black hole case ( fig .
[ fig2 ] ) , the phase space is very complicated , and we find solutions for which @xmath4 has up to 17 nodes .
the corresponding plot in @xcite focused on the small region near @xmath278 , @xmath279 for which there are nodeless @xmath252 solutions . for comparison , in fig .
[ fig9 ] we show a close - up of the parameter space near the @xmath252 region , which is in agreement with the corresponding plot in @xcite .
eym , with @xmath254 .
the different values of @xmath251 are indicated by colour - coding the regions - in black and white the different colours are different shades of grey .
the region labelled `` no solution '' ( the red / mid grey region containing larger values of @xmath262 and @xmath263 ) is the region where we have not found any regular soliton solutions .
there are solutions for which the number of nodes @xmath251 of the magnetic gauge field function @xmath4 is either zero or one.,width=321 ] as @xmath9 increases , the phase space of dyonic soliton solutions simplifies , in the same way as we observed for the black hole solutions .
this can be seen in fig .
[ fig10 ] , where we plot the phase space of solutions for @xmath254 .
we find solutions where the magnetic gauge field function @xmath4 has either zero nodes or one node .
having discussed the phase space of @xmath2 dyonic black holes and solitons in some detail , we now present new @xmath3 dyonic black holes and solitons , and explore the phase space of solutions .
dyonic black hole solutions of @xmath3 eym are described by the following six parameters : @xmath237 , @xmath163 , @xmath238 , @xmath280 , @xmath239 and @xmath281 .
we fix the event horizon radius @xmath240 .
the field equations ( [ eq : ymefinal][eq : eefinal ] ) are symmetric under the transformations @xmath282 , @xmath283 , for each function separately , and so we can consider @xmath284 , @xmath285 , for @xmath286 without loss of generality .
eym in ads .
the parameters are : @xmath240 , @xmath241 , @xmath242 , @xmath287 and @xmath288 .
both electric gauge field functions @xmath12 and @xmath289 are monotonic and nodeless .
the magnetic gauge field function @xmath4 has two zeros and @xmath290 has three.,width=321 ] fig .
[ fig11 ] shows a typical @xmath3 dyonic black hole solution , with @xmath241 and the initial values @xmath291 ,
@xmath292 , @xmath293 . for this solution
the two electric gauge field functions are monotonic and have no zeros ; @xmath12 is monotonically decreasing and @xmath289 monotonically increasing .
the two magnetic gauge field functions @xmath4 and @xmath290 have zeros ; @xmath4 has two zeros and @xmath290 has three . in order to explore the phase space using two - dimensional plots
, it is necessary to fix two of the four parameters @xmath238 , @xmath280 , @xmath239 and @xmath281 as well as the event horizon radius @xmath237 and cosmological constant @xmath163 .
overall , the structure of the phase space of solutions is extremely complicated ; we give a flavour of some of the key features in figs .
[ fig12 ] and [ fig13 ] . as explained in sec .
[ sec : numerics ] , we can classify the solutions according to the numbers of zeros @xmath294 and @xmath295 of the magnetic gauge field functions @xmath4 and @xmath290 respectively . in both figs .
[ fig12 ] and [ fig13 ] , we have fixed the values of @xmath239 and @xmath281 , scanning the values of @xmath238 and @xmath280 such that the constraint ( [ eq : murh ] ) is satisfied . as in the @xmath2 case , there are values of the parameters such that ( [ eq : murh ] ) is satisfied but for which we do not find regular black hole solutions .
eym with @xmath240 , @xmath241 , and @xmath296 .
all shaded points in the plot correspond to black hole solutions .
the solutions are indexed by @xmath297 , the number of zeros of the magnetic gauge field functions @xmath4 , @xmath290 respectively .
the different combinations of values of @xmath298 are indicated by colour - coding the regions - in black and white the different colours are different shades of grey . in general
the number of zeros of the magnetic gauge field functions increases as we move towards the edges of the phase space . for these values of the parameters
, there are no nodeless solutions .
the number of zeros of the magnetic gauge field functions is @xmath299 in the green / lighter grey region containing the origin.,width=321 ] eym with @xmath240 , @xmath254 , @xmath300 and @xmath301 .
the solutions are indexed by @xmath297 , the number of zeros of the magnetic gauge field functions @xmath4 , @xmath290 respectively .
the different combinations of values of @xmath298 are indicated by colour - coding the regions - in black and white the different colours are different shades of grey . for these values of the parameters ,
all nontrivial solutions are nodeless .
we have also shown in red / darker grey ( marked `` no solution '' ) those values of the parameters @xmath302 for which the constraint ( [ eq : murh ] ) for a regular event horizon holds , but we have not found regular solutions.,width=321 ] when the magnitude of the cosmological constant is small , we typically find a rich phase space structure with solutions with many different numbers of nodes .
this is illustrated in fig .
[ fig12 ] , where we set @xmath241 .
we have also fixed the values of the magnetic gauge field functions on the event horizon to be @xmath296 , and scanned the phase space by varying @xmath238 , @xmath280 . with these values of the parameters , for all the solutions both magnetic gauge field functions @xmath4 and @xmath290 have at least two zeros .
as we have set the magnetic gauge field functions to be equal on the event horizon , from ( [ eq : su2gauge ] , [ eq : absu2 ] ) there are embedded @xmath2 solutions along the line @xmath303 . in @xcite the existence of dyonic black hole solutions of @xmath0 eym in ads in a neighbourhood of embedded @xmath2 dyonic black holes
is proven . in fig .
[ fig12 ] we see the neighbourhood of the embedded @xmath2 solutions for which there are nontrivial @xmath3 dyonic black holes . also from fig .
[ fig12 ] , the number of nodes of the magnetic gauge field functions increases as @xmath238 and @xmath280 increase . as the magnitude of the cosmological constant increases , for @xmath3 black hole solutions
we find ( as for the @xmath2 case ) that the phase space simplifies considerably .
this is illustrated in fig .
[ fig13 ] , where we have fixed the cosmological constant to be @xmath254 and the magnetic gauge field functions on the horizon have the values @xmath300 , @xmath301 . for these values of the parameters ,
all nontrivial solutions are such that the two magnetic gauge field functions @xmath4 and @xmath290 have no zeros . since @xmath304 in fig .
[ fig13 ] , there are no embedded @xmath2 solutions in this part of the phase space . for @xmath3 gauge group ,
the electric and magnetic gauge field functions need to be expanded to @xmath305 in a neighbourhood of the origin , as follows ( see the discussion in sec . [
sec : origin ] ) : @xmath306 for @xmath286 , where we have assumed without loss of generality that @xmath307 .
the vectors @xmath308 and @xmath309 are eigenvectors of the matrix @xmath310 ( [ eq : calm ] ) with eigenvalues @xmath311 and @xmath312 respectively . for @xmath191 ,
the matrix @xmath310 simplifies to @xmath313 and the relevant normalized eigenvectors are @xcite @xmath314 in terms of which @xmath125 and @xmath125 are given by ( [ eq : omega23sol ] ) . similarly , the vectors @xmath315 and @xmath316 are eigenvectors of the matrix @xmath317 ( [ eq : caln ] ) with eigenvalues @xmath311 and @xmath312 respectively . for @xmath191 ,
the matrix @xmath317 is @xmath318 with the relevant normalized eigenvectors ( in terms of which @xmath160 and @xmath192 are given by ( [ eq : h23sol ] ) ) being @xmath319 it is straightforward to check that the eigenvectors @xmath133 ( [ eq : u1u2 ] ) , @xmath320 ( [ eq : v1v2 ] ) , @xmath321 , are related by @xmath322 where , for @xmath191 , the matrix @xmath323 ( [ eq : calfdef ] ) takes the form @xmath324 in order to numerically integrate the field equations ( [ eq : ymefinal ] , [ eq : eefinal ] ) with the initial conditions ( [ eq : su3origin ] ) , we seek new variables @xmath225 , @xmath33 , @xmath321 , with the behaviour ( [ eq : betagammaorigin ] ) near the origin .
the @xmath225 depend only on the magnetic gauge field functions and the @xmath33 depend only on the electric gauge field functions . using the relations ( [ eq : betagamma ] ) and the eigenvectors ( [ eq : u1u2 ] , [ eq : v1v2 ] )
, we define the @xmath225 functions so that the magnetic gauge field functions @xmath10 take the form @xmath325 , \nonumber \\
\omega_{2}(r ) & = & \sqrt{2 } + \frac{1}{\sqrt{2}}\left[\beta_{1}(r ) - \beta_{2}(r ) \right ] , \label{eq : omegasu3sol}\end{aligned}\ ] ] and the @xmath33 functions so that the electric gauge field functions @xmath13 take the form @xmath326 the new variables @xmath272 , @xmath327 , @xmath328 and @xmath329 satisfy the following equations , which are derived from ( [ eq : ymefinal ] ) : [ eq : su3soleqns ] @xmath330 , \\
\beta_2 '' & = & -\left(\frac{\sigma'}{\sigma } + \frac{\mu'}{\mu}\right)\beta_2 ' + \frac{1}{4\mu r^2}(7\beta_1 ^ 2 + 28\beta_2 + \beta_2 ^ 2 + 24)\beta_2 \nonumber\\ & \ , & - \frac{1}{\sqrt{2}\sigma^2\mu^2}\left [ \sqrt{6}\gamma_1 \gamma_2 + \frac{\sqrt{3}\beta_2\gamma_1 \gamma_2}{\sqrt{2 } } - \frac{\beta_1}{\sqrt{2}}\left(\frac{9\gamma_1 ^ 2}{16 } + \frac{3\gamma_2 ^ 2}{2}\right)\right ] , \\ \gamma_1 ''
& = & \left ( \frac{\sigma'}{\sigma } - \frac{2}{r}\right)\gamma_1 ' + \frac{2\gamma_1}{\mu r^2 } + \frac{1}{\mu r^2}\left[\frac{1}{2}(\beta_1 ^ 2 + \beta_2 ^ 2 ) + 2(\beta_1 - \beta_2 ) - \beta_1\beta_2\right]\left(\frac{1}{2}\gamma_1 + \frac{\sqrt{3}}{2}\gamma_2\right ) \nonumber \\ & \ , & + \frac{1}{\mu r^2}\left[\frac{1}{2}(\beta_1 ^ 2 + \beta_2 ^ 2 ) + 2(\beta_1 + \beta_2 ) + \beta_1\beta_2\right]\left(\frac{1}{2}\gamma_1 - \frac{\sqrt{3}}{2}\gamma_2\right ) ,
\\ \gamma_2 '' & = & \left ( \frac{\sigma'}{\sigma } - \frac{2}{r}\right)\gamma_2 ' + \frac{6\gamma_2}{\mu r^2 } + \frac{1}{\mu r^2}\left[\frac{1}{2}(\beta_1 ^ 2 + \beta_2 ^ 2 ) + 2(\beta_1 - \beta_2 ) - \beta_1\beta_2\right]\left(\frac{\sqrt{3}}{2}\gamma_1 + \frac{3}{2}\gamma_2\right ) \nonumber\\ & \ , & + \frac{1}{\mu r^2}\left[\frac{1}{2}(\beta_1 ^ 2 + \beta_2 ^ 2 ) + 2(\beta_1 + \beta_2 ) + \beta_1\beta_2\right]\left(\frac{\sqrt{3}}{2}\gamma_1 - \frac{3}{2}\gamma_2\right).\end{aligned}\ ] ] we also substitute for @xmath13 and @xmath10 from ( [ eq : omegasu3sol ] , [ eq : hsu3sol ] ) into the einstein equations ( [ eq : eefinal ] ) , and then numerically integrate the resulting equations , together with ( [ eq : su3soleqns ] ) , using the initial conditions ( [ eq : betagammaorigin ] ) .
the solutions are parametrized by the cosmological constant @xmath163 , and the four parameters @xmath136 , @xmath137 , @xmath154 and @xmath155 . as in the @xmath2 case
, there are no _ a priori _ constraints on the values of these four parameters . in general @xmath136 and @xmath137
can take both positive and negative values . in our phase space plots ,
we have restricted our attention to @xmath331 , @xmath332 since this reveals the key features of the phase space .
eym in ads .
the parameters are : @xmath241 , @xmath333 , @xmath334 , @xmath335 , @xmath336 .
both electric gauge field functions @xmath12 and @xmath289 are monotonic and nodeless .
the magnetic gauge field functions @xmath4 and @xmath290 both have a single zero.,width=321 ] a typical dyonic soliton solution of @xmath3 eym is shown in fig .
[ fig14 ] . the cosmological constant is @xmath241 and the other parameters are @xmath333 , @xmath334 , @xmath335 , @xmath336 . for these values of the parameters ,
the two electric gauge field functions @xmath12 and @xmath289 are monotonic and have no zeros ; @xmath12 is monotonically decreasing and @xmath289 is monotonically increasing .
the two magnetic gauge field functions @xmath4 and @xmath290 are monotonically decreasing , and both have a single zero .
eym with @xmath241 , @xmath333 , @xmath334 .
all shaded points correspond to soliton solutions .
the solutions are indexed by @xmath297 , the number of zeros of the magnetic gauge field functions @xmath4 , @xmath290 respectively .
the different combinations of values of @xmath298 are indicated by colour - coding the regions - in black and white the different colours are different shades of grey .
we find solutions where the magnetic gauge field functions have a wide variety of numbers of nodes .
there is a small region close to the origin where both @xmath4 and @xmath290 have no zeros.,width=321 ] eym with @xmath254 , @xmath337 , @xmath338 .
all shaded points correspond to soliton solutions .
the solutions are indexed by @xmath297 , the number of zeros of the magnetic gauge field functions @xmath4 , @xmath290 respectively .
the different combinations of values of @xmath298 are indicated by colour - coding the regions - in black and white the different colours are different shades of grey .
the phase space has a much simpler structure for this larger value of @xmath8 . for most of the solution space , both magnetic gauge field functions @xmath4 and @xmath290 have no zeros .
there are small regions at the edges of the space of solutions where either @xmath4 or @xmath290 has a single zero.,width=321 ] once again we find a very rich space of solutions , and illustrate some features in figs .
[ fig15 ] and [ fig16 ] .
[ fig15 ] shows the phase space for @xmath241 .
the parameters @xmath333 and @xmath334 which govern the behaviour of the magnetic field functions near the origin are fixed ; we have scanned over positive values of the parameters @xmath154 and @xmath155 describing the electric gauge field functions near the origin . as we have seen previously for both @xmath2 solutions and @xmath3 dyonic black holes , for @xmath3 dyonic solitons with @xmath9 comparatively small
, the phase space is very complicated .
there are many different regions in which the two magnetic gauge field functions @xmath4 and @xmath290 have different numbers of zeros . for these values of the parameters , we find that @xmath4 and @xmath290 can have up to four zeros .
we find a small region close to @xmath339 for which both magnetic gauge field functions have no zeros . embedded @xmath2 solutions
correspond to @xmath340 , and therefore there are no embedded solutions in fig .
[ fig15 ] . as the magnitude of the cosmological constant increases , the phase space simplifies considerably .
this can be seen in fig .
[ fig16 ] , where @xmath254 .
the size of the space of nontrivial solutions also expands as @xmath9 increases . in fig .
[ fig16 ] , we have fixed @xmath337 and @xmath338 , and varied @xmath331 and @xmath332 . with these values of the parameters ,
most of the nontrivial dyonic soliton solutions are such that @xmath341 and both magnetic gauge field functions have no zeros .
there are small regions close to the edge of the solution space where one of @xmath342 ( but not both ) has a single zero .
in this paper we have presented new dyonic soliton and black hole solutions of the @xmath0 einstein - yang - mills ( eym ) field equations in asymptotically anti - de sitter ( ads ) space - time with a negative cosmological constant @xmath7 .
the metric is static and spherically symmetric .
the gauge field has nontrivial electric and magnetic components , and is described by @xmath343 independent gauge field functions , with equal numbers of electric and magnetic gauge field functions .
we have explored the phase space of soliton and black hole solutions for @xmath2 and @xmath3 gauge groups .
the solutions can be categorized by the numbers of zeros , @xmath344 , of the magnetic gauge field functions @xmath10 . in general the phase space
is very rich , with many different combinations of @xmath344 possible .
however , we find the following general features , many of which are in common with the phase space of purely magnetic @xmath0 solutions @xcite : * for small @xmath9 , we find solutions in which the magnetic gauge field functions have large numbers of zeros , particularly for the @xmath2 case ; * for small @xmath9 , the phase space is particularly complicated , with many different combinations of values of @xmath344 ; * as @xmath9 increases , the phase space expands in parameter space and the number of different combinations of values of @xmath344 decreases ; * for large @xmath9 , there are solutions for which all the magnetic gauge field functions have no zeros . * in a neighbourbood of the embedded trivial solution , either pure ads ( for solitons ) or schwarzschild - ads ( for black holes ) ; * in a neighbourhood of embedded nontrivial @xmath2 dyonic soliton and black hole solutions ; * in a neighbourhood of nontrivial purely magnetic @xmath0 solutions ( whose existence is proven in @xcite ) . of particular interest
are the solutions in the intersection of the last items in the two lists above , namely nontrivial nodeless solutions in the intersection of a neighbourbood of embedded @xmath2 solutions and a neighbourhood of embedded purely magnetic @xmath0 solutions .
we conjecture that it may be possible to prove that such solutions are stable under linear , spherically symmetric perturbations .
recently the existence of stable dyonic soliton and black hole solutions of @xmath2 eym in ads has been proven @xcite , and it would be interesting to attempt to extend that analysis to the case of a larger gauge group . in the @xmath2 case , the perturbation equations for dyonic solutions are much more complicated than the corresponding equations for purely magnetic background solutions , and the same will be true for @xmath0 gauge group .
we therefore leave this question open for future work .
the dyonic soliton and black hole solutions studied in this paper are spherically symmetric , with the event horizon being a surface of constant positive curvature .
the existence proof in @xcite is more general , and applies also to topological black holes for which the event horizon has either zero or constant negative curvature
. a natural question would be to investigate the phase space of dyonic topological black hole solutions of the @xmath0 eym equations , extending the recent study of the phase space of purely magnetic topological black hole solutions @xcite .
black holes with a flat event horizon in particular have attracted a great deal of attention in the literature as models of holographic superconductors ( see @xcite for a recent review ) .
planar black holes with a @xmath2 gauge field have been used to model @xmath345-wave superconductors ( see , for example , @xcite for a selection of papers ) , and enlarging the gauge group in these models would also be of interest .
we plan to return to this topic in a future publication .
finally , we anticipate that the thermodynamics of the dyonic black holes presented here would be very interesting .
the thermodynamics of purely magnetic @xmath2 black holes in ads has recently been studied @xcite . in @xcite ,
a complex picture emerges : it is found that purely magnetic , spherically symmetric @xmath2 black holes with unit magnetic charge at infinity are globally thermodynamically unstable ; those with zero magnetic charge at infinity have two branches of solutions , both of which are globally thermodynamically unstable to decay to an embedded schwarzschild - ads black hole ; while those with general magnetic charge at infinity also have two branches of solutions , one of which is thermodynamically stable .
we expect that the additional complexity of both enlarging the gauge group and including an electric as well as a magnetic part of the gauge field will render the thermodynamics of the @xmath0 dyonic black holes studied in this paper to be even more complicated than that presented in @xcite for the purely magnetic @xmath2 case .
accordingly we leave a systematic study of the thermodynamics to future research .
we thank erik baxter for insightful discussions .
the work of b.l.s . is supported by uk epsrc .
the work of e.w . is supported by the lancaster - manchester - sheffield consortium for fundamental physics under stfc grant no .
st / l000520/1 .
r. bartnik and j. mckinnon , phys .
lett . * 61 * , 141 ( 1988 ) .
m. s. volkov and d. v. galtsov , jetp lett .
b * 50 * , 346 ( 1989 ) ; m. s. volkov and d. v. galtsov , sov .
j. nucl .
51 * , 747 ( 1990 ) ; p. bizon , phys .
lett . * 64 * , 2844 ( 1990 ) ; h. p. kunzle and a. k. m. masood - ul - alam , j. math .
* 31 * , 928 ( 1990 ) .
d. v. galtsov and a. a. ershov , phys .
a * 138 * , 160 ( 1989 ) ; a. a. ershov and d. v. galtsov , phys .
a * 150 * , 159 ( 1990 ) .
j. a. smoller , a. g. wasserman , s. t. yau and j. b. mcleod , commun .
phys . * 143 * , 115 ( 1991 ) ; j. a. smoller , a. g. wasserman and s. t. yau , commun .
phys . * 154 * , 377 ( 1993 ) ; j. a. smoller and a. g. wasserman , commun .
phys . * 151 * , 303 ( 1993 ) ; p. breitenlohner , p. forgacs and d. maison , commun .
* 163 * , 141 ( 1994 ) .
r. ruffini and j. a. wheeler , phys . today * 24 * , 30 ( 1971 ) .
n. straumann and z. h. zhou , phys .
b * 237 * , 353 ( 1990 ) ; n. straumann and z. h. zhou , phys .
b * 243 * , 33 ( 1990 ) ; g. v. lavrelashvili and d. maison , phys .
b * 343 * , 214 ( 1995 ) ; m. s. volkov , o. brodbeck , g. v. lavrelashvili and n. straumann , phys .
b * 349 * , 438 ( 1995 ) ; s. hod , phys . lett .
b * 661 * , 175 ( 2008 ) .
p. bizon , acta phys .
b * 25 * , 877 ( 1994 ) .
m. s. volkov and d. v. galtsov , phys .
rept . *
319 * , 1 ( 1999 ) .
d. v. galtsov and m. s. volkov , phys .
b * 274 * , 173 ( 1992 ) ; b. kleihaus , j. kunz and a. sood , phys .
b * 354 * , 240 ( 1995 ) ; b. kleihaus , j. kunz and a. sood , phys . lett .
b * 418 * , 284 ( 1998 ) ; b. kleihaus , j. kunz , a. sood and m. wirschins , phys .
d * 58 * , 084006 ( 1998 ) .
o. brodbeck and n. straumann , phys .
b * 324 * , 309 ( 1994 ) ; o. brodbeck and n. straumann , j. math .
phys . *
37 * , 1414 ( 1996 ) .
e. winstanley , class .
quant . grav .
* 16 * , 1963 ( 1999 ) .
j. bjoraker and y. hosotani , phys .
lett . * 84 * , 1853 ( 2000 ) ; j. bjoraker and y. hosotani , phys .
d * 62 * , 043513 ( 2000 ) .
breitenlohner , d. maison and g. lavrelashvili , class .
* 21 * , 1667 ( 2004 ) .
o. sarbach and e. winstanley , class .
* 18 * , 2125 ( 2001 ) ; e. winstanley and o. sarbach , class .
* 19 * , 689 ( 2002 ) .
j. e. baxter , m. helbling and e. winstanley , phys .
d * 76 * , 104017 ( 2007 ) ; j. e. baxter , m. helbling and e. winstanley , phys .
lett . * 100 * , 011301 ( 2008 ) .
j. e. baxter and e. winstanley , class .
grav . *
25 * , 245014 ( 2008 ) .
j. e. baxter and e. winstanley , j. math .
* 57 * , 022506 ( 2016 ) .
b. l. shepherd and e. winstanley , class .
* 29 * , 155004 ( 2012 ) .
e. winstanley , lect .
notes phys .
* 769 * , 49 ( 2009 ) ; e. winstanley , arxiv:1510.01669 [ gr - qc ] .
b. c. nolan and e. winstanley , class .
grav . * 29 * , 235024 ( 2012 ) .
b. c. nolan and e. winstanley , class .
grav . * 33 * , 045003 ( 2016 ) .
j. e. baxter , j. math .
phys . * 57 * , 022505 ( 2016 ) .
h. p. kuenzle , class .
* 8 * , 2283 ( 1991 ) .
r. bartnik , j. math .
* 38 * , 3623 ( 1997 ) .
r. a. brandt and f. neri , nucl .
b * 186 * , 84 ( 1981 ) .
y. shnir , hep - th/0508210 .
h. p. kuenzle , commun .
phys . *
162 * , 371 ( 1994 ) .
s. karlin and j. l. mcgregor , scripta mathematica * 26 * , 33 ( 1961 ) .
j. e. baxter and e. winstanley , phys .
b * 753 * , 268 ( 2016 ) .
r. g. cai , l. li , l. f. li and r. q. yang , sci . china phys .
* 58 * , 060401 ( 2015 ) .
s. s. gubser , phys .
* 101 * , 191601 ( 2008 ) ; s. s. gubser and s. s. pufu , jhep * 0811 * , 033 ( 2008 ) ; r. manvelyan , e. radu and d. h. tchrakian , phys .
b * 677 * , 79 ( 2009 ) ; c. p. herzog and s. s. pufu , jhep * 0904 * , 126 ( 2009 ) ; k. peeters , j. powell and m. zamaklar , jhep * 0909 * , 101 ( 2009 ) ; s. s. gubser , f. d. rocha and a. yarom , jhep * 1011 * , 085 ( 2010 ) ; m. ammon , j. erdmenger , v. grass , p. kerner and a. obannon , phys .
b * 686 * , 192 ( 2010 ) ; a. akhavan and m. alishahiha , phys .
d * 83 * , 086003 ( 2011 ) ; s. gangopadhyay and d. roychowdhury , jhep * 1208 * , 104 ( 2012 ) ; r. e. arias and i. s. landea , jhep * 1301 * , 157 ( 2013 ) ; z. y. nie , r. g. cai , x. gao and h. zeng , jhep * 1311 * , 087 ( 2013 ) ; c. p. herzog , k. w. huang and r. vaz , jhep * 1411 * , 066 ( 2014 ) ; z. y. nie , r. g. cai , x. gao , l. li and h. zeng , eur .
j. c * 75 * , 559 ( 2015 ) .
o. kichakova , j. kunz , e. radu and y. shnir , phys .
b * 747 * , 205 ( 2015 ) .
z. y. fan and h. lu , jhep * 1502 * , 013 ( 2015 ) ; | we present new spherically symmetric , dyonic soliton and black hole solutions of the @xmath0 einstein - yang - mills equations in four - dimensional asymptotically anti - de sitter space - time .
the gauge field has nontrivial electric and magnetic components and is described by @xmath1 magnetic gauge field functions and @xmath1 electric gauge field functions .
we explore the phase space of solutions in detail for @xmath2 and @xmath3 gauge groups .
combinations of the electric gauge field functions are monotonic and have no zeros ; in general the magnetic gauge field functions may have zeros .
the phase space of solutions is extremely rich , and we find solutions in which the magnetic gauge field functions have more than fifty zeros . of particular interest
are solutions for which the magnetic gauge field functions have no zeros , which exist when the negative cosmological constant has sufficiently large magnitude .
we conjecture that at least some of these nodeless solutions may be stable under linear , spherically symmetric , perturbations . |
proving the existence of black holes remains a key goal of observational high energy astrophysics .
while dynamical evidence ( e.g. charles 1998 ) convincingly demonstrates the existence of compact accreting objects in binary systems which have masses in excess of the highest theoretical limit for a neutron star ( @xmath0 ) , and are therefore strong black hole candidates ( bhcs ) , we can not rule out the possibility that some as - yet - unconsidered state of matter may provide an alternative explanation . as an alternative approach , in recent years
much attention has been focussed on finding evidence for black hole event horizons .
one promising and actively pursued route has been a comparison of the x - ray luminosities of bhc and neutron star ( ns ) x - ray binaries ( xrbs ) in ` quiescence ' . in such states ,
black hole accretion could be advection dominated and considerably fainter than neutron stars .
this is indeed what has been found observationally , with ` quiescent ' bhcs being typically two to three orders of magnitude ( in eddington units ) less luminous than their ns xrb equivalents .
this has been claimed to represent some of the strongest evidence to date for the existence of black hole event horizons ( narayan , garcia & mcclintock 1997 ; menou et al .
1999 ; garcia et al .
however , alternatives to this interpretation have also been discussed ( campana & stella 2000 ; bildsten & rutledge 2000 ; abramowicz , kluzniak & lasota 2002 ) .
abramowicz et al . ( 2002 ) , in particular , stress that ` absence of evidence is not evidence of absence ' , and draw attention to alternatives to black holes .
even if bhcs _ do _ contain black hole with event horizons , it is important to establish how much , if any , of the potential accretion energy may be being advected across their horizons . in a series of important and related observations , in recent years it has been established that jets are an integral and relatively ubiquitous component of the process of accretion in both black hole and neutron star x - ray binaries ( e.g. mirabel & rodriguez 1999 ; fender 2002 ) .
we are now beginning to understand just how powerful these jets may be .
corbel et al .
( 2003 ) discovered that , over four orders of magnitude in x - ray luminosity , the relation between radio and x - ray luminosity for the bhc x - ray binary gx 339 - 4 has the form @xmath1 , where @xmath2 for @xmath3 in the 39 kev range ( @xmath4 increases slightly with the increasing energy of the x - ray band used for comparison ) .
gallo , fender & pooley ( 2003 ) demonstrated that the same relation holds over a comparable range in x - ray luminosity for the transient v404 cyg ( gs 2023 + 338 ) , and furthermore that the data for all measured low / hard state sources is consistent with a such a universal relation holding for all of them .
this power law relation between radio and x - ray luminosity is a key observational discovery providing clues to the underlying physics of the disc jet coupling .
are there differences in jet power between the bhcs and ns xrbs ?
fender & kuulkers ( 2001 ) found that bhc xrbs were , in general between 10100 times more ` radio loud ' ( in the sense of the radio to soft x - ray ratio ) than neutron star binaries .
migliari et al .
( 2003 ) compared the radio strength of the atoll - type neutron star binary 4u 1728 - 34 with the comparable state and x - ray luminosity of bhcs , and found a ratio of radio loudness @xmath5 .
the origin of this difference in radio loudness is not clear ( see fender & kuulkers 2001 for a discussion ) .
in the following all luminosities and accretion rates are in eddington units , where the eddington luminosity is @xmath6 erg s@xmath7 , where @xmath8 is the mass of the accreting compact star .
the eddington accretion rate , defined as that accretion rate at which the eddington luminosity is achieved , is , for an accretion efficiency of @xmath9% ( ie .
@xmath10 is liberated during the accretion process ) approximately @xmath11 g s@xmath7 .
we assume that the total power output @xmath12 from an x - ray binary in a ` low / hard ' or analogous state is a combination of the radiative luminosity of the flow ( @xmath3 , directly observed as x - rays ) and jet power ( @xmath13 indirectly traced by e.g. radio flux density ) : @xmath14 now we already know ( corbel et al . 2002 ; gallo et al . 2003 ) the relation between radio ( @xmath15 ) and x - ray luminosity : @xmath16 how does observed radio flux relate to jet power ; i.e. what is the relation between @xmath15 and @xmath13 ? in models of optically thick jets ( e.g. blandford & knigl 1979 ; falcke & biermann 1996 ; markoff , falcke & fender 2001 ; heinz & sunyaev 2003 ) , the following scaling applies : @xmath17 combining equations ( 2 ) and ( 3 ) : @xmath18 therefore @xmath19 which provides the relation between total power and x - ray luminosity .
the normalisation @xmath20 between can be estimated .
fender ( 2001 ) and corbel & fender ( 2002 ) conservatively estimate @xmath21 for cyg x-1 and gx 339 - 4 at an accretion luminosity of @xmath22 .
fender et al .
( 2001 ) estimated that , at an accretion luminosity of @xmath23 , the black hole transient xte j1118 + 480 had @xmath24 ( see also corbel & fender 2002 for an estimate for gx 339 - 4 ) . conservatively adopting the equality for xte j1118 + 480 corresponds to @xmath25 in eddington units . equivalently the relation between total power and jet power is given by : @xmath26 in the following we shall assume that @xmath12 is proportional to the mass accretion rate @xmath27 ( i.e. all the available accretion power goes either into the x - rays or the jet ) . in eddington units
this corresponds to @xmath28 which is the condition of no advection of accretion energy across the event horizon
. we can then plot the variation of @xmath3 and @xmath13 as a function of mass accretion rate .
these are plotted for black holes in the top panel of fig 1 .
we note that there are two regimes , ` x - ray dominated ' at higher mass accretion rates , and ` jet dominated ' at lower accretion rates . in the x - ray
dominated regime , @xmath29 and @xmath30 .
however in the jet dominated regime @xmath31 and @xmath32 .
the transition between the two regimes occurs at @xmath33 or , equivalently , @xmath34 .
the shaded region in the top panel of fig 1 indicates the observed range of x - ray luminosities of black hole x - ray binaries in ` quiescence ' if our model is correct then _ all _ of these systems are in the jet - dominated regime , with accretion rates @xmath35 , and with jet powers one to two orders of magnitude greater than the observed x - ray luminosity .
a major uncertainty in knowing if the arguments outlined above apply to neutron stars is that the relation @xmath36 has not yet been measured .
migliari et al . (
2003 ) note that the relation seems steeper ( @xmath37 ) for the atoll - type x - ray binary 4u 1728 - 34 , but this is over a small range in x - ray flux compared to that measured for black holes . at present
we must consider that this relation remains unmeasured , due primarily to the relative faintness of atoll - type sources in the radio band compared to black holes ( fender & hendry 2000 ) , which results from the greater ` radio loudness ' of black holes ( fender & kuulkers 2001 ; migliari et al .
however , we will make the assumption in what follows that the same relation does indeed apply for atoll - type ns xrbs .
as already noted , the ratio of ` radio loudness ' between bhc and ns xrbs is @xmath38 . using equation ( [ jetradio ] )
, this translates into a difference in jet power of a factor 10 , i.e. @xmath39 . the x - ray luminosity below which neutron star systems would be jet - dominated is therefore @xmath40 .
this is comparable to the lowest x - ray luminosity measured from a neutron star in quiescence ( garcia et al .
2001 ) implying that , quite unlike black holes , we may have never observed a neutron star in a jet - dominated state . in the lower panel of fig 1
we plot the variation of @xmath3 and @xmath13 as a function of mass accretion rate for neutron star binaries . as for the bhcs
, the shaded region indicates the observed range of ` quiescent ' x - ray luminosities . in the absence of core / crustal emission ( see below ) which is decoupled from the accretion flow on all but the longest timescales , the observed x - ray luminosities of ` quiescent ' ns xrbs correspond to a range in accretion rate of @xmath41 , overlapping with the range in @xmath27 for ` quiescent ' black holes .
therefore , in at least this respect , in the model presented here the data are consistent with both ns and bhc xrbs in ` quiescence ' accreting at the same rate .
brown , bildsten & rutledge ( 1998 ) have argued that , once accretion has halted , neutron stars will have a luminosity in the range @xmath42 @xmath43 erg s@xmath7 from crustal emission .
this model seems to be supported by observations of transient neutron star binaries in quiescence ( e.g. rutledge et al .
2001a ; rutledge et al .
2001b ; rutledge et al .
2002 ; see also wijnands et al .
2001 ) , although garcia et al .
( 2001 ) argue that even at ` quiescent ' levels the x - ray luminosity is dominated by accretion . indeed , the quiescent emission of sax j1808.43658 is uncomfortably low ( 5@xmath44 ) and hard ( power law index 1.5 with a blackbody contribution of less than 10 per cent ) for the neutron star crustal emission model unless the neutron star is more massive than 1.7 m@xmath45 ( campana et al .
2002 ) .
if this crustal emission does exist then it adds a new term to the total observed x - ray emission : @xmath46 in fig 1 ( lower panel ) we also indicate the solutions with the addition of persistent ` crustal ' emission to the observed x - ray flux from a neutron star , at a level of @xmath47 erg s@xmath7 , approximately the lowest luminosity observed from a quiescent neutron star .
this has a significant effect , since this crustal luminosity , at @xmath48 is above that at which neutron stars would enter the jet - dominated regime .
whereas in the case of accretion - only luminosity , while we had not observed neutron stars in jet - dominated regimes they were still possible , if such crustal luminosities are ubiquitous then neutron stars will not enter the jet - dominated regime , unless their time - averaged mass accretion rates are very low ( @xmath49 @xmath50 ) .
however , note that at the lowest accretion rates neutron stars will make just as powerful jets as bhcs ( see next section ) .
this work leads naturally to some interesting consequences if correct .
we outline these below .
we have seen in the above that below a certain mass accretion rate bhc x - ray binaries probably enter a jet - dominated state . because of their higher ` radio loudness ' ,
black holes make the transition to this jet - dominated state at a higher mass accretion rate than neutron stars ( by a factor @xmath51 ) .
consequently , if there are no other effects , once both ns and bh are in the jet - dominated regime , the ns systems will be a factor @xmath51 brighter in x - rays than the bh systems .
since @xmath52 then we expect a ratio of @xmath53 between quiescent x - ray luminosities at the same accretion rate .
in fact , while the expressions and plots given so far are specifically for the condition in equation ( 3 ) and the estimated values of @xmath54 , @xmath55 , there is a more general expression for the ratio of x - ray luminosities when both classes of object are in the jet - dominated regime : @xmath56 since @xmath38 and @xmath57 , we expect a ratio of x - ray luminosities in quiescence of @xmath58 , when both bhcs and ns xrbs are in the jet - dominated regime , and at the same mass accretion rate .
this is consistent with what is observed .
the x - ray luminosities as a function of @xmath27 are illustrated in fig 2 ( top panel ) .
as already noted , the quiescent ns xrbs may not be quite in the jet - dominated regime ; however , the bhc xrbs are clearly in this regime , and the difference in x - ray luminosities at the same accretion rate is already one order of magnitude at @xmath59 , increasing to @xmath60 at @xmath61 ( fig 2 , top panel ) .
we therefore find that the observed @xmath62 scaling , combined with the order of magnitude greater radio loudness of bhc xrbs , naturally results in a signficant difference in the quiescent luminosities of ns and bhc xrbs , as observed .
more precisely , we expect there to be three regimes in which the ratio of x - ray luminosities , @xmath63 has different values : [ cols="<,<,^,^ " , ] where ( a ) corresponds to both classes of objects being ` x - ray dominated ' , ( b ) corresponds to bhcs being jet - dominated and ns not , ( c ) corresponds to both classes being jet - dominated . from this study
it appears that @xmath64 , and that observed ` quiescence ' corresponds to regimes ( a ) or ( b ) , which consistent with the observations without requiring any accretion energy to be advected across an event horizon .
it is interesting to note that in the jet - dominated regime , the scaling of x - ray luminosity with mass accretion rate , @xmath65 is exactly the same as that predicted theoretically by adaf models ( e.g. narayan et al .
1997 ; mahadevan 1997 ) .
it would be a mistake to assume that the persistent difference in x - ray luminosity will result in a difference in jet powers between ns and bhc systems at the lowest luminosities .
in fact , below an accretion rate of @xmath66 both ns and bhc systems are putting the same amount of power into the jet ( fig 2 , lower panel ) , which dominates the power output of the system .
the tiny fraction of the total power released as x - rays is insignificant whether its a bhc or a ns xrb one hundred times brighter .
it is also interesting to note that the ratio in radio loudness , @xmath67 is maintained throughout this scenario , but for somewhat different reasons in the two regimes .
when ` x - ray dominated ' , bhcs are more ` radio loud ' because they match the ns xrbs in x - rays but put out more radio power .
however , at the lowest accretion rates the radio power is the same but the x - ray luminosity of the bhcs is lower , maintaining the ratio . it has been suggested that the hard x - ray spectra observed from low / hard state bhcs may be in some , maybe all , cases optically thin synchrotron emission directly from the jet ( markoff , falcke & fender 2001 ; markoff et al . 2003 ) .
this is at odds with the more standard view of the hard x - ray spectrum as being dominated by thermal comptonisation from electrons with a temperature of @xmath53 kev ( e.g. sunyaev & titarchuk 1980 ; poutanen 1998 , zdziarski et al .
if it is the correct interpretation , how does it affect the analysis performed here ?
since in those models the bhcs are already completely jet dominated at @xmath68 , then @xmath69 ( as @xmath70 [ equation ( 4 ) ] and @xmath32 this is always the case for jet - dominated emission ) .
in fact it can be shown that the same ratio of ` quiescent ' luminosities is achieved as in the previous analysis , as long as the ns xrbs are _ not _ already jet - dominated at @xmath71 ( otherwise @xmath3 in both classes of objects would track each other ) .
however , the transition to the jet - dominated regime would occur at a higher @xmath27 ( by approximately two orders of magnitude ) , meaning the observed ` quiescent ' mass accretion rates would be considerably higher than those indicated in fig 1 .
the results presented in this letter are necessarily a subset of the possible consequences of the empirical relations and model upon which they are based . of the more general results , to be expanded upon in a further work ,
only one is given , namely that once both bhc and ns xrb are in jet dominated states the ratio of x - ray luminosities depends only upon two quantities which have already been measured , i.e. @xmath4 and @xmath67 : @xmath72 we suggest that based upon existing observational data , ` quiescent ' bhcs are in the ` jet - dominated ' regime and that ns xrbs are , if not jet - dominated , close to the transition to this regime .
specifically , if a similar value of @xmath4 holds for ns xrbs as for bhcs ( and this is the key observational uncertainty ) then quiescent ns xrbs are , in the most conservative case , putting @xmath73% of their power into jets .
thus the observed ratio of x - ray luminosities should be close to @xmath60 , consistent with what is observed .
an additional core / crustal contribution to the x - ray emission from ns xrbs will only widen the discrepancy . essentially , we find that the difference in quiescent x - ray luminosities between ns and bhc xrbs can be mostly , if not entirely , explained by a difference in the efficiency of jet production between the two types of sources ( the origin of which remains unclear ) .
therefore we find that the observed difference in quiescent luminosities of neutron star and black hole candidate xrbs _ does not require the presence of black hole event horizons_. this should not be taken as a statement to the effect that we do not believe that black hole candidates contain black holes ( see related discussion in abramowicz et al .
certainly there are differences between the neutron star and black hole candidate xrbs , which may be naturally explained by the fact that black hole candidates do in fact contain black holes . however , since there is no requirement for significant energy to be advected into the black hole in order to explain the ratio of quiescent x - ray luminosities , these luminosities can not in turn be taken as ` proof ' of the existence of event horizons .
furthermore , even if , as seems probable , the bhcs do contain black holes with event horizons , this work shows that there is no requirement for the advection of any significant quantity of accretion energy across the horizon .
rather , the relative radiative inefficiency of bhcs compared to ns xrbs at low @xmath27 is due to the low radiative efficiency of the _ jets _ they are powering , not the accretion flow itself .
rpf would like to thank sera markoff , sebastian heinz and lars bildsten for useful discussions .
pgj would like to thank gordon ogilvie for useful discussions .
poutanen , j. , 1998 , _ accretion disc corona models and x/@xmath74-ray spectra of accreting black holes _ , in : abramowicz , m. a. , bjrnsson , g. , pringle , j. e. ( eds ) , theory of black hole accretion discs , cambridge contemporary astrophysics , cup , 1998 , p.100 | we demonstrate that at relatively low mass accretion rates , black hole candidate ( bhc ) x - ray binaries ( xrbs ) should enter ` jet - dominated ' states , in which the majority of the liberated accretion power is in the form of a ( radiatively inefficient ) jet and not dissipated as x - rays in the accretion flow .
this result follows from the empirically established non - linear relation between radio and x - ray power from low / hard state bhc xrbs , which we assume also to hold for neutron star ( ns ) xrbs .
conservative estimates of the jet power indicate that all bhc xrbs in ` quiescence ' should be in this jet - dominated regime . in combination with an additional empirical result , namely that bhc xrbs are more ` radio loud ' than ns xrbs , we find that in quiescence ns xrbs should be up to two orders of magnitude more luminous in x - rays than bhc xrbs , without requiring any significant advection of energy into a black hole .
this ratio is as observed , and such observations should therefore no longer be considered as direct evidence for the existence of black hole event horizons .
furthermore , even if bhcs do contain black holes with event horizons , this work demonstrates that there is no requirement for the advection of significant amounts of accretion energy across the horizon .
binaries : close radio continuum : stars ism : jets and outflows stars : neutron black hole physics |
since its inception , the large @xmath2 expansion @xcite @xcite has afforded a host of valuable insights into @xmath3 dynamics beyond standard perturbation theory in the coupling constant . besides providing a solvable example of a nonabelian gauge theory in the case of one space dimension @xcite
, it has been significant e.g. in clarifying the relation between the skyrme model and @xmath3 and in qualitatively explaining the zweig rule @xcite . on the other hand ,
the number of colors @xmath2 is not the only parameter of @xmath3 which can be regarded as large .
the same procedure may also be applied to the number of flavors @xmath4 . in the real world , especially at higher energies , where the heavy quarks contribute to the dynamics , the effective @xmath0 ratio becomes larger than unity .
thus it seems natural to investigate in what way the presence of a large number of flavors modifies the large @xmath2 picture and how those parts of large @xmath2 phenomenology which appear to be successes survive the modification .
early work in this direction was done by veneziano @xcite and some more recent explorations into the field of multiflavor gauge theory have been carried out e.g. in connection with bosonization treatments @xcite @xcite . particularly the low - dimensional models such as @xmath5 with a large number of colors become considerably simplified in the framework of light - cone quantization @xcite . at this point
it should be remarked that the light - cone treatment corresponds in some respects to an effective theory with well - defined limitations , as was elucidated in @xcite . in the work presented here
, no attention will be paid to such questions .
the results will be obtained entirely within the usual naive light - cone framework and no firm claim is made as far as their validity in normal coordinates is concerned .
the prevailing experience gives rise to the hope that the light - cone treatment does yield correct excitation spectra .
the present foray into the multiflavor world focuses purely on the mesonic spectrum of the model with zero quark mass matrix .
the hamiltonian is calculated in a ( truncated ) basis of states made up of t hooft mesons , which diagonalizes the hamiltonian for @xmath6 , @xmath7 .
increasing the number of flavors such that @xmath0 remains finite as @xmath7 induces nonvanishing interactions between the t hooft mesons ; the evolution of the low - lying spectrum as a function of @xmath0 can be followed by rediagonalizing the hamiltonian . here , only the flavor singlet sector is discussed ; however , the generalization to finite flavor density presents no conceptual problem .
since the numerical work up to now was carried out with limited computational resources , only the one - meson and two - meson ( i.e. two- and four - parton ) sectors could be included .
the hamiltonian of multiflavor @xmath5 in the light - cone gauge with zero quark mass matrix reads in momentum space @xmath8 here , @xmath9 and @xmath10 are the color indices , @xmath11 and @xmath12 the flavor indices .
note that the fermion fields do not carry a dirac index , since one of the components completely decouples in the light - cone gauge .
the hamiltonian has already been normal - ordered , by applying wick s theorem , with respect to the vacuum of the theory , which is the perturbative one : @xmath13 the singularities in the integrands finally are defined via the principal value prescription .
as already mentioned in the introduction , all subleties associated with the singular nature of the light - cone description are disregarded here ; the reader is referred to @xcite .
the hamiltonian ( [ ham ] ) conserves baryon number , total momentum , and overall color . here
, the focus will be on the color singlet sector @xcite with zero baryon number and total momentum denoted by @xmath14 . due to the confining coulomb potential ,
the physical spectrum consists entirely of color singlet quark - antiquark bound states , i.e. mesons , and conglomerates thereof @xcite @xcite .
it thus seems meaningful to consider the hamiltonian matrix in a fock space spanned by color singlet mesonic excitations above the ( perturbative ) vacuum .
mesons are created by the bilinear operators @xmath15 where the @xmath16 are a complete set of bound state wave functions .
a particularly suitable such set are t hooft s wave functions , @xmath17 where the @xmath18 satisfy the equation @xmath19 with the boundary condition @xmath20 the mass scale @xmath21 is given by @xmath22 and @xmath23 is the invariant mass squared of the @xmath24-th t hooft meson .
apart from the exact zero - mode solution @xmath25 , a suitable orthonormal set in which to expand the @xmath18 in order to solve ( [ hoofeq ] ) numerically is given in this case by cosine functions . with this choice ,
the hamiltonian is already diagonal in the basis defined by the @xmath26 in the case that @xmath4 remains finite and @xmath7 .
note that the operators @xmath27 are not suitable for a rigorous bosonization of the theory since they only obey canonical bosonic commutation relations up to additional terms of the order @xmath28 .
it is now a straightforward , albeit lengthy , procedure to calculate the hamiltonian matrix by commuting through quark operators .
one obtains @xmath29 \nonumber \\ & & \langle k+q/2,n , a , b ; k - q/2,n^{\prime } , a^{\prime } , b^{\prime } | h | k+q^{\prime } /2,m , c , d ; k - q^{\prime } /2,m^{\prime } , c^{\prime } , d^{\prime } \rangle = \nonumber \\ & & \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \delta_{ac } \delta_{bd } \delta_{a^{\prime } c^{\prime } } \delta_{b^{\prime } d^{\prime } } \delta_{nm } \delta_{n^{\prime } m^{\prime } } \delta ( q - q^{\prime } ) \left ( \frac{\mu_{n}^{2 } } { 2k+q } + \frac{\mu_{n^{\prime } } ^{2 } } { 2k - q } \right ) + \frac{m_0 ^ 2 } { 8k^2 n_c } \times \label{me22 } \\ & & \left [ \delta_{a^{\prime } c } \delta_{ac^{\prime } } \delta_{bd } \delta_{b^{\prime } d^{\prime } } f_{mm^{\prime } nn^{\prime } } ^{exc } \left ( \frac{k\ ! \ ! + \
! q/2}{k\ ! \
q^{\prime } /2 } , \frac{k\ ! \ ! -\ !
q/2}{k\ ! \
q^{\prime } /2 } \right ) + \delta_{ac } \delta_{a^{\prime } c^{\prime } } \delta_{b^{\prime } d } \delta_{bd^{\prime } } f_{m^{\prime } mnn^{\prime } } ^{exc } \left ( \frac{k\ ! \ ! + \ !
q/2}{k\ ! \ ! -\ !
q^{\prime } /2 } , \frac{k\ ! \ ! -\ !
q/2}{k\ ! \ ! -\
! q^{\prime } /2 }
\right ) \right
. \nonumber \\ & & + \delta_{a^{\prime } c^{\prime } } \delta_{ab^{\prime } } \delta_{bd } \delta_{cd^{\prime } } f_{mm^{\prime } nn^{\prime } } ^{ann } \left ( \frac{k\ ! \ ! + \ !
q/2}{k\ ! \
q^{\prime } /2 } , \frac{k\ ! \ ! -\ !
q/2}{k\ ! \
q^{\prime } /2 } \right ) + \delta_{ac^{\prime } } \delta_{a^{\prime } b } \delta_{b^{\prime } d } \delta_{cd^{\prime } } f_{mm^{\prime } n^{\prime } n}^{ann } \left ( \frac{k\ ! \ ! -\
! q/2}{k\ ! \
q^{\prime } /2 } , \frac{k\ ! \ ! + \ !
q/2}{k\ ! \
q^{\prime } /2 }
\right ) \nonumber \\ & & + \left .
\delta_{a^{\prime } c } \delta_{ab^{\prime } } \delta_{bd^{\prime } } \delta_{c^{\prime } d } f_{m^{\prime } mnn^{\prime } } ^{ann } \left ( \frac{k\ ! \ ! + \ !
q/2}{k\ ! \ ! -\ !
q^{\prime } /2 } , \frac{k\ ! \ ! -\ !
q/2}{k\ ! \ ! -\
! q^{\prime } /2 } \right ) \ !
+ \delta_{ac } \delta_{a^{\prime } b } \delta_{b^{\prime } d^{\prime } } \delta_{c^{\prime } d } f_{m^{\prime } mn^{\prime } n}^{ann } \left ( \frac{k\ ! \ ! -\ !
q^{\prime } /2 } , \frac{k\ ! \ ! + \ !
q/2}{k\ ! \ ! -\
! q^{\prime } /2 }
\right ) \right ] \nonumber\end{aligned}\ ] ] with the form factors ( note that the @xmath18 are defined to be zero outside the interval @xmath30 $ ] ) @xmath31 here , in the kinetic part of the two - meson two - meson matrix element , a convention of definite ordering of the momenta in the states was assumed , namely @xmath32 .
this avoids double - counting two - meson states .
also , in this kinetic part , @xmath28-suppressed terms in the normalization of the states were neglected .
all interaction parts , on the other hand , are valid for any @xmath33 and no terms were neglected .
the processes described by @xmath34 , @xmath35 , and @xmath36 are associated with a flavor flow represented diagrammatically in graphs ( a),(b ) and ( c ) of figure ( [ graphs ] ) , respectively .
these graphs clarify the large @xmath4 counting associated with each vertex ; this will crucially influence which contributions survive in the limit @xmath0 constant , @xmath7 .
note that the hamiltonian can only change the number of mesons by one . to change it by two
, it would have to contain a contribution consisting of four parton creation or annihilation operators with respect to the vacuum ; this , however , would completely commute through any meson operators to the right or left , respectively , ultimately annihilating the vacuum in relation to which the hamiltonian is already normal - ordered . equations ( [ me11])-([me22 ] ) completely define the hamiltonian in the subspace containing up to two mesons . but for the fact that the meson operators @xmath27 only obey canonical commutation relations up to terms of order @xmath28 , the three - meson and four - meson interactions described by ( [ me11])-([me22 ] ) would already represent the full dynamics in mesonic language .
this is because the coulomb interaction , as is clear from the original quark representation ( [ ham ] ) , can only modify at most two out of an arbitrarily long string of meson operators as it is being commuted through to eventually annihilate the vacuum .
however , the @xmath28-suppressed additional terms in the meson commutators introduce additional interactions in larger clusters which in general must be treated on the same footing as the vertices derived above . since the hilbert space will here be truncated to the one- and two - meson sectors , these terms need not be considered further in the present calculation .
the above considerations still apply to an arbitrary structure in flavor space ; the results will now be specialized to the flavor singlet sector .
a flavor singlet meson can be constructed by the superposition @xmath37 whereas a flavor singlet two - meson state may consist either of two singlet mesons or of two mesons carrying flavor spin one coupled to a singlet : @xmath38 carrying out the flavor algebra , the hamiltonian matrix in these states takes the following form , where only terms which remain finite in the limit @xmath0 finite , @xmath7 are kept : @xmath39 & \label{singn}\end{aligned}\ ] ] here , the symmetry behavior of t hooft s wave functions @xmath40 around @xmath41 was used , making explicit the selection rules for the meson excitation quantum numbers .
physically , this is the manifestation of cp invariance .
these selection rules decouple states containing single - meson components of even excitation number from those containing single - meson components of odd excitation number .
note that this feature does not persist in the flavor non - singlet sector . the hamiltonian matrix obtained above exhibits in very transparent form the physics of the large @xmath4 world : flavor singlets do not interact among each other .
nonvanishing interactions only occur within flavor singlets , which resemble loops built out of alternating color strings and flavor `` strings '' in the sense that the two quarks at the end of a flavor string must be coupled to a singlet . within these loops
, color strings may break into two , creating a flavor string in between .
flavor strings are not necessarily binding ; however , the diagonalization of the hamiltonian will show that there are indeed bound states made up of more than one meson such that the flavor string effectively acts as a bond . note the remarkable similarity of this picture to the model with one flavor , but adjoint color @xcite @xcite . in the original quark hamiltonian
, flavor does not appear as a quantum number related to the interaction ; thus , the notion of a flavor string introduced above does not carry physical meaning in the same sense as the color string , which in fact represents a region of nonvanishing chromoelectric field energy density .
the flavor string `` force '' is generated entirely by the quantum fluctuations , which manifest themselves in the diverging polarizability of the vacuum as @xmath42 .
the weakness of the color forces between the mesons is offset by the ease with which quark - antiquark pairs are created .
formally , this is reflected in the fact that only graphs which imply the existence of free flavor indices to be summed over have a chance to compensate the usual @xmath28-suppression of meson vertices by powers of @xmath4 in the numerator .
these are exactly the graphs in which a ( flavor singlet ) quark - antiquark pair is created or annihilated , i.e. a color string is broken into two or vice versa .
thus , only the graphs ( a ) and ( b ) in figure ( [ graphs ] ) survive ; only the form factors @xmath34 and @xmath35 enter the hamiltonian matrix ( [ sing1])-([singn ] ) .
in contrast to the one - flavor theory , where pair creation is not enhanced , and which is simply solved by t hooft s mesons , the bound state spectrum will now consist of superpositions of states containing different numbers of mesons . in the case of the one - flavor , adjoint color theory @xcite it was found that the low - lying spectrum is still very well described by a truncation to few quark - antiquark pairs .
this gave rise to the hope that the restriction to one- and two - meson states imposed in the present calculation would also give an adequate description of the first few bound states .
ultimately , this point will have to be checked by enlarging the hilbert space .
numerical evaluation and diagonalization of the hamiltonian matrix ( [ sing1])-([singn ] ) gives the low - lying spectrum as a function of @xmath0 . in the actual calculation ,
the two - meson states were supplied with a relative momentum wave function , i.e. the physical basis used was of the form ( cf . the convention introduced further above of definite ordering of momenta in the kets ) : @xmath43 were chosen to be ( properly normalized ) cosines ( with the plus sign ) and sines ( with the minus sign ) .
note that for @xmath44 only even ( cosine ) relative momentum wave functions are possible due to the bose symmetry .
note also that the form factors ( [ formf3 ] ) contain logarithmic divergences for zero quark mass which pose no problem if a continuous relative momentum wave function is introduced as above , whereas they make a direct discretization of the momentum difficult .
the basis used in the computation consisted of eight one - meson and 26 two - meson states .
bound states were identified by subtracting the expectation value of the kinetic energy from the energy eigenvalue and checking for a negative result .
the numerical calculation yielded four nontrivial zero modes besides the zeroth t hooft meson @xmath45 , which is an eigenstate of the hamiltonian for all @xmath0 , as can be easily seen from the three - meson form factor @xmath46 ( cf .
( [ formf3 ] ) ) .
furthermore , a spectrum of massive bound states , plotted in figure ( [ spec ] ) as a function of @xmath0 , was found .
some of the massive states disappear for intermediate values of @xmath0 and reappear with very similar energy and mesonic composition . in such cases the range where the state disappears is bridged in figure ( [ spec ] ) by the dotted lines to indicate the kinship between the spectral trajectories in different regions . the different classes of states obtained can each be described in more detail as follows .
the nontrivial zero modes are of the structure @xmath47 where the restriction to odd @xmath24 comes simply from the selection rules already discussed above in connection with equations ( [ sing1])-([singn ] ) .
it should be noted that the hilbert space used in the present calculation contained the first four relative momentum states made up of two ground state t hooft mesons and also the first four single - meson states of odd excitation number . in this basis , four nontrivial zero modes were found .
it thus seems natural to expect that the full theory in fact contains an infinity of zero modes of the form ( [ zmfo ] ) . for small @xmath0
, the one - particle admixture vanishes as @xmath0 .
such a scaling behavior is to be expected ; for @xmath48 , the hamiltonian becomes diagonal in the mesonic basis and thus the eigenstates are dominated by a single ( either one- or two - particle ) fock state , with @xmath0-suppressed admixtures obtainable in perturbation theory .
the same behavior is exhibited by the binding energies .
this feature of course also arises in the massive part of the spectrum to be discussed further below . at @xmath49 ,
the one - particle content ranges from 40% for states in which these one - particle admixtures are made up of low - lying t hooft mesons , downwards as more excited mesons participate .
the binding energies at @xmath0 are exactly @xmath50 ( up to the numerical accuracy of the calculation ) for three of the zero modes , rising to @xmath51 for the fourth . in the opposite limit of large @xmath0 on the other hand ,
the two - meson admixture is suppressed as @xmath52 and the states become mixtures of odd single t hooft meson states . as before , this scaling is easy to understand : for large @xmath0 , the two - meson two - meson interaction hamiltonian , which is positive ( this will be discussed further below ) , becomes dominant , and thus it is not favorable for a state to contain a high two - meson proportion . in order to maintain a meaningful balance between the different parts of the hamiltonian
, the two - meson amplitude must be proportional to @xmath53 , i.e. the two - meson content of the states must scale as @xmath52 .
again , this property of course persists for the massive part of the spectrum .
the binding energies of the zero modes must accordingly approach constants .
numerically one obtains two states with a binding energy of @xmath54 for large @xmath0 , one with binding energy @xmath55 , and one with @xmath56 .
it has not been possible to give analytical solutions for the nontrivial zero modes ; there must be subtle dynamical mechanism responsible for suppressing transitions from the single - meson admixtures in ( [ zmfo ] ) to the massive part of the spectrum .
it should be noted that some of the massive bound states to be discussed below also contain sizeable admixtures of odd single - meson states .
continuing on to the massive part of the spectrum , the states containing odd single - meson components appear as the third and the fourth massive state in the @xmath57 region ( cf . figure ( [ spec ] ) ) . in detail
, these two states however exhibit some differences : the fourth state only exists at low @xmath0 , and in the region shown consists of about equal parts of @xmath58 and @xmath59 ( the quantum numbers in the kets merely indicate the t hooft meson content , whereas the momentum quantum numbers have been suppressed ) .
for very low @xmath0 , though , the state becomes dominated by the @xmath59 part . by contrast
, the third massive state becomes dominantly a one - meson state in the @xmath48 limit , namely a @xmath60 .
as @xmath0 rises , it acquires admixtures of @xmath61 , @xmath59 and @xmath62 so that the @xmath60-component is still found with 55% probability at @xmath49 .
whereas all other states found are dominated by two to three fock components even in the transition region around @xmath49 , this state becomes a strong mixture .
this is especially pronounced where the state becomes unstable .
remarkably , as @xmath0 rises beyond unity , the two - meson content rises further to reach 96% at @xmath63 ; one third of this is the @xmath61 admixture . in accordance with the argument given further above
, this ultimately upsets the balance between potential and kinetic energies .
thus , the state is very weakly bound above @xmath64 and finally disappears around @xmath65 instead of increasing its one - particle content to give a well - defined @xmath66 limit .
the large range over which this state survives is remarkable , though . on the other hand ,
the rest of the massive states shown in figure ( [ spec ] ) contain even single - meson states . here
again , one finds states which become dominated by a two - meson fock component for @xmath67 ( namely the second massive state , which becomes predominantly a @xmath68 ) and others dominated by a one - meson component in this limit : the first massive state becomes a @xmath69 , the fifth an @xmath70 .
the one - particle fraction of the first massive state falls to about two thirds at @xmath49 due to admixtures of @xmath68 and then rises again to give a sensible @xmath66 limit .
remarkably , in this region , its mass falls to zero .
apart from this last feature , the fifth massive state behaves very similarly , acquiring a @xmath71 admixture of about one third in an intermediate regime of @xmath0 and then becoming again a pure @xmath70 .
the second massive state acquires a steadily rising admixture of @xmath72 ( about 25% at @xmath49 ) to rapidly become pure @xmath72 for large @xmath0 .
the state which is predominantly a @xmath73 for @xmath74 only appears at @xmath75 ; it is the fourth massive state in this region ( cf . figure ( [ spec ] ) ) .
this behavior is most probably due to the truncation of hilbert space , i.e. the state may be dominated at lower @xmath0 by a two - meson component not included in the calculation . summing up ,
the gross features of the spectrum can be characterized as follows : on the one hand , there are the zero modes , which consist predominantly of @xmath61 with growing admixtures of odd single - meson states as @xmath0 rises until these admixtures become dominant for large @xmath0 .
concomitantly , the massive states containing odd single - meson components become unstable as @xmath0 is increased , which is signalled by these states becoming strong mixtures without a dominant fock component . on the other hand ,
the massive states containing even single - meson admixtures display a well - defined large @xmath0 limit , where they are dominated by just these even single - meson states .
the lowest of these states becomes massless for @xmath66 .
the reason for this exceptional behavior is unclear ; all other states display a remarkably stable mass over a wide range of @xmath0 .
also this stability of the masses is a puzzle ; the binding energies of the states on their own vary quite strongly , as is evident from the values indicated in figure ( [ spec ] ) .
some comments are in order regarding the results presented above .
the first of these concerns the low @xmath0 limit .
the reader will have noticed that not all of t hooft s mesons appear in the bound state spectrum in this limit . to be precise , only @xmath76 , @xmath69 , @xmath60 and @xmath70 are present ; @xmath77 ,
@xmath72 , @xmath58 , @xmath73 and @xmath78 are missing .
these states of course do appear in the numerical diagonalization , with small admixtures of two - meson states , but with a positive expectation value of the potential energy .
thus they are not bound and must rather be viewed as resonances .
for very small @xmath0 , this distinction becomes meaningless .
an experimenter in 1 + 1 dimensions , working in a finite laboratory for a finite time , will observe all of t hooft s mesons as @xmath67 .
strictly speaking , however , if for fixed @xmath0 , the experimenter waits long enough , he / she will only find some of these mesons as stable particles . in the case of the higher mesons ,
the picture may be strongly influenced by the truncation of hilbert space , but the absence of the @xmath77 at least seems conspicuous . on the other hand ,
the bound states become very weakly bound for @xmath67 , so that the experimenter would indeed observe the @xmath68 ( this is the second massive state ) as two essentially independent ground state and first excited t hooft mesons , respectively , in a finite experiment . as a further comment
, it should be stressed that , to generate bound states in the model , it is crucial to combine one - meson and two - meson states .
diagonalizing the two - meson sector on its own does not give binding , i.e. the two - meson two - meson interaction matrix is positive , as mentioned above .
this seems plausible , for the following reason : if there was a pure two - meson state with a negative expectation value of the interaction , then one would have a state with arbitrarily negative overall energy by simply choosing @xmath0 suitably large such that the interaction dominates the kinetic part . on the other hand ,
the ground state ( vacuum ) energy of the theory has been defined to be zero , so that a state with negative energy should not be possible .
conversely , including one - meson states can very well generate binding through the three - meson vertex , and the energy is stabilized to remain positive by the two - meson two - meson vertex ; this is indeed precisely what happens , as mentioned above in the detailed discussion of the spectrum at large @xmath0 .
if one writes a mesonic hamiltonian in terms of the operators @xmath27 and their adjoints such that it correctly reproduces the hamiltonian matrix ( [ me11])-([me22 ] ) , it is the cubic term that may become negative , and the quartic term which stabilizes the theory . summing up , the large @xmath2
, large @xmath4 theory contains a considerably wealthier phenomenology than the one - flavor theory .
it represents a more sophisticated model for mesons , incorporating pair creation effects and thus mimicking the presence of a meson cloud , thought to be an important ingredient of three - dimensional hadrons . despite the complex behavior of the model ,
many of its features seem to be susceptible to numerical analysis , given the power of modern - day computers . in this respect ,
a significant improvement on the treatment presented here should be possible , since the latter only made use of two workstations . by using a mesonic basis , which already incorporates a sizeable part of the dynamics
, the bulk of the computation lies in the evaluation of the hamiltonian matrix rather than its diagonalization .
the matrix elements , however , can be evaluated completely independent of each other , making a large improvement in speed possible by simply using parallel processors . including sectors with a higher number of mesons
demands the computation of higher - dimensional integrals ; this will only be feasible beyond a certain point by employing monte carlo methods .
besides improvements of the numerical aspects , there are a number of other extensions of this work which seem to merit attention
. it would be interesting to understand in more detail the dynamics leading to binding .
this endeavor would profit greatly from some analytical understanding of the nontrivial zero modes .
especially the zero mode sector may to some extent be tractable analytically , at least in some approximation , since the form factors simplify considerably if some of the participating mesons are ground - state ones .
furthermore , the approach followed here becomes the most questionable in the high @xmath0 limit .
there , the coupling between sectors of different particle number is the strongest and the inclusion of fock components with three or more mesons could substantially alter the results
. it would be helpful to devise an approximation scheme which is particularly suited for this potentially interesting new limit , i.e. one which makes explicit use of the smallness of the parameter @xmath52 . with a view to approaching more realistic theories , it may be helpful to work out the perturbative corrections in @xmath28 .
finally , note also that the model is easily generalized to non - singlet flavor ( flavor spin one objects will resemble open strings ) as well as non - zero quark masses .
in fact , in the latter case , the hamiltonian matrix retains exactly the form ( [ me11])-([me22 ] ) ; only the t hooft wave functions entering the form factors and the corresponding meson masses are modified .
it is to be expected that quark masses suppress pair creation related effects .
thus , the physical spectrum should remain closer to the t hooft one for given @xmath0 .
the zero quark mass model considered in this work is the case furthest removed from the well - studied one - flavor theory , since it maximally allows for the pair creation processes which represent the essence of the model s nontrivial dynamics .
g.t hooft , nucl .
b72 ( 1974 ) 461 g.t hooft , nucl .
b75 ( 1974 ) 461 i.klebanov , lectures presented at the workshop of the graduiertenkolleg erlangen - regensburg ( germany ) on `` qcd and hadron structure '' , 9.6.-11.6.1992 at kloster banz ; gk - notes 3 - 92 g.veneziano , nucl .
b117 ( 1976 ) 519 d.gepner , nucl .
b252 ( 1985 ) 481 g.d.date , y.frishman and j.sonnenschein , nucl .
b283 ( 1987 ) 365 f.lenz , m.thies , k.yazaki and s.levit , ann .
( n.y . ) 208 ( 1991 ) 1 k.demeterfi , i.klebanov and g.bhanot , nucl .
b418 ( 1994 ) 15 ; + g.bhanot , k.demeterfi and i.klebanov , phys . rev .
d48 ( 1993 ) 4980 d.kutasov , nucl .
phys . b414 ( 1994 ) 33 m.engelhardt and b.schreiber , `` elimination of color in 1 + 1-dimensional @xmath3 '' , to appear in z. phys .
a i.bars and m.b.green , phys .
d17 ( 1978 ) 537 c.g.callan , n.coote and d.j.gross , phys . rev . d13 ( 1976 ) 1649 | qcd in 1 + 1 dimensions is examined in the limit of a large number of colors and flavors .
the hamiltonian matrix is given in a fock space spanned by t hooft meson states and , for the case of zero fermion mass , a submatrix is diagonalized numerically to give the low - lying spectrum as a function of @xmath0 .
pair creation effects generate bound states which are complicated mixtures of components of different meson number .
there are a number of nontrivial zero modes ; in the massive part of the spectrum some states tend to a well - defined @xmath1 limit while others become unstable and disappear . the masses of most states remain remarkably constant over a large range of @xmath0 . |
according to the reverberation mapping model ( e.g. * ? ? ? * ) , the broad emission line variations follow the ionizing continuum variations through the photoionization process . the variation correlations between broad - lines and continua
were observed with time lags in type 1 active galactic nuclei ( agns , see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the disturbances from the central engine in agns are transported with ionizing continua to broad - lines .
theoretical researches show that the jets can be ejected from inner accretion disk in the vicinity of the central black hole ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite indicated a disk - jet symbiosis with comparable power channelled through the disk and the jet .
correlations between radio powers and broad - line luminosities are found for agns and are regarded as evidence for the disk - jet symbiosis ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these previous results indicate that the disturbances in the central engine are likely propagated outwards along the jets .
observations show that dips in the x - ray emission , generated in the central engine , are followed by ejections of bright superluminal radio knots in the jets of agns and microquasars ( e.g. * ? ? ?
* ; * ? ? ?
the dips in the x - ray emission are well correlated with the ejections of bright superluminal knots in the radio jets of 3c 120 @xcite and 3c 111 @xcite .
the outbursts are physically linked to the ejections of superluminal knots ( e.g. * ? ? ?
* ) . then these outbursts of broad - line and jet emission might respond to the stronger disturbances in the central engine .
it is expected that there might be correlations with time lags between variations of broad - line and jet emission .
a method was proposed to connect the time lags , the size of broad - line region ( blr ) , and the location of jet emission for blazar 3c 273 ( * ? ? ?
* hereafter paper i ) .
the blrs are important to gamma - ray emission in blazars .
the gamma rays from blazars are generally believed to be from a relativistic jet with a small viewing angle @xcite .
the diffuse radiation field of blr could have a strong impact on the expected external compton ( ec ) spectrum of the most powerful blazars ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this strong impact rises from two factors .
one of them is the seed photons from the blr in the inverse compton scattering , and the seed photons significantly influence the ec spectrum .
the other is photon - photon absorption between the seed photons and the gamma - ray photons of the ec spectrum .
there is a underlying physical factor that constrains how much the above two factors influence on the gamma - ray spectrum .
the underlying factor is the location of gamma - ray emitting region relative to the blr .
if the gamma - ray emitting region is inside the blr , the gamma - ray spectrum will shift to higher energies and the gamma - ray luminosity will become larger due to the relativistic effects . at the same time , the photon - photon absorption becomes more significant as the emitting region goes deeper into the blr ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as the emitting region is outside the blr , the gamma - ray spectrum will shift to lower energies and the gamma - ray luminosity will become lower due to the relativistic effects . at the same time , the photon - photon absorption becomes insignificant as the emitting region keeps away from the blr .
thus it is valuable to connect the blr size with the location of jet emission
. it will be an important step for this connection to confirm correlation between variations of broad - lines and radio emission of jet and to estimate the relevant time lags . in the paper , we study this issue in the broad - line radio galaxy 3c 120 .
the structure of this paper is as follows .
section 2 presents method .
section 3 presents application to 3c 120 .
section 3 contains three subsections : subsection 3.1 presents constraint on time lag , subsection 3.2 data of 3c 120 , and subsection 3.3 analysis of time lag .
section 4 is for discussion and conclusions . in this work ,
we assume the standard @xmath3cdm cosmology with @xmath4 , @xmath5 = 0.27 , and @xmath6= 0.73 .
according to equation ( 7 ) in paper i , we have a relation between @xmath7 , @xmath8 , and @xmath9 @xmath10 where @xmath7 is the size of the blr , @xmath8 is the radio - emitting location of the jet , @xmath11 is the speed of light , @xmath12 is the travelling speed of disturbances down the jet , equivalent to the bulk velocity of jet @xmath13 , @xmath14 is the viewing angle of the jet axis to the line of sight , and @xmath15 is the measured time lag of the radio emission relative to the broad lines . from the velocity @xmath16 and the viewing angle @xmath14 , we have the apparent speed @xmath17 , which gives @xmath18 . substituting this expression of @xmath0 for the velocity term in equation ( 1 )
, we have @xmath19 from equation ( 2 ) , we have an expression to estimate @xmath20 @xmath21
3c 120 , at redshfit @xmath22 , has the one - sided jet with the apparent superluminal motion in the approaching jet @xcite .
the central supermassive black hole of this object has a mass of a few times @xmath23 @xcite .
the very long baseline array ( vlba ) imaging observations revealed a very complex radio jet in 3c 120 , such as superluminal components , stationary components , and trailing components in inner jet @xcite .
there are two stationary features in the radio jet , d and s1 @xcite .
the feature d is located at the base of the jet , and could be the core of the jet .
the feature s1 is most likely a standing shock formed in the jet . as the moving knots pass through the stationary feature s1
, these knots will produce outbursts .
the process was tested in the optical light curves of 3c 120 with peaks corresponding to the moving component passages through s1 .
it is not possible to verify whether the radio flux density reacted to the passage of moving features in the same fashion as the optical continuum .
the prominent superluminal feature o reaches it maximum flux density around 0.5 mas from the radio core @xcite .
the stationary feature s1 was identified at @xmath2 0.7 mas separated from the vlba 43 ghz core @xcite .
thus the radio outbursts may originate in the region from d to s1 in the inner jet . for 3c 120
, @xcite derived a distance of @xmath2 0.5 pc from the corona to the vlba 43 ghz core region using the average time delay between the start of the x - ray dips and the time of ejection of the corresponding superluminal knots . for 3c 111 ,
a distance of @xmath2 0.6 pc from the corona is estimated with the same method as in 3c 120 @xcite .
the 43 ghz vlba observations give the global parameters of the jet for 3c 120 , such as @xmath24@xcite .
these global parameters are widely accepted by other researches .
the 15 ghz components may correspond to the strongest events in the central engine @xcite . due to the optical depth
, the 15 ghz outburst may follow the 43 ghz outburst .
the emitting location of the 15 ghz outburst is estimated as the sum of the distance of 0.5 pc from the corona to the vlba core and the de - projected distance of 0.00.7 mas from the vlba core . under the standard @xmath3cdm cosmology we considered ,
an angular separation of 1 mas in the sky corresponds to a projected linear distance of 0.66 pc .
the de - projected distance of 0.7 mas from the radio core is equal to [email protected] pc/@[email protected] pc/@xmath27=1.32 pc .
the distance of 15 ghz outburst emitting region from the central engine is @xmath280.5 pc0.5 pc+1.32 pc=0.51.82
pc=1.635.93 ly . for 3c 120 , the h@xmath0 line has a blr size of @xmath29 light - days , i.e. @xmath30 ly @xcite . the apparent speeds of the moving components with well - determined motions are all within a range of @xmath31 @xcite . based on equation ( 3 ) , @xmath32 ly , @xmath33 , @xmath34 and @xmath8=1.635.93 ly , we derive @xmath35 .
this positive time lag means that the broad - line variations lead the radio variations .
we make use of the 15 ghz light curve with a higher sampling rate of 59 times per year .
this radio light curve is published in @xcite . for the h@xmath0 line
, @xcite presents a light curve with a very dense sampling of 20 times per month , and also @xcite presents a light curve of sampling 20 times per month in the reverberation mapping observations .
these light curves are presented in figure 1 , and are used to analyze the cross - correlation between them .
the z - transformed discrete correlation function ( zdcf ; * ? ? ?
* ) is used to analyze time lags characterized by the centroid of the zdcf .
the zdcf method is straightforward to determine whether there is a time lag between different light curves , and firstly it is applied to analyze the time lags .
the centroid time lag @xmath36 is computed by all the points with correlation coefficients not less than 0.8 times the maximum of correlation coefficients in the zdcf bumps closer to the zero - lag .
the uncertainties of each point in the zdcfs only take into account the uncertainties from the measurements by monte carlo simulation , and the uncertainties of time lags are underestimated ( see * ? ? ?
* hereafter paper ii ) .
thus we use the model - independent flux randomization / random subset selection ( fr / rss ) monte carlo method @xcite to re - estimate the time lags and their uncertainties in the cross - correlation results .
the fr / rss method is based on the discrete correlation function ( dcf ) method @xcite for the sparsely sampled light curves , and on the interpolated cross - correlation function ( iccf ) method for the densely sampled light curves .
the h@xmath0 line light curves with a very dense sampling are published in 2012 @xcite .
3c 120 is densely observed from 2008.0 to 2012.5 in the 15 ghz radio monitoring program with the 40 m telescope at the owens valley radio observatory @xcite .
firstly , it is obvious that the h@xmath0 line light curve in @xcite can be well matched with the outburst in the 15 ghz light curve from 2011.0 to 2011.3 as the line light curve is moved right by 0.32 yr ( see figure 1b ) .
the zdcf method and the fr / rss method are performed to investigate the correlation between these h@xmath0 line light curves and the 15 ghz light curve , and estimate the time lags from the correlation .
as these two line light curves are combined into one light curve , the calculated zdcf is presented in figure 2a .
the horizontal and vertical error bars in figure 2a represent the 68.3 per cent confidence intervals in the time lags and the relevant correlation coefficients , respectively .
there are positive and negative correlations ( see figure 2a ) .
the positive correlation has a time lag around 0.3 yr .
the h@xmath0 line variations lead the 15 ghz variations by a factor of @xmath2 0.3 yr .
this lag has the same sign as that lag estimated in section 3.1 .
this indicates that the positive correlation is reliable . for the positive correlation
, the zdcf method gives @xmath37 yr . the fr / rss method gives @xmath38 yr with a mean of peak correlation coefficients @xmath39 in monte carlo simulations of 10,000 runs ( see figure 2b ) .
this time lag is well consistent with that lag derived from the zdcf method .
the 15 ghz light curve shows a simple baseline superimposed with some outbursts and flares ( see figure 3a ) .
this baseline should influence the cross - correlation function between this radio light curve and the h@xmath0 line light curve .
thus we subtract this baseline assumed as a simple gaussian profile from the 15 ghz light curve .
it is obvious that the simple gaussian profile can well account for the underlying baseline in the 15 ghz light curve ( see figure 3a ) .
the cross - correlation function between the residual radio light curve and the h@xmath0 line light curve is calculated with the zdcf method and the fr / rss method .
the calculated zdcf is presented in figure 3b . the positive cross - correlation around 0.3
yr is re - confirmed , and the significance of cross - correlation is significantly improved with this residual light curve .
the zdcf method gives @xmath40 yr . the fr / rss method gives @xmath41 yr with a mean of peak correlation coefficients @xmath42 in monte carlo simulations of 10,000 runs ( see figure 3c ) .
this time lag is in excellent agreement with that lag derived from the zdcf method .
the broad h@xmath0 line variations lead the 15 ghz variations .
hereafter , @xmath36 is equivalent to @xmath9 .
we simplify the ionizing continuum region to be a point .
this simplification indicates that the disturbances will simultaneously be transported outwards with the ionizing continuum and the relativistic jet , i.e. it makes equations ( 1)(3 ) to be valid .
this simplification will influence the time lag @xmath9 .
the disturbances in the accretion disk will take a certain time to travel between their location of origin and the event horizon of the central black hole .
then the disturbances will take some time to pass through the ionizing continuum region to the event horizon . for 3c 120 ,
the uv ionizing continuum region is located at @xmath2 5@xmath43 from the black hole , where @xmath44 is the gravitational radius of the black hole @xcite .
the black hole mass is on the order of @xmath23 @xcite , and the size of @xmath2 5@xmath43 is on the order of @xmath45 ly .
if the disturbances are a thermal fluctuation propagating inward , it should have an effective speed @xmath46 @xcite , to cause a time delay of @xmath47 yr for the distance of @xmath2 5@xmath43 .
this time delay should be negligible compared with the time lag @xmath48 yr .
in estimation of the magnitude of @xmath9 in section 3.1 , we use the distance of @xmath2 0.5 pc from the vlba 43 ghz core region to the corona in the accretion disk - corona system , and ignore the size of corona .
the size of corona may influence @xmath9 .
the coronal radius is @xmath2 40@xmath43 @xcite .
the jet velocity near the central engine will be @xmath2 0.9c ( see paper i ) , and this radius of @xmath2 40@xmath43 will cause a time delay on the order of @xmath49 yr .
the time delay is negligible . from equation
( 2 ) , we have an expression to estimate @xmath8 from @xmath50 , @xmath14 , @xmath7 and @xmath9 @xmath51 for @xmath52 , @xmath53 , @xmath30 ly , and @xmath54 yr , we have @xmath55 ly from monte carlo simulations based on equation ( 4 ) .
thus we have @xmath56 for 3c 120 . @xmath57-large area telescope ( lat ) detected gamma rays from 3c 120 @xcite , and it was suggested that the gev emission of broad - line radio galaxies is most likely dominated by the beamed radiation of relativistic jets observed at intermediate viewing angles .
the radio and gamma - ray emitting regions are closely connected with each other , and there is @xmath58 between the radio - emitting position @xmath8 and the gamma - ray emitting location @xmath59 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is unclear for 3c 120 that @xmath60 or @xmath61 , which will significantly influence the gamma - ray spectrum produced in the ec processes .
the locations of gamma - ray emitting regions relative to the blrs are still an open and controversial issue in the researches on blazars .
there are three options for the issue .
the first option is that @xmath60 for the powerful blazars ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the second one is that @xmath62 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the third option is that the same source can display both behaviors .
that is , most of the time the dissipation region is inside the blr , but there could be some epochs when the gamma - ray emitting region drift outside the blr .
the very first idea was advanced in @xcite , and a very clear case with multi - wavelength coverage was recently found by @xcite .
the gamma - ray light curves and the corresponding broad - line light curves should shed light on this issue .
the chosen parameters are average for the jet , and do not correspond to any of the components identified in @xcite and @xcite .
the jet components will have different orientations and different velocities .
this seems to be an issue for the choice of jet parameters in equations ( 1)(4 ) .
the reverberation mapping model assumes the linear response of broad emission lines to ionizing continuum .
in fact , the line response is not linear .
it is most likely that the line and radio emission respond nonlinearly to the events in the central engine .
thus some weaker events in the central engine might not produce the correlative responses in the radio and broad - line variations .
both radio and broad - line emission may have good responses to the strongest events in the central engine , and their relevant outbursts should have good matching . there is a matching between these outbursts in the h@xmath0 line and 15 ghz light curves sampled densely .
however , the overall complexity of the light curves and the jet structure at radio bands may lead to difficulties of cross - identifying individual events in different bands .
thus it is difficult to identify the radio knots and radio outbursts corresponding to the broad - line outbursts .
we investigated the correlation and time lag between the radio and broad - line light curves by comparing their profiles and cross - correlating them .
the 15 ghz light curve is observed with the 40 m telescope at the owens valley radio observatory .
the 40 m telescope can not resolve the inner jet on the pc scales .
then the 15 ghz fluxes contain all the emission from the inner jet .
it is difficult to identify the component responsible for the 15 ghz outburst .
it is not possible to determine the relevant velocities of this component along the jet from the central engine to the emitting site of outburst .
thus the average velocity of primary components rather than trailing features will be a good proxy of the global velocity of component emitting outburst .
the viewing angle is the same case as the velocity .
the similar choice is accepted for the jet parameters for 3c 273 ( see paper i ) .
there are positive and negative time lags between radio variations and those of broad - lines h@xmath63 , h@xmath0 , and h@xmath64 due to the relative short coverage of these line light curves in 3c 273 .
the longer ultraviolet line light curves show that these broad - line variations lag the radio variations ( see paper ii ) .
thus the broad - line variations lag the radio variations .
a constraint of @xmath65 0.402.62 pc is set by these negative time lags ( see paper i ) .
the gamma - ray flares detected with @xmath57-lat set a limit of @xmath66 1.6 pc for 3c 273 @xcite .
the limit is marginally consistent with that constraint of @xmath65 0.402.62 pc .
the acceptance of the average parameters of inner jet is reasonable . in this paper
, we find correlation between the broad - line and radio variations with the zdcf method and the fr / rss method , and determine a positive lag for the broad - line radio galaxy 3c 120 .
the positive lag means that the 15 ghz variations lag the h@xmath0 line variations .
we derive @xmath67 yr from the fr / rss method .
this time lag is consistent with that estimated from the zdcf method .
monte carlo simulations give the radio - emitting location @xmath68 ly from this time lag , the average parameters of inner jet , and equation ( 4 ) .
it is reasonable to use the average parameters of inner jet in equations ( 1)(4 ) .
this reasonability is supported by the marginal agreement of our previous constraint of @xmath65 0.402.62 pc with the limit of @xmath66 1.6 pc from @xmath69-lat observations of gamma - ray flares in 3c 273 .
the underlying baseline of 15 ghz light curve significantly influences the correlation between broad h@xmath0 line and 15 ghz variations .
the subtraction of this baseline from the 15 ghz light curve can well improve the correlation .
the longer h@xmath0 line light curve well sampled may test the correlation .
the existence of this correlation is a key to connect the blr size with the emitting location of jet , and it is important to the gamma - ray emission of agns .
we are grateful to dr .
l. foschini for constructive comments and suggestions .
htl thanks the national natural science foundation of china ( nsfc ; grant 11273052 ) for financial support .
jmb acknowledges the support of the nsfc ( grant 11133006 ) .
htl thanks the financial support of the youth innovation promotion association , cas and the project of the training programme for the talents of west light foundation , cas .
blandford , r. d. , & mckee , c. f. 1982 , apj , 255 , 419 blandford , r. d. , & payne , d. g. 1982 , mnras , 199 , 883 blandford , r. d. , & rees , m. j. 1978 , in pittsburgh conf . on bl lac objects ,
a. m. wolfe ( pittsburgh , pa : univ .
pittsburgh press ) , p. 328
peterson , b. m. , wanders , i. , bertram , r. , hunley , j. f. , pogge , r. w. , & wagner , r. m. 1998a , apj , 501 , 82 peterson , b. m. , wanders , i. , horne , k. , collier , s. , alexander , t. , kaspi , s. , & maoz , d. 1998b , pasp , 110 , 660 sikora , m. , begelman , m. c. , & rees , m. j. 1994 , apj , 421 , 153 sikora , m. , moderski , r. , & madejski , g. m. 2008 , apj , 675 , 71 sikora , m. , stawarz , l. , moderski , r. , nalewajko , k. , & madejski , g. m. 2009 , apj , 704 , 38 | in the paper , we investigate correlation between broad - line and radio variations for broad - line radio galaxy 3c 120 . by the z - transformed discrete correlation function method and the model - independent flux randomization / random subset selection ( fr / rss ) monte carlo method
, we find that the broad h@xmath0 line variations lead the 15 ghz variations .
the fr / rss method shows that the h@xmath0 line variations lead the radio variations by a factor of @xmath1 yr .
this time lag can be used to locate the position of emitting region of radio outbursts in jet , on the order of @xmath2 5 light - years , from the central engine .
this distance is much larger than the size of broad - line region .
the large separation of the radio outburst emitting region from the broad - line region will observably influence the gamma - ray emission in 3c 120 . |
oscillatory interlayer exchange coupling in metallic magnetic multilayers causes the magnetizations of neighbouring magnetic layers separated by non - magnetic spacers to be spontaneously aligned ferromagnetically or antiferromagnetically depending on the thickness of the spacer material .
this phenomenon has been intensively studied both experimentally and theoretically over the last few years @xcite .
the relation between the periods , by far the most investigated feature of the oscillations , and the electronic structure of multilayered systems is a fundamental aspect of the mechanism responsible for the origin of this phenomenon .
in fact , the periods were originally shown to be dependent on the shape of the spacer fermi surface , where two distinct criteria were proposed for determining the periods with which the interlayer coupling oscillates @xcite . within the quantum well
theory the wave vectors yielding the effective oscillation periods correspond to extremal radii of the fermi surface in the direction perpendicular to the layers , or in other words to half the caliper measurements @xcite . on the other hand , in the rkky theory adapted to the layered geometry ,
the periods are associated with wave vectors perpendicular to the layers linking two points of the fermi surface with antiparallel velocities @xcite . despite the distinction between those geometrical criteria , the periods predicted by both theories , at least in their original formulations , agree in most cases .
when they do not coincide , such as , for example , when the lattice lacks reflection symmetry about the layer planes , suitable transformations of the fermi surface can be performed , which reestablish the correspondence between the quantum well and the rkky periods @xcite . recently however , fundamental oscillation periods not directly associated with the shape of the spacer fermi surface were obtained @xcite .
those periods , which correspond to higher order contributions to the exchange coupling , are no longer determined by the above geometrical criteria but by a set of more general conditions on the electronic structure of the system , although they are still indirectly dependent on the shape of the fermi surface .
these conditions provide a systematic way of determining the oscillation periods involved in the interlayer coupling . in order to test the applicability of those general conditions
, it is worth investigating them in situations where the spacer fermi surface , clearly a relevant factor in this matter , is variable .
one approach is to use an alloy spacer of variable composition although the fermi surface is no longer sharply defined .
also , it is well know that the scattering of electrons by random impurities strongly damps the oscillations .
in fact , the amplitude of the coupling is found to decay exponentially with the concentration of impurities @xcite .
an alternative to an alloy is a metallic superlattice whose fermi surface depends on the proportions and layer thicknesses of the constituent materials but where the oscillations are not damped by incoherent electronic scattering .
moreover , nature provides us with only a limited number of non - magnetic metals .
however , by combining distinct materials in superlattices the number of possible non - magnetic superlattices is infinite .
the purpose of this communication is to investigate the oscillation periods of the interlayer exchange coupling when the spacer is a non - magnetic superlattice .
it is shown that the alteration of the oscillation periods reflects the deformation of the corresponding fermi surface only when the coupling is investigated as a function of the number of superlattice cells separating the magnetic layers . in the simple model treated here ,
the oscillation periods of the coupling are written in terms of controllable parameters which are related to the composition of the superlattices , and which , accordingly , allow the periods to be tuned .
it is hoped that such a simple model reflects what occurs in realistic systems , and this may be confirmed by future experiments .
the sequence adopted in this work is described as follows . in order to make the work
self - contained , a brief summary of the general method for determining the oscillation periods in a multi - orbital tight - binding model @xcite is presented in section [ periods ] .
secondly , the method is shown to be applicable when the spacer is a superlattice , using the clear correspondence between orbital and intracell indices .
subsequently , the oscillation periods are obtained following the analysis of the electronic structure of the system and are explicitly written in terms of parameters associated with the superlattice composition .
a brief summary of the general method for determining the oscillation periods of the interlayer coupling is presented here . see reference @xcite for a comprehensive discussion of the method . for the sake of simplicity , we consider two parallel magnetic planes , labelled @xmath0 and @xmath1 , embedded in an infinite non - magnetic metal .
as far as the coupling as a function of the spacer thickness is concerned , the number of magnetic planes influences only the phase and amplitude of the oscillations and does not affect the periods @xcite .
since the periods are the main concern here , the restriction to two single magnetic planes does not pose any limitations on the results obtained .
the interlayer coupling @xmath2 is given by the difference in total energy between the antiferromagnetic and ferromagnetic configurations or , in other words , by the energy necessary to flip the magnetization in one of the planes .
based on the formalism introduced by dalbuquerque e castro _ et al _ , the coupling @xmath2 is written in terms of the one - electron green functions of the multilayered system by the following expression @xmath3 where @xmath4 . in the equations above @xmath5
is the propagator between planes @xmath0 and @xmath1 for electrons with spin @xmath6 in the ferromagnetic configuration of the system , @xmath7 is the fermi function , @xmath8 is a diagonal matrix in orbital indices representing the strength of the local exchange potential in the ferromagnetic layers , and the sum over @xmath9 is restricted to the two - dimensional brillouin zone . bearing in mind that the exchange splitting of the magnetic material is localized on the two planes @xmath0 and @xmath1 , the propagators @xmath10 can be rewritten as a function of the bulk spacer propagators @xmath11 and @xmath12 , i.e. , @xmath13 moreover , it can be shown that within the single - band model the dependence of @xmath14 upon those propagators is such that @xmath15 .
it is clear that the spacer thickness dependence of the interlayer coupling is entirely determined by the off - diagonal propagators @xmath11 and @xmath12 .
the investigation of the oscillation periods thus requires an analysis of @xmath11 as a function of the interplane distance . in the layered geometry ,
where @xmath9 represents the in - plane wave - vector , the expression for a general matrix element @xmath16 representing the propagation from an orbital @xmath17 at plane @xmath18 into an orbital @xmath19 at plane @xmath20 of an electron at the fermi level @xmath21 is given by @xmath22 where the integration variable @xmath23 is the wave vector perpendicular to the layers , @xmath24 is the interplane distance , @xmath25 describes the bulk spacer band structure , @xmath26 is the band index , and @xmath27 is the projection of the eigenstate @xmath26 into the orbital @xmath17 for a given @xmath28 . by changing the integration contour from the real axis to the boundaries of a semi - infinite rectangle in the complex plane whose
base lies on the real axis between @xmath29 and @xmath30 , the evaluation of the integral above is simplified to the determination of the poles of the integrand inside the appropriate contour and their respective residues .
the rectangle is in the upper - half plane for @xmath31 and in the lower - half plane otherwise .
clearly , the poles are given by the values of @xmath28 satisfying the condition @xmath32 , where @xmath33 . by adding the residues associated with the contributory poles , labelled @xmath34 , an analytical expression for @xmath35
is obtained , _
i.e. _ , @xmath36 where the matrix elements of the matrix @xmath37 are @xmath38 _ { q_\perp = q_j^s}\right\}^{-1}\,\ , .
\label{aa}\ ] ] the coefficients @xmath39 depend neither on @xmath18 nor on @xmath20 and the only dependence on the distance between the planes is inside the argument of the exponentials . it should be stressed that this simple method for obtaining the one - electron propagators is exclusive to the one - dimensional case , and therefore for fixed values of @xmath40 . the exponentials in eq.([glm ] ) are independent of the orbital indices , which means that all matrix elements oscillate with the same periods .
note that in calculating eq.([aa ] ) it was assumed that the coefficients @xmath41 have neither singularities nor branch points inside the integration contour .
in this way , the coefficients are evaluated just at the poles @xmath34 . in general , branch cuts associated with the coefficients
@xmath41 may exist , but because they may be chosen not to intersect the real axis , the factor @xmath42 in the integrand of eq.([integral ] ) ensures that the corresponding contribution is negligible for large @xmath43 .
therefore , the expression above is correct in the asymptotic limit of large interplane distance @xcite .
it is clear from eq.([glm ] ) that @xmath35 is an oscillatory function of the distance between the planes and it oscillates with different periods , each one associated with its respective wave vector @xmath34 .
for this reason it is interesting to look at the physical significance of the poles . the poles may be real or complex , but we note that only the real poles contribute significantly to the coupling .
this is because complex poles , when inserted in eq.([glm ] ) , damp the oscillations and are important only in cases of very thin spacers .
real poles , which are obtained from the electronic structure of the bulk spacer , are just the perpendicular coordinates of the fermi surface for a fixed value of @xmath9 . in other words , they are the perpendicular components of the wave vectors with which electrons of energy @xmath21 propagate across the spacer . having investigated @xmath35 as a function of the interplane distance , it is worth stressing that the functions @xmath44 must oscillate with the same periods . when expanded in a fourier series , taking into account the possibility of quasi - periodicity of @xmath11 @xcite , the function @xmath44 is written as @xmath45 where @xmath46 are the fourier coefficients determining the amplitude of the oscillations and depend on the exact form of the function @xmath44 .
the indices @xmath47 are integers and there are as many as the number of poles @xmath34 . note that the periods are given by @xmath48 and do not depend on the exact form of @xmath49 , since they are obtained from the bulk spacer band structure . because the indices @xmath50 run over the set of integers , there is always an infinite number of periods defined by eq.([f ] ) , even for a small number of poles @xmath34 . in spite of the somewhat congested notation of the general expression above , in practice
there are only a few indices involved due to the restricted number of real solutions crossing the fermi level .
in fact , the number of solutions is related to the number of sheets of the fermi surface . when eq.([f ] ) is inserted into eq.([j ] ) , the two - dimensional integral can be evaluated through the stationary phase approximation , valid in the asymptotic limit of large spacer thickness @xcite . in doing so
, the periods of the coupling are identified by selecting the periods of @xmath44 associated with wave vectors satisfying the following condition , @xmath51 \ , = 0 \,\ , , \label{grad}\ ] ] where @xmath52 is the two - dimensional gradient in @xmath40-space .
the equation above gives the necessary conditions for constructive interference between electrons across the spacer .
the otherwise destructive interference does not contribute to the coupling .
eq.([grad ] ) simply selects the points across the two - dimensional brillouin zone which yield the effective periods with which the coupling oscillates .
as pointed out in reference @xcite , although the wave vectors @xmath34 are directly related to the fermi surface , the general wave vectors @xmath53 are not .
this is a result of the interference between incommensurate wave vectors @xmath34 giving rise to a new set of periods not directly obtained from the structure of the fermi surface .
consequently , geometrical criteria can not be used in general to select the oscillation periods . however , for fermi surfaces with a single sheet , where the off - diagonal propagators oscillate as a function of the interplane distance with only one period , the selection rules given by eq.([grad ] ) coincide with the geometrical criteria of the rkky theory @xcite .
the only difference in this case is that eq.([grad ] ) determines the respective harmonics , in addition to the fundamental rkky periods . in summary ,
the determination of the oscillation periods results from the analysis of the bulk spacer off - diagonal propagators where the relevant real wave vectors , under the stationary phase condition , are obtained from the band structure .
when dealing with multi - sheet fermi surfaces , the possibility of interferences between incommensurate wave vectors may introduce new features in the oscillations , which are not present in the case of single - period oscillations of the propagators .
having summarized the method for determining the oscillation periods of the coupling , it is worth stressing that no particular form for the spacer material was assumed .
the only information required is the bulk band structure which makes possible the investigation of the coupling across arbitrary spacers .
spacers consisting of superlattices of distinct non - magnetic materials will be investigated here .
more specifically , the material into which the magnetic planes are embedded is composed of two non - magnetic metals @xmath54 and @xmath55 with respective thicknesses @xmath56 and @xmath57 periodically alternated , as schematically displayed in figure [ superlattice ] . in order to make a distinction , these structures will be hereafter named superlatticed spacers , in contrast to the conventional single - layered spacers .
the interlayer coupling across superlatticed spacers can be looked upon as a function of different thicknesses , namely @xmath56 , @xmath57 , @xmath58 and @xmath59 , where @xmath59 represents the number of cells separating the magnetic planes .
each cell is formed by a set of @xmath56 planes of metal @xmath54 and @xmath57 planes of metal @xmath55 .
as far as the oscillation periods are concerned , the coupling across superlatticed spacers as a function of @xmath56 or @xmath57 does not present new features when compared to the coupling across conventional single - layered spacers as a function of @xmath56 or @xmath57 , respectively .
in other words , if a single - layered spacer of material @xmath54 yields an interlayer coupling with period @xmath60 and likewise @xmath61 is the period for the coupling across a single - layered spacer of material @xmath55 , the coupling across a superlatticed spacer of materials @xmath54 and @xmath55 will oscillate with period @xmath60 when seen as a function of @xmath56 and with period @xmath61 when seen as a function of @xmath57 .
however , when looked upon as a function of @xmath59 , for fixed values of @xmath56 and @xmath57 , the coupling oscillates with periods different from @xmath60 and @xmath61 .
although the superlattice fermi surface is different from the individual fermi surfaces of materials a and b separately , it is only when investigated as a function of @xmath59 that the deformation of the fermi surface is translated into a change in the periods . following the steps of the previous section
, the bulk spacer band structure must be calculated . for the sake of simplicity
, the single - band tight - binding model is considered assuming a simple cubic lattice structure for the superlattice grown along the ( 100 ) direction . for the two given materials @xmath54 and @xmath55 , the in - plane energies are labelled @xmath62 and @xmath63 , and the interplane hoppings are @xmath64 and @xmath65 .
the hopping between planes @xmath54 and @xmath55 is represented by @xmath66 .
we recall that the tight - binding parameters above depend on the in - plane wave vector @xmath9 through @xmath67 $ ] and @xmath68 , where @xmath69 is either @xmath54 or @xmath55 , @xmath70 is the on - site energy and @xmath71 is the hopping between nearest - neighbour sites in material @xmath69 .
note that @xmath72 is perpendicular to the layers whereas @xmath73 and @xmath74 are the coordinates of @xmath9 .
because each set of @xmath58 planes is periodically repeated , where @xmath75 , the system can be represented by cells of @xmath58 planes equally spaced . consequently , the tight - binding hamiltonian of the infinite multilayered system , when written in the appropriate basis , becomes a @xmath76 matrix given by @xmath77 & \left[{\cal h}_{ab}\right ] \\
\left[{\cal h}_{ba}\right ] & \left[{\cal h}_b(n_b)\right ] \\
\end{array } \right ) \,\ , , \label{h(nxn)}\ ] ] where @xmath78 $ ] and @xmath79 $ ] are square matrices of order @xmath56 and @xmath57 , respectively .
the sub - matrices @xmath78 $ ] and @xmath79 $ ] are explicitly given by @xmath80 = \left .
\left ( \begin{array}{c c c c } \epsilon_\beta\ ; & t_\beta \ ; & 0\ ; & 0\ ; \\ t_\beta\ ; & \epsilon_\beta\ ; & t_\beta\ ; & 0\ ; \\ 0\ ; & t_\beta\ ; & \epsilon_\beta\ ; & t_\beta\ ; \\ 0\ ; & 0\ ; & t_\beta\ ; & \epsilon_\beta\ ; \\ \end{array } \right ) \,\,\ , \right\ } \,\ , { n_\beta } \,\,\,\,.\ ] ] note that @xmath81 $ ] is a tri - diagonal matrix equivalent to the single - band tight - binding hamiltonian of a slab consisting of @xmath82 planes .
the sub - matrices off the diagonal in eq.([h(nxn ) ] ) are @xmath83 and @xmath84 matrices such that @xmath85 = t_{ab } \underbrace { \left ( \begin{array}{c c c c } 0\ ; & 0 \ ; & 0\ ; & e^{- i \,q_\perp \ , n \ , d } \\ 0\ ; & 0 \ ; & 0\ ; & 0\ ; \\ 1\ ; & 0\ ; & 0\ ; & 0\ ; \\ \end{array } \right ) } _ { n_b } \left .
\begin{array}{c } \\ \\ \\ \end{array } \right\ } n_a \label{hab}\ ] ] and @xmath86 = \left[\left({\cal h}_{ab}\right)^\dagger\right]$ ] . in the matrix directly above @xmath28 is the wave vector perpendicular to the layers and @xmath24 is the interplane distance , which is assumed to be the same for both materials . the basis producing such matrices results from a fourier transform of the localized - atomic - orbital basis along the direction perpendicular to the layers .
since the cells of @xmath58 planes are equally spaced by @xmath87 , the associated wave vector @xmath28 is in the range @xmath88 $ ] , that is , the corresponding brillouin zone along this direction is reduced by a factor @xmath58 .
the tight - binding hamiltonian above is equivalent to the @xmath58-band hamiltonian of a one - dimensional system . in other words ,
the @xmath58 intracell planes may be considered as orbitals of a linear chain whose nearest neighbour elements are @xmath89 apart .
having established this correspondence , hereafter intracell and orbital indices will be mentioned with no distinction between them .
following section [ periods ] , the oscillation periods result from the analysis of the bulk spacer off - diagonal propagators as a function of the distance between the planes .
note however that in the case treated here the effective interplane distance must be the distance between the cells , _
i.e. _ , @xmath90 .
hence , the off - diagonal propagators as a function of the number of cells @xmath59 are the ones to be investigated .
these propagators depend on the band structure of the system , as clearly shown in eq.([integral ] ) . the original tight - binding hamiltonian is partially diagonalized by both in - plane and perpendicular fourier transforms , the remaining problem being the determination of the eigenvalues of the @xmath76 matrix in eq.([h(nxn ) ] ) .
therefore , the information on the oscillation periods of the coupling in a superlatticed system depends on the diagonalization of this matrix , which will be presented in the next section . to maintain the correspondence between the intracell planes of the superlattice and orbital indices of the linear chain
, one has to assume that in the case of superlatticed spacers the magnetic planes embedded in the non - magnetic system must be formed by two separate magnetic cells .
however , for the sake of simplicity and with no loss of generality , we consider only two magnetic planes . if we label the magnetic planes with the intracell indices @xmath91 in the cell @xmath0 and @xmath92 in the cell @xmath59 , the relevant green functions to be calculated are the off - diagonal propagators @xmath93 , i.e. , @xmath94 in addition to the fact that @xmath91 and @xmath92 are intracell indices , the only difference between the equation above and eq.([integral ] ) is the interplane distance @xmath95 which modifies the integral limits .
similarly to section [ periods ] , the integral above is evaluated through the residues method , where the determination of the poles follow the diagonalization of the hamiltonian .
making use of the particular symmetry of the matrices displayed above , the determinant of @xmath96 in eq.([h(nxn ) ] ) can be written in terms of the sizes of its sub - matrices @xmath56 and @xmath57 yielding a simple analytical expression for the oscillation periods of the one - electron propagators .
the energy eigenvalues @xmath97 are solutions of the equation @xmath98 , where @xmath99 is the identity matrix . in this case
the sub - matrices along the diagonal become @xmath100 = \underbrace { \left ( \begin{array}{c c c c } \epsilon_\beta - \omega\ ; & t_\beta \ ; & 0\ ; & 0\ ; \\ t_\beta\ ; & \epsilon_\beta - \omega\ ; & t_\beta\ ; & 0\ ; \\ 0\ ; & t_\beta\ ; & \epsilon_\beta - \omega\ ; & t_\beta\ ; \\ 0\ ; & 0\ ; & t_\beta\ ; & \epsilon_\beta - \omega\ ; \\ \end{array } \right ) } _ { n_\beta } \,\,\,\,,\ ] ] whereas the block matrices off the diagonal are still given by @xmath101 $ ] and @xmath86 $ ] .
the determinant of a @xmath76 square matrix is a sum of @xmath102 terms .
those terms are made of products of @xmath58 matrix elements , following all their possible permutations . because the sub - matrices on the diagonal have only two distinct non - zero elements , the evaluation of their determinants by counting the number of combinations allowed by the possible permutations is not particularly difficult and can be expressed in term of the sizes @xmath82 .
the solution of this combinatorial problem is given by @xcite @xmath103 where @xmath104 is an integer , the upper limit of the sum is the integer part of @xmath105 and @xmath106 is the binomial coefficient .
note that the expression above is valid for symmetric tri - diagonal matrices with similar elements along each of the diagonals .
bearing in mind that @xmath107 is equivalent to the slab hamiltonian , the values of @xmath97 satisfying @xmath108 through eq.([det - tri ] ) are just the associated eigenvalues . the hamiltonian @xmath96 in eq.([h(nxn ) ] ) is almost tri - diagonal .
the matrix @xmath96 is not tri - diagonal due to two of its elements , namely @xmath109 and @xmath110 , which are non - zero .
therefore , an expression for the determinant of @xmath96 is not expected to be very different from the one in eq.([det - tri ] ) .
in fact , by considering the only two non - zero elements of the submatrices in eq.([hab ] ) and recounting the number of possible permutations , the determinant of the matrix @xmath96 can be expressed in terms of the determinants of the matrices @xmath111 and @xmath112 . in the similar case of @xmath113 , the determinant is given by @xmath114 note that the equation above is written in terms of lower - order determinants of eq.([det - tri ] ) and is an analytical expression for @xmath115 in terms of its sizes @xmath56 and @xmath57 .
it is instructive to point out that when @xmath116 , that is , when a single - layered spacer is considered , the usual eigenvalues @xmath117 of an infinite linear chain are obtained from eq.([deth ] ) through the condition @xmath118 .
the eigenvalues @xmath97 of the hamiltonian @xmath96 are solutions of the polynomial equation defined by eqs.([deth ] ) and ( [ det - tri ] ) through the condition @xmath98 .
to illustrate this point , figure [ bandn3n6 ] displays the band structure of a superlatticed system consisting of two materials @xmath54 and @xmath55 whose tight - binding parameters are arbitrarily chosen as @xmath119 , @xmath120 and @xmath121 .
two cases are shown in figure [ bandn3n6 ] , namely @xmath122 and @xmath123 . in each case a number of possible values for @xmath56 and @xmath57 is presented . due to the equivalence between the single - band - cellular and multi - band linear hamiltonians , it is clear in all the cases that , for a given @xmath56 , there are @xmath58 solutions per @xmath28 forming the respective bands .
all bands are separated by energy gaps except for the ones in figure [ bandn3n6]a corresponding to @xmath124 and @xmath125 .
those two cases correspond to single materials @xmath55 and @xmath54 , respectively .
since we are interested in determining the oscillation periods of the one - electron propagators , what is needed from the polynomial equation defined by eqs.([deth ] ) and ( [ det - tri ] ) is the wave vector @xmath28 with which electrons with energy @xmath97 propagate across the superlatticed spacer , that is , the poles @xmath28 of the integrand in eq.([off ] ) . whereas the determination of the eigenvalues requires finding the zeros of an @xmath58-th degree polynomial in @xmath97 , the calculation of the wave vector @xmath28 for a given energy @xmath97 is straightforward due to the simple dependence of eq.([deth ] ) on @xmath28 .
the advantage of writing the expression of the determinant in terms of the number of planes @xmath56 and @xmath57 becomes clear , for the oscillation periods can be written in terms of those parameters .
furthermore , bearing in mind that the wave vectors @xmath28 are the perpendicular components of the spacer fermi surface , we can describe how the fermi surface is deformed with the composition of the cells .
the wave vector @xmath28 is such that @xmath126 where , for simplicity , the interplane hopping @xmath66 was assumed to be the geometrical average between @xmath64 and @xmath65 . note that for a given energy @xmath97 ( and @xmath9 ) there is only one value of @xmath127 defined .
this indicates that the off - diagonal propagators oscillate as a function of the number of cells @xmath59 with a single period . in this case
the fermi surface is deformed as a function of @xmath56 and @xmath57 but the number of periods is left unchanged .
this can be understood through the size reduction of the brillouin zone along the perpendicular direction @xmath72 , folding the fermi surface back into the zone when it crosses the boundaries in @xmath128
. here , given two materials @xmath54 and @xmath55 , we shall look at the change of the fermi surface for different thicknesses @xmath56 and @xmath57 . still assuming the tight - binding parameters used in the evaluation of figure [ bandn3n6 ] , the cross sections of the superlatticed spacer fermi surface for @xmath129 are shown in figures [ fsn3 ] and [ fsn6 ] .
two cases are displayed , namely @xmath122 with @xmath130 in figure [ fsn3 ] and @xmath123 with @xmath131 in figure [ fsn6 ] , where @xmath21 represents the fermi level . in all those figures the vertical axes are along the direction perpendicular to the layers whereas the horizontal direction represents @xmath74 .
the fermi surfaces are drawn inside the first brillouin zone represented by the rectangles , which are reduced along the @xmath132 direction by a factor @xmath58 .
note that figures [ fsn3]a ( @xmath124 ) and [ fsn3]d ( @xmath133 ) correspond to the bulk fermi surfaces of the pure materials @xmath55 and @xmath54 , respectively .
the choice of different fermi levels for @xmath122 and @xmath134 results from the analysis of figure [ bandn3n6 ] , using the criterion of selecting a fermi level which crosses the most number of bands .
in fact , in figure [ bandn3n6]a @xmath130 yields solutions for all possible values of @xmath56 and @xmath57 whereas in figure [ bandn3n6]b @xmath131 intersects the curves @xmath135 , @xmath133 and @xmath136 . as discussed in section [ periods ] , in the case where only one period exists in the propagators
, the rkky geometrical picture coincides with the more general stationary phase conditions of eq.([grad ] ) for selecting the effective oscillation periods of the coupling . due to the equivalence between @xmath73 and @xmath74 , it is clear from figures [ fsn3 ] and [ fsn6 ] that all wave vectors @xmath28 satisfying the stationary phase conditions are located at the origin of the 2-dimensional brillouin zone , _
i.e. _ , @xmath137 .
therefore , the oscillation period of the propagator @xmath138 is associated with the value of @xmath28 determined by the solution of eq.([cosq ] ) for @xmath139 . according to eqs .
( [ j ] ) and ( [ gg ] ) , the coupling is a function of the propagators @xmath140 between the magnetic planes . in the case
considered here , @xmath141 and @xmath142 are the propagators involved .
we recall that in this case the coupling depends on the product @xmath143 and bearing in mind that both propagators oscillate with the same periods , the coupling oscillates with an effective wave vector which is twice that for @xmath138 .
the corresponding periods of the coupling are then given by @xmath144 .
the periods depend on the intra - cell composition determined by @xmath56 and @xmath57 as can be seen by the change in caliper of the extremal wave vectors spanning the fermi surface in the centre of the brillouin zones . following eq.([cosq ] ) , the periods are displayed in table [ tabelperiods ] and the product @xmath145 as a function of @xmath59 for different intracell parameters and for @xmath146 is shown in figure [ oscillations ] .
note the agreement between the periods in figure [ oscillations ] and the ones in table [ tabelperiods ] .
the results presented above indicate that the oscillation periods of the interlayer coupling across superlatticed spacers are related to the shape of the fermi surface when the coupling is investigated as a function of the number of cells of the superlattice .
therefore , a change in the composition of the superlattice enables one to deform the corresponding fermi surface and consequently tune the oscillation periods of the coupling .
moreover , although the present analysis is not concerned with the strength of the oscillation periods , it should be mentioned that the change in intra - cell parameters not only alters the relevant wave vectors but also the respective curvatures of the fermi surface .
bearing in mind the dependence of the oscillation amplitude on such curvatures , the contribution of periods may be enhanced by the variation of the parameters @xmath56 and @xmath57 .
non - rkky periods , for instance , which in some cases contribute weakly to the oscillations @xcite , could become more relevant through the variation of those parameters .
it is hoped that the simple case treated here reflects what occurs in real systems , which may be confirmed by future experiments .
alloys that , under certain critical temperatures , present ordered phases may be used to test the veracity of the present claims . in the ordered phase , cu - zn alloys , for instance ,
have a bcc structure consisting of interpenetrating simple cubic sublattices @xcite . in this case , cells formed by one plane of cu and one plane of zn are periodically repeated , corresponding to a possible realization of the system proposed in this communication .
the lattice parameters of the ordered cu - zn alloy and that of fe are similar , which may allow them to be grown in layered structures .
although the oscillation periods of the coupling can not be tuned in this system , it is suggested that the periodicities associated with the corresponding fermi surface will become evident only when plotted as a function of even ( or odd ) number of spacer layers .
alternatively , by growing superlattices of non - magnetic metals , _
e.g. _ noble metals , in wedge - shaped systems @xcite one can change the intracell composition of the system to test whether the periods of the coupling can be tuned , as the present calculations indicate .
i wish to acknowledge professor d. m. edwards , dr . r. b. muniz , dr . j. dalbuquerque e castro and dr .
r. wright for their invaluable contributions to this work . also , i am grateful to cnpq of brazil for financial support . b. heinrich and j. f. cochran , adv . phys . * 42 * 523 ( 1993 ) b. heinrich and j. a. c. bland , ultrathin magnetic structures - vol.2 ( 1994 ) s. s. parkin , n. more and k. p. roche , phys .
lett . * 64 * , 2304 ( 1990 ) d. m. edwards , j. mathon , r. b. muniz and m. s. phan , phys .
67 * , 493 ( 1991 ) .
p. bruno and c. chappert , phys .
lett . * 67 * , 1602 ( 1991 ) p. bruno and c. chappert , phys .
b * 46 * , 261 ( 1992 ) d. m. edwards , j. mathon and r. b. muniz , phys . rev .
b * 50 * 16066 ( 1994 ) m. s. ferreira , j. dalbuquerque e castro , d. m. edwards and j. mathon , j. magn
. magn . mater . *
154 * , l1 ( 1996 ) m. s. ferreira , j. dalbuquerque e castro , d. m. edwards and j. mathon , j. phys . : condens .
matter * 8 * , 11259 ( 1996 ) a. t. costa jr , r. b. muniz , j. dalbuquerque e castro , m. s. ferreira and j. mathon , phys .
b * 55 * , 3724 ( 1997 ) j. dalbuquerque e castro , m. s. ferreira and r. b. muniz , phys . rev .
b * 49 * , 16062 ( 1994 ) s. n. okuno and k. inomata , phys .
lett . * 70 * , 1711 ( 1993 ) j. dalbuquerque e castro , j. mathon , murielle villeret and d. m. edwards , phys .
b * 51 * , 12876 ( 1995 ) j. dalbuquerque e castro , unpublished
. j. dalbuquerque e castro , j. mathon , murielle villeret and a. umerski , phys .
b * 53 * , 13305 ( 1996 ) r. a. brualdi , combinatorial matrix theory ( 1991 ) c. kittel , introduction to solid state physics , 7th ed . , ( 1996 ) i. goldfarb , e. zolotoyabko and d. shechtman , j. appl . phys . * 74 * , 2501 ( 1993 ) f. g. aliev , e. kunnen , k. temst , k. mae , g. verbanck , v. v. moshchalkov and y. bruynseraede , phys .
lett . * 78 * , 134 ( 1997 ) | the oscillation periods of the interlayer exchange coupling are investigated when two magnetic layers are separated by a metallic superlattice of two distinct non - magnetic materials . in spite of the conventional behaviour of the coupling as a function of the spacer thickness , new periods arise when the coupling is looked upon as a function of the number of cells of the superlattice .
the new periodicity results from the deformation of the corresponding fermi surface , which is explicitly related to a few controllable parameters , allowing the oscillation periods to be tuned . psbox |
figures [ fig:1],[fig:2 ] show diagrams of processes in pump femtosecond laser pulse ( fslp ) action on metal .
the three time slices `` ei '' , @xmath3 and @xmath4 in fig .
[ fig:1 ] correspond to the following non - equilibrium processes : ( e - i ) the electron - ion thermal relaxation , ( m ) the melting of an overheated crystal lattice , and ( c ) the cavitation decay of a metastable state .
duration of fslp @xmath5 is shorter than characteristic times of these three processes .
they have very various time scales from subpicoseconds to nanoseconds .
the electron overheating @xmath6 starts from @xmath7 when a fslp arrives @xcite and disappears at @xmath8 when temperatures @xmath9 equilibrate @xmath10 for al at our intensities ) .
the time is reckoned from the maximum of pump fslp in fig . [ fig:1 ] . since arriving of the pump to a target
the conductivity electrons become much hotter than the ions .
two - temperature ( 2 t ) matter with hot electrons transits to a peculiar state with thermodynamic and optical characteristics different from one - temperature ( 1 t ) case . in 2 t
there are appearance of excesses of electron energy and pressure above equilibrium 1 t ones .
also there are changes in elastic moduli and band structure . in semiconductor lattice
the binding forces become weaker with increase of @xmath11 while in metals situation is opposite .
large changes in optics of au at high @xmath12 result from excitation of _ d_-electrons . on account of the ion heat capacity @xmath13 ( thermal `` inertia '' of a lattice ) the beginning of melting @xmath14
is delayed relative to the instant @xmath15 where @xmath16 and @xmath17 are the melting temperature and e - i energy exchange rate .
the pump fslp @xmath18 and a chain of kinetic or transient processes ( 1 ) `` ei '' , ( 2 ) @xmath19 and ( 3 ) @xmath20 initiated by absorption of the pump . ]
it has been known that : \(i ) for metals and semiconductors the fluences @xmath21 near the ablation threshold @xmath22 are significantly higher than the melting threshold @xmath23 @xcite . in this sense
the lennard - jones ( lj ) case is an exception @xcite .
lj @xmath24 vacancy migration to cohesion energy ratio and @xmath25 triple to critical temperature ratio are large in comparison with metals and semiconductors . while @xmath26 is approximately the same for all three groups .
therefore it is not suprising that in the lj case @xmath27 in metals for @xmath28 the molten layer is approximately as thick as heated @xmath29 when stretching stress overcome material strength a cavitation ( fragmentation of liquid ) begins inside molten metal at the instant @xmath30 shown in fig .
[ fig:1 ] . in lj near threshold @xmath31 spallation ( fragmentation of solid ) starts in deformed crystal @xcite .
above threshold @xmath32 the lj spallation transforms to cavitation as molten lj layer becomes thicker and the fragmentation zone transits from solid to molten lj .
\(ii ) the electronic heat conduction wave `` ehc '' in fig .
[ fig:2 ] is _ supersonic _ within the 2 t slice `` ei '' shown in fig .
[ fig:1 ] @xcite .
these facts ( i , ii ) result in the isochoric heating and stress confinement @xcite .
estimates of the ehc speed are : @xmath33 @xmath34 is a thermal diffusivity , @xmath35 is a mean free path , @xmath36 is a collision frequency , @xmath37 is the fermi velocity .
therefore the mach number of ehc wave is high @xmath38 up to a few picosecond . within the time period
@xmath39 the `` ehc '' creates a heated layer @xmath40 thick in al and @xmath41 thick in au @xcite .
the @xmath42 is much thicker than acoustic penetration depth @xmath43 at @xmath44 as illustrated in fig .
[ fig:2 ] . as a result of ( i ) and ( ii )
there is a volume non - equilibrium melting in the slice @xmath45 @xcite with formation of overheated solid grains surrounded by melt . at the slice @xmath46 @xmath47 in fig .
[ fig:1 ] the heat wave velocity becomes much lower than @xmath48 and then the well - defined melting / recrystallization front is formed @xcite . at @xmath49 the melt layer covering a residue @xcite of the target is completely solidified .
electrons are heated up by a pump in a skin layer of metal ( the rectangular `` skin '' ) .
the `` ehc '' supersonically ( comp . the heat @xmath42 and sonic propagations @xmath50 at @xmath51 carries heat from the `` skin '' layer into the bulk of the target along the horizontal arrow ( it shows the heat flux ) forming the heated layer @xmath29 the energy of hot electrons is transferred to ions in fast process of `` e - i heating '' ( the vertical arrows ) @xcite . as a result
the crystal lattice is overheated to @xmath52 above the solidus temperature .
the solidus and liquidus temperatures at @xmath53 are 1.2 kk , 1.48 kk ( al ) , 1.9 kk , 2.05 kk ( au ) .
they are significantly higher than the melting or the triple point temperatures 933 k ( al ) , 1337 k ( au ) corresponding to melting at low pressure . ]
in lagrangian variables the 2 t hydrodynamics ( 2t - hd ) equations @xcite are @xmath54 for kinematics and mass conservation ,
@xmath55 for a force balance , and two thermal equations @xmath56}{\partial t } & = & \frac{\partial}{\partial x^o } \left(\frac{\rho\kappa(\rho , t_e , t_i)}{\rho^o } \frac{\partial t_e}{\partial x^o } \right ) - \nonumber \\ -p_e\frac{\partial u}{\partial x^o } & - & \frac{\rho^o}{\rho } \alpha ( t_e - t_i ) + \frac{\rho^o}{\rho } q , \\
\rho^o\frac{\partial [ e_i(\rho , t_i)/\rho]}{\partial t } & = & - p_i\frac{\partial u}{\partial x^o } + \frac{\rho^o}{\rho } \alpha ( t_e - t_i)\end{aligned}\ ] ] describing instant local electron and ion heat balances , where @xmath57 is an eulerian coordinate defined in fig .
[ fig:2 ] , @xmath58 is the trajectory of material particle with lagrangian coordinate @xmath59 which is equal to @xmath57 before the action of a pump fslp . in 2t - hd simulation
the pump intensity is @xmath60 @xmath61 and @xmath62\exp(-t^2/\tau_l^2),$ ] @xmath63 @xmath64 is the initial material density : @xmath65 ( al ) and 19.3 ( au ) , @xmath66 are pressure and density , @xmath67 is a coefficient of a thermal heat conduction . @xmath68
\exp(-x/\delta),$ ] where @xmath69 is the thickness of the skin - layer shown in fig .
[ fig:2 ] as `` skin '' .
unknowns @xmath70 are functions of the lagrangian variables @xmath71 when knowing the trajectories @xmath58 , the variables @xmath70 can be presented as the functions of the eulerian coordinates @xmath72 @xmath73 are the internal electron and ion energies , @xmath74 is the absorbed power , @xmath75 is the e - i energy exchange term . from the profiles @xmath76 as functions of @xmath57 , obtained from the eqs.(1 - 3 ) , one can find the corresponding profiles of the dielectric permittivity @xmath77 and the complex index of refraction @xmath78 @xmath79 where @xmath80 is a volume fraction of liquid phase discussed below for case of solid - melt mixture .
next , for the @xmath57-profile @xmath81 at fixed time @xmath82 ( velocity of light is taken to be infinite ) , helmholtz equation @xmath83 is solved for amplitude @xmath21 of the probe fslp perpendicular to a target . equation ( [ eq : helmholtz ] ) describes a reflection of the probe light from target .
its solution gives the amplitude and phase of the reflected wave and correspondingly the time evolution @xmath84 of the reflectivity @xmath85 and phase @xmath86 they are compared with the experimental dependencies in figs.[fig:3],[fig:4 ] . by contrast , the pump absorption is taken from the experiment @xcite .
equation ( [ eq : helmholtz ] ) is evaluated by the @xmath87 transfer matrices method @xcite . in our experiments
the chromium - forsterite laser with the pump @xmath88 and probe @xmath89 ( first and second harmonics ) is used @xcite .
a fslp duration is @xmath90 values @xmath91 have been measured by microinterferometric technique described in @xcite .
role of the e - e collision frequency @xmath92 in optics .
the solid curve obtained for @xmath93 the experimental ( squares ) and theoretical reflectivity @xmath94 normalized to initial @xmath95 of al ; @xmath96 @xmath97 @xmath98 from @xcite . ]
the values @xmath99 @xmath100 and @xmath101 in ( 1 - 3 ) are taken as in @xcite from the wide range equation of state @xcite and from fermi model for the conduction electrons , @xmath102 is the total pressure .
the coupling factor @xmath17 and the heat capacity @xmath103 are taken from @xcite .
according to the drude formula the electron heat conductivity @xmath67 in ( 2 ) is @xmath104 @xmath105 @xmath106 @xmath107 is fermi energy . at @xmath108
when electron degeneracy is significant the collision frequency is @xmath109 ( `` deg '' stands for degenerate ) , @xmath110 the factor @xmath111 in ( [ eq : nu = ei+ee ] ) approximates the quantum - mechanical molecular dynamic ( md ) data @xcite showing the drop of the electrical conductivity of al with the density decrease in the temperature range under consideration .
the electron - electron collision frequency is @xmath112 the coulomb collisions @xmath113 in ( [ eq : kappa ] ) is taken as @xmath114 they dominate at very high @xmath115 and limit @xmath116 ( [ eq : kappa ] ) at @xmath117 by an atomic frequency @xmath118 ( saturation of the @xmath116 increases with increasing @xmath119 the electron - ion collisions @xmath120 in @xmath121 ( [ eq : kappa],[eq : nu = ei+ee ] ) is calculated separately for solid and liquid al @xmath122 , \label{eq : nueisol } \\
\nu_{ei}^{liq } & = 1.1 \cdot 10^{14 } t_i / ( 130 + 0.0367 t_i - 66700 / t_i ) \ , [ s^{-1 } ] , \label{eq : nueiliq}\end{aligned}\ ] ] where @xmath123 @xmath124 in k. the coefficients in ( [ eq : nueiliq ] ) approximate the quantum - mechanical md heat conduction of the molten 1 t al @xcite .
the thermal conductivity is calculated with kubo - greenwood formula up to @xmath125 @xcite .
values for @xmath67 from ( [ eq : kappa],[eq : nu = ei+ee],[eq : nueisol ] ) and for electric conductivity @xmath126 with ( [ eq : nueisol ] ) describe well 1 t reference data for solid al between the debye and triple point @xmath127 temperatures .
the drop of the phase shift @xmath128 results from the solid - liquid transition within first few picoseconds , @xmath129 @xmath130 is an initial phase ( a phase of reflected wave from a cold target ) .
diamonds indicate experimental data . ]
the value @xmath131 necessary for ( [ eq : helmholtz ] ) is a sum of the drude and the interband terms @xcite @xmath132 where @xmath133 in al @xmath134 in an al crystal @xmath98 results mainly from the transitions between parallel zones @xcite .
this term dominates at room temperatures .
its contribution increases the absorption more than by order of magnitude in comparison with pure drude absorption . in a molten al @xmath98
disappears @xcite .
electron density of states ( dos ) in al is stable against melting @xcite and against @xmath12 increase as was checked up to @xmath135 in @xcite .
therefore the ion charge @xmath136 and the electron effective mass @xmath137 defining @xmath138 weakly depend on melting and overheating of electrons . a phenomenological dependence of @xmath98 ( [ eq : eps ] ) on a total frequency @xmath116 ( [ eq : kappa],[eq :
nu = ei+ee ] ) has been proposed @xcite . at large @xmath139 the term @xmath98
becomes small as compared to the drude term as it is in liquid .
md simulations show that in al within the early @xmath140- few ps ) stage the solid - liquid mixture fragmentation space scale @xmath0 is small in comparison with @xmath141 therefore the @xmath142 can be defined by a volume fraction @xmath80 of liquid in mixture : @xmath143 for the weak mixtures @xmath144 or @xmath145 we have @xcite @xmath146 an approximate interpolation for intermediate @xmath147 of these linear in @xmath80 solutions ( [ eq : e - mix ] ) is @xmath148 expressions ( [ eq : e - mix],[eq : interpolation ] ) have been used in calculations shown in figs .
[ fig:3 ] and [ fig:4 ] . at an early stage
it is necessary to consider optics of a solid - liquid mixture given by these expressions because at this stage the thickness of a solid - liquid mixture layer is comparable with the thickness @xmath69 of a skin - layer a penetration depth of a probe photon
. values @xmath149 @xmath150 @xmath151 @xmath152 @xcite differ moderately . in this case
there is another approximation @xcite @xmath153 nonlinear in @xmath80 but linear in small difference @xmath154 comparison of this approximation with ( [ eq : e - mix],[eq : interpolation ] ) shows that results differ small : the maximum deviation achieved at @xmath155 is @xmath156 for re@xmath157 and less than 3% for im@xmath158 as was said at early stage optics of solid - liquid mixture is significant .
let us mention that as it will be shown below at this stage there is non - equilibrium melting of an overheated crystal . at intermediate and late stage @xmath4 shown in fig .
[ fig:1 ] existence of vapor - liquid mixture can influence optical reflection if thickness of cavitation layer @xcite is less or comparable with @xmath141 the cavitation layer covers the undersurface vapor - liquid layer against probe photons .
in an early stage the @xmath12 is high then the @xmath92 ( [ eq : nuee ] ) dominates in the total frequency @xmath116 ( [ eq : kappa ] ) as maximum @xmath124 for our fluences are smaller than 10 kk and the @xmath120 collisions are less frequent .
there are different coefficients @xmath159 and @xmath160 for the @xmath92 ( [ eq : nuee ] ) in expressions for @xmath67 ( [ eq : kappa ] ) and for optics in the drude and in the interband terms ( [ eq : eps ] ) . in crystals
the @xmath159 includes normal and umklapp processes while the @xmath160 in the solid al may differ from zero as a result of the umklapp effect ( situation is different for gold , see below ) . in liquid
al the coefficient @xmath161 the e - e collisions do not contribute into optical absorption as the umklapp is impossible . in al
a fermi sphere is larger than in au while the brillouin zones are approximately equal ( the lattice constants are @xmath162 for al and au both ) . therefore the umklapp effect is more significant in al .
e.g. , it results in order of magnitude increase of the electron - ion energy exchange rate @xmath17 @xcite .
value of @xmath163 in ( [ eq : nuee ] ) is a subject of discussions . below the debye temperature
the @xmath164 tends to zero strongly and becomes less than the @xmath165 then specific electrical resistance @xmath166 for very pure crystals is @xmath167 the measurements @xcite give @xmath168 here we suppose that @xmath92 does not depend on frequency of electromagnetic field and the coefficient @xmath163 in ( [ eq : nuee ] ) for resistance @xmath169 is equal to @xmath170 at the same time the theory @xcite gives @xmath171 for high @xmath12 the calculations @xcite @xmath172 is accepted .
estimate of the upper limit for @xmath159 follows from the check of the wiedemann - franz law @xmath173 done for melt al in @xcite , where @xmath174 is lorentz number . in drude approximation
we have @xmath175 because in melt umklapp is absent and @xmath92 ( 7 ) does not contribute to electric conductivity @xmath176 relative deviation @xmath177 of the ratio @xmath178 from @xmath174 in @xcite is less than 10% .
this means that electrons in al remain degenerate up to temperature @xmath125 achieved in @xcite .
the limit @xmath179 impose restriction on value @xmath180 it is obtained from @xmath181 where expressions ( 7 ) and ( 9 ) was used . on the other hand ,
the value @xmath159 is important at early stage because it influences heat propagation into bulk when the propagation is supersonic . to achieve thickness @xmath182 at al ablation threshold the values @xmath159
should be near this restriction @xmath183 the thickness @xmath182 is necessary to reproduce experimentally defined crater depth @xmath184 for smaller @xmath159 the simulated crater is deeper and fluence threshold is higher than the experimental ones .
deviation of a phase of reflected light resulting from melting , @xmath21 is a complex amplitude of a reflected wave .
@xmath185}}4\pi/\lambda_{prob}$ ] and @xmath186}}4\pi/\lambda_{prob}.$ ] ] performed optical measurements shown in figs .
[ fig:3 ] , [ fig:4 ] together with simulations give additional information about @xmath170 the interesting sharp narrow `` well '' at the theoretical @xmath94 in fig .
[ fig:3 ] corresponds to @xmath187 its minimum is achieved when @xmath12 and hence @xmath92 [ eq : nuee ] have the largest amplitudes ( at the end of the pump ) .
this well might be very useful for diagnostics .
the left wing of the well follows the history of electron heating while the right wing reflects the kinetics of melting because the gradual phase transformation from crystal state to melt in a skin - layer progressively suppresses optical contribution of the e - e collisions ( [ eq : nuee ] ) . unfortunately the well is not observed .
density of experimental points at the time axis in figs .
[ fig:3 ] , [ fig:4 ] is large enough to exclude missing of the well between the two successive points .
the accuracy of experimental measurements of a relative reflection @xmath188 is sufficient to catch the well .
analysis of our simulation runs in a fluence range @xmath189 shows that it is necessary to have the @xmath160 below 0.2 - 0.5 to meet the measurements ; for al the calculated and measured thermomechanical ablation threshold is @xmath190 @xmath191 @xcite .
rise of @xmath124 in 2t - hd and md ( the fluctuating profiles ) simulations , al , @xmath192 ]
simulations show that the phase evolution @xmath193 presented in fig .
[ fig:4 ] contains important information concerning the kinetics of melting .
the base for this is the difference commented in fig .
[ fig:5 ] between the defined in section 4 @xmath194 and @xmath195 as a result of attenuation of the band - band transition during melting the value @xmath196 becomes smaller .
this is why the @xmath197 in al is 2.2 nm differing from solid in the direction of the phase rotation to the crater .
the value 2.2 nm corresponds to the case of a fresnel reflection from homogeneous semispace . remarkably that this small difference is measurable by the pump - probe interferometry .
the sign of rotation directions in fig .
[ fig:5 ] defines the sign of the phase difference @xmath128 in fig .
[ fig:4 ] .
detection of ultrafast melting of semiconductors @xcite is possible due to the same liquid - solid phase difference @xmath198 but in this case the melting transforms semiconductor into a metallic state
therefore @xmath128 is significantly larger ( e.g. , @xmath199 for gaas , @xmath200 and this transformation can be detected easily .
the phase @xmath201 obtained from 2t - hd equations is compared with experimental data in fig .
[ fig:4 ] . expansion movement of reflecting boundary should increase @xmath128 but at the early time it decreases as a result of gradual melting of skin - layer .
agreement between data and theory indicates that theory given below properly describes the melting .
figure [ fig:6 ] illustrates the heating of ions by hot electrons in 2t - hd model . in md simulation atoms
are heated by the space - time distributed thermostat power source with the temperature distribution taken from 2t - hd .
this md approach is similar to one developed in @xcite , see also @xcite where phenomenological terms describing non - equilibrium melting are added to 2t - hd equations .
particular mechanism of heating ( heat flow from electrons or thermostat ) has no action upon the kinetics of melting if we suppose that elastic moduli do not depend on @xmath12 as in the case of al @xcite . in fig .
[ fig:6 ] the 2t - hd and md @xmath124 profiles are approximately the same . some difference results from equilibrium and non - equilibrium description of melting .
the arrows in fig .
[ fig:6 ] mark the slice of melting from equilibrium 2t - hd .
the kinks at the ends of this slice result from hidden fusion energy . in md
the degradation of crystal symmetry during the fast heating and melting is distributed in wider range beyond the liquidus / solidus positions .
crystal beyond solidus is in overheated state @xcite . phase transformation and propagation of melting into bulk
is shown in fig .
[ fig:7 ] .
the profiles of the symmetry index @xmath202 are presented .
the index @xmath202 is defined as a number of crystal axis passing through an atom and averaged over atoms within a x - slab , and @xmath203 in a cold fcc lattice .
one can see how quickly the rather thick ( thicker than @xmath204 layer of mixture is formed .
later the layer of pure melt ( the plateau at the instant @xmath205 @xmath206profile in fig . [ fig:7 ] ) appears . much later the narrow melting front with small overheating separating a melt from a crystal is formed .
maximum thickness of the molten layer for the fluence @xmath207 is @xmath208 evolution of the phase composition as a result of the increase of @xmath209 the e - i heating gradually rises @xmath124 as it is shown in previous figure .
transversally averaged @xmath202-profiles obtained from md simulations are presented .
the arrows `` liquidus '' and `` solidus '' at @xmath205 are taken from previous figure .
the values of the md symmetry index @xmath210 taken at the 2t - hd solidus fronts are approximately the same for different instants . ]
absorption of pump fslp causes sharp changes in optical properties at very early time during the pulse @xmath211 they are shown in fig . [ fig:8 ] .
bulk gold targets are used .
this response is caused by fast heating of electrons .
if we compare figs .
[ fig:3],[fig:4 ] ( al ) and fig .
[ fig:8 ] ( au ) having similar relative temperatures @xmath212 at 2 t stage and similar final @xmath213 temperatures after e - i relaxation we will see obvious large differences .
they are related to the differences in the band structures of al and au @xcite .
estimates of heating history from the maximum @xmath12 to e - i thermalization give for the main seven thermal parameters @xmath214 @xmath215 @xmath216 @xmath217 @xmath218 @xmath219 @xmath220 the values : ( 0.5 , 0.7 , 50 , 1 , 1.5 , 1.2 , 0.8 ) , ( 1 , 1.3 , 140 , 3 , 2.5 , 1.5 , 2.5 ) , ( 2 , 2.9 , 500 , 10 , 5 , 2.4 , 8) for the three cases shown in figs.[fig:8],[fig:9 ] , where @xmath221 is a thermomechanical ablation threshold , @xmath222 @xmath223 are electron thermal energy and temperature at the end of a pump , @xmath224 is number of electrons exited from the 5d to the 6s,6p band at the maximum @xmath225 @xmath226 is maximum ion temperature achieved after e - i relaxation .
an expression @xmath227 for electron thermal energy at fixed density @xmath228 approximates data obtained from the abinit @xcite simulations up to @xmath229 here @xmath230 is in mj / kg , and @xmath12 in k. our curve is above the standard parabola @xmath231 because we include exciting of the _ d_-electrons .
but it is below the curve @xcite where a red - shift of the _ d_-band with @xmath12 was not considered .
experimental and theoretical values for the depth of a crater at a threshold @xmath232 is 110 nm @xcite . the first case with the smallest @xmath233 is near a melting threshold for bulk au .
our three cases cover a range of energy densities obtained in @xcite for ultrathin ( 25 - 30 nm ) freestanding au films . here
we consider bulk targets and use different technique of measurements .
fast growth of @xmath131 as result of electron heating .
the @xmath235 was calculated from data shown in fig .
[ fig:8 ] and the fresnel formulae valid while the @xmath131-profiles may be approximated by a step function .
the solid curves presents data from @xcite at @xmath236 they corresponds to our intermediate case @xmath237 two bars give @xmath131 for liquid au at 1337 k@xcite . ]
the values of @xmath224 presented above have been defined using the abinit @xcite code , the normalization condition for the number of electrons , and the expression for the amount of electrons in the 6sp band .
the abinit has been used as in @xcite for calculation of the @xmath12 dependent dos at different @xmath12 supposing that density is equal to @xmath53 ( isochoric heating ) . for given @xmath12 the dos of the 5d band obtained from abinit
has been approximated by a rectangular and the dos of the _
6s,6p _ band has been described as the function @xmath238 in order to calculate @xmath239 at a given @xmath12 a root @xmath240 of the normalization condition @xmath241 where @xmath242 defines a chemical potential @xmath243 here @xmath244 are edges of the 5d band relative to the bottom point of the _
6s , p _ bands .
the first and the second terms correspond to the numbers of electrons in the _ 6s , p _ and _ 5d _ bands , resp . at room temperature
these numbers are @xmath245 and 10 .
it is known that for au difference between this approach and calculation of @xmath246 with exact dos is small @xcite .
the exact function @xmath247 is obtained together with the dos in the abinit simulation .
the excitation degree @xmath224 is given by an expression @xmath248 where @xmath174 stands for the second term in ( [ eq : norm ] ) but now with known @xmath249 to estimate possible influence on the value of @xmath224 of position of the bottom point of the 6sp band this position was varied in the range @xmath250 relative to its position at @xmath251 the values of @xmath224 and collision frequency @xmath116 are necessary for the drude estimates of @xmath252 growth of them is responsible for the rise of @xmath253 in fig . [ fig:9 ] . at room temperature @xmath254
is small : 1.2% from electrical and thermal conductivities or 3.3% from optical data @xcite , @xmath255 heating of an electron subsystem in our conditions rises @xmath116 to large values : @xmath256 there are three candidates responsible for the growth of @xmath257 ( i ) an enhancement of @xmath120 in solid or liquid au with @xmath258 in comparison with @xmath259 case , ( ii ) @xmath92 and umklapps in crystal au , and ( iii ) the e - e collisions between electrons from the _ d _ and the _ sp _ bands in solid or liquid au
. let s consider these cases .
\(i ) the frequency @xmath120 and the e - i exchange rate @xmath17 are coupled ,
@xmath260 @xmath258 @xcite , where @xmath261 is an energy transfer from a light electron to a heavy ion in one e - i collision , @xmath262 @xmath263 is an interatomic distance in condensed matter , @xmath264 is debye temperature .
therefore @xmath265 if @xmath266const as in al @xcite , then @xmath267 as usual .
but in gold there is significant dependence on @xmath11 @xmath268 an approximation of calculation of @xmath269 with dos @xmath270 from abinit @xcite ( including the shift of the _ d_-band with @xmath271 is @xmath272 it is valid up to @xmath229 here @xmath273 is fermi distribution @xcite .
the @xmath274 begins to grow at @xmath275 from @xmath276 and saturates at the value ten times larger @xmath277 @xcite .
this means that @xmath278 is enhanced if @xmath279 @xmath280 there is @xmath281 @xmath282 in 1 t gold , @xmath283 in k. for @xmath284 we have @xmath285 if @xmath286 as in two cases shown in figs.[fig:8],[fig:9 ]
then the enhancement due to increase of @xmath12 gives large frequency @xmath287 \(ii ) in al @xmath288 is small , sec.5 .
as was said , the umklapp processes are weaker in au as the result of smaller fermi / brillouin ratio .
therefore it seems plausible that this candidate is less important than two other candidates .
\(iii ) electrons from _ d _ versus _ sp _ bands have different angular momentum and effective mass .
therefore photons are absorbed in collisions between them .
corresponding frequency @xmath289 is given by ( [ eq : nuee ] ) with possible dependence @xmath290 the @xmath289 may be @xmath291 at @xmath292 from imaginary and real parts of expression ( [ eq : eps ] ) we obtain @xmath293 $ ] and @xmath294 the mass @xmath295 at room temperatures @xcite remains approximately the same with increase of @xmath296 this follows from our abinit simulations .
according to @xcite in 2 t au @xmath297 where @xmath298 is an absorption edge of the _ d_-band .
if we neglect the band - band term @xmath299 in ( [ eq : eps ] ) , @xmath300 @xmath301 then @xmath302(0.9 ; 0.5 ) , ( 1.7 ; 0.9 ) , ( 3.3 ; 1.3 ) at the maximum @xmath253 in the three cases shown in fig .
[ fig:9 ] . our data shown in fig .
[ fig:9 ] agree with data from @xcite in @xmath303 but give larger @xmath304 and have a maximum at the time dependence @xmath305 while the dependence @xmath306 from @xcite saturates .
perhaps the last difference results from the conductive cooling absent in ultrathin films .
let s mention that data @xcite contain the useful dependence @xmath307 but may be less accurate at a particular frequency .
@xmath302(1.2 ; 0.4 ) for the maximum of the dependence @xmath308 from @xcite shown in fig .
[ fig:9 ] as the solid curves . as a result of smaller @xmath304 these values
are below than our values ( 1.7 ; 0.9 ) for the case with approximately the same energy @xmath309 nevertheless there are appreciable excitation @xmath224 and frequent collisions .
therefore we can conclude that measurements confirm the theoretical findings presented above that a pump irradiation creates an excited population @xmath310 rising @xmath138 ( [ eq : eps ] ) and transfers gold into the state with strongly collisional widened energy levels .
the cavitation or run - away layer is thermally weakly coupled with a future crater bottom and therefore may remains molten while the bottom becomes frozen . at the @xmath4 slice in fig .
[ fig:1 ] the cavitation layer is mechanically coupled with the bottom through the fragmentation or nucleation zone , @xcite , agranat et al .
* 253 * ( 2007 ) 6276 .
this value is only 1.5 times smaller than the @xmath311 ( at low @xmath12 @xmath312 may be this explains why the @xmath42 and the crater depth at the ablation threshold differ only twice in au versus al .
the saturation @xmath12 is @xmath313 | the theory and experiments concerned with the electron - ion thermal relaxation and melting of overheated crystal lattice constitute the subject of this paper .
the physical model includes two - temperature equation of state , many - body interatomic potential , the electron - ion energy exchange , electron thermal conductivity , and optical properties of solid , liquid , and two phase solid - liquid mixture .
two - temperature hydrodynamics and molecular dynamics codes are used . an experimental setup with pump - probe technique
is used to follow evolution of an irradiated target with a short time step 100 fs between the probe femtosecond laser pulses .
accuracy of measurements of reflection coefficient and phase of reflected probe light are 1% and @xmath0 , respectively .
it is found that , _
firstly _ , the electron - electron collisions make a minor contribution to a light absorbtion in solid al at moderate intensities ; _ secondly _ , the phase shift of a reflected probe results from heating of ion subsystem and kinetics of melting of al crystal during @xmath1 where @xmath2 is time delay between the pump and probe pulses measured from the maximum of the pump ; _ thirdly _ the optical response of au to a pump shows a marked contrast to that of al on account of excitation of _ d_-electrons .
femtosecond laser ablation , pump - probe , optics of hot al and au |
the increasing environmental awareness , together with the progressive reduction of production and installation costs@xcite , leads to a considerable growth in the amount of renewable energy sources ( res ) that is installed worldwide .
moreover , the increasing propensity to reduce greenhouse gas emissions requires an increment of the energy produced by clean , accessible energy sources such as wind and photovoltaic ( pv ) generation . despite the great advantages of these energy sources , their intrinsic variability in power production badly fits to the very hierarchical structure and the strictly dispatch rules of actual power systems .
the limited accuracy of the prediction of their energy production profiles makes the management of these intermittent power sources difficult and limits the amount of res generation that the power system can tolerate .
after the network liberalization over the last 15 years , the system balancing in real time is performed via the electricity balancing market ( ebm ) , which is a subphase of the ancillary services market ( asm ) .
this market phase shall ensure the correct balanced state over the system at the transmission level , providing the security of the supply at the lowest possible costs .
however , the short time - scale and the volatility of this market phase produce higher energy costs when compared with the day - ahead market phase .
therefore an increase in the ebm volume can lead to very high system maintenance costs .
the growing amount of the installed res generation introduces a high number of partially correlated fluctuations in the power production . along with that , it becomes more difficult to predict the amount of energy that is needed for balancing the system . in general ,
an increase in production fluctuations could lead to both an increase in market average volumes and a more frequent occurrence of extremely high values of the volume . whereas an increase in average volumes could be cured by strengthening the reserve capacity ,
the occurrence of extreme volumes is more difficult to control .
moreover , given the fact that the relation between price and demand , also known as power stack function @xcite , is highly nonlinear , large volume events can lead to very high energy prices .
such extreme and unwanted price events have been observed already by nicolosi in @xcite for the german system : if they happen too often , the total costs of the market session increase and undermine the principles on which the electricity market was designed . therefore , the forecast of the fluctuations impact on the balancing market can be vital for an optimal planning of the network growth , and for uncovering possible critical situations of the network .
so it is not surprising that the evaluation of volumes and prices of ebms , and in general , of electricity markets , has attracted much interest in the last years .
however , the proposed solutions are mostly based on historical data for the market volume together with learning procedures of agents and game theory @xcite , @xcite .
they need an update to more recent data and an extrapolation towards future increased contributions from res . in the next section
we shall show how a combination of an approach from statistical physics with agent based modeling overcomes the need for using historical data and allows for the prediction of energy prices , the emergence of price peaks , volumes of balancing markets and overall daily costs for different contributions from res , to cover up to 60% of the load .
this combination of methods was proposed and validated for the italian ebm in @xcite .
our results together with the methodology should enter planning procedures of how to further increase and control the amount of res in the future .
before going into detail , let us summarize the procedure , which consists of three steps : ( i ) based on real data for production and consumption at a certain representative day in the winter period of 2011 - 2012 in the italian grid , we generate a certain set of starting configurations , each one describing a combination of production and load at nodes of the italian transmission grid , which lead to a stable performance by construction .
the real data were taken every 15 minutes over a whole day for all 6 price zones of the power grid in italy .
we then extrapolate these data towards a higher contribution of res , ranging from the real value of 24% to 60% in the extrapolates sets . in all extrapolated cases
we guarantee a stable performance by running optimal dc - power flow equations to adjust the production by conventional generators so as to guarantee an overall balanced power in the grid .
\(ii ) each of the configurations of the resulting set ( 6 zones x 96 time instants per day for 24 , 30 , 40 , 50 , and 60% of res ) serves as starting point for generating an ensemble of configurations in the spirit of statistical physics . in statistical physics one usually describes a macrostate say of a gas of molecules by an ensemble of microstates ; microstates differ by small deviations of the generalized coordinates and momenta , so that an average over many microstates leads to representative macroscopic observables . in analogy here ,
members of each ensemble differ from the starting configuration and therefore also mutually by small deviations in the load and the renewable energy production , chosen from a gaussian or weibull distribution , as explained below .
mean and width of the gaussian distribution as well as the parameters of the weibull distribution are chosen from real data and depend on the load and the type of renewable energy .
so each configuration j of the ensemble represents a certain realization of fluctuations in load and production , for which quantities like the resulting mismatch @xmath0 of power at node i , induced by the various fluctuations , can be measured .
@xmath0 can be summed over all nodes of the grid to obtain the total mismatch @xmath1 in power production for a given configuration j , which the balancing market is supposed to compensate for . to obtain representative values of the mismatch ( later called the market volume ) , also here ( as for the microstates of a gas ) a sufficient number of configurations
should be included in the ensemble .
\(iii ) the energy balancing market the energy balancing market is modeled by a so - called market authority and a set of agents .
the market authority knows the required amount of power @xmath1 , which is needed for balancing consumption and production ; it accepts or rejects bids from the agents until an amount of @xmath1 is obtained at the lowest possible price .
it then informs the agents about the outcome of their placed bids as to whether they were accepted or not .
the agents are assigned to conventional generators in a one - to - one relation ( for simplicity ) .
they choose their bids from a distribution of propensities to offer a certain amount of energy at a certain price .
the propensity distribution changes with time during the learning phase , using a modified roth - erev algorithm@xcite . for the learning phase
we choose 3000 updates of the propensity distribution , in which the agents are trained on the same number of different configurations , chosen from the ensemble around a fixed starting configuration , so differing just by fluctuations among each other . this number of updates turned out to be sufficient for the propensity distribution to converge towards an optimized distribution , resulting from the learning experience of the agents after feedback from the market authority .
for the next thousand configurations of a given ensemble the propensity distribution of agents is then kept fixed , and the energy price in this market session can be calculated , for each configuration , leading to a distribution of energy prices over all configurations of the ensemble .
the distribution of energy prices refers to a certain time during a day for a given zone in italy . repeating the whole procedure for different starting configurations , corresponding to different instants of time at the reference day and different price zones , we can measure histograms of
how often a price from a certain price interval was achieved over the day or for a restricted time interval of an hour etc .. we are particularly interested in the shape of these histograms as a function of the percentage of renewables , which contributed to the power production .
details are presented in the following sections .
let us first estimate the effect of res and load power fluctuations on the system s power balance . according to the literature@xcite , wind , pv and load forecasting errors in the power production
are often treated as normal - distributed .
so their power production or consumption can be modeled in a statistical way , assuming truncated gaussian - like forecast errors with standard deviations @xmath2 , where the errors represent the expected power variations at each single fluctuating element @xmath3 of the power grid at a given time .
the associated variables are the following : * load power demand @xmath4 , the corresponding standard deviation @xmath5 of the forecasting error and minimum and maximum values of the distribution , @xmath6 and @xmath7 , respectively , corresponding to the load - power constraints ; * wind power production @xmath8 and the corresponding @xmath9 , together with the power constraints @xmath10 and @xmath11 of the generators ; * photovoltaic power production @xmath12 , the corresponding @xmath13 , and the production limits @xmath14 and @xmath15 , corresponding to the power constraints of the pv - generators . according to these constraints , a possible state of the system can be sampled numerically by adding a random value to the expected power production and consumption at every node and res generator @xmath3 of the grid .
the random variable is extracted from the truncated normal distribution , whose probability density function ( pdf ) is defined in equations [ eq : normal_def ] and [ eq : normal_pdf ] .
@xmath16 @xmath17 the outcome of this procedure is one of the possible configurations or states @xmath18 , in which the system in zone k ( @xmath19 ) can be found in real time , due to assumed unavoidable fluctuations of res and load . in order to check the impact of other than gaussian - type fluctuations , we complemented the normal distribution of ( [ eq : normal_pdf ] ) in case of wind production by a weibull distribution , whose pdf is defined in equation [ eq : weibull ] , while the fluctuations for photovoltaics production and load were kept being chosen from gaussian distributions . since the literature gives values between 1.5 and 3 for the value of the parameter @xmath20 @xcite , we have chosen @xmath21 and @xmath22 , where @xmath23 here is chosen as the wind production from the reference configuration and @xmath24 denotes the gamma - function .
the pdf is then given as @xmath25 starting from the so generated configuration or network state , we apply the optimal dc - power flow algorithm @xcite to calculate the power @xmath1 that is needed for balancing the mismatch in power , induced by the deviations from the starting configuration , for each state @xmath18 of the ensemble , related to zone k and to each of the 96 time instants a day . due to the stochastic nature of res and load fluctuations , also @xmath1 is a random variable .
therefore , to sample sufficient statistics of the market behavior , a significant number of possible balancing requirements is needed .
it is obtained by numerical sampling a large number of possible perturbed configurations @xmath18 , each one with an associated balancing requirement @xmath1 .
its distribution over the ensemble and over the day is then used for describing the daily expected volume of the balancing market in the system .
energy prices and total costs in the ebm are determined by an agent based modeling approach , for which we use a modified roth - erev algorithm , introduced by nicolaisen et al .
@xcite in 2001 and used by rastegar et al.@xcite already for the simulation of the italian oda electricity market .
the electricity - market operators are represented by agents , who learn how to place optimal bids in competitive auctions with the aim of buying ( or selling ) in the most profitable way . in order to simulate
how real market operators acquire knowledge about the market in the course of time and adapt their decisions , roth - erev algorithms simulate this learning process by adjusting the offer propensities of agents in a self - consistent way with the goal to maximize profits .
market operators pursue economic guidelines , when they represent power plants ( or groups of them ) in the ebm auction phase .
they are allowed to place bids into the ebm auction , in which they must specify how much the corresponding power plants can vary their amount of power supplied to the system , and at what price they will offer this service . for simplicity
we represent each conventional power - plant generator by a single agent , although the exact relationship among market operators and brokers may vary over time and can be more involved .
each agent @xmath26 is allowed to offer an amount of power @xmath27 that must meet the physical constraints of the power generator k : * @xmath28 , where @xmath29 and @xmath30 are the minimum and maximum power production constraints of the generator , respectively , and @xmath31 is its actual power production .
* @xmath32 , where @xmath33 is the generator ramping constraint .
( depending on the technology , each generator has ramping constraints , which limit its maximum change in power production @xmath33 in time . ) in order to define the bids price , we use the concept of agent propensities , representing the willingness of each agent to place a bid at a certain price on the market .
the offer propensities of each operator @xmath26 are described in terms of a discrete set of probabilities , @xmath34 , to be defined below , corresponding to possible bidding strategies @xmath35 , which roughly speaking differ by how to deal with risks in offering higher prices .
the index i , @xmath36 , labels the strategy , @xmath37 is the number of possible strategies , and @xmath38 is the k - th operator s propensity to make an offer at a given ( so - called markup ) value @xmath39 ( @xmath40 for upward bids , @xmath41 for downward bids ) .
the number of strategies equals the number of intervals into which the range of @xmath39 is divided .
here we have chosen @xmath42 , so that one has to assign 50 propensities to values of @xmath39 .
the markup value determines the bidding price according to @xmath43 , where @xmath44 is the production cost ( per mwh ) of each generator k , given by its technology type , labeled by the subscript @xmath45 , and @xmath46 is the actual chosen value from the discrete distribution for the bid of agent k. so the operators behavior is modeled stochastically , where the probability of placing a bid at a given price @xmath43 is given by the normalized propensity @xmath47 with @xmath48 and @xmath49 with g the number of conventional producers equal to the number of agents .
it is then the set of propensities , which get optimized when updated in an iterative reinforced learning algorithm .
initially , all propensities @xmath38 are set to the same value @xmath50 .
each learning iteration step is divided into three phases : 1 .
bid presentation : every agent k presents a bid @xmath51 , both for the upward and downward market .
this bid is given by a feasible quantity of offered energy @xmath52
( i.e. satisfying the physical constraints ) and by a price @xmath53 , which will be drawn from the agents propensities .
2 . market session : given the knowledge of the total balancing needs of the system @xmath1 , all the bids , which are needed to ensure sufficient energy supply , are checked with respect to their economic profit and the physical constraints of the system .
3 . agent update : market outcomes are communicated by the market authority to each agent , who updates his propensities in relation to the profit , which he made in the previous session .
the agents propensities at iteration step @xmath54 are updated as follows : @xmath55 where @xmath56 $ ] is a memory parameter and @xmath57 is obtained from the relation : @xmath58 for all k , where @xmath59 $ ] is an experimental parameter that assigns a different weight to played and non - played actions . to the best of our knowledge ,
roth - erev algorithms were previously applied to training agents based on the exclusive use of historical data with their limited relevance for current and future power distributions in electricity grids . in this paper , following the guidelines presented in @xcite , we overcome the usage of outdated data by performing the training on realistic system configurations , which are synthetically generated as we explain in the next section .
for a correct representation of the market phase , we need a detailed description of the power transmission system in space and time , in terms of network nodes , branches and generators .
the reference configuration of the power system is obtained by combining three datasets .
the first dataset is related to the characteristics of the power system ( from the terna website @xcite ) and includes the geo - referenced position of every 220 and 380 kv substations together with their electrical characteristics , the geo - referenced position of the conventional generators , together with their power rates and power ramp limits , and the electrical characteristics of the power network .
a geographical representation of these data is depicted in figure [ fig : ita ] .
the second dataset ( from the gme website @xcite ) reports the detailed time evolution of production / consumption every 15 minutes of a reference day in the winter period 2011 - 2012 , so that 96 data sets per day are available .
the third dataset is obtained from atlasole and atlavento ( see the website @xcite ) .
these sites were made available by the italian energy services authority @xcite .
they contain the full georeferenced information on each italian pv and wind generator , such as the installed power and technology . combining these datasets
, we reconstructed the time evolution of power production and consumption in steps of 15 minutes over a full day in the winter period 2011 - 2012 for all six market zones ( virtual ones are excluded ) .
these data ensured already a balanced grid performance , based on optimal dc - power balancing , when the conventional power production was adjusted accordingly .
moreover , starting from the real distribution of installed res capacity , we extrapolated the data to starting configurations with a different percentage of res production .
we separately assumed aggregated wind and photovoltaic productions to take desired values , given by @xmath60 , with @xmath61 being the total load that is kept fixed and @xmath62 being the percentage of load covered by res .
biomass production as a third type of res was not considered to contribute to fluctuations , as energy production from biomass is easily controlled .
we then estimated the required adapted production by conventional generation to cover the load by means of an optimal dc - power flow .
as result we obtained initial configurations that lead to a stable grid performance for daily peak shares @xmath62 of 24% ( which was the actual one used in @xcite ) up to and including 60% , which is nowadays already temporarily achieved . for values larger than 60%
it became increasingly difficult to adjust the conventional power production to compensate for the increased amount of renewables .
based on the approach as outlined in the previous section [ sec : methods ] , we estimate the impact of an increasing share of wind and pv generation on the italian balancing market .
in particular , we identify a change in the daily market volumes and costs , due to an increased percentage of fluctuating production , introduced by these sources .
we tested the balancing market phase in five scenarios , characterized by a res power production with a share of @xmath62 , chosen as 24% ( as of the reference day @xcite ) , 30% , 40% , 50% and 60% of the total load @xmath63 .
it should be noticed that the size and variation of fluctuations in the power grid are heterogeneous , as they depend on the nature of the fluctuating quantity being load , wind or photovoltaics .
their respective strengths are described by gaussian distributions with fluctuation parameters @xmath64 , @xmath65 and @xmath66 @xcite , that is , chosen from real data , and in case of the assumed weibull - distributed fluctuations of wind , chosen as @xmath21 and @xmath22 with @xmath23 the wind production from the reference configuration .
next it is of much interest how these fluctuations sum up over all nodes of the grid and over a whole day . the resulting overall deviation from the total balanced power of the starting reference configuration can be positive or negative .
figure [ fig : marketvolumes ] shows a distribution of cases when it turns out to be negative , so that the up - energy balancing market has to compensate for this missing power .
its total volume in gwh ( sum over all nodes of mismatched energy ) measured over all configurations of an ensemble , over the six zones of the italian grid , and over the reference day yields the histogram of figure [ fig : marketvolumes ] , .
the peak value of each distribution is stable , but the distributions for larger @xmath67-values are more skewed to high values than they are for low percentages of res . ] normalized over the total number of events , to predict the frequency of the various volume events .
different colors code the different scenarios , characterized by the varying contributions of res .
we present the distributions in two ways , as barplots , and as boxplots @xcite . while the median ( expected volume ) only slightly increases for a higher amount of res , the number of extreme market sizes of more than 10 gwh considerably increases by roughly 50% between 24% and 60% of res .
the fact that the median is relatively insensitive to the amount of res may be due to the fact that the strongest fluctuations in size are due to load fluctuations .
however , the whole distribution gets more skewed for a larger amount of res .
therefore rare but large market sizes should lead to high energy prices in the corresponding market sessions .
this must be expected even more in view of the nonlinear relation between price and demand , which is ruled by the power stack function @xcite , emerging from the market session ( agent based modeling part ) , so that moderate increases in the demand can lead to high changes in the price .
here a remark is in order about the choice of the ensemble size , over which the observables are measured : we have chosen thousand configurations , after 3000 configurations were reserved for the learning phase of the agents .
this size seems to be sufficient , see figure [ fig : marketvolumescomparison ] , in which we compare the market volume for two samples of an ensemble with the same reference configuration and with 60% of renewables ( for which the contribution of fluctuations is most pronounced ) .
therefore the differences here are of purely statistical nature and due to the ensemble size .
as it is seen from figure [ fig : marketvolumescomparison ] , the median and quartiles of the distributions vary then less than 1% , differences are seen in the outliers . .
the plot is organized as barplot showing the two distributions in the form of histograms , and as a boxplot in the inset .
the two boxes of the boxplot show the distribution s median ( red line ) , the first and third quartiles ( blue lines next to the red one ) , and its statistical minimum and maximum , together with the outliers . ] moreover , the same analysis of the market volume was performed by assuming the wind fluctuations to be weibull - distributed .
the results are shown in figure [ fig : marketvolumeweibull ] .
the results for the market do not sensibly change as long as @xmath68 , while for @xmath69 the fat tail of wind fluctuations has an impact on the volume distribution . in particular , the mode of the distribution is shifted from about 9.75 gwh to 10.5 gwh .
so for larger contributions of renewables one should be aware of the possibility that the non - gaussian fluctuations in power generation finally may shift the prices and costs towards higher values , which we here have not pursued . distribution . ] the actual results for the market prices and costs during one day under the assumption of gaussian fluctuations are displayed in the following figures . in figure
[ fig : costs_daily_stacked ] we show the probability distribution of the market costs per day , for different values of @xmath62 . again
different colors represent the different scenarios , the height of the columns is a measure for the probability to have a daily cost in the covered interval . as shown in the boxplot in the inset , the median and quartiles of these distributions do not sensibly differ , similarly to figure [ fig : marketvolumes ] .
this does not come as a surprise , as the balancing market is only sensitive to the power mismatch @xmath1 , where the amount of res mainly influenced the tails of the distribution rather than the median .
so , also here the effect of a larger contribution of renewables is visible in the tail of the histogram of the main figure : high values of res generation cause more likely very high costs .
for example , the number of events , for which the costs exceed 3 million euro is roughly 4% for @xmath70 as compared to 5% for @xmath69 renewables .
similar features are observed if the distribution of daily costs are resolved in energy prices averaged over four data per hour over all hours a day , see figure [ fig : prices_histogram ] , where the third quartile and upper whiskers reflect a fat tail in the price distribution . as a further result of the simulations , figure [ fig : gen_profit ] shows the average profit made by generators of different technologies during the day , for different values of @xmath62 .
it was calculated from the outcomes of the market sessions for each configuration , sorted with respect to the used technology of the generators , represented by the agents , and then averaged over the ensembles , the zones and the day .
we list generators based on coal , combined cycle , turbogas , and oil technologies .
it is worthwhile to stress how the average profit changes by changing @xmath62 .
an increase in @xmath62 causes an increase in the profit made by fast generators like turbogas or oil .
when high market volumes become likely due to strong fluctuations , the market authority will be forced to accept very high bids , this way encouraging a higher risk propensity for this type of generators . .
although the prices medians and quartiles do not sensibly increase with the increase in the share of res , more events with extreme costs are observed . ] .
an increase in @xmath62 causes an increase in the amount of energy that is sold at high prices . ] .
generators with high production costs are more likely called in the market when a high amount of renewables is present . ]
following the methodology as proposed in @xcite , we combined agent based modeling with a description from statistical physics by calculating the power mismatch over an ensemble of configurations that differ by small deviations from a stable power grid configuration .
the power mismatch entered as market volume in the energy balancing market , whose agents offer energy from a certain learned distribution of prices , until the missing amount of energy is covered by their offers .
the original fluctuations in load and renewables lead to fluctuations in the daily costs , handled in the balancing market , and the fluctuating energy prices .
while the average values of prices and actual costs only slightly increased for a higher percentage of shared renewables , the shape of the distributions became more skewed , and the number of accepted offers at extreme prices considerably increased with an increasing amount of renewables .
so far we have reduced the production of conventional generators as required by the artificially increased production of res , while keeping the load and geographical location of production places fixed , that is , without closing any conventional production places .
moreover , power reserve in the background was kept for sale only from online conventional generators , while different technologies may share the reserve in future grid extensions .
also the underlying datasets should be further updated in view of subsequent real extensions of res in the grid .
the methodology can be extended towards including other market phases than the short - time balancing market , different strategies of the agents in placing their bids , different price policies beyond the mere minimization of the energy - balancing market costs ; also additional players may be included in the market games .
so far they were restricted to agents from conventional production places . at the same time
high fluctuations of res may be damped by installing storage devices : before the fluctuations are fed into the grid , high peaks may be cut - off via storing some part of the production peaks . in any case , our methodology together with these various possible extensions allow for a quantification of the risk of high price peaks .
so they aid decisions on whether countermeasures against such peaks are worth the effort , or the rare peaks are so rare to be safely ignored .
the authors gratefully acknowledge the support from the german federal ministry of education and research ( bmbf grant no . 03sf0472d ( net-538 - 167 ) ) .
10 url # 1`#1 ` urlprefix href # 1#2#2 # 1#1 s. g. goyena , o. a. sdaba , sizing and analysis of big scale and isolated electric systems based on renewable sources with energy storage , 1st ieee - pes / ias conference on sustainable alternative energy , sae 2009 - proceedings ( 2009 ) 17http://dx.doi.org/10.1109/sae.2009.5534837 [ ] .
m. nicolosi , http://www.sciencedirect.com/science/article/pii/s0301421510005860[wind power integration and power system flexibility an empirical analysis of extreme events in germany under the new negative price regime ] ,
energy policy 38 ( 11 ) ( 2010 ) 72577268 . http://dx.doi.org/10.1016/j.enpol.2010.08.002 [ ] . http://www.sciencedirect.com/science/article/pii/s0301421510005860 f. j. nogales , j. contreras , a. j. conejo , r. espinola , http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=1007902[forecasting next - day electricity prices by time series models ] , ieee transactions on power systems 17 ( 2 ) ( 2002 ) 342348 .
j. contreras , r. espinola , f. j. nogales , a. j. conejo , http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=1216141[arima models to predict next - day electricity prices ] , ieee transactions on power systems 18 ( 3 ) ( 2003 ) 10141020 .
a. weidlich , d. veit , http://www.sciencedirect.com/science/article/pii/s0140988308000170[a critical survey of agent - based wholesale electricity market models ] , energy economics 30 ( 4 ) ( 2008 ) 17281759 .
http://www.sciencedirect.com/science/article/pii/s0140988308000170 a. c. tellidou , a. g. bakirtzis , agent - based analysis of capacity withholding and tacit collusion in electricity markets , ieee transactions on power systems 22 ( 4 ) ( 2007 ) 17351742
. http://dx.doi.org/10.1109/tpwrs.2007.907533 [ ] .
g. x. g. xiong , t. hashiyama , s. okuma , an electricity supplier bidding strategy through q - learning , ieee power engineering society summer meeting , 3 ( 2002 ) 15161521 . http://dx.doi.org/10.1109/pess.2002.1043645 [ ] .
m. mureddu , g. caldarelli , a. chessa , a. scala , a. damiano , green power grids : how energy from renewable sources affects networks and markets , plos one 10 ( 9 ) ( 2015 ) e0135312 .
http://dx.doi.org/10.1371/journal.pone.0135312 [ ] .
j. nicolaisen , v. petrov , l. tesfatsion , http://ieeexplore.ieee.org/articledetails.jsp?arnumber=956714 http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=956714[market power and efficiency in a computational electricity market with discriminatory double - auction pricing ] , ieee transactions on evolutionary computation 5 ( 5 ) ( 2001 ) 504523 . http://dx.doi.org/10.1109/4235.956714 [ ] .
http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=956714[http://ieeexplore.ieee.org/articledetails.jsp?arnumber=956714 http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=956714 ] e. d. castronuovo , j. a. p. lopes , on the optimization of the daily operation of a wind - hydro power plant , ieee transactions on power systems 19 ( 3 ) ( 2004 ) 15991606 . http://dx.doi.org/10.1109/tpwrs.2004.831707 [ ] .
v. pappala , i. erlich , k. rohrig , j. dobschinski , http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=4813198[a stochastic model for the optimal operation of a wind - thermal power system ] , ieee transactions on power systems 24 ( 2 ) ( 2009 ) 940950 .
n. lu , r. diao , r. p. hafen , n. samaan , y. v. makarov , a comparison of forecast error generators for modeling wind and load uncertainty , ieee power and energy society general meeting ( 2013 ) 04http://dx.doi.org/10.1109/pesmg.2013.6672978 [ ] .
j. v. seguro , t. w. lambert , modern estimation of the parameters of the weibull wind speed distribution for wind energy analysis , journal of wind engineering and industrial aerodynamics 85 ( 1 ) ( 2000 ) 7584 .
http://dx.doi.org/10.1016/s0167-6105(99)00122-1 [ ] .
r. d. zimmerman , c. murillo sanchez , r. thomas , http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=5491276 http://ieeexplore.ieee.org/ielx5/59/5695072/05491276.pdf?tp=&arnumber=5491276&isnumber=5695072[matpower : steady - state operations , planning , and analysis tools for power systems research and education ] , power systems , ieee transactions on 26 ( 1 ) ( 2011 ) 1219 .
http://dx.doi.org/10.1109/tpwrs.2010.2051168 [ ] .
http://ieeexplore.ieee.org/ielx5/59/5695072/05491276.pdf?tp=&arnumber=5491276&isnumber=5695072[http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=5491276 http://ieeexplore.ieee.org/ielx5/59/5695072/05491276.pdf?tp=&arnumber=5491276&isnumber=5695072 ] m. rastegar , e. guerci , s. cincotti , http://ieeexplore.ieee.org/articledetails.jsp?arnumber=5207128[agent-based model of the italian wholesale electricity market ] , in : 2009 6th international conference on the european energy market , ieee , 2009 , pp .
statistico - energia da fonti rinnovabili ] , tech .
, gse - gestore dei servizi energetici ( 2014 ) . | an increase in energy production from renewable energy sources is viewed as a crucial achievement in most industrialized countries .
the higher variability of power production via renewables leads to a rise in ancillary service costs over the power system , in particular costs within the electricity balancing markets , mainly due to an increased number of extreme price spikes .
this study focuses on forecasting the behavior of price and volumes of the italian balancing market in the presence of an increased share of renewable energy sources .
starting from configurations of load and power production , which guarantee a stable performance , we implement fluctuations in the load and in renewables ; in particular we artificially increase the contribution of renewables as compared to conventional power sources to cover the total load .
we then forecast the amount of provided energy in the balancing market and its fluctuations , which are induced by production and consumption . within an approach of agent based modeling
we estimate the resulting energy prices and costs .
while their average values turn out to be only slightly affected by an increased contribution from renewables , the probability for extreme price events is shown to increase along with undesired peaks in the costs .
renewable energy , electricity markets , statistical physics , agent based modeling |
it is commonly believed that the cosmological _ singularity _ problem @xcite may be resolved in a theory which unifies gravity and quantum physics .
recent analyses done within the _ loop _ quantum cosmology ( lqc ) concerning homogeneous isotropic universes of the friedmann - robertson - walker ( frw ) type , strongly suggest that the evolution of these universes does not suffer from the classical singularity . strong _
quantum _ effects at the planck scale cause that classical big bang is replaced by quantum big bounce ( bb ) @xcite .
the resolution of the cosmic singularity problem offered by lqc requires the existence of a fundamental _ length _ , which effectively implies the discreteness of quantum geometry .
however , the _ size _ of this length has not been determined satisfactory yet , i.e. derived within lqc .
presently , it is an _ ad - hoc _ assumption of standard lqc @xcite .
our paper is an extended version of the _ classical _ part of @xcite .
its goal is the demonstration that the resolution of the initial cosmological singularity is due to the _ modification _ of the classical theory by the loop geometry .
quantum effects seem to be of a secondary importance , but should be examined since the energy scale specific to bb has not been identified yet @xcite .
the big bounce may occur deeply inside the planck scale where it is commonly expected that quantum effects can not be ignored .
the difference between _ standard _
lqc @xcite and our _ nonstandard _ lqc is the following : ( i ) we determine an _ algebra _ of observables on the kinematical phase space ; ( ii ) we _ solve _ the hamiltonian constraints at the classical level to identify _ physical _ phase space ( i.e. the space of dirac s observables ) ; ( iii ) we express functions on the physical phase space , like the matter density and geometrical operators ( length , area , volume ) , in terms of dirac s observables and an _ evolution _
parameter ; and ( iv ) by _ quantization _ we mean : finding a self - adjoint representation of the algebra of the dirac observables and solution to the eigenvalue problem for operators corresponding to functions specified in ( iii ) . roughly speaking , in standard lqc
one does not identify an _
algebra _ of physical observables and one imposes the hamiltonian constraints _ only _ at the quantum level .
the present paper concerns the _ classical _ level so we are mainly concern with items ( i)-(iii )
. our next paper @xcite will be an extended version of the _ quantum _ part of @xcite , i.e. it will be devoted to the realization of item ( iv ) . for simplicity of exposition
we restrict ourselves to the flat frw model with massless scalar field .
this model of the universe includes the initial cosmological singularity and has been intensively studied recently within lqc . in order to have our paper
self - contained , we recall in sec .
ii the form of classical hamiltonian in terms of holonomy - flux variables . in sec .
iii we solve the constraint and analyze the relative dynamics in the _ physical _ phase space .
section iv is devoted to the algebra of _
we consider the energy density of the scalar field and geometrical operators in sec . v.
higher order holonomy corrections are shortly discussed in the appendix .
we conclude in the last section .
the gravitational part of the classical hamiltonian , @xmath0 , in general relativity is a linear combination of the first - class constraints , and reads @xcite @xmath1 where @xmath2 is the spacelike part of spacetime @xmath3 , @xmath4 denote lagrange multipliers , @xmath5 are the gauss , diffeomorphism and scalar constraint functions . in our notation @xmath6 are spatial and @xmath7 internal @xmath8 indices .
the constraints must satisfy a specific algebra .
having fixed local gauge and diffeomorphism freedom we can rewrite the gravitational part of the classical hamiltonian ( for the flat frw model with massless scalar field ) in the form ( see , e.g. @xcite ) @xmath9 where @xmath10 is the barbero - immirzi parameter , @xmath11 is an elementary cell , @xmath2 is spacelike hyper - surface , @xmath12 denotes the lapse function , @xmath13 is the alternating tensor , @xmath14 is a densitized vector field , @xmath15 , and where @xmath16 is the curvature of an @xmath8 connection @xmath17 .
the resolution of the singularity , obtained within lqc , is based on rewriting the curvature @xmath18 in terms of holonomies around loops .
the curvature @xmath18 may be determined @xcite by making use of the formula ( see the appendix ) @xmath19 where @xmath20 is the holonomy of the gravitational connection around the square loop @xmath21 , considered over a face of the elementary cell , each of whose sides has length @xmath22 ( where @xmath23 ) with respect to the flat fiducial metric @xmath24 ; fiducial triad @xmath25 and cotriad @xmath26 satisfy @xmath27 ; the spatial part of the frw metric is @xmath28 ; @xmath29 denotes the area of the square ; @xmath30 is the fiducial volume of @xmath31 .
figure 1 shows geometrical setup for determination of @xmath32 .
are chosen to diagonalise the connection @xmath33.,scaledwidth=55.0% ] the holonomy along straight edge @xmath34 of length @xmath35 reads @xmath36 where @xmath37 ( @xmath38 are the pauli spin matrices ) and @xmath39 denotes the path ordering symbol .
equation ( 5 ) presents the holonomy calculated in the fundamental , j = 1/2 , representation of su(2 ) .
the _ connection _ @xmath33 and the density weighted _ triad _ @xmath40 which occurs in ( [ identt ] ) is determined by the conjugate variables @xmath41 and @xmath42 as follows : @xmath43 and @xmath44 , where @xmath45 and @xmath46 .
equation ( [ hol ] ) presents the holonomy calculated in the fundamental , @xmath47 , representation of @xmath8 . making use of ( [ hamg ] ) , ( [ cur ] ) and the so - called thiemann identity @xcite @xmath48 leads to @xmath0 in the form @xmath49 where @xmath50 and where @xmath51 is the volume of the elementary cell @xmath52 .
the classical total hamiltonian for frw universe with a massless scalar field , @xmath53 , reads @xmath54 where @xmath0 is defined by ( [ hamr ] ) .
the hamiltonian of the scalar field is known to be : @xmath55 are the elementary variables satisfying @xmath56 .
the relation @xmath57 defines the _ physical _ phase space of considered gravitational system with constraints . making use of ( [ hol ] ) we calculate ( [ haml ] ) and get the _ modified _ total hamiltonian @xmath58 corresponding to ( [ ham ] ) in the form @xmath59 where @xmath60 are the canonical variables proposed in @xcite .
the variable @xmath61 so it corresponds to the hubble parameter @xmath62 , whereas @xmath63 is proportional to the scale factor @xmath64 .
the relationship between the coordinate length @xmath65 ( which depends on @xmath42 ) and the physical length @xmath66 ( which is a constant ) reads @xmath67 the complete poisson bracket for the canonical variables @xmath68 is defined to be @xmath69 + \frac{\partial \cdot}{\partial \phi } \frac{\partial \cdot}{\partial p_\phi } - \frac{\partial \cdot}{\partial p_\phi } \frac{\partial \cdot}{\partial \phi}.\ ] ] the dynamics of a canonical variable @xmath70 is defined by @xmath71 where @xmath72 , and where @xmath73 is an evolution parameter .
the dynamics in the _ physical _ phase space , @xmath74 , is defined by solutions to ( [ dyn ] ) satisfying the condition @xmath75 .
the solutions of ( [ dyn ] ) ignoring the constraint @xmath75 are in the _ kinematical _ phase space , @xmath76 . in what follows
we apply the dirac method of dealing with hamiltonian constraints @xcite : the poisson bracket is worked out before one makes use of the constraint equations .
in the case a hamiltonian is a _ constraint _ which may be rewritten in the form of a product of a simpler constraint and a function on @xmath76 which has no zeros , the original dynamics may be reduced ( to some extent ) to the dynamics with the simpler constraint . equation ( [ regh ] ) can be rewritten as @xmath77 where @xmath78 where @xmath79 .
it is clear that @xmath80 only in the case when @xmath81 .
such case , due to ( [ 1a])-([5a ] ) , implies no dynamics .
an equation of motion for a function @xmath82 defined on _ physical _ phase space , due to([dyn ] ) , reads @xmath83 since @xmath84 by analogy , for other function @xmath85 we have @xmath86 therefore we have the relation @xmath87 which we rewrite in the form @xmath88 equation ( [ integ ] ) shows that in the case of the _ relative _ dynamics we have : ( i ) one phase space variable may be used as an ` evolution parameter ' of all other variables , ( ii ) dynamics is gauge independent in the sense that there is no dependance on the specific choice of the lapse function @xmath12 , and ( iii ) suitable choice of @xmath12 may lead to a simpler form of hamiltonian .
since the relative dynamics is gauge independent , it is reasonable to chose the gauge which simplifies the calculations .
our choice is @xmath89 . in this gauge
the equations of motion read @xmath90 combining ( [ 3a ] ) with ( [ 4a ] ) gives @xmath91 rewriting ( [ pp1 ] ) ( and using @xmath92 ) gives @xmath93 making use of the identity @xmath94 and ( [ 5a ] ) gives @xmath95 combining ( [ zcos ] ) with ( [ nacos ] ) , for @xmath960 , \pi/2 \lambda[$ ] , leads to @xmath97 since @xmath98 is just a constant ( due to ( [ 1a ] ) ) we can easily integrate ( [ dynamika ] ) and get @xmath99 rewriting ( [ int ] ) leads to @xmath100 the solution for the variable @xmath101 may be easily determined from ( [ 5a ] ) rewritten as @xmath102 the final expression reads @xmath103 where the domain of the variable @xmath101 has been extended to the interval @xmath1040,\pi/\lambda[$ ] .
equations ( [ res2 ] ) and ( [ res4 ] ) present the dependence of the canonical variables @xmath105 and @xmath101 on the evolution parameter @xmath53 , which is a monotonic function due to ( [ 3a ] ) .
a function , @xmath106 , defined on phase space is a dirac observable if @xmath107 since we have @xmath108 it is clear that on the constraint surface , @xmath109 , the dirac observable satisfies ( independently on the choice of @xmath12 ) a much simpler equation @xmath110 thus , we put @xmath111 and solve ( [ dirac ] ) in the whole phase space , i.e. we solve the equation @xmath112 a function @xmath113 satisfies ( [ dir ] ) if @xmath114 where @xmath115 is the dimension of the _ kinematical _ phase space .
it is so because one has @xmath116 in what follows we consider only _ elementary _ observables .
the set of such observables , @xmath117 , is defined by the requirements : ( i ) each element of @xmath117 is a solution to ( [ dir ] ) , ( ii ) elements of @xmath117 are functionally independent on the constraint surface , @xmath109 , ( iii ) elements of @xmath117 satisfy a lie algebra , and ( iv ) two sets of observables satisfying two algebras are considered to be the same if these algebras are isomorphic . in our case @xmath118 and solutions to ( [ dir ] )
are found to be @xmath119 where @xmath120 .
one may verify that the observables satisfy the lie algebra @xmath121 because of the constraint @xmath109 ( see ( [ res3 ] ) ) , we have @xmath122 thus , we have only two elementary dirac observables which may be used to parameterize the physical phase space @xmath74 . to identify the poisson bracket in @xmath74 consistent with the poisson bracket ( [ re2 ] ) defined in @xmath76 , we find a symplectic twoform corresponding to ( [ re2 ] ) .
it reads @xmath123 the twoform @xmath124 is degenerate on @xmath74 due to the constraint @xmath109 . making use of the explicit form of this constraint ( [ res3 ] ) and
the functional form of @xmath125 and @xmath126 , leads to the symplectic form @xmath127 on @xmath74 .
direct calculations give ( see app .
b ) @xmath128 where @xmath129 denotes the reduction of @xmath124 to the constraint surface .
the poisson bracket corresponding to ( [ sym2 ] ) reads @xmath130 so the algebra satisfied by @xmath125 and @xmath126 has a simple form given by @xmath131 our kinematical phase space , @xmath76 , is four dimensional . in _ relative
_ dynamics one variable is used to parametrize three others .
since the constraint relates the variables , we have only two independent variables .
this is the reason we have only two elementary physical observables parametrizing @xmath74 .
in this section we discuss the functions on the _ constraint _ surface that may describe singularity aspects of our cosmological model .
considered functions are not observables , but they can be expressed in terms of observables and an evolution parameter @xmath53 .
they do become observables for each fixed value of @xmath53 , since in such case they are only functions of observables .
an expression for the energy density @xmath132 of the scalar field @xmath53 reads @xmath133 in terms of elementary observables we have @xmath134 for fixed @xmath98 the density @xmath132 takes its maximum value at the minimum value of @xmath105 . rewriting ( [ res2 ] ) in the form @xmath135 we can see that @xmath136 takes minimum value equal to one at @xmath137 . thus , the maximum value of the density , @xmath138 , corresponds to @xmath139 and reads @xmath140 we can determine @xmath138 if we know @xmath66 .
however , @xmath66 is a free parameter of the formalism .
thus , finding the critical energy density of matter corresponding to the big bounce is an open problem .
it is tempting to apply ( [ cr1 ] ) to the planck scale . to make use of planck s length @xmath141 and planck s energy density @xmath142 , we multiply ( [ cr1 ] ) by @xmath143 and recall that @xmath79 .
thus ( [ cr1 ] ) reads @xmath144 substituting @xmath145 into ( [ ps1 ] ) gives @xmath146 . resolving ( [ ps1 ] ) in terms of @xmath66 makes possible finding @xmath66 corresponding to @xmath147 .
we get @xmath148 .
( we have used @xmath149 determined in black hole entropy calculations @xcite . )
surprisingly , the classical expression ( [ ps1 ] ) fits the planck scale .
a natural next step is an examination of the energy scale of the big bounce at the quantum level .
preliminary calculations , @xcite , suggest that the energy scale would be described by the classical expression ( [ ps1 ] ) . in the case
the volume @xmath150 of an elementary cell @xmath52 is a cube , geometrical operators have simple dependence on canonical variables . since the volume operator , @xmath151 , is given by @xmath152 the area operator , @xmath153 , reads @xmath154 and the length operator , @xmath155 , is found to be @xmath156 as far as we know , the geometrical operators have been considered so far only in the _ kinematical _ hilbert space of the loop quantum _ gravity _ @xcite .
we propose the examination of these operators in the _ physical _ hilbert space of the loop quantum _
cosmology_. it results from ( [ res3 ] ) that we have @xmath157
so the geometrical operators are bounded from below by zero ( as @xmath158 ) .
an examination of the spectra of the geometrical operators at the quantum level is the next natural step .
an interesting question is : do these operators have the nonzero minimum and discrete eigenvalues ?
it is expected that the story will turn out to be similar to the case of the linear harmonic oscillator , where we have the nonzero ground - state energy and discrete energy levels .
we present an answer to this intriguing question in our forthcoming paper @xcite .
we have shown that the resolution of the initial singularity of the flat frw model with massless scalar field is due to the modification of the model at the _ classical _ level by making use of the loop geometry .
the modification is parametrized by the _ continuous _ parameter @xmath159 .
each value of @xmath66 specifies the critical energy density of the scalar field corresponding to the big bounce . as there is no specific choice of @xmath66
, the bb may occur at any low and high densities .
the former case ( big @xmath66 ) contradicts the data of observational cosmology ( there was no bb in the near past ! ) and leads to weakly controlled modification of the expression for the curvature @xmath18 , i.e. gravitational part of the hamiltonian ( see the appendix ) .
the latter case ( small @xmath66 ) gives much better approximation for the classical hamiltonian ( see the appendix ) , but may easily lead to densities much higher than the planck scale density , where the classical formalism is believed to be inadequate . finding specific value of the parameter @xmath66 ,
i.e. the _ energy scale _ specific to bb is an open problem .
our approach is quite different from the so - called effective or polymerization method ( see , e.g. @xcite ) , where the replacement @xmath160 in the hamiltonian finishes the procedure of quantization . in our method this replacement has been done entirely at the _ classical _ level : eq .
( 10 ) results from using an explicit form of the holonomy ( 5 ) in ( 8) .
quantization consists in finding a self - adjoint representation of observables on the physical phase space and an examination of the spectra of these observables @xcite .
the elementary observables @xmath125 and @xmath126 constitute a complete set of constants of motion in the constraint surface .
they are used to parametrize the physical phase space and are `` building blocks '' for the compound observables like the energy density of the scalar field and the geometrical operators .
so they have deep physical meaning .
the role of @xmath125 and @xmath126 becomes even more important at the quantum level as they enable finding quantum operators corresponding to the classical compound observables @xcite .
our theoretical framework may be used for examination of the _ discreteness _ aspects of geometrical _ quantum _ operators , which may help in the selection of @xmath66 .
an extension of our formalism to the quantum level is straightforward .
the algebra of observables is defined in the physical phase space ( hyper - surface in the kinematical phase space determined by the constraint equation ) .
the carrier space of the self - adjoint representation of the algebra defines the _ physical _ hilbert space .
examination of the eigenvalue problem of the length , area and volume operators ( for fixed value of an evolution parameter they are observables ) may lead to the specification of an unique @xmath66 ( or an interval of allowed values ) .
our next paper @xcite is devoted to examination of these problems .
it may happen , however , that the value of the parameter @xmath66 can not be determined , for some reason , theoretically .
the story may turn out to be similar to the case of the short - range _ repulsive _ part of the potential of the nucleon - nucleon interaction introduced to explain the scattering data @xcite and the nuclear matter saturation of energy @xcite .
in such a case @xmath66 will become a phenomenological variable parameterizing our ignorance of microscopic properties of the universe .
an independent source of information on discreteness aspects of geometry is the observational cosmology .
the cosmic projects for the detection of gamma ray bursts may reveal that the velocity of cosmic photons depend on their wave lengths , which may be ascribed to the _ foamy _ nature of spacetime @xcite .
the detection of the primordial gravitational waves created at the big bounce may bring valuable information on the geometry of this phase @xcite .
the observational cosmology data may help to determine the phenomenological value of the parameter @xmath66 .
we are grateful to janusz dabrowski , jacek jezierski , and andrzej krlak for helpful discussions .
we would like to thank the anonymous referee for the suggestion of including @xcite into our list of references .
the curvature of @xmath8 connection @xmath161 , entering the expression ( [ hamg ] ) for the gravitational part of the hamiltonian , can be expressed in terms of holonomies . using the mean - value and stokes theorems we have@xmath162 where @xmath163 is the boundary of a small surface @xmath164 with center at @xmath165 , and where @xmath166 .
the expression for @xmath18 is exact but in the limit when we shrink the area enclosed by the loop @xmath163 to zero .
if we choose @xmath163 in the form of the square @xmath167 with sides length @xmath65 , the expression for a _ small _ value of @xmath168 has the form @xcite @xmath169 and we have @xmath170 in the standard lqc the @xmath171 holonomy corrections are ignored ( see , e.g. @xcite ) . it was found in @xcite that including higher order corrections leads to new curvature singularities different from the initial singularity and increases an ambiguity problem of loop cosmology .
however , the holonomy corrections do not change the result that the big bounce is a consequence of the loopy nature of geometry @xcite . taking only the first term of ( [ finite ] ) leads to the simplest modification of gravity , but may be insufficient for the description of the inflationary phase .
the choice of @xmath172 based on the expectation that big bounce should occur at the planck scale @xcite has little justification @xcite .
the significance of planck s scale for quantum gravity seems to be rather a belief than proved result ( see , e.g. @xcite ) .
heuristic reasoning playing game at the same time with heisenberg s uncertainty principle , schwarzschild s radius and process of measurement can not replace a proof ( see , e.g. @xcite ) .
the symplectic form on the physical phase space @xmath127 may be obtained from the symplectic form on the kinematical phase space @xmath124 by taking into account the constraint ( [ res3 ] ) .
the symplectic form corresponding to ( [ re2 ] ) reads @xmath173 making use of ( [ res3 ] ) we get @xmath174 insertion of ( [ ff2 ] ) into ( [ form1 ] ) gives @xmath175 since @xmath176 we have @xmath177 on the other hand , due to ( [ obser1 ] ) , we have @xmath178 therefore , @xmath179 thus , the physical phase space may be parametrized by the variables @xmath180 and @xmath181 , and the corresponding poisson bracket is given by ( [ sym3 ] ) .
j. albert _ et al .
_ `` probing quantum gravity using photons from a flare of the active galactic nucleus markarian 501 observed by the magic telescope '' phys . lett .
b * 668 * ( 2008 ) 253 [ arxiv:0708.2889 [ astro - ph ] ] .
j. grain , a. barrau and a. gorecki , `` inverse volume corrections from loop quantum gravity and the primordial tensor power spectrum in slow - roll inflation , '' phys
. rev .
d * 79 * ( 2009 ) 084015 [ arxiv:0902.3605 [ gr - qc ] ] . | the big bounce ( bb ) transition within a flat friedmann - robertson - walker model is analyzed in the setting of loop geometry underlying the loop cosmology .
we solve the constraint of the theory at the classical level to identify physical phase space and find the lie algebra of the dirac observables .
we express energy density of matter and geometrical functions in terms of the observables .
it is the modification of classical theory by the loop geometry that is responsible for bb .
the classical energy scale specific to bb depends on a parameter that should be fixed either by cosmological data or determined theoretically at quantum level , otherwise the energy scale stays unknown . |
coronal mass ejections ( cmes ) are one of the leading causes of space weather , especially when they have organized southward directed magnetic field .
therefore , it is of great importance for space weather forecasting to understand how their direction of propagation and their orientation at 1 au relate to properties on the solar disk ( location of source region , orientation of the polarity inversion line , etc ... ) .
the deflection of cmes in the latitudinal direction has been observed and reported since the launch of the first space coronagraphs in the 1970s and 1980s .
for example , @xcite reported an average deflection of 2.2@xmath1 towards the equator for 29 cmes during solar minimum ( 19731974 ) , while they found no systematic deflection for 19 cmes during solar maximum ( 1980 ) . with the launch of the _ solar - terrestrial relations observatory _ ( stereo ) in 2006 , there has been a renewed focus on the deflection of cmes in the corona and the heliosphere , thanks to stereoscopic measurements from the two spacecraft ( e.g. , see : * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
the presence of coronal holes is known to affect the direction of propagation of cmes @xcite .
however , the exact cause of this influence is uncertain . @xcite and @xcite have used an _ ad hoc _ , fictitious force exercised by coronal holes on cmes , which is directly dependent on the coronal hole area and the distance of the cme source region from the coronal hole
. this force might be due to a `` pushing '' of the cme by the fast wind from the coronal hole @xcite .
@xcite proposed that the effect of coronal holes is due to strong magnetic fields , @xcite proposed that the non - radial motion is due to the guiding action of the coronal magnetic field , while @xcite and @xcite recently proposed that magnetic pressure gradient acting on a cme results in a net force directed along the gradient of the magnetic pressure . obviously , magnetic forces , such as the lorentz force could also cause the coronal hole to exercise a force on the cme .
magnetic forces are expected to affect strongly the evolution and propagation of a cme in the corona while the effect of the fast solar wind from the coronal hole would continue into the heliosphere and be more gradual .
similar effects may result in the rotation of an eruption .
one of the leading causes of cme rotation is the kink instability @xcite but recent numerical efforts have focused on other sources of rotation , such as reconnection with the background magnetic field @xcite , the effect of the lorentz force @xcite and its association with shearing motions @xcite .
these numerical works point towards a near universal rotation for cmes not initially aligned with the heliospheric current sheet ( hcs ) . while statistically speaking , cmes rotate towards the hcs @xcite , it appears from observations that some cmes do not rotate , even though they are originally not aligned with the hcs ( see some cases in * ? ? ?
* ; * ? ? ?
eruptions from active regions which are inside a coronal hole present a perfect configuration to study the effect of coronal holes on cmes .
for example , the fictitious force of @xcite is inversely proportional to the square of the distance between the source region and the coronal hole , and it results in a infinite force for a cme originating from such an active region !
active regions can form inside coronal holes and , then , they often develop into anemone active regions @xcite .
anemone active regions were originally observed in soft x - ray and named x - ray fountains @xcite due to the structure looking like a fountain with loops emerging in all directions from the center .
almost identical structures are observed in extreme ultraviolet ( euv ) and in chromospheric lines @xcite . in a survey of active regions observed by the soft x - ray telescope ( sxt ,
see : * ? ? ?
* ) onboard _ yohkoh _
@xcite , @xcite found that as many as 25@xmath3 of active regions observed in 19911992 appeared as anemone active regions at one time during their evolution , and that there is a near equivalence between an active region having an anemone structure and being inside a coronal hole .
anemone active regions have been primarily studied as the source of x - ray jets since the emergence of a bipolar active region inside unipolar open magnetic fields naturally yield such phenomena @xcite . of greater importance for space weather
are the instances of cmes and filament eruptions from anemone active regions @xcite .
@xcite found that eruptions from active regions inside open magnetic field ( which contain anemone active regions ) are statistically faster on average than other eruptions ( even originating from under the heliospheric current sheet ) .
most authors have explained this result by invoking two facts .
since the eruption originates from a region of low - density , fast solar wind , the fast wind can `` push '' the cmes to a faster speed .
closed field in the low corona can hinder the eruption of the cme because of the lorentz force ( see for example the model of * ? ? ? * ) and in the absence of closed field lines , the cme can erupt without being strongly decelerated . in the present work ,
we investigate the effect of the particular magnetic structure of anemone active regions on the evolution and coronal propagation of a cme .
we focus on the first of two eruptions in 2005 august 22 from active region noaa 10798 .
this cme was associated with the eruption of a southward - directed h@xmath0 filament and reached a speed of about 1200 km s@xmath4 ( for a complete overview of the observations , see * ? ? ?
the second eruption was faster ( @xmath5 2400 km s@xmath4 ) and it occurred about 16 hours after the first one from the same active region .
the eruption of a preceding cme from the same active region is expected to significantly change the magnetic topology of the source region @xcite and also modify the background solar wind into which the second cme propagates @xcite , which would complicate a numerical investigation of the second cme .
for this reason , we focus on the first cme . our investigation is based on a numerical magnetohydrodynamic ( mhd ) model part of the space weather modeling framework ( swmf , see : * ? ? ? * ) . in order to validate our simulation with euv observations and to reproduce more accurately the lower part of the corona , we use the low corona ( lc ) model recently developed by @xcite , which captures the energy balance of the coronal plasma .
the organization of the paper is as follows : in section [ sec : models ] , we briefly summarize the main features of the lc and flux rope models important for this study .
then , we discuss the magnetic topology of the anemone active region before presenting a comparison of the pre - event corona as observed in euv with the modeled corona in synthetic euv . in section [ sec : td ] , we follow the initial phases of the eruption and explain its interaction with the adjacent magnetic flux systems . in this section , we also discuss the cme aspect in real and synthetic coronagraphic images .
we study the cme rotation , deformation and deflection in section [ sec : rotation ] .
we discuss our results and conclude in section [ sec : conclusion ] .
the simulation is done using the low corona ( lc ) component of the swmf .
the lc model includes radiative losses , field - aligned electron heat condition and empirical heating terms in the energy equation as detailed in @xcite .
it has been recently used to study the nature of the extreme ultraviolet ( euv ) wave observed by the two _ stereo _ spacecraft in 2008 march 25 @xcite .
the initial magnetic field and boundary conditions are provided by a finite difference solution of the potential field source surface method ( pfssm : * ? ? ?
* ) for the synoptic magnetogram of carrington rotation 2033 as observed by the _ soho_/mdi instrument @xcite .
as noted in @xcite , ar 10798 was rapidly evolving as it crossed the sun central meridian and therefore , the synoptic map does not fully capture the complexity of the active region as revealed by observations on august 21 and 22 .
however , on these two days , the ar was already too close to the eastern limb of the sun , so that the daily line - of - sight magnetogram ( shown in the left panel of figure 1 ) can not be adequately used to constrain our model . as discussed below , even using the synoptic map
, the mhd model successfully reproduces the important features of the anemone active region .
as described in @xcite , we followed the previous works of @xcite and @xcite to widen the transition region by modifying the ratio of heat conduction to radiative cooling at temperatures below 300,000 k. the heating model and boundary conditions are the same as the ones described in @xcite , i.e. we use the chromospheric boundary conditions .
the grid is spherical with non - uniform radial scales near the transition region ( minimum uniform @xmath6 230 km ) with a resolution on the solar surface of @xmath7 .
in addition , we further refine around ar 10798 by 2 levels of refinement as follows .
the solar surface is refined twice in a box of about 20@xmath1 latitude and @xmath8 longitude around the active region over a height of 0.4 r@xmath2 and the region around the top of the anemone structure ( null point ) is refined by an additional factor of 2 in a twice smaller box centered at 1.15 r@xmath2 , resulting in a smallest angular resolution of @xmath9 .
the grid around ar 10798 is shown in the top left panel of figure 2 .
this version of the lc model includes a uniform parker - like solar wind model but not a bimodal solar wind model as in the models of @xcite and @xcite . here ,
the solar wind accelerates and becomes faster than the fast magnetosonic speed between 6 and 14 r@xmath2 and reaches a value of 300350 km s@xmath4 at the outer boundary of the domain ( 24 r@xmath2 ) . while it is not realistic for the fast wind originating from coronal holes , it is the approximate speed and acceleration profile expected for the slow solar wind .
we only focus on the coronal evolution of the cme , therefore resolving a bimodal solar wind is not critical to our study .
since the solar wind is uniform , we can expect a negligible amount of deflection and rotation due to hydrodynamical effect ( velocity shear , non - uniform drag ) .
it allows us to focus solely on the effect of magnetic forces and reconnection on the cme evolution .
compared to previous studies of the evolution of cmes in the corona with the swmf @xcite , using the lc model has a number of advantages .
the inclusion of the thermodynamics makes the treatment of the lower corona more realistic . using the chromospheric boundary conditions ,
the density and temperature of the plasma are based solely on the magnetic topology and the particular choice of heating function . also , as discussed in @xcite , it results in a less potential steady - state solution for the background coronal magnetic field .
last , it gives us the opportunity to validate our model with a direct comparison to extreme ultraviolet ( euv ) observations . to model the cme
, we employ a semi - circular flux rope prescribed by a given total toroidal current , as in the models by @xcite and @xcite .
an azimuthal current is also added at the surface of the flux rope in order to construct a force - free magnetic field inside the flux rope .
the toroidal current is largely dominant , creating a very highly twisted flux rope .
a more complete description of this implementation of the flux rope model can be found in @xcite and @xcite .
because in this version of the @xcite flux rope , there is neither sub - surface line current nor sub - surface magnetic charges to generate strapping field for the flux rope in the corona , the flux rope is not expected to rotate as found for example in @xcite .
additionally , the flux rope is constructed so that it is not kink unstable .
the flux rope solution once superimposed onto the coronal magnetic field leads to an immediate eruption because of force imbalance with the ambient magnetic field .
+ a filament was observed on august 21 ( see middle panel of figure 1 ) and the eruption of the northern section of the filament was the cause of the studied eruption @xcite .
we place the flux rope at a position and with an orientation in agreement with the observation of the filament ( see figure 1 ) .
the flux rope ( and the line current inside ) makes an angle of @xmath10 with the @xmath11 axis ( the sun rotation axis ) .
it results in a right - handed flux rope with an axial southward magnetic field , in agreement with the observations of the filament in h@xmath0 and with the reconstruction of the associated ejecta at 1 au @xcite .
the flux rope is making an angle of about 20@xmath1 with the polarity inversion line of ar 10798 with its negative ( resp .
positive ) footprint in the main part of the negative ( resp .
positive ) polarity spot of the ar .
the direction of the axial field of the flux rope is chosen in agreement with the observations of a southward ( sinistral ) filament @xcite which is also consistent with the prefilament structure ( see figure 4 from * ? ? ?
the orientation of the axis of the flux rope is consistent with the observations of a structure oriented from the northwest to the southeast .
the exact orientation is chosen so that the positive ( resp .
negative ) footprint of the flux rope are in a region of positive ( resp .
negative ) polarity , while the axis of the flux rope remains close to the polarity inversion line .
it should be noted that in our mhd model , because we use a synoptic magnetogram , the positive polarity spot of the active region ( on the west side of the negative spot ) is not yet fully developed while the positive polarity on the east side is stronger that it is on august 2122 .
the flux rope is chosen as right - handed to agree with the overall magnetic field in the overlying arcades once the direction of the axial field ( sinistral ) of the flux rope is chosen .
this is also consistent with what was reported in @xcite .
finally , the amount of twist is determined by the prescribed toroidal current .
the current is set by a trial - and - error procedure to match the coronal speed of the cme as observed by lasco ( 1200 km s@xmath4 ) . because of the way our flux rope is created ( with a dominant toroidal current ) , the chosen value of the current results in a flux rope with a much larger amount of twist as compared to what is derived from observations ( see review by * ? ? ? * ) .
previous time - dependent simulations with the lc model were performed with the cme initiation model of @xcite . here ,
we use , for the first time , a flux rope model in the lc model because the presence of a filament strongly suggests that a flux rope was present prior to the eruption .
this cme model has been used before to study the evolution of cme in the corona @xcite .
a view of the flux rope at time @xmath12 as it is superposed onto the steady - state corona is shown with blue field lines in the bottom panels of figure 2 .
the magnetic topology before the eruption is very important because it determines , to a great extent , the location of the reconnection of the flux rope with adjacent magnetic field structures during the eruption .
the steady - state corona around ar 10798 is relatively simple with : 1 ) a large - scale streamer connecting the active region with a region of the quiet sun about 40@xmath1 east of the active region , 2 ) open field lines with positive polarity , and , 3 ) active region closed field lines .
we found no magnetic connectivity between ar 10798 and adjacent active regions , even though ar 10797 is only about 2025@xmath1 away .
it is not surprising since ar 10798 is a reverse polarity active region ( negative - positive ) while ar 10797 follows the @xcite polarity law ( positive - negative ) .
therefore the positive parts of ars 10798 and 10797 are next to each other and , moreover , they are embedded into positive polarity open magnetic fields from the coronal hole .
this type of magnetic field configuration make it very unlikely that there was a direct magnetic connectivity between these two active regions .
the topology is much simpler than other cases near solar maximum , where @xcite found connectivity between as many as three active regions via multiple null points and quasi - separatrix layers .
there is one null point above the negative polarity part of ar 10798 separating the 3 flux systems as illustrated in the figure 2 : ar field ( pink ) , streamer field ( green ) and open field ( white ) .
because the negative polarity spot is surrounded by positive polarities , it develops into an anemone active region ( see top right panel of figure 2 ) as described before in @xcite and @xcite .
this null point is originally at a height of 0.09 r@xmath2 above the solar surface .
this value is probably lower than that on the sun at the time of the eruption because we use a synoptic map of the sun and ar 10798 was not fully developed .
the existence of close field lines as part of a unipolar region has been reported before @xcite .
the steady - state configuration is an embedded dipole , identical to that of the asymmetrical breakout model @xcite and similar to that of coronal jet studies ( e.g. , see * ? ? ?
* ) but with a closed outer spine . here
, the inner and outer spines are initially closed and there is a simple fan surface around the negative polarity spot of the active region .
the inner and outer spines as well as one of the lines of the fan surface are shown in red in figure 2 . as we superpose the flux rope onto the active region , the null point
is pushed by about 4@xmath1 toward the north - east and by 0.04 r@xmath2 upward , but the active region retains its anemone structure ( see bottom panels of figure 2 ) . in addition , a bald patch forms below the flux rope ( see bottom panels of figure 2 ) . because ar 10798 develops and maintains its anemone structure in the simulation as was observed on the sun , we believe our model captures the most important features of the solar corona before the 2005 august 22 eruptions .
+ one of the advantages of using the lc model is the ability to simulate extreme ultraviolet ( euv ) images following the procedure described in @xcite .
it allows us to validate our steady - state model by comparing synthetic images with real ones prior to the event .
the euv signal depends on the density and temperature of the plasma and it is greatly influenced by the magnetic structure of the lower corona .
therefore , this type of comparisons validates not only the plasma properties of our simulated corona but also its detailed magnetic structure .
we show a comparison of an 195 and 171 images from the euv imaging telescope ( eit , see : * ? ? ?
* ) onboard _ soho _ 18 hours before the eruption with synthetic images for the same filters from our simulation in figure 3 . synthetic and real images are plotted using the same scale .
the 195 filter response function peaks around 1.4 mk , while the 171 peaks around 1 mk , illustrating different heights in the corona . in synthetic and real images , the three active regions are clearly visible as regions of enhanced emission ( hot and dense ) : the large ar 10800 on the eastern side of the sun near disk center , ar 10797 near the western limb and ar 10798 near w50 .
in addition to the northern polar coronal hole , there are a number of equatorial coronal holes , including two on the south - western side of ars 10798 and 10800 .
the southern polar coronal hole is almost absent in the simulated and real images .
overall , there is good agreement between synthetic and real images , which gives us confidence that our model of the the corona is a relatively realistic representation of the actual corona at the time of the 2005 august 22 eruption .
the most important features for our study that the model reproduces are the appearance of ar 10798 and the presence and aspect of open field regions ( dark ) around it .
the main difference between simulated and real images is the eastern limb of the sun where the modeled emission is too weak as compared to the real one .
this is relatively unimportant because these regions are far from the source region of the cme ( more than 120@xmath1 separation ) and are not involved in the eruption process .
as soon as we superpose the flux rope onto the steady - state coronal magnetic field , it erupts due to force imbalance . as in other simulations with the same cme model , the exact kinematics of the cme early on in the corona are not realistic as the cme reaches its maximum speed ( @xmath13 km s@xmath4 ) about 1.5 minutes after the superposition of the flux rope .
however , the simulated cme kinematics past 3 r@xmath2 ( after 15 minutes ) are in good agreement with the height - time profile as observed in lasco / c2 . additionally , the cme speed after 1 hour is about 1200 km s@xmath4 , similar to what is measured with lasco ( @xmath14 km s@xmath4 ) .
it confirms that the total energy of the simulated cme is comparable to that of the real cme .
the model does not intend to capture the slow rise phase before the loss of equilibrium nor the acceleration phase .
this is why when comparing synthetic and real images , it is best not to use the onset time of the flare as the starting time of the numerical simulation but a later time when the cme has already significantly accelerated .
previous studies have found that the cme acceleration happens during the x - ray rise phase @xcite .
we use goes-12 @xcite and rhessi @xcite data to investigate the flare time in soft x - ray ( sxr ) and hard x - ray ( hxr ) , respectively . for the ejection ,
the flare onset in sxr was 00:44ut , the hxr flare started at 01:02ut and the flare peaked at 01:22ut .
we believe the onset of the hxr flare is the best time to use for the start time of our simulation since the cme was already observed by lasco at 4 r@xmath2 10 minutes after the flare peak .
figure 4 shows a line - of - sight image of the cme observed by lasco / c2 and processed with the method of @xcite and a synthetic line - of - sight image from our simulation .
the lasco image is taken at 01:54 ut , 52 minutes after the onset of the hxr flare , while the synthetic image is made 45 minutes after the superposition of the flux rope onto the solar surface .
the synthetic image is processed as explained in @xcite using a synthetic 27-day minimum image created from the steady - state simulation .
the main difference between the synthetic and real images is the latitudinal direction of propagation of the cme . in the synthetic image ,
the fastest moving part of the cme is at a pa of 210@xmath1 while it is about 220225@xmath1 for the real image ( this region is dimmer than the main part of the cme which is closer to pa 270 ) . because the cme appears as a halo
, it is not straight - forward to determine its central pa but the synthetic image shows a cme propagating about 1520@xmath1 more towards the south as what appears in the real image .
the active region was at the carrington latitude south 11 ( s11 ) when the ejection occurred and in our simulation , the cme propagates without a north - south deflection , i.e. along a latitude of s11 .
as noted in @xcite , only the northern part of the filament erupted during the flare / ejection studied here , while the southern part erupted during the second cme on august 22 .
it is therefore likely that the real cme propagated more towards the north than what our simulation shows since only part of the filament erupted . with our relatively simple model to start the eruption , it is not yet possible to study partial filament eruptions . as it rises and expands , the flux rope interacts and reconnects with the adjacent magnetic flux systems .
because this is a numerical study with finite resolution and based on numercial diffusion , the timing of the reconnection might not be realistic in our simulation . however , we believe that the reconnection process itself reflects what occurred on the sun . it is because : 1 ) our steady - state corona is a good representation of the pre - event sun ( see previous section ) , 2 ) the cme was due to a filament , which was originally at approximatively the same position with a similar length , and with the same orientation as our flux rope , and , 3 ) the cme kinematics are in good agreement with what was observed by lasco , meaning that the total energy in the flux rope is similar to that of the cme . in the section
below , we analyze the interaction of the flux rope with the corona and in section [ sec : dimmings ] , we present some observational consequences of the reconnection process , which further validate the discussed scenario .
only a few recent numerical works have focused on the interaction of a cme with a realistic magnetic structure obtained from magnetogram measurements @xcite .
other simulations have usually relied on an ideal representation of the corona with a single dipolar or quadrupolar active region in a dipolar sun ( as in * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for example ) . here
, the interaction of the flux rope with the background magnetic field is different from both these ideal cases and the realistic cases previously studied because of the anemone nature of the active region .
the topology is different from the solar minimum cases studied before with an isolated ar not surrounded by equatorial coronal holes @xcite and also different from the two solar maximum case studies of @xcite and @xcite with complex connectivity between multiple ars .
recently , @xcite studied in great detail the magnetic topology of the corona before and during the evolution phase leading to the 1997 may 12 cme .
their study was also for a relatively simple magnetic topology without influence from equatorial coronal holes .
+ the evolution of the cme is altered by the anemone nature of the active region , i.e. by the presence of unipolar magnetic field around it .
we find that the positive footpoint of the flux rope does not reconnect away from the active region since there is a very limited negative polarity magnetic flux around it . as a consequence ,
the flux rope remains line - tied at its positive footpoint . in contrast , the negative footpoint reconnects quickly and extensively with the positive magnetic flux , partly from the positive spot of ar 10798 and partly from the neighboring coronal hole .
the reconnection is a 2-phase process as illustrated in figure 5 . in the first phase ( top panels ) , all of the twisted closed field lines of the flux rope ( dark blue ) reconnect with the streamer magnetic field ( green ) .
it results in erupting field lines connecting the positive footpoint of the flux rope with the streamer , shown in yellow and post - flare loops in fushia . in the second phase ( bottom panels ) , the newly formed ( yellow ) erupting field lines reconnect with open field lines to form ( orange ) open and twisted erupting field lines .
below , we describe the detailed reconnection process and the change in magnetic topology during the eruption .
the reconnection involves two phases : the first one , when the negative footpoint of the flux rope reconnects outside the fan surface ( top panels of figure 5 ) and the second one , when some of the erupting field lines open up ( bottom panels of figure 5 ) .
the detailed reconnection process of this first phase goes as follows : first , the ( dark blue ) field lines of the flux rope reconnect with overlying anemone field ( pink ) to form twisted ( light blue ) field lines connecting the positive foopoint of the flux rope with the negative main spot of the active region ( top left panel of figure 5 )
. this reconnection happens at the bald patch . in a second step , the newly formed field lines ( light blue ) reconnect with closed ( green ) field lines outside of the fan surface to form the ( yellow ) field lines connecting the positive footpoint of the flux rope with the streamer ( top right panel ) .
this reconnection occurs at the null point .
the two steps of the first phase are nearly simultaneous and overlying closed field lines from the streamer belt flux system ( green field lines ) reconnect away . during the reconnection process , a null point forms below the flux rope and there is a separator linking the bald patch and this null point .
this separator is similar to the bald patch - bald patch separator or the quasi - separatrix layer discussed for example in @xcite and @xcite and it is shown in metallic blue in the middle left panel of figure 5 . a current sheet forms in association with the separator and
it is shown as a purple surface in the bottom left panel of figure 5 .
the current sheet takes a sigmoidal shape , which is very common from flux emergence and shearing motions @xcite but it can also be the result of reconnection with adjacent flux systems and due to the pre - event topology as is found here . during the first phase , which lasts about 510 minutes , the flux rope expands to a height of 1 r@xmath2 above the solar surface . due to the cme expansion , the null point is pushed eastward along the outer spine by about 3@xmath1 and its height increases by about 0.04 r@xmath2 . at the end of this phase , the flux rope remains composed of closed field lines but it appears totally disconnected from its original negative footpoint ( see top right panel of figure 6 ) .
open field lines ( white ) from the leading positive spot reconnect at the separator below the flux rope with anemone field lines to form new open field lines ( shown in brown ) passing at proximity of the null point .
separator reconnection is described in @xcite for example .
this is illustrated in the middle left panel of figure 5 .
a similar type of reconnection following the eruption of a flux rope has been previously discussed in @xcite . due to the reconnection of the anemone flux system and the streamer belt and due to the formation of the separator , the magnetic topology changes quite drastically from the pre - event topology .
the null point below the flux rope separates the anemone flux system and open field lines originating from the leading ( white ) and trailing ( brown ) positive spots ( middle left panel of figure 5 ) .
the outer spine of this system is open .
such an opening of a similar topology during a time - dependent process was shown in @xcite .
the eastern null point separates the erupting ( yellow ) magnetic field and the closed field originating from the positive trailing spot .
the outer spine connects the positive trailing spot and the positive footpoint of the flux rope through the flux rope .
the full magnetic topology with all the different type of field lines discussed here is shown in the middle right panel of figure 5 .
the presence of two null points separating different but overlapping flux systems , as shown here , is only possible in a fully 3-d simulation as is the case here .
the second phase of reconnection happens when the flux rope is not connected anymore to the negative spot of ar 10798 .
the second phase of the evolution starts when the negative footpoint of the flux rope has fully reconnected outside of the fan surface ( light and dark blue field lines no longer exist)s .
then , the flux rope has raised , expanded and deflected enough so that * side * reconnection happens with open magnetic field lines at the null point ( similar to what was found in * ? ? ?
the reconnection of closed erupting field lines with open field lines from an equatorial coronal hole is a type of interchange reconnection .
here , contrary to the case described in @xcite , the open field lines are only transported over a very small distances as they reconnect with closed field lines , because the entire active region is surrounded by equatorial coronal holes .
this special form of interchange reconnection is due to the anemone nature of the active region and one can expect that it happens in most eruptions from an active region inside a coronal hole .
it has been previously discussed in an observational study of an eruption from an anemone active region in 2007 october 17 @xcite . at the time
the second phase of reconnection starts , about six minutes after the cme initiation , the apex of the flux rope is at 1 r@xmath2 above the solar surface and the null point is only at a height of 0.2 r@xmath2 . in this phase ( bottom panels of figure 5 ) , the yellow erupting field lines connecting the positive footpoint of the flux rope to the streamer reconnect with white open field lines to form open , twisted field lines ( shown in orange color ) . during this phase ,
the null point rises and is pushed eastward .
its evolution as well as that of the apex of the flux rope is shown in the right panels of figure 7 .
note that at all time , the null point is below and to the east side of the flux rope .
an online animation shows the first ten minutes of the cme evolution with the 2-phase reconnection .
the end result of the reconnection described in the previous subsection is typically an opening of the flux rope magnetic field but it is a relatively complex process and , importantly , some of the magnetic field lines of the flux rope remain closed until the end of our simulation ( see bottom right panels of figure 5 and 6 ) . as seen in the bottom right panel of figure 5 , at time @xmath15 minutes , the cme is composed of a mix of closed ( yellow ) and open ( orange ) field lines .
the closed field lines are highly twisted and they have , in general , one footpoint ( positive polarity ) near the source region of the cme and the other footpoint ( negative polarity ) in the quiet sun due to reconnection with the streamer belt .
the open field lines are also highly twisted because they are due to the reconnection of the magnetic field of the flux rope with that of the background originating from the coronal hole just south - west of ar 10798 ( positive polarity ) .
this picture is similar to the sketches of @xcite with open plasmoid field line embedded inside closed twisted field lines .
figure 8 shows a base difference of the 195 image 95 minutes after the start of the eruption with a pre - event image for the simulation ( left ) and from eit ( middle ) . as highlighted in this figure ,
the presence of erupted field lines with a footprint at the edge of the streamer belt near disk center results in coronal dimmings from this region in our synthetic euv images .
theses coronal dimmings were clearly visible in eit images on the same region of the solar surface from about 01:26 ut on august 22 to the end of august 23 .
dimming regions have been analyzed as the location of the footprint of the magnetic cloud @xcite and/or the region where interchange reconnection happens @xcite . here , we find that the dimming region corresponds to the footprint of the magnetic cloud since erupting closed field lines have their negative footprint in this region .
we believe that the appearance of the dimming region is only possible if the first phase of the reconnection process indeed happened as described in the previous section .
it is hard to understand otherwise how the filament could reconnect with a distant region when reconnection with the adjacent coronal hole should happen preferentially .
our simulation also shows how a flux rope originating from an active region inside a coronal hole can still contain closed field lines all the way to the outer corona ( 10 r@xmath2 ) .
on 2005 august 24 , ace measured the passage of a small magnetic cloud , which has been associated with this eruption @xcite . while the smooth rotation only lasted for about 2 hours , it implies that twisted field lines associated with one of the two eruptions on august 22 made it to 1 au .
figure 8 also includes a base difference image about 7.5 hours after the start of the eruption to illustrate the persistence of the dimming region .
the dimming region was visible until the end of august 23 and it is likely that it persisted until around 12:00ut on august 24 but the observing geometry makes it hard to assess ( the dimming region had moved close to the western limb of the sun on august 24 ) .
the fact that dimming from the location of the negative footprints of the cme persisted until then would corroborate the measurement of a magnetic cloud at 1 au at this time .
in the top right panel of figure 6 , we show the flux rope at time @xmath16 = 5 minutes as a pink surface as well as the initial streamer field lines ( green ) and the initial flux rope position ( white surface ) corresponding to @xmath16 = 0 . from this figure , it is clear that the flux rope has been deformed from a straight tube into one with a sigmoid - like shape .
the reason for the deformation is the reconnection with the streamer magnetic field lines ( as discussed in the previous section ) .
the reconnection happens at the negative footpoint of the flux rope ( the one on the southern part of the sunspot ) and it moves the footprints of the erupting field westward to the initial position of the streamer field lines , as is clear from this figure .
the reconnection results in a rotation of the part of the flux rope which reconnects .
it is similar to the rotation found in @xcite but , here , we find that it only affects part of the cme . also , the net results is not to align the cme with the current sheet but to deform it to make part of the cme aligned with the streamer belt .
because only part of the flux rope `` rotates '' , it leads to a deformation , which makes the flux tube looks similar to a sigmoid with a marked skewness .
this type of skewness has long been thought to be associated with shearing motions and it is produced by many cme initiation mechanisms involving shearing motions @xcite and flux emergence @xcite . here , we find that a straight flux rope may naturally develop a skewness as it interacts with the background magnetic field at a height as low as 0.7 r@xmath2 .
this finding is in addition to the sigmoidal current sheet developing under the flux rope as discussed in section 3.2 .
we calculated the orientation of the central axis of the flux rope as it propagates from a distance of 1 r@xmath2 to 8.5 r@xmath2 ( first hour ) and we found no indication of rotation with an axis making an angle of -15 to -20@xmath1 with the @xmath11 axis .
the bottom right panel of figure 6 shows the cme after 1 hour as it has reached 8.5 r@xmath2 .
the axis of the cme as illustrated by the magnetic field lines is found to be almost aligned with the initial axis of the flux rope .
it is also illustrated in the first panel of figure 7 with 5 views of the flux rope at different times from 1 to 20 minutes ( corresponding to distances from 1 to 3.5 r@xmath2 ) . in this figure ,
the flux rope is plotted at different times as isosurfaces of current density .
the color contours show the cme longitude .
the angle between a given color and the cme axis gives an indication of the cme orientation .
the angle is approximatively constant over the 20 minutes covered by this figure , illustrating the lack of rotation .
the magnetic field measurements at 1 au ( see figure 9 of @xcite ) are consistent with a right - handed flux rope with a dominant southward axial component .
it is the same as the orientation of the filament on the solar surface and , therefore , implies that there was no large - scale rotation during the cme propagation .
@xcite proposed that the rotation of a cme in the corona is related to its skewness .
it would be due to the force balance required to maintain this skewness which is lost as the cme erupts .
it generates a torque created by the now unbalanced lorentz force .
however , here , the skewness is not the result of shearing motions and it is not associated with force imbalance during the cme propagation .
therefore , our conclusion that the cme does not rotate is also consistent with this scenario .
this result provides a caution on the association of skewness with cme rotation as it was found , here , that cme skewness can developed during the eruption , in which case no rotation is expected .
@xcite found a strong rotation ( @xmath17 ) associated with the reconnection of a cme with the background magnetic field , yielding a cme whose axis is aligned with the heliospheric current sheet ( as in * ? ? ? * ) . in our case
, we found that only part of the flux rope reconnected , resulting in a deformation instead of a rotation of the flux rope axis .
although reconnection continues , it occurs primarily with open magnetic field lines , resulting in no change of the cme axis . because there is no observational evidence that the cme rotated , it is important to understand why it is the case , when most recent simulations have had a tendency to show that cme rotation is a very frequent phenomena . here
, we find that the reconnection of one footprint of the cme results in a deformation of the cme but not a rotation and that the effect of the magnetic forces is to deflect the cme not to rotate it ( see next section ) . figure 7 illustrates the deflection of the flux rope . as shown in the right panels , the apex of the flux rope is found to deflect by about 10@xmath1 in 35 minutes as it propagates to 5 r@xmath2 . after this time , the current inside the flux rope becomes so small that it is impossible to track the flux rope the same way as we did for the first 5 r@xmath2 .
instead , we track the position of the velocity disturbance associated with the cme .
we find that the cme further deflects by about 1.5@xmath1 from @xmath18 minutes to @xmath19 hour , at what time the cme has reached 8.3 r@xmath2 . in the absence of a velocity shear in the solar wind which could `` push '' the cme away from the coronal holes , there are two possible causes for the deflection : i ) the side reconnection with the null point which happens at the east side of the cme and ii ) the effect of the lorentz force .
the flux rope is created by a southward line current and the global magnetic field above ar 10798 is of unipolar positive polarity .
the lorentz force acting on the flux rope is , therefore , almost purely eastward .
the lorentz force here is similar to that of the models of @xcite and @xcite but in these studies , it results in acceleration and rotation , respectively due to different magnetic topologies . here , due to the largely unipolar polarity , the lorentz force can deflect the cme .
side reconnection resulting in cme deflection can be seen in the work of @xcite for example .
it is not straight - forward to determine the cause of the cme deflection .
however , during the first phase of the reconnection ( first six minutes ) , the main part of the flux rope appears to be relatively unaffected by the reconnection as its negative footpoint reconnects ( see top left panel of figure 7 , for example ) .
in addition , the reconnection is a breakout type reconnection ( above the flux rope ) and not a side reconnection .
no deflection has yet been reported for asymmetric breakout simulations ( e.g. , see * ? ? ?
in addition , the deflection rate of the null point does not correspond to that of the apex of the flux rope , as shown in the right panel of figure 7 . during the second phase of the reconnection , the side reconnection results in an opening of the magnetic field of the flux rope and
it is unclear whether it affects or not the cme direction . for these reasons
, we believe that it is likely that the lorentz force is the main contributor to the deflection of the flux rope .
the lorentz force is expected to decrease as the cme rises since both the magnetic field strength and the current decrease with height .
we find that the deflection rate decreases with height and the cme only deflects by about 2@xmath1 in the second 30 minutes of its propagation .
it is consistent with the lorentz force being the cause of the deflection .
the deflection of the actual filament at the sun could be greater than what is found in the simulation because its speed in the low corona , where the lorentz force is stronger , was certainly lower than that in our simulation . as noted above
, the simulation does not reproduce the slow rise of the cme for the 18 minutes from the start of the sxr flare to the start of the hxr flare .
it might create an additional deflection of about 510@xmath1 in the low corona , assuming that the actual deflection rate is more or less consistent with what we find in our simulation .
another reason why our simulation does not reproduce the observed deflection is also due the initiation mechanism . as the force - free field of the flux rope
is superimposed onto the coronal magnetic field , it results in a magnetic configuration which is not force - free .
therefore , there are additional lorentz forces on all directions , which partially compensate the main eastward lorentz force discussed before . in the real sun
, it is likely that the filament was in a force - free state before the eruption and it would only be subject to the main eastward lorentz force , which would yield a larger deflection .
deflection of cmes from the same active region during the following carrington rotation has also been reported @xcite .
we have investigated , using a numerical simulation , the first of two eruptions on 2005 august 22 from anemone active region 10798 .
we have used a recently developed component of the swmf to include realistic thermodynamics into the representation of the low corona ( lc ) . using the lc component
has enabled us to produce euv images of the solar corona which we can compare to eit observations .
in addition , previous works have shown that including thermodynamics is important to reproduce the plasma and magnetic properties of the entire corona .
we have investigated the pre - event topology of the active region and found that it indeed developed into an anemone active region with a single null point separating the active region flux system to unipolar positive magnetic field .
we have found that some of these unipolar magnetic field lines are open while others are part of the streamer belt connected to a region near disk center .
pre - event 171 and 195 images of the corona are in good agreement with simulated ones , validating our steady - state model . in agreement with observations of a h@xmath0 filament ,
which was well observed prior to the eruption , we have superposed a highly twisted flux rope onto the steady - state solar corona at the same location and with the same orientation and chirality ( right - handed and southward ) .
due to force imbalance , the flux rope immediately erupts and we have focused our study on its interaction with the adjacent magnetic flux systems and in particular with the open flux from the surrounding coronal holes .
the anemone nature of the source region has two major consequences on the evolution of the cme : an eastward deflection due to the lorentz force and a reconnection of the negative footpoint of the flux rope , which results in the appearance of a long - duration dimming region and a mix of open and closed field lines in the cme . since the flux rope was southward and right - handed , there was a strong southward axial current .
the lorentz force generated by this current and the unipolar positive magnetic field in the vicinity of the active region is east - directed and , in our simulation , it deflects the cme by about 1015@xmath1 as the cme propagates to 8 r@xmath2 .
we have not address the reason why cmes from anemone active regions tend to be fast cmes @xcite .
it should be however noted that , in the simulation , the cme encounters closed field lines in the low corona , so the magnetic topology can not be fully responsible for the fast speed .
it is therefore likely that the fast speed of cmes from anemone active regions is at least partially due to the fact the cme propagates into a low density , fast solar wind , where the hydrodynamical drag is low @xcite .
the negative footpoint of the flux rope quickly reconnects with the positive field part of the streamer belt .
it leads to a cme with one footpoint ( the positive one ) that is line - tied , and the second one which connects to a region of the sun about 40@xmath1 away from the active region .
a dimming region was found to develop in the region of the second footpoint , both in simulated euv images and in eit images and to persist at the sun for as much as a day after the eruption .
additionally , some of these closed field lines reconnect with open field from the equatorial coronal hole in a type of interchange reconnection previously discussed in @xcite .
it yields a mix of open and closed field lines in the erupting ejecta .
while we have conducted our simulation only to 10 r@xmath2 , this type of mixed nature of the erupting field is expected from observations of periods of unidirectional strahl inside periods of bi - directional electrons as a magnetic cloud passes over a spacecraft @xcite . to the best of our knowledge ,
this is the first time such a mixture is obtained from a numerical simulation of a real cme event .
the fact that we obtain interchange reconnection in our simulation is not surprising , since it has been noted before that anemone active regions are the most suitable environment for interchange reconnection @xcite .
we have also found that the cme does not rotate as it propagates in the corona ( up to a distance of 10 r@xmath2 ) .
this result is found although there is large - scale reconnection , the cme axis is initially not aligned with the hcs and the flux rope deforms into a skewed tube at a distance of less than 1.7 r@xmath2 .
it shows that the rotation of cme axis is not a universal process and that some cmes maintain their orientation in spite of large - scale reconnection .
this result is a pendant to previous numerical studies @xcite which found that rotation happens when a cme is initially misaligned with the hcs , or when it is skewed .
the deflection of the cme , the persistence of closed field lines into the inner heliosphere and the lack of rotation are also consistent with the observation of a short magnetic cloud at 1 au about 2 days after the first eruption .
@xcite associated this magnetic cloud with the first of the two eruptions of 2005 august 22 . while our simulation captures some of the most important observational features of the cme , there are some discrepancies . in our simulation ,
the cme does appear to propagate along the same latitude as that of its source region ( s11 ) while lasco images show a cme propagating close to the ecliptic .
one likely explanation is that only the northern part of the filament erupted ( as reported in * ? ? ?
our method to obtain an eruption is too simple to study partial filament eruption or to reproduce the exact dynamics and kinematics of the eruption at distances less than 3 r@xmath2 .
another difference between the simulation and the observations is the amount of deflection . in our simulation ,
the cme deflects by about 10@xmath1 eastward whereas a total deflection of about 4050@xmath1 is expected from observations .
a smaller deflection might be present in our simulation compared to what happened on the real sun because of different kinematics and force balance in the low corona . however , we believe that a significant portion of the deflection can not be accounted by magnetic forces , which are dominant in the low corona .
this finding implies that a significant part of the deflection is due to hydrodynamical effect . for this cme from an anemone active region , followed by a twice faster ejection 18 hours later , these effects may take two forms : deflection by the fast wind from the coronal hole as discussed , for example , in @xcite and @xcite and also deflection by the second , faster cme pushing the eruption from behind , as discussed in @xcite . in this case
, the additional deflection can be expected to happen more gradually in the heliosphere .
future simulations will investigate the heliospheric evolution of cmes from anemone active regions and overtaken by faster eruption .
the research for this manuscript was performed during and supported by a jsps short - term post - doctoral fellowship at kwasan observatory in kyoto university .
we would like to thank the reviewer and terry forbes for their helpful comments , which helped us clarify the discussion of the reconnection process .
n. l. would like to thank everyone from the kwasan observatory for their welcome during his 7-month stay there .
n. l. and i. r. acknowledge additional financial support from nsf grants atm0639335 and atm0819653 and nasa grant nnx08aq16 g . c. d. was supported by a nasa nessf08-helio08f-0007 . | we present a numerical investigation of the coronal evolution of a coronal mass ejection ( cme ) on 2005 august 22 using a 3-d thermodynamics magnetohydrodynamic model , the swmf .
the source region of the eruption was anemone active region ( ar ) 10798 , which emerged inside a coronal hole .
we validate our modeled corona by producing synthetic extreme ultraviolet ( euv ) images , which we compare to eit images .
we initiate the cme with an out - of - equilibrium flux rope with an orientation and chirality chosen in agreement with observations of a h@xmath0 filament . during the eruption , one footpoint of the flux rope reconnects with streamer magnetic field lines and with open field lines from the adjacent coronal hole .
it yields an eruption which has a mix of closed and open twisted field lines due to interchange reconnection and only one footpoint line - tied to the source region .
even with the large - scale reconnection , we find no evidence of strong rotation of the cme as it propagates .
we study the cme deflection and find that the effect of the lorentz force is a deflection of the cme by about 3@xmath1 / r@xmath2 towards the east during the first 30 minutes of the propagation .
we also produce coronagraphic and euv images of the cme , which we compare with real images , identifying a dimming region associated with the reconnection process .
we discuss the implication of our results for the arrival at earth of cmes originating from the limb and for models to explain the presence of open field lines in magnetic clouds . |
the recent discovery of the higgs boson at the large hadron collider ( lhc ) @xcite confirms that the higgs mechanism is responsible for spontaneously breaking electroweak symmetry in the standard model ( sm ) .
however , the question of whether the higgs sector is natural or not remains to be answered .
the two most appealing solutions for stabilising the weak scale are supersymmetry ( see ref .
@xcite for a review ) and compositeness @xcite .
both are now constrained by measurements of the higgs mass and its couplings , which have led to consequences for the spectrum of exotic states predicted in these two scenarios . in supersymmetric models
the radiative corrections from coloured states , such as the stop and gluino , have been extensively studied in the literature and shown to have an important effect on tuning in the higgs sector .
combined with the lower limits on sparticle masses from the lhc , the conclusion is that supersymmetric models are now tuned to below the 5@xmath0 level @xcite .
exotic coloured states also exist in composite higgs models and can similarly play an important role in determining the higgs mass . in many of these models
the sm higgs doublet is identified with nambu - goldstone modes , arising from the spontaneous breaking of a global symmetry group @xmath1 to a subgroup @xmath2 due to some strong dynamics .
however , the original global symmetry is also explicitly broken via mixing between operators in the strongly coupled sector and elementary fields in the sm sector , as the latter need not come in complete representations of @xmath1 .
hence a coleman - weinberg type potential is generated , leading to dynamical electroweak symmetry breaking and a mass for the higgs boson @xcite .
a key component in such models is the existence of composite , fermionic top partner resonances in the low energy spectrum .
they are required to facilitate a strong coupling between the top quark and the composite higgs ( through substantial mixing between composite and elementary states in the top sector ) and , often , to break electroweak symmetry in the first place .
the scale of the higgs mass is typically set by these top partner masses so , to provide a higgs mass of around 126 gev @xcite , the top partners can not be too heavy . including such coloured fermionic resonances
generically implies that there will also be coloured vector meson resonances in the low energy spectrum .
this follows because , even though a complete description of the underlying dynamics of the strong sector remains unknown , the constituents of the strong sector must be coloured in order to produce top partner bound states charged under @xmath3 colour .
consequently , the strong sector is also expected to produce coloured vector mesons that necessarily couple to any top partner bound state .
we will refer to these states as gluon partners .
indeed , if the strong sector contains a fermionic operator @xmath4 , in the fundamental representation of @xmath3 so as to produce top partners , one can always write down a vector operator @xmath5 in the adjoint representation of @xmath3 .
equivalently , in the five - dimensional ( 5d ) version of these composite higgs models @xcite the gluon partners are simply the kaluza - klein gluons , which are required by 5d gauge invariance if the sm fermions are located in the bulk . thus far the contribution of gluon partners to the higgs mass has been neglected .
since they only couple to the higgs through the top partners , the size of the correction to the higgs mass will be proportional to the gluon partner - top partner coupling , @xmath6 .
this coupling can be estimated either through direct calculations in the 5d theory @xcite or via holographic techniques @xcite . in both instances
the coupling is large , hence the contribution to the higgs mass is expected to be sizeable . in this paper
we explicitly compute the leading order correction and do indeed find it to be important . specifically
, we calculate the one - loop correction to the two - point function of ( dirac ) top partners coming from a massive gluon partner .
this first result is model independent but , to quantify the effect on the higgs mass , we apply it to the specific example of the mchm@xmath7 .
interestingly , we find that the higgs mass is decreased in this particular model , thereby easing some of the tension between the tuning in composite higgs models and the non - observation of coloured top partners . for a gluon partner of mass 3
tev , a spontaneous global symmetry breaking scale of about 750 gev and with the higgs mass fixed at 126 gev , the top partners are about 10% heavier than in the uncorrected model .
the rest of this paper is organised as follows . in section 2
we consider gluon partners in a general composite higgs model .
the one - loop radiative correction from a gluon partner to the two - point function is first estimated in the large @xmath8 limit and then the exact calculation is performed with the final result expressed in integral form .
the exact result is used in section 3 to compute the contribution to the composite higgs mass in the mchm@xmath7 .
the concluding remarks are presented in section 4 . in appendix
a we use the holographic basis to estimate the size of the radiative correction , and in appendix b we list the passarino - veltman integral expressions utilised earlier in the paper .
many composite higgs models consists of a strong sector that is responsible for producing a set of nambu - goldstone bosons , a combination of which is identified with the higgs boson , @xmath9 .
the strong sector is joined by an elementary sector containing the sm fermions and gauge bosons , and the two sectors are assumed to mix only through bilinear couplings [ eq : lmix ] _
mix=|q_l_r+|t_r_l+a_g^_+ where the elementary fields @xmath10 and @xmath11 are the left handed top quark doublet , right handed top and gluon respectively . we have only included the fermionic operators @xmath12 and the strong sector @xmath3 colour current @xmath13 ; the dots represent other bilinear couplings which will not be considered here , such as those for the light fermions . integrating out the strong sector results in an effective lagrangian for the top - higgs sector @xcite
@xmath14}t_l+\bar{t}_r\slashed{p}{\left[\pi_r^0(p^2)+{\cal y}_r(h / f)\pi_r^h(p^2)\right]}t_r+ { } \nonumber\\ & { \left[\bar{t}_l{\cal y}_m(h / f)m(p^2)t_r+\mbox{h.c.}\right]},\end{aligned}\ ] ] where @xmath15 is the nambu - goldstone boson decay constant .
the functions @xmath16 and @xmath17 are determined by two - point functions of the fermionic operators @xmath18 up to a factor of @xmath19 in @xmath20 coming from the usual sm kinetic term ( i.e. @xmath21 ) . assuming the strong sector is a large @xmath8 gauge theory and working to leading order in @xmath22 , one can write these two - point functions as a sum over narrow , top partner resonances , @xmath23 , to find @xmath24 for masses @xmath25 , constant form factors @xmath26 and where the coefficients @xmath27 are derived from the group structure of any particular model . their precise form and that of their prefactors , the functions @xmath28 of the higgs fields , are determined by the details of the global symmetry breaking pattern and the representations into which the top quarks are embedded
. however , they can always be split into those components that are sensitive to the higgs vacuum expectation value ( vev ) and those that are not , hence @xmath29 .
since the top quarks do not make up full representations of the global symmetry group in the strong sector , the symmetry is explicitly broken and a potential can be generated for the erstwhile nambu - goldstone bosons . at one loop
the potential from the top sector is given by @xmath30}{\left[\pi^0_r(p_e^2)+{\cal y}_r(h / f)\pi_r^h(p_e^2)\right]}\right .
\nonumber\\ \left.{}-{\left|{\cal y}_m(h / f)m(p_e^2)\right|}^2\right),\end{aligned}\ ] ] where @xmath31 is the qcd colour factor and the integral is performed over euclidean momentum @xmath32 .
expanding the logarithm and discarding the constant term gives an approximate form for the potential @xmath33,\end{aligned}\ ] ] which can be considered an expansion in @xmath28 ( corresponding to a small higgs vev ) or an expansion in @xmath34 and @xmath17 ( corresponding to weak mixing between elementary and composite degrees of freedom ) .
the operators @xmath35 and @xmath36 in the strong sector , which are required to mix with the top quark , guarantee the existence of top partners in this framework .
however , as argued in the introduction , they are typically accompanied by massive , coloured vector meson resonances , or gluon partners .
these are associated with the strong sector current operator @xmath13 as , at leading order in @xmath22 , the two - point function @xmath37 can be written as a sum over narrow gluon resonances , @xmath38 , much like the two - point functions @xmath16 and @xmath17 .
because the higgs is colour neutral , any correction to its mass from gluon partners must enter , at two - loop order , through the two - point functions of @xmath35 and @xmath36 .
three - point function ( which already comes with an extra @xmath39 suppression ) must vanish and the vertex correction is suppressed by several additional composite - elementary mixing factors .
] this means the effect can be accounted for by calculating the gluon partner corrections to the functions @xmath17 and @xmath16 , defined in eq . in the large @xmath8 limit .
a naive large @xmath8 analysis suggests that the correction is not important because it depends on a @xmath40 coupling in the strong sector , which scales like @xmath39 .
however , this is only a scaling dependence and ignores the prefactor for the coupling , which we will now show is large . to estimate the importance of the correction
more accurately we will use the ads / cft correspondence to estimate the strength of couplings in the strong sector , then use large @xmath8 results to estimate the amplitudes of the relevant diagrams . in theories of warped extra dimensions
one can relate the @xmath8 appearing in the large @xmath8 cft expansion with the 5d gauge couplings using @xcite [ eq : ngauge ] = =( ) , where @xmath41 is the curvature scale of the 5d warped ads space , @xmath42 a bulk gauge coupling , @xmath43 a four - dimensional ( 4d ) gauge coupling and @xmath44 a numerical factor distinguishing between different gauge groups .
we have also made use of the relation between 5d and 4d gauge couplings @xmath45 , which includes the logarithmic running between the uv and the ir scales .
the expression can be used to provide quantitative information about the strength of the couplings between resonances in the strong sector . to estimate the coupling between the top partner and the gluon partner , we first use the @xmath3 gauge coupling to obtain the value for @xmath46 .
using one finds [ eq : gn ] = ( ) , where @xmath47 is the qcd gauge coupling strength and @xmath48 .
thus the coupling between the fermionic and gluon resonances can be estimated as [ eq : alphag ] _ g3 .
this compares well with the exact 5d calculation @xcite , where the overlap integral between the first kaluza - klein gluon and the first kaluza - klein top gives @xmath49 . , which implies @xmath50 where @xmath51 is the strong coupling evaluated at the scale @xmath15 .
in this region of parameter space one recovers the usual kaluza - klein fermion spectrum for unmixed boundary conditions . ]
note that in the estimate we have neglected sm loop and brane kinetic term contributions to the 4d coupling .
these contributions are model dependent and can increase or decrease the coupling @xmath6 @xcite .
for example , if the sm loop contributions are not cancelled by uv brane kinetic terms then the value of the coupling @xmath6 is reduced by approximately 1/4 . to avoid model dependency
we will assume the value for concreteness in the rest of this paper .
limit.[fig : selfenergy],title="fig:",scaledwidth=40.0% ] + limit.[fig : selfenergy],title="fig:",scaledwidth=40.0% ] now the coupling can be used to estimate the size of the one - loop gluon partner contribution to the two - point functions relative to that of the tree - level contribution : [ eq : ratio ] = . the expression on the left hand side
is obtained using the large @xmath8 result @xmath52 for the tree - level contribution in the denominator ( the top diagram in figure [ fig : selfenergy ] ) .
for the one - loop gluon partner contribution in the numerator ( the bottom diagram in figure [ fig : selfenergy ] ) there are two vertices coupling top partners to a gluon partner , each contributing a factor of @xmath53 , an additional loop factor @xmath54 , as well as the quadratic casimir @xmath55 ( coming from @xmath56 , where @xmath57 are the generators of @xmath3 in the fundamental representation ) .
finally , we substitute in the estimated value for @xmath58 from eq . to get a numerical value . despite being a higher order effect
, we find that the contribution due to the gluon partners could be of order @xmath59 . an alternative derivation , based on the mixing between the holographic and mass bases and yielding the same result , is presented in the appendix .
note that these estimates neglect a momentum dependent loop function , which will be explicitly computed in the next section . from a gluon partner @xmath1.[fig : selfenergyc],scaledwidth=27.0% ] to fully quantify the effect of gluon partners on the higgs mass we must calculate the contribution from figure [ fig : selfenergyc ] explicitly .
this assumes that the gluon partners associated with the current @xmath13 can be modelled as narrow resonances , like their fermionic brethren .
each top partner propagator is renormalised to [ eq : sr ] s^-1(p)=s^-1_0(p)+i_1(p ) , at one - loop order , where @xmath60 is the unrenormalised propagator and @xmath61 is the one - loop renormalised self energy .
including appropriate counterterms for the wave function and mass renormalisation , this self energy is given by [ eq : sigmar ] _
1(p)=(p)-(-m_q)z_2-m_q , where [ eq : sigmal ] i(p)=_g^^ , is the integral expression obtained from the diagram in figure [ fig : selfenergyc ] .
renormalising the propagator in the on - shell scheme gives the two renormalisation conditions @xmath62 which are used to determine the two counter terms @xmath63 and @xmath64 . by writing ( p)=m_qa(p^2)+b(p^2 ) , then solving the above equations and substituting into eqs . and , we find the one - loop expression [ eq : prop ] is^-1(p)=(1-(p^2)+.m_q|_=m_q)(-m_q ) , where we have further defined @xmath65 the first factor in is absorbed into an overall rescaling of the top partner fields and is not important for our purposes . the second term can be expressed in the propagator as an effective correction to the top partner mass m_q(p^2)=m_q .
we now need to evaluate eq .
to determine the functions @xmath66 and @xmath67 .
the answer is succinctly expressed in terms of passarino - veltman integrals @xcite ( see the appendix ) @xmath68[(k - p)^2-m_g^2 ] } \nonumber\\ & = \frac{2i\alpha_g}{3\pi}{\left[2m_qb_0(p^2,m_q^2,m_g^2)+\slashed{p}b_1(p^2,m_q^2,m_g^2)\right]},\end{aligned}\ ] ] and therefore @xmath69 } , \\
\widehat{b}(p^2 ) & = \frac{2\alpha_g}{3\pi}{\left[b_1(p^2,m_q^2,m_g^2)-b_1(m_q^2,m_q^2,m_g^2)\right]}.\end{aligned}\ ] ] in an explicit integral form we find the final result [ eq : deltam ] m_q(p^2)=m_q_0 ^ 1^x(x-2 ) .
this one - loop result can be easily generalised to include contributions from multiple gluon partners by replacing the above expression with a sum of identical terms , each using different values for @xmath6 and @xmath70 .
however , we expect that the lightest gluon partner will always dominate because the loop integral can be seen to decrease as the gluon partner mass is increased , and the coupling is expected to behave similarly ( this latter effect can be seen explicitly in the 5d calculation ) .
in addition to the coloured contribution we can estimate the equivalent non - coloured contribution arising from the electroweak resonances . using the relation we obtain [ eq : gnw ] = ( ) , where @xmath71 is the electroweak coupling strength and , for simplicity , we have set @xmath72 , corresponding to a redefinition of @xmath8 .
this enables us to estimate the coupling @xmath73 of the top partner fermions with the electroweak vector mesons . using eq .
we obtain @xmath74 .
this is a factor of three smaller than the gluon coupling .
the ratio of the electroweak resonance correction relative to that of the gluon partner correction is then @xmath75/[\alpha_gc_2(3 ) ] \simeq 3/16 $ ] , where we have further used the @xmath76 quadratic casimir @xmath77 .
thus the correction coming from electroweak vector mesons is much less important than the gluon partner correction , and will henceforth be neglected .
, for example , the mass limit on the electroweak resonances is not that stringent @xcite so there is no real limit on the gluon partner masses coming from this observation . ]
having gone beyond leading order in @xmath22 one may anticipate that other corrections to the higgs potential , beyond the simple mass shift of eq .
, should be considered . while such corrections do exist , and are at the same order in @xmath22 , they will always be subdominant .
the reason is that the correction calculated above is the only one proportional to the top partner - gluon partner coupling @xmath6 , in contrast to other corrections which go like @xmath78 or just @xmath79 .
the couplings in the coloured sector satisfy @xmath80 , as shown in the previous subsection , hence any other corrections coming from this sector can be neglected .
the situation is even more acute in the non - coloured sector , as these corrections suffer a similar suppression from the reduced coupling strength , and do not even benefit from qcd multiplicity factors when evaluating loops .
another correction one might consider beyond leading order in @xmath22 is a wave function renormalisation of the higgs .
however , in this framework the higgs always appears in the ratio @xmath81 , meaning that any wave function renormalisation is simply absorbed in a rescaling of the symmetry breaking scale . while the above arguments are robust in the models we are considering here , where the strong sector only communicates with elementary fields through bilinear couplings like those in eq .
, they do not immediately apply in more general models . then
, there may be other corrections to the higgs mass of similar importance .
however , it should be noted that the effective lagrangian will no longer apply either .
such models are beyond the scope of this paper , but could result in interesting deviations from the usual behaviour found in composite higgs models .
( top ) and @xmath82 ( bottom ) .
the functions are evaluated using eq . for a single top and gluon partner resonance , fixing @xmath83 and @xmath84 for simplicity .
the normalisation is chosen to respect the top mass constraint in eq . .
the dashed lines show the uncorrected functions and the solid lines show the effect of gluon partner correction for @xmath85 and @xmath86 tev.[fig : mplot],title="fig:",scaledwidth=48.0% ] ( top ) and @xmath82 ( bottom ) .
the functions are evaluated using eq . for a single top and gluon partner resonance , fixing @xmath83 and @xmath84 for simplicity .
the normalisation is chosen to respect the top mass constraint in eq . .
the dashed lines show the uncorrected functions and the solid lines show the effect of gluon partner correction for @xmath85 and @xmath86 tev.[fig : mplot],title="fig:",scaledwidth=48.0% ] + ( top ) and @xmath82 ( bottom ) .
the functions are evaluated using eq . for a single top and gluon partner resonance , fixing @xmath83 and @xmath84 for simplicity .
the normalisation is chosen to respect the top mass constraint in eq . .
the dashed lines show the uncorrected functions and the solid lines show the effect of gluon partner correction for @xmath85 and @xmath86 tev.[fig : mplot],title="fig:",scaledwidth=48.0% ] ( top ) and @xmath82 ( bottom ) .
the functions are evaluated using eq . for a single top and gluon partner resonance , fixing @xmath83 and @xmath84 for simplicity .
the normalisation is chosen to respect the top mass constraint in eq . .
the dashed lines show the uncorrected functions and the solid lines show the effect of gluon partner correction for @xmath85 and @xmath86 tev.[fig : mplot],title="fig:",scaledwidth=48.0% ] the leading order effect of gluon partners on the higgs mass in the large @xmath8 limit is accounted for by shifting the mass parameters in eq . using the expression in eq . .
one then uses the renormalised functions @xmath87 to calculate the higgs potential in eq . .
leaving these functions as they are , i.e. not expanding in @xmath6 again , ensures that the approximation remains good at high momentum , and corresponds to consistently resumming all contributions generated by the 1pi diagram in figure [ fig : selfenergyc ] .
this is easily seen to be the leading order effect .
note also that , since the correction shows up as a mass shift only , all of the group structure is preserved and we can simply apply existing results to calculate the higgs mass with no need to worry about the theory becoming non - renormalisable .
this may not have been the case if the renormalisation procedure had introduced momentum dependence into the @xmath88 s .
to get an idea how the form factors are changed , we evaluate eq . with a single top partner , a single gluon partner , and with all unknown constants fixed using a normalisation respecting eq .
this results in form factors behaving as in figure [ fig : mplot ] . the corrected form factor @xmath89 ( that provides the top quark yukawa ) is seen to decrease relative to the uncorrected form factor at low momentum , but it then crosses zero and continues to decrease to a magnitude bigger than the uncorrected function .
this does not imply that the one - loop result has become unreliable , as the expansion parameter when renormalising the propagator goes like @xmath90 , which is still much less than one at large momentum .
meanwhile the corrected form factor @xmath91 displays the opposite behaviour ; it is seen to increase relative to the uncorrected form factor at low momentum , but it decays more quickly so ends up smaller than the uncorrected function at high momentum . the way that these form factors appear in the higgs potential is model dependent so we can not make a universal statement about how the gluon partner correction changes the higgs mass . even for a specific model
it is not obvious what will happen .
only the magnitude of @xmath89 appears in eq . , so there is always some cancellation when performing the momentum integral and it is not immediately apparent whether the higgs mass will be increased or decreased .
a similar cancellation occurs when integrating @xmath91 , and the situation is further complicated by including more top parters , whereupon there can be cancellations in the expressions even before integrating over momentum .
nonetheless , eqs . and are all that is needed to calculate the leading order gluon partner correction to the higgs mass in any model of this form . as a specific example we now consider the minimal composite higgs model ( mchm ) .
this is based on the symmetry breaking pattern @xmath92 and supports numerous embeddings for the top quarks .
if the left handed doublet and right handed singlet are both embedded into @xmath93 s ( the mchm@xmath94 ) the higgs mass is given by @xcite m_h^2 , where @xmath31 and @xmath95 gev is the higgs vev .
of the top partner contribution .
] this expression simplifies when the mixing between elementary and composite degrees of freedom is small , such that @xmath96 ( equivalently @xmath97 ) , to become [ eq : mhint ] m_h^2 . using weinberg sum rules @xcite refs .
@xcite show that at least two top partners are required for the higgs potential to be convergent in this model . for concreteness we will consider only two low mass states .
denoting their masses by @xmath98 and @xmath99 ( corresponding to their @xmath100 representations ) the uncorrected functions in eq .
then take the specific forms @xmath101},\end{aligned}\ ] ] where @xmath102 is the phase difference between form factors , i.e. @xmath103 and the subscripts have been omitted from the @xmath104 ( which have been set to be equal by a field redefinition ) . substituting into eq . and integrating , one arrives at the uncorrected expression [ eq : mh20 ] [ m_h^2]^(0 ) ( ) . to eliminate both form factors and the phase @xmath102 from the above expression two relationships have been used , which will continue to be assumed throughout the rest of this paper .
first one takes @xmath105 .
this is helpful in arranging for electroweak symmetry breaking , as there is an additional , positive contribution to the quadratic term in the potential proportional to @xmath106
. then one uses the expression for the top mass [ eq : mt ] m_t^2|m(0)|^2 , which follows from reading off the yukawa coupling from the effective lagrangian in the low energy limit @xmath107 .
if the renormalised expressions are substituted into eq . there is no simple analytic expression for the functions @xmath16 and @xmath17 .
nonetheless , we can still apply the two relationships used above to solve for the @xmath88 s , upon which the integrals can all be performed and the effect of the gluon partner quantified .
gev for @xmath85 , @xmath108 and @xmath109 gev .
blue , dashed lines represent the result without massive gluons from ref .
@xcite and solid , red lines the result including gluon partners calculated here .
black dotted lines represent approximate experimental limits for the top partners and gluon partners of 770 gev @xcite and 2.5 tev @xcite respectively .
_ left _ : two different mass top partners with a 3 tev gluon partner .
_ right _ : two equal mass top partners with variable gluon partner mass.[fig : contour],title="fig:",scaledwidth=48.0% ] gev for @xmath85 , @xmath108 and @xmath109 gev .
blue , dashed lines represent the result without massive gluons from ref . @xcite and solid , red lines the result including gluon partners calculated here .
black dotted lines represent approximate experimental limits for the top partners and gluon partners of 770 gev @xcite and 2.5 tev @xcite respectively .
_ left _ : two different mass top partners with a 3 tev gluon partner .
_ right _ : two equal mass top partners with variable gluon partner mass.[fig : contour],title="fig:",scaledwidth=48.0% ] in figure [ fig : contour ] we show contours of @xmath110 gev in the @xmath111 plane for two distinct top partner masses .
both the uncorrected result from ref .
@xcite ( blue , dashed ) and our result , that includes the gluon partner correction ( red , solid ) , is shown . for distinct top partner masses the phase @xmath102 can no longer be completely absorbed because the top mass is evaluated at a fixed momentum but the gluon partner correction , which changes the phase dependence , varies with momentum .
however , we have checked that the @xmath102 dependence is only mild , so we fix @xmath112 for definiteness .
in addition , we include approximate experimental limits for the top partners ( 770 gev @xcite ) and gluon partners ( 2.5 tev @xcite ) .
the main message of figure [ fig : contour ] is that the top partner masses that result in @xmath110 gev are universally shifted to larger values .
the shift is significant , as predicted by the discussion in section [ sec : nda ] , and provides additional breathing space for the model with respect to collider searches .
^{(1)}-[m_h^2]^{(0)})$ ] to the uncorrected higgs mass squared as a function of @xmath113 for two equal mass top partners .
_ right _ : the absolute higgs mass as a function of @xmath114 with @xmath86 tev in the same scenario .
the dashed line represents the result without massive gluons from ref .
@xcite and the solid line the result including gluon partners calculated here . in both plots @xmath85 , @xmath108 and @xmath109
gev.[fig : deltamh2],title="fig:",scaledwidth=48.0% ] ^{(1)}-[m_h^2]^{(0)})$ ] to the uncorrected higgs mass squared as a function of @xmath113 for two equal mass top partners .
_ right _ : the absolute higgs mass as a function of @xmath114 with @xmath86 tev in the same scenario .
the dashed line represents the result without massive gluons from ref .
@xcite and the solid line the result including gluon partners calculated here . in both plots
@xmath85 , @xmath108 and @xmath109 gev.[fig : deltamh2],title="fig:",scaledwidth=48.0% ] it is also instructive to consider the special case in which the two resonances are degenerate in mass : @xmath115 .
this gives an uncorrected higgs mass value of @xmath116^{(0)}=n_cm_t^2m_q^2/\pi^2f^2 $ ] .
in figure [ fig : contour ] and [ fig : deltamh2 ] we show the effect of the gluon partner correction in this scenario . on the left of figure [ fig : deltamh2 ]
is the size of the gluon partner correction relative to the uncorrected result .
the correction is always negative ( i.e. the higgs mass is decreased ) and decreases in magnitude as the ratio @xmath113 is increased and the gluon partner decouples .
on the right of figure [ fig : deltamh2 ] is the absolute higgs mass as a function of @xmath114 for @xmath86 tev , explicitly showing the reduction of the higgs mass relative to the uncorrected result .
the change in the value of the top partner mass required for a 126 gev higgs can be found more precisely in this case ; it is increased from 1 tev to 1.1 tev a relative change of 10% .
gluon partners are generically present in composite higgs models that include top partners in their low energy spectrum .
the gluon partners are expected to couple strongly to the top partners , and this expectation is quantitatively confirmed by arguments based on holography .
they can therefore provide a significant correction to the composite higgs mass .
the leading order correction is parameterised by a momentum dependent top partner mass shift in the two - point functions of the strong sector , which we have explicitly calculated in this paper . the final effect on the higgs mass is model
dependent , but we find a decrease in the higgs mass in the mchm@xmath94 .
this means that the mass of the top partners required to yield a 126 gev higgs is increased by about 10% ( for a 3 tev gluon partner and a spontaneous global symmetry breaking scale of about 750 gev ) easing constraints from direct collider searches for top partners .
we thank christophe grojean and alex pomarol for useful discussions .
this work was supported in part by the australian research council .
tg thanks the sitp at stanford for support and hospitality .
jb and tg thank the galileo galilei institute for theoretical physics for hospitality and the infn for partial support during the completion of this work .
am thanks the kitp at santa barbara for hospitality .
tsr thanks cern theory division for hospitality during the final stage of this work .
this research was supported in part by the national science foundation under grant no .
nsf phy05 - 51164
we can also use the mixing between the holographic and mass bases to estimate the size of the one - loop correction resulting from massive gluons . at next to leading order in @xmath22 there are gluon partner contributions to the two - point functions of @xmath35 and @xmath36 . for simplicity ,
let us assume that the strong sector only produces one gluon partner , @xmath117 , that mixes with the elementary gluon , @xmath118 , so that the mass eigenstates can be written as @xmath119 where @xmath11 is the ( massless ) physical gluon and @xmath1 is the massive gluon .
there is then one diagram for each mass eigenstate propagating around the loop .
each diagram gets a factor of @xmath6 from the @xmath120 coupling in the strong sector , and a @xmath121 or @xmath122 from the @xmath117-@xmath118 mixing to give @xmath123 where @xmath55 is the quadratic casimir for the fundamental representation of @xmath3 and @xmath124 is multiplied by an order one loop function .
the factor of @xmath8 comes from factors of @xmath125 on each external leg , themselves originating from vacuum creation amplitudes . although @xmath6 scales like @xmath22 it
can still be large if this scaling comes with a large prefactor . in the case at hand we know from gauge invariance that @xmath126 .
we also know from holographic arguments that @xmath127 ( see e.g. ref . @xcite or ref .
@xcite for an explicit calculation ) .
hence @xmath128 , clearly overcoming any @xmath22 suppression , and we find _ g~. the ratio of the one - loop result with the tree - level result @xmath129 gives a factor @xmath130 , which is the same as obtained in section [ sec : nda ] .
expressions for the passarino - veltman integrals used are @xmath131[(k - p)^2-m_g^2 ] } \nonumber\\ & = \delta_\epsilon-{\int_{0}^{1}\mathrm{d}^{}x\,}\ln{\left[\frac{xm_g^2+(1-x)m_q^2-x(1-x)p^2}{\mu^2}\right ] } , \\
b^\mu(p^2,m_q^2,m_g^2 ) & = \frac{1}{i\pi^2}{\int\mathrm{d}^{4}k\,}\frac{k^\mu}{[k^2-m_q^2][(k - p)^2-m_g^2 ] } \nonumber\\ & = p^\mu b_1(p^2,m_q^2,m_g^2 ) \nonumber\\ & = p^\mu{\left(-\frac{1}{2}\delta_\epsilon+{\int_{0}^{1}\mathrm{d}^{}x\,}x\ln{\left[\frac{xm_g^2+(1-x)m_q^2-x(1-x)p^2}{\mu^2}\right]}\right)},\end{aligned}\ ] ] where @xmath132 is the renormalisation scale .
collaboration , s. chatrchyan _ et .
_ , _ observation of a new boson at a mass of 125 gev with the cms experiment at the lhc _ , _ phys.lett . _ * b716 * ( 2012 ) 3061 [ http://arxiv.org/abs/1207.7235[1207.7235 ] ] .
collaboration , s. chatrchyan _ et .
_ , _ observation of a new boson with mass near 125 gev in @xmath133 collisions at @xmath134 and 8 tev _ , http://arxiv.org/abs/1303.4571[1303.4571 ] .
collaboration , g. aad _ et .
_ , _ observation of a new particle in the search for the standard model higgs boson with the atlas detector at the lhc _ , _ phys.lett . _ * b716 * ( 2012 ) 129 [ http://arxiv.org/abs/1207.7214[1207.7214 ] ] . s. p. martin , _ a supersymmetry primer _ , http://arxiv.org/abs/hep-ph/9709356[hep-ph/9709356 ] .
d. b. kaplan and h. georgi , _
@xmath135 breaking by vacuum misalignment _ , _ phys.lett . _
* b136 * ( 1984 ) 183 .
h. georgi and d. b. kaplan , _ composite higgs and custodial @xmath76 _ , _ phys.lett . _
* b145 * ( 1984 ) 216 .
t. gherghetta , b. von harling , a. d. medina and m. a. schmidt , _ the scale - invariant nmssm and the 126 gev higgs boson _ , _ jhep _ * 1302 * ( 2013 ) 032 [ http://arxiv.org/abs/1212.5243[1212.5243 ] ] .
r. contino , y. nomura and a. pomarol , _ higgs as a holographic pseudo - goldstone boson _ , _ nucl.phys .
_ * b671 * ( 2003 ) 148174 [ http://arxiv.org/abs/hep-ph/0306259[hep-ph/0306259 ] ] .
o. matsedonskyi , g. panico and a. wulzer , _ light top partners for a light composite higgs _ , _ jhep _ * 1301 * ( 2013 ) 164 [ http://arxiv.org/abs/1204.6333[1204.6333 ] ] .
d. marzocca , m. serone and j. shu , _ general composite higgs models _ , _ jhep _ * 1208 * ( 2012 ) 013 [ http://arxiv.org/abs/1205.0770 [ 1205.0770 ] ] .
a. pomarol and f. riva , _ the composite higgs and light resonance connection _ , _ jhep _ * 1208 * ( 2012 ) 135 [ http://arxiv.org/abs/1205.6434[1205.6434 ] ] .
g. panico , m. redi , a. tesi and a. wulzer , _ on the tuning and the mass of the composite higgs _ , _ jhep _ * 1303 * ( 2013 ) 051 [ http://arxiv.org/abs/1210.7114[1210.7114 ] ] .
d. pappadopulo , a. thamm and r. torre , _ a minimally tuned composite higgs model from an extra dimension _ , _ jhep _ * 1307 * ( 2013 ) 058 [ http://arxiv.org/abs/1303.3062[1303.3062 ] ] .
k. agashe , r. contino and a. pomarol , _ the minimal composite higgs model _ , _ nucl.phys .
_ * b719 * ( 2005 ) 165187 [ http://arxiv.org/abs/hep-ph/0412089[hep-ph/0412089 ] ] .
a. d. medina , n. r. shah and c. e. wagner , _ gauge - higgs unification and radiative electroweak symmetry breaking in warped extra dimensions _ , _
_ * d76 * ( 2007 ) 095010 [ http://arxiv.org/abs/0706.1281[0706.1281 ] ] .
b. gripaios , a. pomarol , f. riva and j. serra , _ beyond the minimal composite higgs model _ , _ jhep _ * 0904 * ( 2009 ) 070 [ http://arxiv.org/abs/0902.1483[0902.1483 ] ] .
j. mrazek , a. pomarol , r. rattazzi , m. redi , j. serra _ et .
_ , _ the other natural two higgs doublet model _ , _ nucl.phys . _ * b853 * ( 2011 ) 148 [ http://arxiv.org/abs/1105.5403[1105.5403 ] ] .
e. bertuzzo , t. s. ray , h. de sandes and c. a. savoy , _ on composite two higgs doublet models _ , _ jhep _ * 1305 * ( 2013 ) 153 [ http://arxiv.org/abs/1206.2623 [ 1206.2623 ] ] .
m. carena , a. d. medina , b. panes , n. r. shah and c. e. wagner , _ collider phenomenology of gauge - higgs unification scenarios in warped extra dimensions _ , _ phys.rev . _ * d77 * ( 2008 ) 076003 [ http://arxiv.org/abs/0712.0095[0712.0095 ] ] .
t. gherghetta and a. pomarol , _ a distorted mssm higgs sector from low - scale strong dynamics _ , _ jhep _ * 1112 * ( 2011 ) 069 [ http://arxiv.org/abs/1107.4697[1107.4697 ] ] . c. csaki , a. falkowski and a. weiler , _ the flavor of the composite pseudo - goldstone higgs _ , _ jhep _ * 0809 * ( 2008 ) 008 [ http://arxiv.org/abs/0804.1954[0804.1954 ] ] .
k. agashe , a. azatov and l. zhu , _ flavor violation tests of warped / composite sm in the two - site approach _ , _ phys.rev . _ * d79 * ( 2009 ) 056006 [ http://arxiv.org/abs/0810.1016[0810.1016 ] ] .
g. passarino and m. veltman , _ one loop corrections for @xmath136 annihilation into @xmath137 in the weinberg model _ , _
_ * b160 * ( 1979 ) 151 .
b. bellazzini , c. csaki , j. hubisz , j. serra and j. terning , _ composite higgs sketch _ , _ jhep _ * 1211 * ( 2012 ) 003 [ http://arxiv.org/abs/1205.4032[1205.4032 ] ] . s. weinberg , _ precise relations between the spectra of vector and axial vector mesons _ , _ phys.rev.lett . _ * 18 * ( 1967 ) 507509 . ,
cms - pas - b2g-12 - 006 , cern , geneva , apr , 2013 . | in composite higgs models light fermionic top partners often play an important role in obtaining a 126 gev higgs mass .
the presence of these top partners implies that coloured vector mesons , or massive gluon partners , most likely exist .
since the coupling between the top partners and gluon partners can be large there are then sizeable two - loop contributions to the composite higgs mass .
we compute the radiative correction to the higgs mass from a gluon partner in the minimal composite higgs model and show that the higgs mass is in fact reduced .
this allows the top partner masses to be increased , easing the tension between having a light composite higgs and heavy top partners . |
lcccccccc number & sequence & & & & orbital + & number & date & time & date & time & xis & hxd & phase@xmath1 + first & 100012010 & 2005 - 08 - 29 & 01:48 & 2005 - 08 - 30 & 01:30 & 50 & 53 & 1.39 + second & 100045010 & 2006 - 02 - 03 & 09:59 & 2006 - 02 - 03 & 22:45 & 21 & 17 & 1.47 + colliding wind binaries ( cwbs ) , a binary system comprised of two early - type stars with stellar winds at high mass loss rates and velocities , are the brightest class of stellar hard x - ray emitters .
theoretical interpretations are given for their luminous hard x - rays as the thermal emission from high - temperature plasma produced by strong shocks due to the colliding winds @xcite .
the interpretation is supported by the x - ray modulation observed along the orbital motion in some cwbs @xcite .
cwbs are also expected to be an agent of cosmic particle acceleration , which is evidenced by the presence of non - thermal radio emission from e.g. , wr140 @xcite .
charged particles with an energy distribution deviating from the maxwellian distribution give rise to the x - ray emission harder than the thermal emission dominant below 10 kev , thus the presence and the process of particle acceleration in cwbs can be best studied in the super - hard x - ray band . for convenience
, we refer to the 512 kev range as `` hard '' x - ray band , while the @xmath215 kev range as `` super - hard '' x - ray band .
the investigation of the super - hard emission from cwbs is important to unveil their enigmatic nature , and possibly gives a clue to understand the origin of @xmath3-ray emission reported from some massive star clusters containing cwbs @xcite .
we anticipate a great leap in our comprehension of non - thermal phenomena in cwbs with the advent of x - ray imaging telescopes in the super - hard band , which will be deployed in the next generation x - ray satellites such as astro - h @xcite . at present , however , our tool is limited to non - imaging or coded - mask techniques in this band , leaving @xmath0 carina the only practical target for this subject .
@xmath0 car is a luminous blue variable in the carina nebula at a distance of 2.3 kpc @xcite .
it is considered to constitute a binary system with another massive star with an orbital period of [email protected] d @xcite in a highly eccentric orbit @xcite . in this paper , we use the orbital phase such that the start of the cycle 11 minimum is phase 1.0 on the julian date of 2452819.8 d @xcite or 2003 june 29.3 .
@xmath0 car is one of the brightest stellar hard x - ray sources , and the hard x - ray modulation along the proposed ephemeris endorses the colliding - wind origin of the emission @xcite .
the development of the x - ray light curve through an orbital phase is as follows ; it experiences an eclipse from phase 0 to 0.03 ( we call `` minimum '' ) , recovers its brightness spending @xmath50.02 phases ( `` recovery '' ) , and gradually increases the flux for the rest of the phase ( `` steady increase '' ) to reach the brightest point ( `` maximum '' ) , which is terminated by the repeated eclipse in the next phase @xcite .
the periastron and the apastron are supposed to occur around phase 0.0 and 0.5 , respectively ( e.g. , @xcite ) . during the development ,
@xmath0 car shows episodic flares @xcite .
the spectrum of @xmath0 car below 10 kev has been described by complex models comprised of several thin - thermal plasma components with different temperatures , but a single plasma component of a 34 kev temperature dominates the hard band emission @xcite .
x - ray studies of @xmath0 car in the super - hard band were conducted by two observatories to date .
one is by bepposax using a combination of a gas scintillation proportional counter ( mecs ; medium energy concentrator spectrometer ) below 10 kev and a non - imaging phoswich scintillators ( pds ; phoswich detection system ) above 15 kev . @xcite and @xcite constrained the parameters of the 34 kev plasma component using the hard band data by the mecs , and extrapolated the model to claim that the pds data contain excess emission upon the 34 kev thermal emission .
the other is by the international gamma - ray astrophysics laboratory ( integral ) using a coded - mask imager ( isgri ; the integral soft gamma - ray imager ) in the 15 kev1 mev range . @xcite
unambiguously confirmed the super - hard emission claimed by the bepposax studies , showing that the isgri image has a significant signal at the position of @xmath0 car . when it comes to characterizing the spectrum of the super - hard excess emission , the results produced by the two observatories have a limited use with some concerns .
the pds is a non - imaging instrument with a wide field of view of in the full width at the half maximum ( fwhm ) , thus the contamination by other sources is inevitable . indeed ,
as was pointed out by @xcite , two neighboring super - hard sources ( igrj104476027 and 1e1048.15937 ) are within the pds field of view , which significantly contribute to the pds spectrum of @xmath0 car .
the isgri resolved super hard emission from @xmath0 car and is free from the contamination by these two sources .
however , its spectrum does not have rich statistics with only three spectral bins for a total of a 1 ms exposure .
the flux variation can not be tested .
also , the excess emission was examined against the extrapolated model of the 34 kev thermal emission obtained by the mecs at a different epoch , which may not be very appropriate given the variable nature of the hard x - rays of @xmath0 car .
the detection and the characterization of the excess emission in the super - hard band heavily depend on the accurate characterization and subtraction of the 34 kev thermal emission .
the examination of the excess emission is straight - forward , but it requires a careful treatment since the extrapolation of the 34 kev thermal model needs to be stretched for nearly an order of magnitude in the energy range . with these concerns ,
it is not surprising that the bepposax and the integral data resulted in a serious disagreement with each other for the super - hard excess emission . while the former claims the power - law model of a photon index of [email protected] , the latter claims [email protected] .
moreover , the flux estimates are inconsistent significantly by a factor of 3 .
it is unclear , in the first place , whether the super - hard excess emission is of a thermal or a non - thermal origin . to mitigate some of these concerns in the previous studies , we present the 550 kev results of suzaku observations conducted in 2005 august and 2006 february .
the non - imaging detector in the super - hard band on board suzaku has a narrower field of view than that of the pds and is free from the contamination by the two neighboring super hard sources . the x - ray charge coupled devices ( ccds ) sensitive below 12 kev have good and well - calibrated responses in the hard band .
their spectroscopic capability outperforms that of the mecs , resolving emission lines in the fe and ni k complex in the 6.08.5 kev band and thus enabling an accurate temperature determination of the 34 kev plasma sufficient to extrapolate toward the super - hard band .
we present the spectral models to explain the 550 kev emission of @xmath0 car at two epochs around the apastron .
we examine the thermal or non - thermal nature of the super hard excess emission and investigate its flux variability .
the results will set a firm basis to study the development of the super - hard emission in an entire orbital cycle , which we plan to conduct in a separate paper with the new data obtained at the end of 2008 during the latest x - ray maximum and minimum .
( 180mm,85mm)f1.eps the suzaku satellite @xcite observed the carina nebula multiple times since its launch .
we present the results of @xmath0 car in the first two observations on 2005 august 29 and 2006 february 3 in this paper , during which the object was around the apastron at the orbital phases of 1.39 and 1.47 , respectively ( table [ t1 ] ) .
the two observations were centered at ( ra , dec ) @xmath6 ( , ) in the equinox j2000.0 with nearly the opposite position angles . using the first data set
, @xcite studied the soft diffuse emission around @xmath0 car .
suzaku has a low earth orbit with an altitude of @xmath5570 km .
the observatory has two working instruments ; the x - ray imaging spectrometer ( xis ; @xcite ) and the hard x - ray detector ( hxd ; @xcite ) , which respectively covers the energy range of 0.212 kev and 10600 kev .
the xis is an imaging spectrometer using x - ray ccds with a format of 1024@xmath71024 pixels .
four ccd cameras are installed in the focal planes of four x - ray telescope modules @xcite coaligned with each other .
three of the four cameras ( xis 0 , 2 , and 3 ) are front - illuminated ( fi ) devices and the remaining one ( xis1 ) is back - illuminated ( bi ) .
the fi and the bi ccds have superior sensitivity to each other in the hard and the soft bands , respectively .
the energy resolution , which is subject to degradation in the orbit , is @xmath5140 ev and @xmath5160 ev at 5.9 kev in fwhm respectively for the first and the second observations .
the absolute energy gain is accurate to @xmath50.2% at 5.9 kev .
together with the xrt , the xis covers a field of view of 18@xmath718 with a pixel scale of 1 pixel@xmath8 .
the half - power diameter of the xrt is 1823 , which is almost independent of energy .
the hxd is a non - imaging spectrometer , consisting of several components responsible for different energy ranges . in this paper
, we use the pin component sensitive at 1070 kev with an fwhm energy resolution of 3.0 kev .
the pin detector is composed of 64 si pin diodes installed at the bottom of 8@xmath78 well - type collimators surrounded by gso scintillators .
the fine collimators within each well restrict the field of view .
the effective area monotonically decreases as an increasing distance from the field center , which is characterized by the fwhm view of @xmath534 or the full width at zero intensity ( fwzi ) view of 70 square ( figure [ f1 ] ) .
the center of the pin field is 35 offset from that of the the xis field .
thanks to the surrounding anti - coincidence scintillators , the narrow field of view , and the low background environment in the orbit , the pin achieves unprecedented sensitivity in the super - hard band . in both observations , the xis is operated in the normal clocking mode with a frame time of 8 s. we used the cleaned event lists produced by the pipeline processing version 2.0 .
the net exposure times are compiled in table [ t1 ] .
we used the heasoft version 6.5.1 for the data reduction and xspec version 11.3.2 for the x - ray spectral analysis .
figure [ f1 ] ( b ) shows the xis image in the first observation .
three point sources are recognized besides @xmath0 car : hd93250 , hd93205 , and wr25 , all of which are massive stars in this region . the underlying diffuse emission , which is a combination of unresolved point sources and the soft extended emission @xcite , is noticeable particularly in the western half of the image .
the second observation produced a similar image . in order to construct the xis spectra , we extracted the source and the background events from the solid and the dashed circles in figure [ f1 ] ( b ) , respectively .
the source extraction region is a 3 radius circle containing @xmath595% of photons from a point - like source at the circle center .
the background region is selected from a region devoid of bright emission .
the pile - up is negligible .
the wings of other point sources and the underlying diffuse emission contribute to the background , which is negligible in our targeted energy range .
this is illustrated by the fact that the emission above 5 kev at the x - ray minimum of @xmath0 car , which gives the upper limit to the contamination by irrelevant emission , is roughly two orders of magnitude smaller than that in the x - ray bright phase @xcite .
we constructed hard band light curves of the count rate during the 50 and 21 ks exposures in the first and the second observations , respectively , and found no significant flux changes such as flare - like variability ( figure [ f8 ] ) .
for the spectral modeling , we merged the spectra obtained by the three fi devices with nearly identical responses .
the merged fi spectrum has a good signal up to @xmath512 kev , while the bi spectrum is effective up to @xmath59 kev .
we generated the redistribution matrix functions and the ancillary response files using the ` xisrmfgen ` and ` xissimarfgen ` tools @xcite , respectively .
( 180mm,85mm)f8.eps figure [ f1 ] ( a ) shows the field of view of the pin detector .
one of the two integral sources ( igrj104476027 ) is outside of the fwzi view .
the other ( 1e1048.15937 ) is close to the edge of the view in both observations , at which the effective area decreases to less than 5% of the value at the field center .
@xcite claimed the integral isgri count rate of 0.16 s@xmath8 for @xmath0 car and 0.09 s@xmath8 for 1e1048.15937 . assuming that 1e1048.15937 does not vary drastically among the integral and the two suzaku observations , the contamination by this source is negligible in the pin spectrum .
we constructed the background spectrum , which is composed mostly of the non x - ray background ( nxb ) with some contribution by the cosmic x - ray background ( cxb ) .
the galactic ridge x - ray emission is negligible toward @xmath0 car at a galactic longitude of @xmath5290@xmath9 . for the nxb
, we used the simulated background events distributed by the instrument team , which reproduces the nxb spectrum at an accuracy of 1.3% . for the cxb
, we assumed the spectral model derived from heao-1 observations @xcite .
we produced a cxb spectrum by convolving the detector angular and energy responses with the simulated emission of the assumed model and the spatially uniform distribution .
the resultant nxb and cxb data were merged to construct the pin background spectrum . for the spectral modeling , we used the latest detector response for a point source at the xis field center ( ` ae_hxd_pinxinome1_20080129.rsp ` ) . at this position ,
the relative flux among the xis devices and the pin is calibrated at an accuracy of @xmath51% .
( 85mm,85mm)f2.eps with the background - subtracted spectra and the response files both for the xis and the pin , we are now ready for the spectral modeling .
it is not a wise decision to start fitting the two spectra simultaneously ; the statistics of the xis spectrum overwhelms that of the pin spectrum , making the pin data ignored when fitted at the equal weight .
we therefore took a more step - wise approach described below .
figure [ f2 ] shows the xis spectra in the 510 kev range , in which we can resolve emission lines from highly ionized fe and ni as well as hard continuum emission extending beyond 10 kev .
these are the characteristics of the emission arising from a thin - thermal plasma of several kev temperatures .
in addition to the thermal features , we can also find an emission line at @xmath56.4 kev due to the k@xmath10 emission of quasi - neutral fe ( hereafter called fe ) and an absorption feature at 78 kev due to the k - shell photoelectric absorption edge by fe .
the rich set of thermal features in our xis spectra enables us to constrain the plasma temperature stringently , in particular , by using the intensity ratio between the fe k@xmath10 line at 6.7 kev and fe k@xmath11 line at 7.9 kev and that between fe k@xmath10 line at 6.7 kev and fe k@xmath10 line at 7.0 kev .
lcc parameters@xmath12 & first@xmath13 & second@xmath13 + @xmath14 ( kev ) & 3.9@xmath15 & 4.1@xmath16 + @xmath17 ( solar ) & 0.47@xmath18 & 0.45@xmath19 + @xmath20 ( solar ) & 1.4@xmath21 & 0.9@xmath22 + @xmath23 ( @xmath24 erg s@xmath8 @xmath25 ) & @xmath26 & @xmath27 + @xmath28 ( kev ) & 6.42@xmath29 & 6.43@xmath29 + @xmath30 ( @xmath31 s@xmath8 @xmath25 ) & @xmath32 & @xmath33 + @xmath34 ( kev ) & 7.11@xmath35 & 6.97@xmath36 + @xmath37 & 0.12@xmath38 & 0.12@xmath39 + @xmath40/d.o.f .
& 346.2/323 & 143.9/138 + ( 180mm,85mm)f3.eps to fully exploit the xis capability , we fitted the 6.08.5 kev range spectra , in which we can find most of the emission lines responsible for the plasma temperature determination .
we applied the model composed of a thin - thermal plasma component ( ` vmekal ` ; @xcite ) for the thermal features and a gaussian line component for the fe k@xmath10 line .
we attenuated the model by an interstellar extinction model ( ` wabs ` ; @xcite ) as well as an additional photoelectric absorption edge to account for the fe k edge feature .
the free parameters in the fitting are the plasma temperature ( @xmath14 ) , the fe and ni abundance ( @xmath17 and @xmath20 , respectively ) , and the flux ( @xmath23 ) in the 510 kev band for the thin - thermal plasma component , the line energy ( @xmath28 ) and the intensity ( @xmath30 ) for the gaussian line component , and the edge energy ( @xmath34 ) and the optical depth ( @xmath37 ) for the photoelectric absorption edge component . here
, the superscripts of the parameters represent the spectral component that they belong to .
the amount of extinction was fixed to 1.7@xmath710@xmath41 @xmath25 , which is a typical value derived in the hard band for the colliding - wind plasma of @xmath0 car @xcite .
the intrinsic width of the gaussian line was fixed to 0 ev .
the linearity coefficient between the pulse height and the energy was adjusted to compensate for a possible systematic uncertainty .
we obtained statistically - acceptable fits both for the first and the second observations ( table [ t2 ] ) .
the plasma temperature ( @xmath14 ) and the flux ( @xmath23 ) were determined at an accuracy of @xmath55% .
the flux between the two observations changed significantly by @xmath520% .
it is interesting to note that , while @xmath17 and @xmath20 are consistent between the two observations , @xmath20 is larger than @xmath17 by a factor of a few ; its cause is not clear .
we extrapolated the best - fit 6.08.5 kev model beyond 10 kev and compared to the pin spectrum in order to examine the excess emission in the super - hard band .
the top panels in figure [ f3 ] show the xis and the pin spectra below 12 kev and above 15 kev , respectively , in comparison with the extrapolated best - fit model in the 6.08.5 kev range .
the middle panels show the residuals , in which we can see a consistent positive residual for the pin spectrum in both observations .
this illustrates the presence of the excess emission beyond 15 kev .
we tested the claim for two major sources of systematic uncertainty .
one is the systematic uncertainty in the nxb model of the pin data , which constitutes the majority of the background signal .
we constructed the background - subtracted pin spectra for the scaled nxb models by @xmath41.3% , which is the accuracy of the model , and plotted them against the extrapolated model in different darkness of green colors in the top panels of figure [ f3 ] .
for the increased nxb model , we still find consistent positive residuals beyond 10 kev .
the other is the statistical uncertainty in the best - fit 6.08.5 kev model . in order to examine the upper bound of the extrapolated model
, we replaced the best - fit @xmath14 and @xmath23 values with their maximum allowed values within 90% statistical uncertainty .
the modified model was extrapolated and compared to the data .
the residuals are shown in the bottom panels in figure [ f3 ] , which again show consistent positive residuals of the data above the extrapolated model .
( 180mm,85mm)f4.eps ( 180mm,85mm)f5.eps we established the presence of the excess emission in the super - hard band and now proceed to the spectral modeling in the entire 550 kev range using both the xis and the pin data .
for the hard emission , we used the same model for the 6.08.5 kev fitting procedure in subsection [ s321 ] with a slight modification that the @xmath17 and @xmath20 are tied between the two observations . upon this thermal emission
, we added another component to account for the super - hard excess emission .
two models were employed ; a power - law model and a thermal bremsstrahlung . the free parameters in the additional components are the power - law photon index ( @xmath42 ) and the flux @xmath43 in the 1050 kev range for the power - law model and the electron temperature ( @xmath44 ) and the 1050 kev flux ( @xmath45 ) for the bremsstrahlung model . both yielded an acceptable fit .
the best - fit values are tabulated in tables [ t3 ] and [ t4 ] , while the data and the best - fit model are shown in figures [ f4 ] and [ f5 ] , respectively , for the additional power - law and bremsstrahlung models .
we assessed the systematic uncertainty in the derived parameters for the two major sources of such uncertainty .
first , for the pin nxb model uncertainty , we repeated the 550 kev fitting procedure using the scaled nxb models by @xmath41.3% .
for the power - law model , we found that this brings systematic uncertainties of ( @xmath46 , @xmath43/10@xmath47 erg s@xmath8 @xmath25 ) @xmath5 ( 0.2 , 0.3 ) for the first and ( 0.1 , 0.3 ) for the second observations .
for the bremsstrahlung model , the systematic uncertainty for @xmath48/@xmath24 erg s@xmath8 @xmath25 is @xmath50.6 and @xmath5 0.5 in the first and the second observations , respectively .
the resultant systematic uncertainty of @xmath49 is so large that it makes @xmath49 practically unconstrained in the first observation and @xmath216 kev in the second observation .
next , for the statistical uncertainty in the best - fit 34 kev thermal model parameters , we replaced the best - fit set of @xmath14 and @xmath23 values with those of their maximum and minimum allowed values within 90% statistical uncertainty .
this brings systematic uncertainties of ( @xmath46 , @xmath43/10@xmath47 erg s@xmath8 @xmath25 ) @xmath5 ( 0.02 , 0.2 ) and ( 0.01 , 0.2 ) respectively for the first and the second observations in the power - law model . for the bremsstrahlung model ,
the @xmath49 kev is unconstrained with minima of @xmath222 kev and @xmath210 kev for the first and the second observations .
the resultant systematic uncertainty for @xmath48/@xmath24 erg s@xmath8 @xmath25 is 0.3 and 0.2 , respectively .
lcc parameters@xmath12 & first@xmath13 & second@xmath13 + @xmath42 & 1.30@xmath50 & 1.51@xmath51 + @xmath43 ( @xmath24 erg s@xmath8 @xmath25 ) & 1.4@xmath52 & 1.9@xmath21 + @xmath14 ( kev ) & 3.61@xmath38 & 3.29@xmath38 + @xmath17 ( solar ) & + @xmath20 ( solar ) & + @xmath23 ( @xmath24 erg s@xmath8 @xmath25 ) & 2.76@xmath53 & 2.07@xmath54 + @xmath28 ( kev ) & 6.43@xmath29 & 6.44@xmath18 + @xmath30 ( @xmath31 s@xmath8 @xmath25 ) & 7.9@xmath55 & 5.5@xmath56 + @xmath34 ( kev ) & 7.12@xmath57 & 7.14@xmath58 + @xmath37 & 0.16@xmath54 & 0.14@xmath57 + @xmath40/d.o.f . & 717.0/590 & 286.1/226 + lcc parameters@xmath12 & first@xmath13 & second@xmath13 + @xmath44 ( kev ) & @xmath59 & 18@xmath60 + @xmath45 ( @xmath24 erg s@xmath8 @xmath25 ) & 1.7@xmath61 & 1.2@xmath62 + @xmath14 ( kev ) & 3.58@xmath63 & 3.14@xmath64 + @xmath17 ( solar ) & + @xmath20 ( solar ) & + @xmath23 ( @xmath24 erg s@xmath8 @xmath25 ) & 2.74@xmath65 & 1.95@xmath66 + @xmath28 ( kev ) & 6.43@xmath29 & 6.43@xmath18 + @xmath30 ( @xmath31 s@xmath8 @xmath25 ) & 7.8@xmath67 & 5.3@xmath68 + @xmath34 ( kev ) & 7.12@xmath69 & 7.1@xmath52 + @xmath37 & 0.16@xmath70 & 0.14@xmath71 + @xmath40/d.o.f . &
714.1/590 & 276.2/226 +
lccc observatory & @xmath72 & @xmath73/10@xmath47 & orbital + & & ( erg s@xmath8 @xmath25 ) & phase@xmath74 + bepposax & 2.42@xmath57 & 3.24 & 0.46 + integral & 1.0@xmath22 & [email protected] & 0.981.36 + suzaku ( 1st ) & 1.30@xmath50 & 2.7@xmath75 & 1.39 + suzaku ( 2nd ) & 1.51@xmath51 & 1.9@xmath76 & 1.47 + suzaku ( both ) & 1.38@xmath77 & 2.3@xmath78 & 1.39 , 1.47 + we compare the results for the super - hard excess emission among the two suzaku observations and the two previous studies .
we use the spectral parameters for the power - law model , which are available both for the bepposax and the integral results ( table [ t5 ] ) .
the flux values of bepposax @xcite and suzaku ( table [ t3 ] ) were reevaluated in the 22100 kev range to facilitate comparison with the integral result @xcite .
we see no significant change between the two suzaku results both for the flux and the power - law index if we consider both the statistical and systematic uncertainties . therefore , we fitted the two 550 kev spectra simultaneously to obtain a much constrained fit . in the fitting ,
the best - fit parameters of the thin - thermal plasma components were fixed to the best - fit values for the independent fits .
the parameters of the statistically - acceptable best fit are appended in table [ t5 ] .
the suzaku results were obtained at different epochs during the x - ray steady increase phase near the apastron with no apparent flare - like flux amplification in the hard band .
this infers that the super - hard excess emission is not caused by episodic events , but is a constant emission with little change both in the flux and the spectral hardness .
the best - fit suzaku model is compared to the integral spectrum in figure [ f6 ] .
@xmath46 is the same within the statistical and the systematic uncertainties , but @xmath79 is significantly higher in suzaku than integral ( table [ t5 ] ) .
the integral result was obtained in a wide range of orbital phases including the minimum , recovery , and steady increase by accumulating a 1 ms exposure .
we speculate that the lower flux in the integral data is a consequence of the fact that the integral spectrum was accumulated in periods including the orbital eclipse , during which the super - hard emission may be absent .
the emission mechanism of the super - hard excess x - rays and the production mechanism of the charged particles ( mostly electrons ) to produce the radiation are two different problems .
for the emission mechanism , we consider the bremsstrahlung and the inverse compton radiations .
the bremsstrahlung emission is either thermal stemming from thermally - relaxed high - temperature plasma or non - thermal from accelerated charged particles with a non - thermal energy distribution .
first , we consider the thermal bremsstrahlung by a thermally - relaxed plasma . in an adiabatic strong shock ,
the post shock temperature is @xmath80 , where @xmath81 is the average particle mass in the unit of the proton mass ( @xmath82 ) and @xmath83 is the shock velocity . substituting @xmath843000 km s@xmath8 @xcite , we obtain @xmath85 18@xmath81 kev .
@xmath81 increases for h - deficient gas widely seen among stellar winds from evolved early - type stars such as @xmath0 car , and reaches 4/3 for the most extreme case that the gas is composed only of fully - ionized he .
the thermal bremsstrahlung model was abandoned in the bepposax @xcite analysis and was unemployed in the integral analysis @xcite .
the combined suzaku spectrum indicates @xmath86 35 kev , but given a large systematic uncertainty , we consider that this emission mechanism remains a viable explanation .
second , we argue against the bremsstrahlung emission by accelerated charged particles colliding upon ambient cold matter .
this process is proposed for the super - hard emission seen in solar and stellar flares @xcite .
@xmath0 car has a rich ambient matter to brake the accelerated particles , which is evidenced by a large amount of circumstellar extinction and the presence of the fe fluorescence line @xcite .
however , the x - ray yield of this process is very low of @xmath8710@xmath88 even for the most efficient bremsstrahlung by electrons of a 10 kev to 1 mev kinetic energy against a thick target .
most of the kinetic energy is lost by the ionization of cold target .
the measured luminosity is @xmath89 erg s@xmath8 in the super - hard band for the two suzaku observations , which implies that a constant energy input is necessary at a rate of @xmath8710@xmath90 erg s@xmath8 .
this is hardly possible given the total kinetic energy of the stellar wind in this cwb being @xmath510@xmath91 erg s@xmath8 @xcite .
finally , we consider the inverse compton emission of stellar uv photons by a non - thermal population of electrons scattering the ambient uv photons up to the super - hard band @xcite .
this is the most popular explanation adopted in @xcite and @xcite , and our suzaku result agrees with this interpretation .
several mechanisms of acceleration to produce the non - thermal population are proposed , which includes the first - order fermi acceleration at the shock , electric field acceleration in the current sheet in a magnetized wind @xcite , and magnetic reconnections .
the x - ray photon index of @xmath51.4 indicates that the injected non - thermal electrons have an energy distribution with an index of @xmath52 or harder , depending on how the compton cooling balances the injected energy .
the diffusive shock acceleration predicts the value of 2 for the electron energy index , and some modifications may need to be considered @xcite .
the authors acknowledge y. hyodo for his help in the xis data reduction , j. pittard for his comments on particle acceleration , and j. c. leyder and r. walter for providing the integral spectrum .
we thank the suzaku science working group team for their effort in the initial phase of the operation , during which the present data set was obtained .
support for this work is provided by the research center of the advanced measurement at rikkyo university ( a.s . ) , the grants - in - aid for scientific research by the ministry of education , culture , sports , science and technology of japan ( grant numbers 19654032 and 19340047 for s.k .
and 20540237 for y.t . ) , a chuo university grant for special research ( y.t . ) , and the national aeronautics and space administration through chandra postdoctoral fellowship award number pf6 - 70044 ( m.t . )
issued by the chandra x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of the national aeronautics space administration under contract nas8 - 03060 .
aharonian , f. , et al .
2007 , , 467 , 1075 albacete colombo , j. f. , mndez , m. , & morrell , n. i. 2003 , , 346 , 704 boldt , e. 1987 , observational cosmology , 124 , 611 chen , w. , & white , r. l. 1991 , , 366 , 512 corcoran , m. f. , et al .
1998 , , 494 , 381 corcoran , m. f. , et al . 2004 , , 613 , 381 damineli , a. , et al .
2008 , , 384 , 1649 davidson , k. , & humphreys , r. m. 1997 , , 35 , 1 evans , n. r. , seward , f. d. , krauss , m. i. , isobe , t. , nichols , j. , schlegel , e. m. , & wolk , s. j. 2003 , , 589 , 509 evans , n. r. , schlegel , e. m. , waldron , w. l. , seward , f. d. , krauss , m. i. , nichols , j. , & wolk , s. j. 2004 , , 612 , 1065 fernndez - lajs , e. , et al .
2009 , , 493 , 1093 hamaguchi , k. , et al . 2007a , , 663 , 522 hamaguchi , k. , et al .
2007b , , 59 , 151 ishibashi , k. , corcoran , m. f. , davidson , k. , swank , j. h. , petre , r. , drake , s. a. , damineli , a. , & white , s. 1999 , , 524 , 983 ishisaki , y. , et al .
2007 , , 59 , s113 jardine , m. , allen , h. r. , & pollock , a. m. t. 1996 , , 314 , 594 kaastra , j. s. 1992 , an x - ray spectral code for optically thin plasmas ( internal sron - leiden report , updated version 2.0 ) kokubun , m. , et al . 2007 , , 59 , 53 koyama , k. , et al .
2007 , , 59 , 23 leyder , j .- c . , walter , r. , & rauw , g. 2008 , , 477 , l29 liedahl , d. a. , osterheld , a. l. , & goldstein , w. h. 1995 , , 438 , l115 luo , d. , mccray , r. , & mac low , m .-
1990 , , 362 , 267 mewe , r. , gronenschild , e. h. b. m. , & van den oord , g. h. j. 1985 , , 62 , 197 mewe , r. , lemen , j. r. , & van den oord , g. h. j. 1986 , , 65 , 511 mitsuda , k. , et al .
2007 , , 59 , 1 morrison , r. , & mccammon , d. 1983 , , 270 , 119 nielsen , k. e. , corcoran , m. f. , gull , t. r. , hillier , d. j. , hamaguchi , k. , ivarsson , s. , & lindler , d. j. 2007 , , 660 , 669 osten , r. a. , drake , s. , tueller , j. , cummings , j. , perri , m. , moretti , a. , & covino , s. 2007 , , 654 , 1052 pittard , j. m. , & corcoran , m. f. 2002 , , 383 , 636 pittard , j. m. , & dougherty , s. m. 2006 , , 372 , 801 serlemitsos , p. j. , et al .
2007 , , 59 , 9 stevens , i. r. , blondin , j. m. , & pollock , a. m. t. 1992 , , 386 , 265 stevens , i. r. , corcoran , m. f. , willis , a. j. , skinner , s. l. , pollock , a. m. t. , nagase , f. , & koyama , k. 1996 , , 283 , 589 takahashi , t. , et al .
2007 , , 59 , 35 takahashi , t. , et al .
2008 , , 7011 , 18 tsuboi , y. , koyama , k. , sakano , m. , & petre , r. 1997 , , 49 , 85 usov , v. v. 1992 , , 389 , 635 viotti , r. f. , et al .
2002 , , 385 , 874 viotti , r. f. , antonelli , l. a. , rossi , c. , & rebecchi , s. 2004 , , 420 , 527 white , r. l. , & becker , r. h. 1995 , , 451 , 352 zhekov , s. a. , & skinner , s. l. 2000 , , 538 , 808 | we present the suzaku results of @xmath0 carinae in the 550 kev range conducted twice around the apastron in 2005 august for 50 ks and in 2006 february for 20 ks .
the x - ray imaging spectrometer ( xis ) produced hard ( 512 kev ) band spectra , resolving k shell lines from highly ionized fe and ni .
the hard x - ray detector yielded a significant detection in the super - hard ( 1550 kev ) band , which is uncontaminated by near - by sources .
we constrained the temperature of the optically - thin thermal plasma emission dominant in the hard band as 34 kev using the k - shell line features with the xis .
we found significant excess emission above the thermal emission in the super - hard band with the pin , confirming the previous integral isgri report .
the entire 550 kev spectra were fitted by a combination of a thermal plasma model plus a flat power - law or a very hot thermal bremsstrahlung model for the excess emission .
no significant change of the excess emission was found at different epochs within the systematic and statistical uncertainties and no flare - like flux amplification was seen in the hard band , indicating that the excess emission is a steady phenomenon .
we argue that the super - hard emission is attributable to the inverse compton of stellar uv photons by non - thermal electrons or to the thermal bremsstrahlung of very hot plasma , and not to the bremsstrahlung by non - thermal electrons colliding with cold ambient matter . |
notwithstanding simple underlying physics , percolation transition are _ bona fide _ second order phase transitions , with all of usual the hallmarks of singular behavior at a critical point @xcite .
percolation transitions , phase transitions mediated by a disordering influence such as the random removal of sites or bonds on a regular lattice geometry , are amenable to numerical study in the context of monte carlo calculations used to treat disorder .
however , while the critical behavior in discrete systems may be characterized to a high level of precision in this manner , there are salient examples of percolation phenomena which can not be reduced to discrete lattices . in realistic porous media ,
the flow of liquid on a macroscopic basis typically entails the passage of liquid around irregularly shaped granules which comprise the material and represent barriers to fluid flow . for the movement of fluid or charge in porous materials ,
there is a dichotomy for volume concentrations of barrier particles above and below a critical concentration @xmath2 .
if inclusions are sufficiently sparse , voids join together in a macroscopic connected network spanning the entire system , allowing for the transmission of fluid at the bulk level . on the other hand , for large enough concentrations of randomly placed barrier particles ,
contiguous navigable voids are limited in size to a finite length scale @xmath3 , barring the transmission of fluid at a system wide level ; the critical density @xmath2 of inclusions which separates the two scenarios is thus of practical significance . for the sake of convenience ,
we rescale coordinates in such a way that the spherical inclusions are each of unit radius .
although in principle one may record the critical density @xmath2 where the percolation transition occurs , a more frequent practice in discussing percolation through voids is to report instead the excluded volume @xcite , given by @xmath4 .
determining if voids may be traversed at the macroscopic level is a continuum percolation problem for which a topologically equivalent discrete counterpart is in general not readily accessible , a condition very distinct from that of a complementary percolation problem ( with the permeability pattern reversed ) where the flux occurs through networks formed of overlapping included particles with interstitial regions impervious to fluid flow .
the geometry of the randomly placed barriers , known _ a priori _ , facilitates the calculation of characteristics related to percolation when the flow is through included particles instead of the interstitial spaces . on the other hand ,
the fact that the shapes of voids between barrier particles are only evident _ a posteori _ hampers a rigorous identification of networks formed by connected voids and enhances the challenge of finding the percolation threshold . nevertheless , with the aid of a finite size scaling analysis , we determine not only the critical volume fraction @xmath5 , but also the critical exponent @xmath6 associated with the correlation length @xmath3 .
previous theoretical studies include calculations which use an approximate discrete scheme and extrapolate to the continuum case @xcite . in a related effort
, the discretization scheme has been applied to randomly placed ellipsoids @xcite and oblate spheroids @xcite . in particular cases , such as systems comprised of randomly placed spheres , discrete networks equivalent to void spaces @xcite may be constructed .
voronoi tessellations , and the appropriate generalizations have been applied @xcite to systems comprised of randomly located spheres . in a recent set of calculations , there is no discretization scheme and the simulation entails following tracer particles in an effectively infinite sized system @xcite ; the percolation threshold is identified by determining the density of spheres for which the tracers cease to move diffusively . in this work , we use a dynamical approach where systems of finite size are examined . in this way
, we are able to apply standard machinery of finite size scaling analysis while providing rigorous treatment to the geometric intricacies of voids between barriers which serve as conduits for fluid or charge in the percolative regime .
an order parameter , an intensive thermodynamic variable finite when the phase is intact , and decreasing continuously to zero at a phase boundary ( e.g. magnetization per unit volume in the case of ferromagnetism disrupted by thermal fluctuations above the curie temperature and nonzero below it ) , may be pressed into service as a tool to locate second order phase transitions and thereby determine the phase diagram . in percolation phenomena where connected navigable clusters are readily identified ,
the `` strength '' of the spanning cluster , or the number of sites in the lattice it encompasses relative to the total number of sites in the system , is conveniently sampled in monte carlo simulations using techniques such as the hoshen - kopelman algorithm for the identification and characterization of connected clusters @xcite . in the scenario of interest , where percolation occurs through
irregularly shaped interstices among randomly placed barrier particles of a prescribed shape , the connectivity of void volumes to adjacent navigable regions may be difficult to establish in more general situations ( e.g. non - spherical inclusions such as tetrahedra , cubes , or octahedra ) . to address this challenge
, we use dynamical simulations to probe and determine the extent of voids .
although stochastically driven exploration is validated in the case of randomly placed spheres , the approach may be used in scenarios where variants of voronoi tessellation are more difficult to apply . in the spirit of a _ gedanken _ experiment ,
we envision applying a reflective coating to exposed surfaces of overlapping spheres such that light striking the included particles is reflected with no penetration of the incident rays .
whereas the transmission of light through the matrix of inclusions would signal that the system would admit the flow of fluid or electrical charge , a configuration which blocks the transmission of incident light also lacks a percolating path .
dynamical trajectories are initiated with an unbiased choice for the point of origin in the interior of a void and exterior to all barrier particles . to select a void point ,
randomly chosen candidate locations are accepted only if the point is exterior to all neighboring spheres .
some economy is gained in partitioning the 3d system into a cubic grid , where the unit cell size is slightly greater than the radius @xmath7 of the spheres .
the subdivisions reduce the task of checking for penetration into nearby spheres to simply checking the cubic unit cell occupied by the prospective ( randomly chosen ) starting point as well as each of the adjacent cubes . to set the virtual ray in motion , a `` velocity '' vector @xmath8 is randomly chosen in a spherically symmetric way to eliminate any directional bias , with @xmath8 of unit length since `` speed ''
has no physical or computational significance in the context of the order parameter calculation .
the ray trajectory is @xmath9 with the starting point @xmath10 and the orientation of the velocity vector @xmath8 stochastically determined .
the path of the ray must be traced until contact is made with the nearest sphere ; with a barrier particle centered at @xmath11 , the condition for intersection with a sphere of radius @xmath7 is @xmath12 .
standard geometric arguments lead to @xmath13 for the distance to a point of intersection , where @xmath14 and @xmath15 . choosing the negative sign corresponds to the closer point of entry and the positive sign to the more distant point of departure as the ray propagates through and exits from the sphere . finding the closest included
particle intersected by the ray is tantamount to minimizing @xmath16 , the shortest distance to a sphere along the trajectory . to reduce computational overhead ,
the system is subdivided into a cubic grid in the same fashion as used to determine if a randomly chosen point is located in a void ; rays are propagated from one small unit cell to the next until a collision with a barrier particle is recorded . at the site of a collision @xmath17 , to avoid penetration of the impermeable particles comprising the matrix , the direction of motion must change in such a way that the incident ray is projected away from the interior of the barrier particle .
specular reflection is implemented by reversing the component of the velocity in the direction of @xmath18 , the radius vector extending from the sphere center to the site of the collision .
although mirror reflection encapsulates the physics of geometric optics appropriate to a ray of light moving in a matrix of silvered spheres , propagation by simple specular reflection may be improved on , particularly in the case of neck - like voids where many collisions may be required to traverse the narrow corridor if the ray is directed normal to the primary axis of the void . to improve the efficiency of navigation by roughly an order of magnitude , we incorporate moves where , instead of mirror reflection , the ray trajectory shifts by a right angle . in 3d ,
even with a perpendicular shift imposed , there remains a degree of freedom , which is selected at random to provide stochastic input at each collision .
the balance of the moves is such that stochastically infused right angle moves outnumber specular moves two to one . to interrogate the system and determine whether a void is part of a spanning cluster , we calculate the excursion parameter @xmath19 , structured to allow for the incorporation of statistical information gleaned from traversals along each of the three cartesian axes . in calculating the rms deviations , special consideration
must be given the periodicity of the system .
when the density of randomly placed spheres is high enough to block percolation , where @xmath20 , the scale @xmath3 of contiguous void networks is in most cases small relative to the system size , and accordingly , @xmath21 . on the other hand , if inclusions are sparse enough that @xmath22 , system wide traversal becomes feasible .
moreover , with periodic boundary conditions imposed the rms deviations along one or more axes could in principle grow without bound as the ray propagates from one replica of the system to the next .
in such situations , the result would be an unrealistically large order parameter , a circumstance which may be remedied by truncating @xmath23 and counterparts along other axes at the system size @xmath24 .
thus , in the extreme case of a system devoid of inclusions , @xmath25 would attain its maximum value of 1 .
the excursion quantity @xmath25 , which saturates at unity in the absence of obstacles and falls to zero at the percolation threshold for bulk systems , serves as an order parameter giving a measure of the size of a randomly selected void . in the thermodynamic limit ,
the configuration averaged @xmath25 also gives the probability that a randomly selected void is a system spanning cavity . in a diffusive excursion , though the disorder and time averaged vector displacement vanishes , the root mean square distance traveled scales as @xmath26 , where @xmath27 is the number of ray interactions ( and concomitant trajectory deflections ) with barrier particles . unless one is precisely at the percolation threshold with rms deviations from the trajectory starting point scaling as @xmath28 , where @xmath29 in the non - diffusive critical regime @xcite , configuration averaged ray excursions will scale diffusively with the number of sphere interactions , if the system is sufficiently large .
if the void is small in size , a ray which propagates in the cavity will probe its extent relatively quickly , yielding a finite @xmath25 value . on the other hand , with a system spanning void scaling as the system size @xmath24
, the ray traversal time varies as @xmath30 on average .
hence , to avoid a spuriously low result for the percolation threshold @xmath5 , trajectories are propagated with collisions accumulating until @xmath31 where @xmath32 is an integer prefactor . by successively doubling @xmath32 , a saturation point
is eventually found ( i.e. for @xmath33 and @xmath34 ) where statistically significant changes in critical quantities cease to occur .
the calculation of the @xmath25 parameter entails averaging over many realizations of disorder ( i.e. at least @xmath35 ) to suppress statistical fluctuations and permit the observation of finite size scaling trends , which are exploited to calculate critical quantities such as @xmath5 and the correlation length exponent @xmath6 . to generate disorder configurations ,
we sample the grand canonical ensemble as described in the appendix to determine the number of spherical barriers encompassed in the simulation volume .
the mersenne twister algorithm , which generates high quality random numbers with a modest computational cost , is used for stochastic input .
( color online ) the excursion order parameter @xmath25 is plotted with respect to excluded volume @xmath36 for various system sizes ; the inset is a magnified view of the intersection of the @xmath25 curves . ]
percolation transitions , whether in the context of continuum or in discrete cases , are second order transitions with singular behavior in thermodynamic variables at the phase transition .
single parameter finite size scaling analysis posits a form @xmath37 l^{1/\nu})$ ] for thermodynamic variables , a singular form anticipated for the excursion order parameter @xmath25 . since @xmath25 is normalized and thus dimensionless , @xmath38 , and the singular dependence reduces to @xmath39 l^{1/\nu})$ ] .
an important implication is the crossing of @xmath25 curves at @xmath5 for different system sizes , provided the system size is large enough .
similar intersections are seen in the case of the binder cumulant @xcite at the curie temperature @xmath40 in the context of ferromagnetic transitions , and the crossing phenomena in both the former and the latter represent a robust avenue for locating second order phase transitions . fig .
[ fig : fig1 ] shows @xmath25 curves for different system sizes ( i.e. @xmath41 , @xmath42 , and @xmath43 ) with a well defined crossing at a common point . to determine @xmath5 in a systematic fashion
, we calculate the order parameter for pairs of system sizes ( e.g. @xmath44 and @xmath45 ) with the larger member of each pair 1.5 times the size of the smaller .
although in the thermodynamic limit crossings would occur exactly at @xmath5 , the locations of intersections for finite size systems [ e.g. @xmath46 differ slightly from @xmath5 due to finite size effects , and must be considered pseudocritical points @xcite .
the pseudo - critical excluded volume @xmath46 for each pair are recorded in table [ tab : phicvalues ] and also appear in the graph in fig .
[ fig : fig2 ] , with the system size reciprocal @xmath47 on the abscissa . for small system sizes
, @xmath46 rises , eventually saturating at @xmath48 , the extrapolated result for the percolation threshold .
.[tab : phicvalues ] @xmath25 curve intersection results with for various @xmath49 values [ cols="^,^,^,^",options="header " , ] ( color online ) pseudocritical excluded volumes ( with calculated error bars ) plotted with respect to the reciprocal of the mean system size @xmath49 for each pair ] ( color online ) data collapse with @xmath25 data shown for various system sizes .
the solid red line is a polynomial curve determined from a nonlinear least squares fit , and symbols represent data from monte carlo simulations . ] a complementary finite size scaling technique for calculating the percolation threshold @xmath5 , based on the data collapse phenomenon in the critical region , simultaneously takes into consideration results for a broad range of system sizes and also yields the correlation length exponent @xmath6 .
data collapse graphs may be prepared by plotting on the vertical axis the quantity @xmath25 with @xmath50 ( with an optimal collapse for quantitatively correct values for @xmath6 and @xmath5 ) on the abscissa for various systems sizes .
the coincidence of data on the same curve is a manifestation of the existence of a universal singular scaling function for @xmath25 as well as an indication that @xmath6 and @xmath5 are specified correctly . to use the data collapse as a quantitative tool ,
one notes that the `` v '' shaped curve in fig .
[ fig : fig3 ] is essentially linear with a small degree of curvature . to accommodate this quasi - linear variation in a perturbative fashion
, we use a form @xmath51 ( the results are not impacted by the inclusion of the quartic term , and the series is truncated at fourth order ) with @xmath52 . to assess the degree to which the points from monte carlo calculations lie on the same curve , we use @xmath53^{2}}.\ ] ] the quantity @xmath54 provides an overall measure of the deviation of monte carlo data from the scaling curve . hence
, optimizing the data collapse entails minimizing @xmath54 with respect to @xmath5 , @xmath6 , and the @xmath55 parameters with a two tiered fit .
first , for a particular choice of @xmath6 and @xmath5 , the parameters @xmath55 are calculated in a linear least squares fit .
the appropriate values of the critical indices @xmath5 and @xmath6 are then determined by a stochastically driven least squares fit , where a sequence of attempts are made to introduce small random perturbations to @xmath5 and @xmath6 , with the shifts accepted only if the sum of squares term @xmath54 is thereby lowered .
[ fig : fig4 ] shows @xmath54 with respect to @xmath6 and @xmath5 .
the global minimum for @xmath56 and @xmath57 is , within the bounds of monte carlo statistical error , consistent with the calculation of @xmath5 based on crossings of the @xmath25 quantity .
moreover , the calculated value of @xmath6 is in accord with results gleaned from high precision monte carlo calculations for discrete 3d systems ( e.g. 0.875[1 ] for discrete percolation models on 3d lattices with cubic symmetry @xcite ) .
( color online ) @xmath54 plot shown with respect to the correlation length critical exponent @xmath6 and the excluded volume @xmath5 with global minimum at @xmath57 and @xmath58 . ]
we have used a large - scale monte carlo calculation based on a dynamical exploration of void spaces to determine the critical properties ( i.e. finding @xmath0 and @xmath1 ) for the percolation through voids surrounding randomly placed impermeable spheres .
two complementary finite size scaling analyses yield the same percolation threshold @xmath5 . our dynamically based calculation is distinct in that it makes no approximation for the structure of void spaces , while also incorporating finite size effects which are exploited to calculate critical indices .
the result for the correlation length critical exponent @xmath6 is compatible with the critical behavior of discrete 3d counterparts .
to sample realizations of disorder from the appropriate statistical distribution , it is important to generate random configurations of spherical inclusions in an unbiased way ; since we operate in the grand canonical ensemble , the number @xmath59 of spheres in the simulation volume must in general vary from one sample to the next due to statistical fluctuations . the integer value closest to the mean occupancy @xmath60 is a convenient initial choice , and random variations in the number of spheres in the system are taken into consideration with a sequence of stochastically driven attempts either to raise or lower @xmath59 .
the latter are part of an importance sampling scheme similar to that used to derive the metropolis criterion @xcite at the heart of monte carlo simulations that sample the boltzmann distribution in the calculation of equilibrium thermodynamic variables . to determine the probability of having @xmath59 spheres in the simulation volume @xmath61
, we divide @xmath62 into @xmath63 subvolumes of equal size where @xmath64 . for large @xmath63
the likelihood of multiple occupancy in any of the subvolumes is very small relative to the chance of having one or zero sites in a subdivision ; in the small @xmath65 limit , the single occupancy probability is @xmath66 , with @xmath67 being the complementary likelihood of null occupancy .
hence , the probability that the entire system is devoid of hopping sites is @xmath68 , which becomes @xmath69 for @xmath70 . for single occupancy , adopting a prefactor @xmath63 to take into consideration that the site may reside in any of the @xmath63 subvolumes , yields @xmath71 , which becomes @xmath72 in the @xmath73 limit .
similar logic gives @xmath74 for the general case of exactly @xmath59 spheres in the simulation volume , where @xmath75 compensates for multiple occupancy . to generate a disorder realization , a succession of attempts ( a number of moves in the vicinity of @xmath76 is sufficient to achieve ergodicity ) is made to raise or lower the occupancy number @xmath59 , where the choice to increase or decrease @xmath59 is randomly determined .
for increments from @xmath59 to @xmath77 , the change is accepted if @xmath78 , where @xmath79 is a random number sampled uniformly from the interval [ 0,1 ] .
similarly , decreasing @xmath59 to @xmath80 occurs if @xmath81 .
finally with @xmath59 properly sampled , three cartesian coordinates for each sphere are chosen independently ( and at random with uniform probability density ) from the interval @xmath82 $ ] . | the percolation threshold for flow or conduction through voids surrounding randomly placed spheres is rigorously calculated . with large scale monte carlo simulations
, we give a rigorous continuum treatment to the geometry of the impenetrable spheres and the spaces between them . to properly exploit finite size scaling , we examine multiple systems of differing sizes , with suitable averaging over disorder , and extrapolate to the thermodynamic limit .
an order parameter based on the statistical sampling of stochastically driven dynamical excursions and amenable to finite size scaling analysis is defined , calculated for various system sizes , and used to determine the critical volume fraction @xmath0 and the correlation length exponent @xmath1 . |
the central challenge in finding or engineering thermoelectric materials suitable for thermal energy conversion beyond niche applications remains to combine good electrical properties and a very low thermal conductivity in a single material .
very recently , the compound snse has been reported to exhibit exceptionally good thermoelectric properties at high temperatures above @xmath2800 k , including a very low thermal conductivity @xcite .
nanostructuring of materials has been pursued extensively in recent years as a possible route towards achieving a sufficiently low thermal conductivity @xcite , but the discovery of extremely low thermal conductivity in bulk snse has markedly improved the prospects of obtaining bulk thermoelectrics suitable for general commercial applications . as snse consists only of earth - abundant elements of comparatively low toxicity , it represents also an important advance in avoiding toxic and rare elements like lead and tellurium in thermoelectric materials . at ambient conditions
, snse adopts a layered orthorhombic crystal structure ( ges structure type ; space group @xmath3 , no .
62 ) as illustrated in fig .
[ fig : crystal_structure ] . the structure can be regarded as a heavily - distorted variant of the nacl structure type , and it comprises nacl - type double layers that are stacked along the @xmath4 direction .
one signature of the distortion are the unequal se sn interatomic distances a1 and a2 , which gives rise to corrugated layers as seen in projection along the @xmath5 direction , see fig .
[ fig : crystal_structure ] .
another is that adjacent nacl - type double layers are displaced by approximately @xmath6 with respect to each other . with increasing temperature ,
the structure was observed to evolve continuously into a higher - symmetry variant ( tli structure type ; space group @xmath7 , no .
63 ) , with a critical temperature of 807 k @xcite .
it is this high - temperature , higher - symmetry phase of snse that has the excellent thermoelectric properties .
unfortunately , snse appears to be thermally rather unstable under these conditions due to sublimation @xcite , which severely limits the use of this material in devices . here
, we investigate the effect of high hydrostatic pressure on the crystal structure of snse .
our study was motivated ( i ) by recent reports that some thermoelectric materials exhibit dramatic improvements in the thermoelectric power factor under compression @xcite , ( ii ) by the fact that earlier studies ( as detailed below ) have yielded only a rather fragmentary and partially contradictory picture of the behaviour of snse under high pressure and ( iii ) by the scarcity of state - of - the - art experimental crystal structure investigations on group - iv chalcogenides in relation to the abundance of computational studies .
using synchrotron powder x - ray diffraction experiments , we quantify the structural changes in snse under pressure and demonstrate that there is a pressure - induced transition to the higher - symmetry ges structure type , and we comment on the possible implications of the structural changes for the thermoelectric properties of snse .
-snse at 0 gpa ( top ) and @xmath0-snse at 27 gpa ( bottom ) .
the light rectangles indicate the unit cell.,title="fig : " ] +
the properties of snse under high pressure have been the subject of a number of experimental and computational studies , largely motivated by snse and related orthorhombic iv - vi compounds being intermediate between two- and three - dimensional solids in terms of their anisotropic physical properties .
@xcite investigated the crystal structure of snse under high pressure to 34 gpa by energy - dispersive x - ray diffraction .
judging by the absence of discontinuities in the interplanar distances , they found no evidence for a structural phase transition . however , they noted the limitations of the energy dispersive diffraction technique , suggesting that a continuous and relatively subtle transition such as the @xmath8 transition known from high - temperature experiments might be difficult to detect with this technique . in a high - pressure mssbauer study @xcite , peters and mcneil identified the low - pressure range below 3.0 gpa as the range in which the interlayer bonds are changed , transforming snse from a predominantly two - dimensional to a more three - dimensional material .
parenteau and carlone @xcite measured the temperature and pressure dependence of the direct and indirect optical gaps of snse up to 4.0 gpa .
they found its electronic properties to be three - dimensional in nature , already at ambient pressure , as opposed to the two - dimensional nature suggested by the easy cleavage of snse crystals , their lattice dynamics @xcite and the interpretation of the chemical bonding in this compound @xcite .
et al_@xcite observed abrupt changes in the electrical resistivity and thermopower near 6 gpa , but their exact origin has remained unclear .
interestingly , the resistivity decreases monotonically with increasing pressure , whereas the thermopower decreases up to 6.3 gpa , followed by large enhancement between 6.5 and 7.6 gpa .
two recent computational studies also investigated the structural and electronic properties of snse under pressure .
alptekin @xcite predicted a pressure - induced second - order transition to a phase with space group @xmath7 , isostructural with the high - temperature phase .
two possible phase transition pressures of 2 or 7 gpa were identified , depending on the method of calculation . in this study , little evidence of pressure - induced electronic changes
was found ; in particular , snse was reported to remain semiconducting in the @xmath7 phase at 7 gpa .
in contrast , makinistian and albanesi @xcite reported a reduction of the band gap with increasing pressure and a transition to a semimetallic state with indirect band overlap .
as the calculations underestimate the zero - pressure band gap by @xmath250% in comparison with experimental results , the calculated metallisation pressure of @xmath25 gpa should be regarded as a lower bound .
polycrystalline snse was synthesised using a standard solid state chemistry reaction .
stoichiometric amounts of 100-mesh sn powder ( 99.9% stated purity ) and shots of se ( 99.999% ) were ground together and cold - pressed into a pellet . the pellet was sealed in an evacuated quartz tube and heated at 900 c for 24 hours ( 10 c / min heating ramp ) , followed by cooling at 3 c / min to 600 c , after which the oven was switched off .
this heating step was repeated once with an intermediate homogenisation .
phase purity was checked using a bruker d8 advance diffractometer with monochromated cu - k@xmath9 radiation .
post - synthesis hot - pressing using a home - built apparatus ( 500c and 80 mpa for 20 minutes ) was used to obtain a fully dense snse sample .
the electrical resistivity and seebeck coefficient were measured using a linseis lsr3 instrument and found to be 5 @xmath10 cm and @xmath11 vk@xmath12 at 300 k. the seebeck coefficient agrees with recently reported values for spark - plasma - sintered polycrystalline snse @xcite , whereas our resistivity is a factor of 10 larger .
the synthesised material was manually ground to a fine powder and loaded into a merrill - bassett - type diamond anvil cell ( dac ) for high - pressure generation .
condensed helium was used as the pressure - transmitting medium , and pressures were determined with the ruby fluorescence method using the calibration of ref .
monochromatic powder x - ray diffraction patterns were measured in the debye - scherrer geometry on beamline id09a of the european synchrotron radiation facility ( esrf ) .
monochromatic x - rays of wavelength @xmath13 were focused to a spot size of @xmath220 m at the sample , and the two - dimensional diffraction images were recorded with a mar555 area detector .
the dac was rotated by @xmath14 during the exposure to improve the powder averaging .
the diffraction images were integrated azimuthally with the fit2d software @xcite to yield conventional intensity vs @xmath15 diffraction diagrams , which were analysed with the rietveld method using the program gsas @xcite .
figure [ fig : xray_patterns ] illustrated the evolution of the diffraction pattern of snse for increasing pressures up to 27 gpa .
no abrupt changes were observed , but there is a gradual disappearance of several reflections as shown in the inset of fig .
[ fig : xray_patterns ] for the ( 011 ) and ( 201 ) reflections .
this suggest a transition from the structure with space group @xmath3 to a higher - symmetry variant , and this transformation is completed at 11.0 gpa .
+ the powder diffraction patterns were analysed with the rietveld method .
starting at low pressure , the crystal structure was first analysed based on the ges - type structure model with space group @xmath3 .
the refined parameters were the lattice parameters , the @xmath16 and @xmath17 atomic coordinates of se and sn , profile parameters of the stephens peak function @xcite and the coefficients of a chebyshev polynomial for the background .
as the diffraction images showed textured powder rings , a correction for preferred orientation of the powder particles , based on spherical harmonics @xcite , was applied .
even a conservative two - parameter correction for preferred orientation improved the agreement between measured and fitted intensities noticeably .
all rietveld refinements were performed with spherical - harmonics corrections of orders 2 , 4 , and 6 ( 2 , 5 , 9 refinable parameters , respectively ) .
the resulting atomic coordinates from the three fits generally agreed within the estimated uncertainties , and the results reported here are the averages of the three refinements for each diffraction pattern .
figure [ fig : rietveld](a ) shows the rietveld fit for @xmath18-snse at 4.2 gpa as an example .
a spherical - harmonics correction of order 2 was used in this case , and the weighted residual of the fit without background was @xmath19 .
-snse at 4.2 gpa and ( b ) the high - pressure phase @xmath0-snse at 20.0 gpa .
the tickmarks indicate the calculated peak positions.,title="fig : " ] + figure [ fig : atom_coords ] shows the variation of the atomic coordinates with pressure as obtained from the rietveld refinements .
most notable is the clear evolution of the sn @xmath17-coordinate towards zero , a process which is again completed at 11 gpa .
attempts to refine the structures for pressure above 11 gpa based on the @xmath3 structure model led to rather unstable fits , but the results for the sn @xmath17-coordinate were consistent with @xmath20 .
the results for the pressure dependence of the se @xmath17-coordinate are less conclusive because it is close to the high - symmetry value of @xmath21 already at ambient pressure , and the lower scattering power of se compared to sn leads to larger uncertainties in the measured parameters for the former . the data in fig .
[ fig : atom_coords](b ) suggest that the se @xmath17-coordinate approaches 1/2 at the critical pressure , possibly after passing through a minimum near 5 gpa .
in addition , the absence of any traces of the ( 011 ) and ( 201 ) reflections at pressures above 11 gpa indicates that the @xmath17-coordinates of both sn and se adopt the high - symmetry values of 0 and @xmath21 , respectively .
this changes the space group of the crystal structure from @xmath3 ( # 62 ) to @xmath22 ( # 63 ) .
the latter is a non - standard setting of space group @xmath7 that keeps the crystal axes in the same orientation as for @xmath3 .
this pressure - induced change in symmetry is the same as that observed upon heating at ambient pressure . for reasons explained below along with a discussion of the relation between the temperature- and pressure - induced structural changes
, we denote the high - pressure phase as @xmath0-snse . for the data collected at pressures of 11 gpa and higher , rietveld refinements of the crystal structure
were performed with space group @xmath22 .
figure [ fig : rietveld](b ) shows the rietveld fit for @xmath0-snse at 20.0 gpa as an example .
as before , a preferred - orientation correction of order 2 was applied , and the weighted residual of the fit without background was @xmath23 .
the atomic coordinates obtained from the @xmath22-based refinements are shown with diamond - shaped symbols in fig .
[ fig : atom_coords ] . within the experimental uncertainty , the @xmath16-coordinates of se and sn change continuously over the whole pressure range and thus across the transition near 11 gpa .
the pressure dependence of the sn @xmath17-coordinate below the transition pressure can be described well with a power law , @xmath24 with @xmath25 , a critical pressure @xmath26 gpa and a critical exponent @xmath27 .
and @xmath22 structures , respectively .
dashed lines are guides to the eye ; the solid line for the sn @xmath17-coordinate represents a power - law fit as described in the text .
the grey vertical line marks the critical pressure of 10.5 gpa.,title="fig : " ] + + figures [ fig : lattice_parameters](a ) and ( b ) show the lattice parameters as a function of pressure . in fig .
[ fig : lattice_parameters](a ) , the @xmath4 parameter has been scaled by @xmath28 , as the orthorhombic lattice parameters would become @xmath29 in a hypothetical nacl - type structure @xcite . at low pressures ,
snse is most compressible along its @xmath30-axis , but this direction becomes significantly less compressible at around the critical pressure .
in contrast , the @xmath5-direction exhibits initially a relatively low compressibility , and it becomes more compressible just above the critical pressure .
the difference between the @xmath5 and @xmath30 parameter diminishes under pressure up to the critical pressure and then increases again for higher pressures .
the axial ratios shown in fig .
[ fig : lattice_parameters](c ) provide an alternative illustration of these observations .
all three axial ratios have extrema near the critical pressure of 10.5 gpa .
the lattice of snse approaches a tetragonal metric at the critical pressure , but remains orthorhombic at all pressures .
the pressure dependences of three characteristic interatomic distances within the double layers are shown in figure [ fig : distances](a ) . with increasing pressure , the se sn bond a1 and contact a2 , which are oriented approximately parallel to the layers ( fig .
[ fig : crystal_structure ] ) , equalise in length and become equivalent bonds in the @xmath0 phase above 10.5 gpa .
interestingly , the a1 bond lengthens with increasing pressure up to the transition , which can be attributed to the increasing influence of the a2 contact . also noteworthy
is the smooth transition from the average a1/a2 distance , @xmath31 , in the low - pressure phase to bond length a in the high pressure phase . the short se
sn bond b , perpendicular to the layers , remains the shortest bond at all pressures , and decreases in length at approximately the same rate as bond a above the transition pressure .
the symmetrisation of the structure manifests itself also in the variation of the spacings of the atomic sheets within and between the double layers , @xmath32 and @xmath33 , as shown in figure [ fig : distances](b ) .
these two layer spacings become indistinguishable above 10 gpa .
the exact transition pressure can not be determined from these data because of the relatively large scatter of the data points around 10 gpa , which can be traced to the extracted @xmath16 atomic coordinates ( see fig . [
fig : atom_coords ] ) , and we consider these fluctuations to be an artefact .
overall , the symmetrisation of the crystal structure is predominantly driven by the changes in atomic positions , with only secondary contributions from the variations of the lattice parameters with pressure . .
solid lines are guides to the eye.,title="fig : " ] + when referring to previous work on snse and related systems , one should note that different settings have been used for the orthorhombic space groups , in particular the combination of space groups @xmath34 and @xmath7 for the structures at ambient conditions and high temperatures , respectively . the transformation from @xmath3 to @xmath34 ( and also from @xmath22 to @xmath7 ) changes the axes as @xmath35 = c , a , b\ , [ pnma]$ ] .
we will continue to use the setting of @xmath36 and convert the results of previous studies to this setting where needed .
the pressure - induced structural changes in snse are similar to , and yet different from those at high temperature and normal pressure @xcite .
with increasing temperature at zero pressure , the se and sn atoms move continuously to a higher - symmetry ( @xmath22 ) position and the difference between the @xmath5 and @xmath30 lattice parameters decreases just as observed here as a function of pressure .
however , above the critical temperature of 807 k , the @xmath5 and @xmath30 parameters become indistinguishable and the lattice pseudo - tetragonal , whereas the high - pressure form described here remains orthorhombic at all pressures .
even though the high - temperature phase , @xmath1-snse , and the high - pressure form are equivalent in terms of their crystallographic symmetry , they differ in terms of their lattice metric and may thus be considered distinct phases .
therefore , we denote the ambient - temperature , high - pressure modification as .
+ the variation of the unit cell volume of snse with pressure is shown in fig .
[ fig : eos ] .
it does not exhibit any discontinuities or other anomalies .
initial attempts to describe the pressure - volume data with the birch or vinet equations of state yielded unsatisfactory agreement between the fitted curves and the measured data , and the thus determined bulk modulus and its derivative at zero pressure depended significantly on the range of pressures that was used in the least - squares fit .
the cause of these problems is the rather large change in compressibility at low pressures .
therefore , we employed the less commonly used davis - gordon equation of state @xcite , which was first applied to the pressure
volume relation of mercury .
it is a simple quadratic expansion of pressure @xmath37 in terms of the relative change in density , @xmath39 .
when expressed in terms of volume @xmath38 , @xmath40 where @xmath41 , and @xmath42 and @xmath43 are the bulk modulus and its pressure derivative at zero pressure , respectively .
equation ( [ eq : dg_p(v ) ] ) can be rearranged to yield @xmath44 the davis - gordon equation of state describes the pressure volume data for snse better than the vinet and birch relations , and a fit with the @xmath45 form gave @xmath46 gpa and @xmath47 with @xmath48 fixed at the measured zero - pressure volume , @xmath49 @xmath50 .
we find snse to be significantly more compressible than reported by chattopadhyay _
et al_@xcite , who found @xmath51 gpa ( and @xmath52 ) .
this discrepancy persists even if we use , like chattopadhyay _ et al _ , the birch equation of state to describe our data , which gives @xmath53 gpa ( and @xmath54 ) . on the other hand ,
our measured bulk modulus is in fair agreement with alptekin s as well as makinistian and albanesi s computational results for @xcite .
the orthorhombic ges and tli structure types are well - known as possible intermediate structures in reconstructive transformations between the nacl and the cscl structure types in group - iv chalcogenides @xcite , but the understanding of the systematics of the transitions in this group is still fragmentary .
a systematic review of the vast literature on the high - pressure behaviour of group - iv chalcogenides is beyond the scope of this work , but we would like to highlight a few interesting cases . a pressure - induced transition from ges- to tli - type phases
is a recurring theme in recent computational work on ges @xcite , gese @xcite , sns @xcite , and snse itself @xcite .
x - ray diffraction studies on ges @xcite and gese @xcite did not yield evidence of such a transition , but the quality of the experimental data would probably not have permitted detection of such a relatively subtle transition . in x - ray diffraction experiments on sns , ehm observed an abrupt phase transition between 16 and 18 gpa @xcite . a monoclinic lattice and space group @xmath55
were tentatively assigned to the high - pressure phase , and on this basis a volume jump of @xmath56 was inferred to occur at the transition . even though the quality of the x - ray diffraction data was insufficient to fully determine the crystal structure of the high - pressure phase , these findings are incompatible with the computational results @xcite .
it is worth noting that in the computational investigations the crystal symmetry often appears to have been restricted to space group @xmath3 ( which includes the higher - symmetry @xmath7 as a special case ) so that no conclusions can be drawn with regards to transitions to other structure types beyond ges and tli .
examples of such alternative structure types are the orthorhombic structures reported for pbte @xcite and snte @xcite at pressures of 616 gpa and 418 gpa , respectively .
we have determined the structural evolution of snse under hydrostatic pressure up to 27 gpa and observed a continuous transition from the ges to the tli structure type with a critical pressure of 10.5(3 ) gpa .
there are no indications of any discontinuities in the structural parameters as a function of pressure .
the high - pressure phase @xmath0-snse is closely related to the high - temperature phase at ambient pressure , @xmath1-snse , but whereas the former remains orthorhombic at all pressures , the latter has a tetragonal metric of the crystal lattice .
a transition of this kind , with an increase of the crystal symmetry from space group @xmath3 to @xmath7 , has been a recurring outcome of computational studies on group - iv chalcogenides . to the best of our knowledge , the present study provides the first experimental evidence for such a transition .
the similarity between the high - pressure and the high - temperature phases leads us to expect that the stability field of the high - temperature , higher - symmetry modification will extend to lower temperatures as pressure is applied .
this suggests that under the application of pressure or compressive strain , snse may exhibit very good thermoelectric properties at temperatures lower than the @xmath2800 k that are required for bulk snse at normal pressure .
it appears therefore worthwhile to explore whether using strained films of snse may permit to lower the operating temperature of snse - based devices while maintaining excellent thermoelectric performance , which would open a route to overcoming the problem of sublimation and hence limited thermal stability of snse at high temperatures of @xmath2800 k.
we thank m. hanfland ( esrf ) and r. howie ( edinburgh ) for their assistance .
facilities were made available by the european synchrotron radiation facility , and part of this work was supported by the leverhulme trust ( rpg-2012 - 576 ) . | the crystal structure of the thermoelectric material tin selenide has been investigated with angle - dispersive synchrotron x - ray powder diffraction under hydrostatic pressure up to 27 gpa . with increasing pressure ,
a continuous evolution of the crystal structure from the ges type to the higher - symmetry tli type was observed , with a critical pressure of 10.5(3 ) gpa . the orthorhombic high - pressure modification , @xmath0-snse , is closely related to the pseudo - tetragonal high - temperature modification at ambient pressure .
the similarity between the changes of the crystal structure at elevated temperatures and at high pressures suggests the possibility that strained thin films of snse may provide a route to overcoming the problem of the limited thermal stability of @xmath1-snse at high temperatures . |
the schwinger effect , involving pair creation by a classical electromagnetic field out of the vacuum , is a well - known representative example of quantum field theory in an external classical background @xcite . the usual treatment , in the case of a spatially uniform and constant ( in time ) electric field , proceeds by evaluating the gauge - invariant effective action which is directly related with the transition amplitude between asymptotically defined _ in - out _ vacuum states @xcite .
the real part of the effective lagrangian describes vacuum polarization by the electric background , and can be added to the classical electromagnetic lagrangian to compute the modified equations of motion for the background incorporating back - reaction at the semiclassical level .
the imaginary part encodes information about particle creation , or more precisely , provides a direct measure of the asymptotic particle content in the _ in _ vacuum with respect to the _ out _ vacuum state . an alternative route to deriving this result is based on a particular choice for the gauge in which the vector potential is time - dependent .
the field modes can be treated as a bunch of quantized time - dependent oscillators , and sensible vacuum states can be defined in the asymptotic regions @xcite .
these states are however in - equivalent due to the non - trivial time dependence of the vector potential though the electric field is constant with the consequence that a positive frequency mode defined at early times evolves into a combination of positive and negative frequency modes with respect to late times .
this allows one to compute the bogolyubov coefficients for the mixing , and hence the mean number of particles in any given mode as seen by an observer at late times .
both the above approaches however involve only the asymptotically defined vacuum states , and do not really provide a direct picture of the time evolution of the quantized field .
moreover , this approach to account for back - reaction is conceptually somewhat unsatisfactory from the point of view of causality , since one is , in effect , getting an effective lagrangian which depends implicitly on the _ out _ vacuum state .
further , one might have thought that physically realistic electromagnetic fields would exist for finite duration and it makes more sense to have a situation where the field is kept switched on for a finite period of time .
in fact a constant , eternally existing electric field would be a rather unphysical notion .
this is because if the field is causing particles to be created continuously , the produced pairs would ` short ' the externally applied field , and cause it to progressively decay over time .
an eternal electric field would thus be unstable and can not really be maintained when particle creation is taken into account .
but if one considers the case of a field which exists for only a finite duration , it will naturally be expected to break the time translation symmetry of the eternal field case , introducing a finite window in time for quantum effects to come into play and leading to a time - dependent picture of the particle creation process .
one can make an analogy between these two cases and those of ( i ) an eternal black hole and ( ii ) a black hole formed from gravitational collapse . in the case of ( i ) , for a quantized field in the hartle - hawking state @xcite , an observer at spatial infinity would not see a flux of particles because of the time translation symmetry of the gravitational background . when one considers a collapsing scenario , however , one is introducing a time dependence into the gravitational field , and in such a background , there would be an outgoing flux of particles ( in the unruh vacuum ) , carrying away energy which would cause the black hole to evaporate over time . in @xcite we adopted a physically reasonable , time - dependent approach to analyzing field modes in the schrodinger picture , and used an ` instantaneous ' definition of particle .
the schwinger effect is amenable to this treatment , as we analyzed in considerable detail for an eternal electric field .
but such a scenario is time translation - invariant , and we would like to adopt the same approach and repeat the analysis for a spatially uniform but time - dependent electric field . here we consider the case of a field which exists for a _
finite _ duration of time @xmath0 , tracking the behavior of the time - dependent particle number and interpreting it appropriately .
for an electric field of magnitude @xmath1 , the relevant time scale ( in natural units ) for quantum effects is provided by @xmath2 . on physical grounds , the results for the late - time particle content etc . for the eternal electric field scenario
would be expected to be recovered when @xmath0 is made very large , i.e. @xmath3 , and we would like to verify that this is indeed the case .
we will also attempt to connect up the instantaneous particle content with the power spectrum of the quantum field and indicate the distinction between the two .
for contrast , we also discuss the other extreme of an electric field sharply localized in time modeled by a dirac delta function .
the question of having a notion of classicality of the evolving quantum state , which was extensively dealt with in @xcite , will also be touched upon , because it has potential implications for the study of back - reaction at the semiclassical level and for the validity of a semiclassical analysis .
this paper is structured as follows . in section
[ sec : formalism ] we sketch an outline of the schrodinger picture formalism which was elaborated on in @xcite , to analyze a quantized field , treated as a bunch of oscillators , in the presence of a time - dependent vector potential . in section
[ sec : analysis ] the analysis of time evolution of any single fourier mode of the quantized field in the presence of a background electric field is carried out for the two specific cases mentioned above : a field which exists for a finite period of time ( and is maintained constant over that period ) , and a spiked @xmath4-function electric field .
finally , we conclude with a recapitulation in section [ sec : conc ] . in what follows
, we shall set @xmath5 .
we start with a complex scalar field @xmath6 which is quantized in a spatially uniform electric field background .
the electric field @xmath7 is assumed to be produced by a ( uniform ) current density @xmath8 , and can be described in a time - dependent gauge corresponding to @xmath9 .
we are also assuming that the magnetic field is zero everywhere , and so is the electric charge .
then , from maxwell s equations , it follows that @xmath10 .
the action for the scalar field in this background is expressible as @xcite [ ] = ( _ + iqa _ ) ( ^ - iqa^)^ * + = dt ( q _ * k*^2 - ( k_^2 + m^2 + ( k_z - q a(t))^2 ) q_*k*^2 ) [ ef_actn ] with @xmath11 and @xmath12 . the field theory thus reduces to quantum mechanics of a bunch of uncoupled oscillators each with unit mass and a time - dependent frequency given by _ *
k*(t)=. ( since the scalar field is complex - valued , each mode @xmath13 is associated with two degrees of freedom , corresponding to the real and imaginary parts of @xmath14 . )
having cast the problem in a time - dependent form , we now briefly outline the schrodinger picture formalism ( based on @xcite ) which will be adopted here .
the starting point is the wave function , chosen to be a _ form - invariant _ gaussian state ( motivated by the requirement of having a state that can be interpreted as the ground state of the oscillator in a time - independent scenario ) parameterized by the complex variable @xmath15 : ( q_*k*,t ) = n_*k*(t ) [ - ( ) q_*k*^2 ] .
[ gwfn ] the time dependent schrodinger equation reduces to a first order ( but nonlinear ) differential equation for @xmath15 : z_*k * + 2 i _ * k * z_*k * + ( z_*k*^2 - 1 ) = 0 .
[ eqz ] the problem of determining the quantum evolution thus boils down to directly solving for @xmath16 ( with an appropriate choice of initial condition ) .
the dimensionless adiabaticity parameter @xmath17 , provides the relevant measure to characterize the nature of the time evolution ; @xmath18 corresponds to the adiabatic limit , in which region a reasonable definition of vacuum state is possible . in @xcite
, we introduced a time - dependent notion of particle .
this is done by comparing the quantum state moment by moment with the complete set of _ instantaneous _ eigenstates of the oscillator defined at every instant .
this provides a number distribution of ` quanta ' , and can be used to compute the mean , which provides a rather general notion of particle content .
it is given by n_*k * = ; _ * k * = ( n_*k * + ) _ * k*(t ) and is thus directly related to the expectation value of the time - dependent hamiltonian . once @xmath15 is specified , the evolution of this particle number can be determined .
it may be noted that the mean particle number as well as the higher moments of the number probability distribution are functions of only the magnitude of @xmath16 , and do not encode information about its phase .
the phase can be fixed by considering the variance of @xmath19 , generally called the power spectrum , which is the fourier transform of the two - point correlation function of the field in coordinate space .
the expression for the power in the @xmath20th mode has the form q_*k*^2 = + ( z_*k * ) .
[ ps_ef ] which makes it clear that the power spectrum is not , in general , _ completely _ specifiable in terms of the mean number of created particles .
this point will be taken up later and a comparison with the case of an inflationary cosmological model made .
returning to the quantum state in , we can discuss another question , that of identifying a suitable criterion for classicality of the state .
the wigner function @xcite provides one reasonable approach . in @xcite
, we looked at two features as possible indicators of classical - like behavior : peaking of the wigner distribution in phase space and correlation ( in the sense of separability of the wigner function in dependence on @xmath21 and @xmath22 ) .
the general conclusion drawn was that the former is not an unambiguous indicator of classical behavior ; although indicative of a correspondence between position and momentum , it can happen even in a near - vacuum state .
the classicality parameter defined as @xmath23 , on the other hand , is directly related to the mean of @xmath24 and shows a close correspondence with particle production , being zero for an instantaneous ground state but saturating at unity when the mean particle number diverges .
we will consider its variation too in the analysis which follows .
we now move on to describing the setting in which the analysis has been carried out , that of a spatially uniform but time - dependent electric field .
we consider a situation in which the electric field is kept switched on for a finite period of time @xmath0 , from @xmath25 to @xmath26 .
the time - dependent vector potential is chosen to have the following form : a(t ) & = & et/2 ( t < t/2 ) + & = & -et ( t t/2 ) .
[ sc_ef ] such a field would be produced by a spatially uniform current density @xmath27 which remains zero at all times except at @xmath28 , being made up of two @xmath4-function spikes separated by a time interval @xmath0 . making the transformation to dimensionless parameters @xmath29 , @xmath30 and @xmath31
, the frequency function corresponding to can be written as _ * k*(x ) & = & ( |x|x_0 t/2 ) + & = & _ * k * ^ ( x < - x_0 , x > x_0 ) .
the oscillator equation for the variable @xmath32 can be solved analytically for this choice of frequency in terms of the parabolic cylinder function @xcite ; assuming that the field is in the vacuum state initially ( for @xmath33 ) , the solution for @xmath16 ( which is continuous across the interval where @xmath34 ) is given by z_*k*(x ) = for @xmath35 , where f_*k*(x ) = d_-1/2 - i_*k*/2[-(1+i)(_*k * + x ) ] + _ * k * ( d_-1/2 - i_*k*/2 [ -(1+i)(_*k * + x)])^ * and _ * k * = . for @xmath36 ,
the solution is z_*k*(x ) = ( ) e^-2 i _ * k*^+ ( x - x_0 ) .
[ n_scef ] from the above expression , it is immediately obvious that @xmath37 for @xmath38 , indicating a non - zero particle content in the quantum state which remains constant once the electric field has been switched off , since all the time dependence is in the phase of @xmath16 .
we are , however , particularly interested in the _ time dependence _ of the particle creation process .
( although the electric field is constant within the time interval @xmath0 , the particle content is expected to exhibit time dependence since we have cast the problem in a time - dependent form by a particular gauge choice . )
tracking the variation of the mean number @xmath39 can provide an idea about this .
[ n - vs - x ] shows the variation of @xmath39 as a function of the variable @xmath40 for the cases @xmath41 3 , 5 and 8 .
the red portion corresponds to the interval over which the electric field is kept switched on . as expected
, the particle number rises from zero to a finite value at @xmath42 and remains constant thereafter , once the field is turned off .
this rise is however not monotonic , and is characterized by oscillations during the phase when the field is present .
this is of course at odds , to some extent , with one s intuitive notion of a ` particle ' , once produced , remaining produced ; an appropriate interpretation of @xmath39 in the intermediate phase is closer to a measure of quantum fluctuations of the scalar field rather than a measure of physical particles .
if the time scale of the electric field is increased , it is evident from a comparison of the plots that these oscillations get progressively ( and rather quickly ) suppressed , and the particle number very soon reaches a saturation .
in fact , the particle creation is seen to really happen only over a rather short interval . for sufficiently large values of @xmath43 ,
the value of @xmath39 at @xmath42 is more or less independent of @xmath43 ( and is equal to the asymptotic value for an eternal electric field , usually determined from the effective action or using the method of bogolyubov coefficients @xcite ) .
this can be gleaned from the plot in fig .
[ n - vs - x0 ] of @xmath44 as a function of @xmath43 . to restate the obvious ,
an electric field which is kept turned on long enough gives essentially the same result for the late - time particle content as the standard constant field background .
[ n_x0 ] at @xmath42 as a function of @xmath43 for @xmath45 and @xmath46 ( blue ) and 0.2 ( red ) .
@xmath43 is proportional to the interval of time over which the electric field is present .
this confirms what the earlier plots suggested .
the saturation beyond a point indicates that in the picture adopted here , for large enough @xmath43 , the final particle content becomes independent of how long the field is kept switched on.,title="fig:",width=377 ] as for the classicality parameter , shown in fig .
[ s - vs - x ] , it starts from zero corresponding to the initial ground state , increases over a short interval of time and then settles down into a steady oscillatory pattern of constant amplitude .
one can interpret this as a growth in the phase space correlation over time starting from an initially uncorrelated vacuum state , although @xmath47 ( not its magnitude ) is oscillatorily zero at late times . from the previous plots
, it can be seen that for large enough @xmath48 , the behavior for @xmath49 closely resembles that in the time - independent case .
we shall now take the @xmath50 limit , and look at the behavior for @xmath51 . the power spectrum in this limit
@xcite is given by the expression q_*k*^2 ( ) + ( ) ( ) which can be written as q_*k*^2 (e^-_*k * + + ( x^2 + _ * k*x + r_*k * ) ) [ psef ] where _ * k * _ * k * ( ir_*k * ) = - i . if the third term in the brackets in , which is rapidly oscillating , is dropped , then the variance at the leading order is just q_*k*^2 ( n_*k*+ ) and the power in the large time limit can be related to the asymptotic mean particle number .
this connection is not quite precise though , since it requires our ignoring the additional oscillatory term which amounts to ignoring the phase information contained in @xmath16 .
we can go a bit further and try computing the spatial two - point correlation function . in the large @xmath40 limit ,
if the oscillatory term in is ignored , the fourier transform of the first term can be found , and one has ( * r*)(0 ) ( * r * ) + e^- - _ ^2 ( z ) where @xmath52 .
the second term in the above expression suggests that on any 2d plane normal to the direction of the applied electric field , the correlation between any two points falls off for distances beyond @xmath53 ; the spatial two - point function , in the late time limit , thus picks out the characteristic length scale for correlations .
this behavior of the power spectrum may be compared with the corresponding evolution of a massless real scalar field in an expanding frw cosmological background @xcite ; the general expression for the power happens to be @xcite : q_k^2 = + ( z_k ) @xmath54 being the scale factor and @xmath55 . at super hubble scales ,
any mode evolves highly non - adiabatically , so @xmath56 and @xmath57 , giving @xmath58 , and the power in any given fourier mode provides a direct measure of its particle content . in particular , for the de sitter model , q_k^2 = ( 1 + ) . at late times ( @xmath59 ) , @xmath60 becomes _ constant _ and proportional to @xmath61 , which corresponds to a scale - invariant spectrum .
the scale invariance , moreover , implies that the spatial two - point function here does not pick out a well - defined correlation length scale , and this is in contrast with the electric field case . in the foregoing analysis we considered an electric field of finite strength
which is kept switched on for a finite duration , before shifting attention to the limiting case of large timescales ( keeping the magnitude of the field fixed ) .
it would be of interest to also look at the other extreme , that of a field sharply peaked in time .
let us consider the case of a spatially constant electric field modeled by a dirac delta function in time : @xmath62 .
this can be described by the vector potential @xmath63 , where @xmath64 denotes the heaviside theta function .
the potential and hence the frequency @xmath65 have a discontinuity at @xmath66 , and it is this step which will create particles . in order to compute the particle creation at the jump , we will work with the mode variable @xmath67 , which satisfies the standard oscillator equation : + _ * k*^2 ( t ) f_*k * = 0 .
@xmath67 is related to the variable @xmath16 by the relation z_*k * = . in this case
, the oscillator frequency makes a sharp transition from @xmath68 for @xmath69 to @xmath70 for @xmath71 .
if it is assumed that the mode starts off in the vacuum state at some @xmath72 , which corresponds to setting @xmath73 , then it follows that @xmath74 .
imposing continuity of @xmath67 and @xmath75 at the jump at @xmath66 , one can then evaluate the mode function for @xmath71 .
this turns out to be given by f_*k * ( t>0 ) = ( a_*k * e^i _ * k*^f t + b_*k * e^-i _ * k*^f t ) where = ( ) .
using this , the form of @xmath16 for @xmath71 is determined to be z_*k*(t ) = ( ) e^-2 i _ * k*^f t , which yields the following expression for the mean number of particles created at the jump : n_*k * = = ( - ) ^2 . [ n_dfn ]
( one can alternatively impose continuity of the wavefunction given by at @xmath66 to directly obtain the jump in @xmath16 . ) from this expression , it is clear that in the limit of large impulse ( @xmath76 ) , we have @xmath77 and the mean diverges .
the variation of the particle content as a function of the spike strength for a particular mode is shown in fig.([n_df ] ) .
the mean particle production is clearly a monotonically increasing function of @xmath78
. created by the delta function electric field at @xmath66 as a function of @xmath79 .
the plot corresponds to a mode with @xmath80 .
the particle content monotonically increases with @xmath78 , diverging in the @xmath81 limit .
, width=302 ] it may be noted that when working directly with the wavefunction ( or with @xmath67 ) , one does not need to choose a value for @xmath82 in deriving the result for @xmath15 , and it can be left unspecified .
if one takes the alternative route of working with the evolution equation ( [ eqz ] ) instead , then the nature of the differential equation brings in a dependence on @xmath82 .
but the requirement of consistency with the earlier method uniquely fixes its value .
this may be seen as follows : let us start with the .
this equation can be integrated over a small interval @xmath83 around @xmath66 followed by taking the @xmath84 limit .
the second term in would give no contribution in this limit , but the presence of the delta function in the @xmath85 term would give a discontinuity in the value of @xmath16 across the jump , which is related to its value at @xmath66 : z_*k * ( t 0^+ ) = c_*k * [ 1 - z_*k * ^2 ( 0 ) ] [ eq : z_df ] where c_*k * = ( k_z + q ( 0 ) ) . in the above relation , @xmath86 and @xmath87 are unspecified . in order for this relation to give the ` correct ' expression for the average particle number ( i.e. one which matches with that derived earlier by imposing continuity of the wavefunction at the jump ) , one needs to externally specify a particular value for @xmath88 , which turns out to be given by z_*k * ( 0 ) = .
[ z_0 ] with this choice for @xmath88 , gives the same expression as ( [ n_dfn ] ) for the mean particle number .
of course , @xmath87 is still arbitrary here , but if we equate the rhs in with the expression for @xmath86 obtained directly from the continuity of @xmath89 at @xmath66 , then the resulting consistency condition also fixes the value of @xmath87 implicitly through the following relation : 2 _ * k*^i ( ) ( k_z + q ( 0 ) ) = _ * k*(0 ) (_*k*(0 ) + _ * k*^i ) ^2 . [ theta_0 ]
thus , if one works with the evolution equation for @xmath16 , then by specifying the value of @xmath82 through the relation ( [ theta_0 ] ) and the value for @xmath88 as given by , the form of @xmath15 gets completely fixed and identical to the result which was derived earlier by working with the variable @xmath67 .
it may also be mentioned that , from the fact that the magnitude of @xmath16 is a constant for @xmath90 , the trajectory of the system in the complex @xmath16 plane will be a circle for @xmath91 .
this in turn corresponds to a finite and oscillatory phase space average of @xmath24 , so compared with the uncorrelated vacuum state prior to the application of the peaked electric field at @xmath92 , the field modes for @xmath71 can be regarded as more classical in this sense .
this feature is , of course , also shared in common by the cases we have considered earlier in section [ box ] .
our analysis of a quantized scalar field in the background of a varying electric field ( i.e. one which exists for a finite duration ) viewed in a time - dependent gauge provides a nice way to visualize the time evolution of the field modes , and not just their asymptotic behavior .
the concept of a time - dependent mean particle number that was adopted here goes over to the standard definition based on adiabatically definable vacuum states in the static _ in _ and _ out _ regions where no field is present , but displays oscillatory variation in a region where the evolution deviates from adiabaticity , presumably reflecting the transition in the field from quantum fluctuations to physical particles .
the particle creation process in this picture is thus not steady , or even monotonic . for sufficiently large @xmath0
, it is found that the particle content very nearly reaches a saturation beyond a finite time scale ( while the electric field is still present ) , and the value of this steady state particle number is independent of the duration @xmath0 over which the field exists .
although the time dependence of the particle production here arises as a consequence of our choice of a time - dependent gauge to describe the electric field in , the effective lagrangian which can be computed using this quantity would be gauge invariant .
when the @xmath93 limit is taken , the late time behavior of the power spectrum of the quantum fluctuations and the corresponding two - point function indicate the setting up of field correlations over length scales of order @xmath94 .
the behavior of the time - dependent classicality parameter , encapsulating one aspect of the phase space evolution of the field modes , shows that the quantum state grows more correlated over time in comparison with the initial vacuum state it starts off in , and can be interpreted as turning more classical , in this sense , as particle production progresses .
for contrast we have also looked at the case of a sharply localized electric field , which in some sense can be thought of as the @xmath95 limit of the earlier model with the impulse @xmath96 held fixed .
the mean number of particles produced per mode can be computed by imposing continuity of the wavefunction at the ` jump ' in the value of the frequency @xmath97 .
it is found to monotonically increase with the spike strength , i.e. the impulse @xmath98 , and diverges when the @xmath76 limit is taken . :
the author wishes to thank prof .
t. padmanabhan for useful discussions and comments on various aspects of this work .
the author is supported by a fellowship from the council of scientific & industrial research , india .
t. padmanabhan , pramana , * 37 * , 179 ( 1991 ) ; g. v. dunne , in ` from fields to strings : circumnavigating theoretical physics ' , shifman , m. ( ed . )
[ arxiv : hep - th/0406216 ] ; s. p. kim et al .
, phys . rev .
d * 75 * , 045013 ( 2007 ) ; s. p. kim , d. n. page , phys .
d * 73 * , 065020 ( 2006 ) ; y. kluger et al .
* 67 * , 2427 ( 1991 ) ; f. cooper and g. c. nayak , arxiv : hep - th/0611125 . n. d. birrell and p. c. w. davies , _ quantum fields in curved space _ ( cambridge univ .
press , cambridge , 1982 ) ; s. a. fulling , _ aspects of quantum field theory in curved spacetime _ ( cambridge univ . press , cambridge , 1989 ) .
e. wigner , phys .
rev . * 40 * , 749 ( 1932 ) ; a. l. matacz , class .
quantum grav .
* 10 * , 509 ( 1993 ) ; l. p. grishchuk and y. v. sidorov , phys . rev .
d * 42 * , 3413 ( 1990 ) ; t. p. singh and t. padmanabhan , annals of physics * 196 * , 296 ( 1989 ) ; d. polarski and a. starobinsky , class .
quantum grav .
* 13 * , 377 ( 1996 ) [ arxiv : gr - qc/9504030 ] ; s. habib and r. laflamme , phys
. rev .
d * 42 * , 4056 ( 1990 ) . | we adopt the general formalism for analyzing evolution of gaussian states of quantized fields in time - dependent backgrounds in the schrodinger picture ( presented in detail in @xcite ) to study the example of a spatially uniform electric field background ( in a time - dependent gauge ) which is kept turned on for a finite duration of time . in particular , we study the _ time - dependent _ particle content , defined in terms of the concept of instantaneous eigenstates , and describe how it captures the time evolution of the quantized field modes .
the actual particle creation process occurs over a relatively short interval in time , and the particle content saturates rather quickly .
we also compare the _ power spectrum _ of the field modes , computed in the asymptotic limit , with the corresponding situation in a cosmological de sitter background .
particle creation under the influence of a spiked electric field localized in time , as a particular limiting case of the above general model , is also considered . |
in ashtekar s theory of gravity , a @xmath2-connection captures the extrinsic curvature of a space - like leaf in a time - transversal foliation , and the intrinsic geometry of the leaf is given by a tetrad @xcite .
it is shown that the einstein - hilbert functional and einstein equations can be written in terms of @xmath2-connections and tetrads , thus these variables together recast einstein s theory of gravity . by rewriting einstein s gravity with the connection variables and tetrad variables
, one could attempt to quantise gravity via the hamiltonian formalism , and obtain a theory of quantum gravity @xcite
. the reader may refer to thiemann s introductory @xcite .
this article offers an alternative view of the space of connections to ones that appear in other loop quantum gravity literatures , and proposes a semi - classical limit using strict @xmath1-algebraic deformation quantisation formalism @xcite . to take the connections as dynamic variables in a quantum theory ,
one studies wave functions , also known as probability amplitudes , on the space of connections .
unfortunately , the lack of a measure on such a space poses the first challenge , since in such case probabilities can not be defined .
one typical solution to obtain a measure on the connection space comes from spaces of progressively refined cylindrical functions . in another description ,
one uses a finite set of curves to probe the space of @xmath0-connections to obtain a finite dimensional manifold that depend on the sets of curves . by successively refining the finite sets of curves , such as successive triangulation of the manifold , one obtains a pro - manifold that extends the original space of connections to the so - called space of generalised connections @xcite . as a step to quantising gravity in the ashtekar framework , there are recent developments of describing such an extended space of connections using a spectral triple in noncommutative geometry @xcite , which captures the geometry of the space of generalised connections as operators on a hilbert space . while the geometries of the base manifold and the space of @xmath0-connections on it are in theory retained , the construction of the spectral triple is considered too discrete to practically allow one to recapture the geometry of the base manifold and its @xmath0-connections .
to rid the discrete description of using finite sets of embedded curves , this article proposes an alternative approach to smoothly probe the space of connections using the tangent groupoid . and then followed by an application of strict deformation quantisation in @xmath1-algebraic formalism , a possible semi - classical limit of can be obtained .
the purposes of this article are to present an alternative idea to the studies of loop quantum gravity with the goal of obtaining a semi - classical limit , which is what loop quantum gravity still lacks today .
this article contains six small sections , they are arranged as follows : section [ sec2 ] discusses the traditional way of putting a measure on the space of connections in loop quantum gravity literature , and the short coming of such a method . in section [ sec3 ] , we review the definitions of a lie groupoid , a lie algebroid , and tangent groupoid of a lie groupoid in an elementary way .
section [ sec4 ] shows that how one uses the tangent groupoid as a tool to model the space of connections in a smooth way , and discuses the corresponding gauge action .
section [ sec5 ] is the first attempt of deforming connections into noncommutating operators acting on a hilbert space .
section [ sec6 ] introduces the notion of strict deformation quantisation in @xmath1-algebra formalism , and an important theorem by landsman @xcite , which states that a tangent groupoid defines a strict deformation quantisation .
hence , using this formalism , we can deformation quantise @xmath0-connections .
finally , section [ sec7 ] is an outlook that summarises the article and the current state of work toward obtaining a semi - classical limit in loop quantum gravity .
we start by requiring that @xmath0 to be a semi - simple , simply connected , compact lie group , such as @xmath2 . and @xmath3 to be the corresponding lie algebra of @xmath0 .
@xmath4 is the exponential map from @xmath3 to @xmath0 .
the space of smooth @xmath0-connections @xmath5 over a manifold @xmath6 is the space of @xmath3-valued @xmath7-forms over @xmath6 . given a smooth manifold @xmath6 and a smooth ( compact ) curve @xmath8 in @xmath6 , one obtains a map @xmath9 from the space of smooth @xmath0-connections @xmath5 to @xmath0 @xmath10 given by taking the holonomy @xmath11 of each connection @xmath12 along the curve @xmath8 .
here it is assumed that there is a fixed local trivialisation of the principal bundle , so that @xmath5 is realised as a vector space .
suppose that there is a smoothly embedded finite graph @xmath13 in @xmath6 ( edges are smooth , and the number of them is finite ) , one can repeat the holonomy evaluation to obtain a map from @xmath5 to multiple copies of @xmath0 s , one for each edge .
thus , one obtains @xmath14 where @xmath15 denotes the number of edges of the graph @xmath13 .
[ agn : prop : denserange ] for any finite graph @xmath13 , the map @xmath16 is a surjection . while the space of smooth @xmath0-connections @xmath5 lacks structures , such as a measure , it surjects to @xmath17 , a compact measure space .
it is done by forgetting the values of the connections outside the edges of the graph .
hence , a lot of information is lost .
however , one can imagine that there is a collection of finite graphs with one finer than the other , such that the collection of graphs is dense in the manifold @xmath6 in a certain sense .
therefore , at every small neighbourhood in the manifold , there is an edge from the collection of graphs in the neighbourhood to probe the holonomies of the connections .
we do not define the notion of a * directed system * of finite graphs , but to say that it is defined naturally by the associated groupoid of a directed finite graph @xcite .
from which , there is associated a directed system of compact measure spaces @xmath18 . to put the above description in a mathematical context
, we state denote by @xmath19 the projective limit @xmath20 .
suppose that there is a system of of smoothly embedded finite graphs @xmath21 , such that the set of vertices is dense in @xmath6 and every neighbourhood of a point @xmath22 contains edges that span the vector space @xmath23 .
then there exists an embedding @xmath24 from the property of the product ( tychonoff ) topology , one has that [ agn : prop : denserange ] let @xmath25 be a directed system of finite graphs in @xmath6
. then the image of @xmath5 under @xmath26 is dense in @xmath19 .
that is @xmath27 this compactification procedure depends on the system of graphs used in probing the connection space .
we give some examples of graph systems that provide good compactifications . 1 .
[ agn : ex : triangulation ] let @xmath28 be a triangulation of @xmath6 and @xmath29 be the graph consisting of all the edges in this triangulation with any orientation .
let @xmath30 denote the triangulation obtained by barycentric subdivision of each of the simplices in @xmath28 @xmath31 times .
the graph @xmath32 is the graph consisting the edges of @xmath30 with consistent orientation . in this way
@xmath33 is a directed system of finite graphs , and @xmath5 densely embeds into @xmath34 .
[ agn : ex : dlattice ] let @xmath29 be a finite , @xmath35-dimensional lattice in @xmath6 and let @xmath36 denote the lattice obtained by subdividing each cell in @xmath29 into @xmath37 cells .
correspondingly , let @xmath32 denote the lattice obtained by repeating @xmath31 such subdivisions of @xmath38 .
in this way @xmath39 is a directed system of finite graphs , and @xmath5 densely embeds into @xmath40 .
moreover , @xmath19 is a compact measure space .
therefore , such a procedure of surjecting @xmath5 to a coarse approximation @xmath41 , then consider the limit of the approximations provide an extension of the space @xmath5 to a compact measure space @xmath19 .
the space @xmath19 is called the space of generalized connections .
now it is possible to consider probability amplitudes on the space of generalized connections , and proceed to canonical quantisation .
the drawback of this compactification is that the original space of connections @xmath5 is forever lost in @xmath19 , consequently obtaining a semi - classical limit from quantisation on @xmath19 is impossible .
the source of the problem comes from probing the space @xmath5 with finite graphs , which are very rigid objects that can not be perturbed .
one can understand or abstract this process of compactifaction of @xmath5 as probing @xmath5 with a groupoid , where for the case of a graph @xmath13 , the groupoid is the associated fundamental groupoid .
it is a finitely generated groupoid , which is very discrete and can not be perturbed .
however , one can replace this discrete groupoid with a smooth groupoid , say a lie groupoid .
this article proposes the use of the tangent groupoid .
[ groupoid ] a * groupoid * @xmath42 is a ( small ) category in which every morphism is invertible .
that is , a set of morphisms @xmath42 together with a set of objects @xmath43 such that * there exist surjective structure maps , called the source and range maps @xmath44 * there exists an injection , called the identity inclusion , @xmath45 * there exists a partially - defined associative composition @xmath46 with identities @xmath47 , and * there exists an inversion map @xmath48 with the usual properties .
[ liegroupoid ] a * lie groupoid * is a groupoid @xmath49 with smooth manifold structures on @xmath42 and @xmath43 such that @xmath50 are submersions , the inclusion of @xmath43 in @xmath42 as the identity homophism and the composition @xmath51 are smooth .
[ liegroupoidex ] 1 .
any lie group @xmath0 is a groupoid @xmath52 over the identity @xmath53 .
2 . the tangent bundle @xmath54 of a manifold @xmath55 forms a lie groupoid @xmath56 with the _ source _ and _ range _ maps @xmath57 given by @xmath58 for @xmath59 , the inclusion @xmath60 given by the zero section , and composition @xmath61 given by @xmath62 .
3 . the product @xmath63 forms a lie groupoid @xmath64 with the _ source _ and _ range _ maps @xmath65 given by @xmath66 , @xmath67 , the inclusion @xmath68 is given by the diagonal embedding , and the composition @xmath69 is given by @xmath70 .
given two lie groupoids @xmath49 and @xmath71 , their direct product @xmath72 is also a lie groupoid .
a * lie algebroid * on a manifold @xmath55 is a vector bundle @xmath73 over @xmath55 , which is equipped with a vector bundle map @xmath74(called the anchor ) , as well as with a lie bracket @xmath75_e$ ] on the space @xmath76 of smooth sections of @xmath73 , satisfying @xmath77_e = [ \rho \circ x,\rho \circ y],\ ] ] where the right - hand side is the usual commutator of vector fields on @xmath78 , and @xmath79_e = f[x , y]_e + ( ( \rho \circ x)f)y\ ] ] for all @xmath80 and @xmath81 .
a lie algebroid is also a lie groupoid with groupoid product given by fibre - wise addition .
[ algebroidex ] 1 .
a lie algebra @xmath3 with its lie bracket is a lie algebroid over a point .
2 . the tangent bundle @xmath54 of a manifold @xmath55 defines a lie algebroid under the lie bracket of vector fields , and the anchor map @xmath82 is the identity .
[ algebroidex3 ] for a lie groupoid @xmath83 , let @xmath84 be the normal vector bundle defined by the embedding @xmath85 , with bundle projection given by @xmath86 .
identify the normal bundle by @xmath87 , the anchor map is given by @xmath88 .
finally , by identifying @xmath89 with @xmath90 , equip @xmath84 with the lie bracket coming from @xmath91 .
@xmath84 is a lie algebroid .
4 . following the construction in the first example .
@xmath92 is the lie algebroid of the lie groupoid @xmath93 .
the anchor map @xmath94 consist of bundle projection on @xmath54 and zero on @xmath3 .
in example [ algebroidex ] above , @xmath84 is called the associated lie algebroid of the lie groupoid @xmath83 .
it is in itself a groupoid over @xmath6 , similar to @xmath95 .
given the associated lie algebroid of @xmath84 of the lie groupoid @xmath83 .
there exists a unique local diffeomorphism @xmath96 .
let @xmath83 be a lie groupoid with lie algebroid @xmath84 .
the * tangent groupoid * @xmath97 of @xmath42 is the lie groupoid @xmath97 over the base @xmath98 $ ] , such that * as a set , @xmath99 $ ] ; * and @xmath84 and @xmath42 are glued together by the local diffeomorphism @xmath100 . we do not elaborate the definition of @xmath100 here , but rather attempt to illustrate it with an example below .
[ tangentgroupoid ] 1 .
the tangent groupoid @xmath101 of a lie group @xmath0 is just @xmath102 $ ] glued together with the exponential map .
the tangent groupoid @xmath103 of a manifold @xmath55 is @xmath104 $ ] as a set . and
the groupoids @xmath54 and @xmath105 are glued together such that for @xmath106 , then @xmath107 $ ] .
we think of an element @xmath108 of @xmath109 $ ] as a geodesic starting at @xmath110 and ending at @xmath111 , and @xmath112 is the time it takes to travel from @xmath110 to @xmath111 in a given velocity .
hence , it is considered a one dimensional object .
in this section , we propose using a smooth groupoid to probe the connection space with , and the holonomies will be encoded by the smooth @xmath0-valued functions over the smooth groupoid .
denote by @xmath113 the space of smooth functions from @xmath63 to @xmath0 and @xmath114 the space of smooth functions from @xmath54 to @xmath3 such that @xmath115 is linear for @xmath116 and each @xmath117 .
+ here we think of an element @xmath118 of @xmath113 as a holonomy presentation of a connection for paths described by @xmath109 $ ] . [ qconnection ] define the space of * q - connections * @xmath119 to be @xmath120 $ ] as a set . and @xmath114 and @xmath113 are glued together as @xmath121 for all @xmath122 $ ] .
the definition is inspired by the following intuition .
the two points @xmath110 and @xmath123 are connected by the geodesic @xmath124 for @xmath125 $ ] .
and the holonomy along @xmath8 is some group element .
as @xmath112 approaches to zero and the end point of @xmath8 shrinks to its starting point @xmath110 , the holonomy contribution gets closer to the identity element in the group .
therefore , the infinitesimal of the geodesic @xmath8 at @xmath110 gives the infinitesimal change in the group ; so for a point @xmath126 in @xmath54 , one associates it an element in @xmath3 .
one observes the following remarks .
the gluing condition of @xmath119 implies that @xmath127 and @xmath128 where @xmath129 is the identity of @xmath0 .
each @xmath130 gives rise to a @xmath3-valued 1-form in a unique way .
hence , @xmath114 is naturally identified as the space of @xmath0-connections @xmath5 . as a result , there is an embedding @xmath131 @xmath113 forms a group under point - wise multiplication of @xmath0 , and @xmath114 forms a group under point - wise addition of @xmath3 .
the product @xmath119 inherits from @xmath114 and @xmath113 is smooth .
the proof follows from @xmath132 the q - connection space @xmath119 is a package that captures information about probing the @xmath0-connection space with the tangent groupoid .
that is , an element @xmath118 of @xmath113 evaluated at @xmath133 gives the holonomy of a connection along the geodesic from @xmath110 to @xmath111 .
this holonomy presentation has a natural gauge action given by applying a symmetry at the starting point @xmath110 , then evaluate the holonomy along the geodesic from @xmath110 to @xmath111 , and finally apply a reverse symmetry at @xmath111 .
we make it formal by the following definition . denote by @xmath134 the set of smooth functions from @xmath55 to @xmath0 .
define the * gauge action * of @xmath134 on @xmath113 by @xmath135 for @xmath136 and @xmath137 . and define the gauge action of @xmath134 on @xmath114 by @xmath138 for @xmath136 and @xmath116 .
equation is the usual gauge action on @xmath0-connections .
the following proposition shows that the gauge actions defined on @xmath139 and @xmath114 are compatible .
the @xmath134 action on the q - connection space @xmath119 induced from equations , is smooth .
suppose that @xmath140 .
then @xmath141 the proof is complete .
let @xmath142 denote the diffeomorphism group of @xmath55 .
@xmath143 $ ] carries a smooth @xmath142 action given by @xmath144 subsequently , @xmath142 acts on the q - connection space @xmath119 smoothly via the induced action @xmath145 for @xmath146 .
suppose that @xmath147 unitarily represents on some finite dimensional vector space , say without loss of generality @xmath148 .
then @xmath0 is included in the matrix algebra @xmath149 as unitary matrices .
let us fix an orientation on @xmath55 , hence a volume form .
then one obtains the hilbert space @xmath150 that @xmath113 acts on by convolution @xmath151 where @xmath137 , @xmath152 , the dot @xmath153 is the unitary representation of @xmath0 .
denote by @xmath154 the space of @xmath149-valued smooth functions on @xmath63 , thus @xmath155 and it acts on @xmath156 .
elements of @xmath154 will again be denoted by @xmath118 .
@xmath154 comes equipped with an involution given by the point - wise conjugate transpose of @xmath149 .
the action of @xmath154 on @xmath156 gives rise to a noncommutative product on @xmath157 given by the convolution @xmath158 for @xmath159 .
the space @xmath154 forms a @xmath160-algebra , and it is identified with the ideal of trace - class operators on the hilbert space @xmath150 .
let @xmath161 denote the operator trace on @xmath156 , and @xmath162 is explicitly given by @xmath163 where @xmath164 is the matrix trace of @xmath149 .
the linear functional @xmath165 is invariant under the gauge action of @xmath134 , as @xmath166 such a property is called * gauge invariant*. the group element @xmath167 represents the holonomy of a connection around a loop with base point @xmath110 .
the functional @xmath162 being gauge invariant is parallel to the fact that loop variables being gauge invariant in loop quantum gravity .
the package of @xmath154 acting on @xmath156 with gauge group @xmath168 resembles the noncommutative standard model , where the algebra is given by the the gauge group , the hilbert space is unchanged , and the resolvent of the dirac operator gives rise to an element of @xmath154 .
a * system of haar measures * for a groupoid @xmath83 is a family of measures @xmath169 , where each @xmath170 is a positive , regular , borel measure on @xmath171 . for every lie groupoid @xmath83
, there exists a smooth system of haar measures @xmath169 , so that the convolution product with respect to the haar system , @xmath172 together with the star structure , @xmath173 give rise to a @xmath1-algebra structure on the space of continuous functions @xmath174 on @xmath42 .
here we are not being specific on the @xmath1-norm and its closure . for our case of a lie groupoid ,
the different closures coincide .
1 . @xmath175 of a lie group @xmath0 is the group @xmath1-algebra .
2 . given a smooth system of haar measures @xmath169 , where each @xmath176 is a haar measure on @xmath177 , and two smooth functions @xmath178 on @xmath54 .
the product @xmath179 defines , by continuity , a @xmath1-algebra structure on the space of continuous functions on @xmath54 , denoted @xmath180 .
3 . given a smooth system of haar measures @xmath169 , where each @xmath176 is a haar measure on @xmath6 , and two smooth functions @xmath181 on @xmath182 .
the product @xmath183 defines , by continuity , a @xmath1-algebra structure on the space of continuous functions on @xmath63 , denoted @xmath184 .
similarly , there is associated a groupoid @xmath1-algebra @xmath185 to @xmath103 , which can also be seen as gluing the groupoid @xmath1-algebras @xmath180 and @xmath186 together . by fourier
transform , the groupoid @xmath1-algebra @xmath180 of @xmath54 is isomorphism to the continuous function algebra @xmath187 on the cotangent bundle under point - wise multiplication .
this algebra contains the poisson algebra @xmath188 .
therefore , one has a poisson structure on ( a sub - algebra of ) @xmath180 .
similarly , one has a poisson structure on ( a sub - algebra of ) @xmath189 .
a * continuous field of @xmath1-algebras @xmath190 } \right)$ ] * over @xmath191 $ ] consists of a @xmath1-algebra @xmath192 , @xmath1-algebras @xmath193 , @xmath122 $ ] , with surjective homomorphisms @xmath194 and an action of @xmath195)$ ] on @xmath192 such that for all @xmath196 : 1 .
the function @xmath197 is continuous ; 2 .
@xmath198 } \lvert \varphi_\hbar(c ) \rvert$ ] ; 3 . for @xmath199)$ ] , @xmath200 .
[ fieldalgebra ] for the tangent groupoid @xmath97 , we define @xmath201 for @xmath202 and @xmath203 .
the pullback of the inclusion @xmath204 induces a map @xmath205 , which extends by continuity to a surjective @xmath206-homomorphism @xmath207 .
the @xmath1-algebras @xmath208 and @xmath209 with the maps @xmath210 for a continuous field over @xmath191 $ ] .
a strict deformation quantisation of a poisson manifold @xmath211 consists of 1 . a continuous field of @xmath1-algebras @xmath212})$ ] , with @xmath213 ; 2 .
a dense poisson algebra @xmath214 under the given poisson bracket @xmath215 on @xmath211 ; 3 .
a linear map @xmath216 that satisfies ( with @xmath217 ) @xmath218 for all @xmath219 and @xmath220 , and for all @xmath221 satisfies the dirac condition @xmath222 - { \mathcal{q}}_\hbar ( \{f , g\ } ) \right\rvert = 0 .\ ] ] [ liequantisation ] let @xmath42 be a lie groupoid and @xmath84 its associated lie algebroid .
the continuous field of @xmath1-algebras @xmath223}\right)$ ] , as defined in example [ fieldalgebra ] , defines a strict deformation quantisation of the poisson manifold @xmath224 .
the poisson structure on @xmath224 is induced dually by the lie bracket on @xmath84
@xmath225 is a poisson algebra with poisson bracket induced from the lie bracket of @xmath3 .
take @xmath192 to be the @xmath1-algebra generated by the tangent groupoid @xmath226 , @xmath227 to be @xmath189 , and @xmath193 to be the group @xmath1-algebra @xmath175 for @xmath228 .
the quantisation map @xmath229 is the inclusion of @xmath225 into @xmath230 , and @xmath231 is @xmath229 followed by map induced by the inclusion of @xmath3 or @xmath0 into @xmath226 .
2 . @xmath232 is a symplectic manifold .
its symplectic structure defines the poisson algebra @xmath188 , which includes into continuous function algebra @xmath233 . by fourier transform
, @xmath233 is isomorphic to @xmath180 .
the groupoid @xmath54 exponentiates to @xmath63 , thus one obtains the inclusion of @xmath188 into the @xmath1-algebra @xmath185 .
3 . take the lie groupoid @xmath93 , it has the associated lie algebroid @xmath92 .
@xmath234 is a dense poisson algebra in @xmath235 .
the continuous field of @xmath1-algebras , @xmath236 is given by @xmath237 , @xmath238 for @xmath202 , and @xmath239 .
the quantisation map @xmath240 is the inclusion , and @xmath241 is @xmath229 followed by the restriction map .
recall that a @xmath242-connection @xmath243 is precisely a @xmath3-valued 1-form when @xmath244 . by considering ( the characteristic function supported on ) the graph of the function @xmath118 , one obtains a distribution on the tangent groupoid @xmath245 of the lie groupoid @xmath246 .
therefore , one has an action of the space of connections @xmath5 on the @xmath1-algebra @xmath247 , for @xmath248 . by a smearing the characteristic function , that is integrating the distribution with some smooth function ,
one turns the characteristic function into an element in @xmath247 .
therefore , a map from from @xmath119 , which includes the space of ordinary connections @xmath5 , to @xmath247 . then theorem [ liequantisation ] allows us to strictly deformation quantise the connections .
therefore , what one has obtained in this procedure is a strict deformation quantising of the space of connections @xmath5 , which is obtained by mapping @xmath5 into a @xmath1-algebra that constitutes the properties of a strict deformation quantisation .
the goal of this formalism is to provide a deforming parameter @xmath112 , so that when @xmath244 , ordinary connections are retrieved .
this line of work is to provide loop quantum gravity a semi - classical limit , so that the theory of classical gravity , general relativity , returns as one takes the limit @xmath249 .
the lack of a semi - classical limit in loop quantum gravity has been a long standing problem .
the main obstacle of obtaining such a limit is by nature how one traditionally constructs a measure on the space of connections probing the connection space with a collection of finite graphs as described in section [ sec2 ] .
finite graphs are very rigid and discrete , they do not provide a parameter that one usually encounters in quantum theory to adjust . as a result , obtaining
a semi - classical limit simply becomes impossible if a `` smoother '' way of probing the connection space is not introduced . to come up with a smooth way of probing the connection space , we understand this probing procedure as evaluating some one - dimensional objects , which is a groupoid .
thus , by using a smooth groupoid , or more specifically a lie groupoid , one has a hope of circumventing the discreteness problem .
the proposal we give here is the tangent groupoid .
as it turns out , when the right tangent groupoid is used , the space of connections includes into the groupoid .
from there , one could deform the connections to convolution operators acting on a hilbert space .
this formalism recreates some elements appear in the noncommutative standard model @xcite in the way that , the gauge group is the unitary part of the algebra in the spectral triple of noncommutative standard model , the hilbert space remains the same , and the connections are realized as trace - class operators , with the trace being a gauge invariant quantity that resembles the loop variables in loop quantum gravity literature .
the tangent groupoid provides another important feature concerning deformation , which is the strict deformation quantisation result of landsman @xcite .
deformation quantisation can naturally be formulated in terms of @xmath1-algebraic language , a theorem due to landsman shows that a tangent groupoid gives rise to a deformation quantisation in the strict sense . by combining landsman s result and our realization of connections using tangent groupoids , we obtain a deformation quantisation of @xmath0-connections . and this quantisation formalism permits the existence of a parameter @xmath112 , so that when @xmath244 , one retrieves the classical connections .
however , this is not the end of the story .
since connection variables are only half of the variables in gravity in arnowitt - deser - misner formulation , one still has to look into the other half of the variables that the connections are conjugate dual to , tetrads or metrics .
it is known that tetrads are quantised to degree one differential operators @xcite , and its semi - classical limit can be obtained from the @xmath249 limit of integral kernel of the differential operator multiplied by @xmath112 , which gives nothing but the symbol of the differential operator @xcite . in the case of the manifold @xmath55 being three dimensional , the symbol is an @xmath250-valued function the poisson manifold @xmath232 .
hence the symbol of the differential operator lives in the same space as a connection does . following up the work of deformation quantisation of connections here , the next step is to examine the interaction of the tetrads with the connections at both the classical level @xmath244 and the quantum level @xmath202 . at @xmath251 ,
one needs to examine the poisson bracket of a connection and a symbol ( of a differential operator ) , and determine which symbol is conjugate dual to a given connection . at @xmath202 ,
one repeats the same procedure except now that the poisson bracket is replaced with a commutator .
we will leave those considerations to another article .
finally , the author would like to stress that the line of work here is toward obtaining a semi - classical limit in loop quantum gravity , and this tangent groupoid application has not been seriously considered before , thus the work here may appear incomplete .
nonetheless , the development so far shows that a semi - classical limit in loop quantum gravity is more within reach than before . | motivated by the compactification process of the space of connections in loop quantum gravity literature .
a description of the space of @xmath0-connections using the tangent groupoid is given .
as the tangent groupoid parameter is away from zero , the @xmath0-connections are ( strictly ) deformation quantised to noncommuting elements using @xmath1-algebraic formalism .
the approach provides a mean to obtaining a semi - classical limit in loop quantum gravity . |
the nature of dark matter , which accounts for the majority of the mass in the universe , is one of the major outstanding problems of modern astrophysics .
although it is often assumed that dark matter is collisionless , there is no a priori reason to believe that this is the case , and it has been noted by other authors that a non - zero self - interaction cross - section can have important astrophysical implications ( e.g. , spergel & steinhardt 2000 ) . in particular , self - interacting dark matter ( sidm ) has been invoked to alleviate some apparent problems with the standard cold dark matter ( cdm ) model , such as the non - observation of cuspy mass profiles in galaxies ( e.g. , moore 1994 ; flores & primack 1994 ; cf .
navarro et al .
1997 ; moore et al .
1999b ) and the overprediction of the number of small sub - halos within larger systems ( e.g. , klypin et al . 1999
; moore et al .
previous simulations and theoretical studies suggest that a self - interaction cross - section per unit mass of @xmath6 @xmath3 g@xmath0 is needed to explain the observed mass profiles of galaxies ( e.g. , dav et al . 2001 ; ahn & shapiro 2003 , though see also ahn & shapiro 2005 ) .
earlier studies have found stringent upper limits on @xmath1 , inconsistent with the above range ( e.g. , yoshida et al .
2000a ; hennawi & ostriker 2002 ; miralda - escud 2002 , though see also sand et al .
however , in general these studies require non - trivial assumptions or statistical samples of clusters and full cosmological simulations .
furlanetto & loeb ( 2002 ) pointed out that if one observes an offset between the gas and dark matter in a merging cluster , arising because of the ram pressure acting on the gas but not the dark matter , it can be used to constrain the collisional nature of dark matter .
markevitch et al .
( 2002 , hereafter m02 ) found just such a cluster , 1e 0657 - 56 , which in the _ chandra _ image shows a bullet - like subcluster exiting the core of the main cluster , with prominent bow shock and cold front features , and a uniquely simple merger geometry ( markevitch et al . 2002 , hereafter m02 ) .
this gas bullet lags behind the subcluster galaxies , which led m02 to suggest that this cluster could be used to determine whether or not dark matter is collisional .
if dark matter were collisionless , one would expect the subcluster dark matter halo to be coincident with the collisionless subcluster galaxies .
a map of the dark matter distribution was subsequently derived by clowe et al .
( 2004 ) using weak lensing observations , which showed that the subcluster dark matter clump lay ahead of the gas bullet , close to the centroid of the subcluster galaxies ( see also clowe et al . 2006a , hereafter c06 ) . the x - ray image of 1e 0657 - 56 is shown in fig
[ fig : image ] with the most recently derived weak lensing mass contours of c06 overlain .
the weak lensing contours are shown instead of the strong lensing contours since they are better for showing the overall structure , as they are derived from a wider field of view .
the positions of the total mass peaks from strong and weak lensing analyses are consistent with one another , and the general structures are similar .
the more massive main cluster is on the left and the high - velocity merging bullet subcluster is on the right .
the main and subcluster mass peaks are clearly visible in the mass map , as is the offset between the gas bullet and the corresponding dark matter ( dm ) peak .
c06 argued that this offset is direct evidence for the existence of dark matter . the weak lensing mass map of clowe et al .
( 2004 ) was used by markevitch et al .
( 2004 , hereafter m04 ) , in conjunction with the x - ray and optical observations available at the time , to analytically estimate upper limits on the self - interaction cross - section per unit mass of dm , @xmath1 , using three independent methods .
these methods were based on the observed offset between the gas bullet and the dm subclump , the high merger velocity of the subcluster , and the survival of the dm subclump ( more precisely , the subcluster s @xmath7 ratio being equal to that observed in other clusters and in the main cluster ) .
m04 assume a king mass profile , based on the original weak lensing mass map , and that the subcluster has passed only once through the main cluster , close to the main cluster core , as indicated by the x - ray image .
their most stringent limit comes from the observed survival of the dm subclump , from which they infer that @xmath8 @xmath9 g@xmath0 .
although the analytic estimates performed by m04 provide useful upper limits on @xmath1 , several conservative simplifying assumptions were necessary .
for instance , the effects of dynamical friction as the subcluster disturbs the main cluster mass distribution were ignored , as was the possibility of multiple scatterings per particle .
although these effects are relatively small , their inclusion may lead to tighter constraints .
furthermore , the analytic estimates can not address any structure that may be found in a high - resolution mass map ( e.g. , tails in the dm distribution , similar to the gas tails seen in the bullet , due to collisional stripping of dm , as described by m04 ) .
this argues for full n - body simulations that would include the effects of sidm with varying cross - sections .
additionally , new data ( both x - ray and lensing ) have become available for 1e 0657 - 56 .
analysis of data from 450 ks of total exposure with _
chandra _ gives a more accurate shock mach number of @xmath10 ( all uncertainties 68% ) , which corresponds to a shock ( and bullet ) velocity of 4700@xmath11 km s@xmath0 ( markevitch 2005 ) .
recent weak and strong lensing analyses of a much larger optical dataset , which includes _ hst _ observations , give a higher quality mass map and a more accurate determination of the subcluster dark matter and galaxy centroids ( brada et al .
2006 , hereafter b06 ; c06 ) .
in particular , the accuracy of the total mass and galaxy centroids is now sufficient for an additional method of constraining @xmath1 . in this paper
, we will concentrate on the most sensitive method from m04 , which is based on the observed mass - to - light ( @xmath7 ) ratios , and on this new test .
the best way to interpret the new high - quality data is through comparisons with detailed numerical simulations of the merger which allow for sidm with varying cross - sections .
we present results from such simulations and give constraints on the self - interaction cross - section of dark matter particles .
we assume @xmath12 , @xmath13 , and @xmath14 km s@xmath0 mpc@xmath0 , for which @xmath15 kpc at the cluster redshift of @xmath16 .
all simulations were performed using a modified version of the publicly available treesph code gadget2 ( springel 2005 ) . to model the self - interaction of the dm particles ,
we adopted a monte carlo method used previously by other authors ( e.g. , burkert 2000 ; yoshida et al .
at each simulation time step , the scattering probability for the @xmath17 particle is given by @xmath18 where @xmath19 is the local density , @xmath20 is the relative velocity between the @xmath17 particle and its nearest neighbor , and @xmath21 is the time step size .
the local density is determined using gadget2 s smoothed particle hydrodynamic ( sph ) capabilities .
collisions are assumed to be elastic and scattering isotropic in the center - of - mass frame .
in order for this relation to be valid , @xmath21 must be chosen such that @xmath22 .
we ran a series of merger simulations with @xmath1 varying between 0 and 1.25 @xmath3 g@xmath0 .
each simulation run included 10@xmath23 dm particles ( gas was not included in the simulations , see discussion in [ sec : gas ] ) .
additionally , we performed a convergence test run with 10@xmath24 dm particles and @xmath25 @xmath3 g@xmath0 , which agreed well with the lower resolution run for all tests we performed .
we interpret this agreement as indicating that the effects of individual self - interacting dm particles are well modeled by the large computation particles used in the simulations , and that the results we present here are not seriously affected by numerical resolution effects .
the ratio of dm particles in the main cluster and subcluster was set equal to the initial total mass ratio of the clusters , which is known analytically from the king models used to build the clusters . in this work ,
we apply a new method for constraining @xmath1 based on the absence of an offset between the subcluster total mass and galaxy centroids . for this
, we added another family of particles to the simulations to represent the collisionless galaxies .
we choose 10@xmath26 `` normal '' galaxy particles for the main cluster and 2.5@xmath27 for the subcluster throughout .
the ratio of the number of normal galaxy particles was estimated based on galaxy counts given by barrena et al .
the galaxies were initially distributed like the dm in each run .
the mass in galaxies was assumed to be roughly 5% of the total mass for each cluster , which , combined with the number of galaxy particles , gives a low mass per galaxy ( 2.15@xmath28 m@xmath29 ) . using a large number of light - weight galaxies was chosen over using a more realistic mass per galaxy so that accurate galaxy centroids could be determined .
test simulations run with the more realistic average mass per galaxy of @xmath30 m@xmath29 , also determined from results given by barrena et al .
( 2002 ) , showed similar results to those given below in [ sec : results ] , though , as expected , with a larger scatter .
two cd galaxies , one at the center of each cluster , were also included in the simulations ( though we note that three cd galaxies are observed , two associated with the main cluster and one with the subcluster ) .
their inclusion leads to conservative estimates of the effects of dm self - interaction , since a lower central dm density is required to reproduce the observed total mass profile ( and the scattering probability depends on the local dm density via eqn .
[ eqn : prob ] ) .
the cd galaxies were each given a mass of 10@xmath31m@xmath29 .
the gravitational softening length was chosen to be 2 kpc throughout , which is on the order of the mean inter - particle separation in the densest region in each simulation ( i.e. , at core passage ) . the softening length for the cd galaxy particles
was set to 60 kpc throughout .
this large softening length was chosen since on the scale of the simulations cd galaxies are significantly extended objects and treating them as concentrated point - like masses would be unrealistic . since the lensing observations do not give an accurate mass profile for the subcluster , king models with density profiles of @xmath32 , where @xmath33 is the central density and @xmath34 is the core radius , were conservatively chosen for the mass profiles of each cluster .
such a choice gives conservative limits for the effects of self - interacting dm , since king models do not have strongly concentrated `` cuspy '' cores , as compared to nfw ( navarro et al .
1995 ) and hernquist ( 1990 ) profiles .
thus the central density is lower and the total number of dm particle collisions is conservatively reduced . further discussion of the impact of the bullet mass profile on our results can be found in [ sec : massprof ] . as suggested by the x - ray morphology ( m04 ) , all simulated mergers were head - on collisions with zero impact parameter and an initial separation of 4 mpc .
the effects of a possible non - zero impact parameter are discussed in [ sec : impact ] . for each run ,
the initial conditions were chosen such that the projected mass profiles of the main cluster and the bullet subcluster , after core passage and at the observed separation ( 720 kpc ) , roughly matched those from the most recent combined strong and weak lensing results derived by b06 , which are given in the last row of table [ tab : obscond ] . a relatively small contribution from the observed distribution of gas mass
was subtracted from the b06 total mass , so that the resulting values could be directly compared with the simulations .
the gas masses were computed from the x - ray observations .
the details of the derivation of the gas mass map will be given in a future paper .
a summary of the parameters for each simulation run is given in table [ tab : ics ] , which gives the initial central density and core radius for the main cluster ( @xmath35 , @xmath36 ) and the bullet subcluster ( @xmath37 , @xmath38 ) , and @xmath1 for each run .
the mass profiles were truncated at 20 @xmath39 . as pointed out by c06
, weak lensing is expected to underestimate the mass of the lens by 10 - 20% in the dense central regions .
furthermore , weak lensing can underestimate masses due to mass - sheet degeneracy , where the mass map is affected by the non - detection of mass at the edges of the field of view .
the effect can be seen by comparing the total ( i.e. , without the gas mass subtracted ) weak lensing mass estimates of c06 to the mass profiles derived by b06 , which combine strong and weak lensing observations .
we chose to match the projected mass profiles in the simulations to those given by b06 , since strong lensing is expected to give better results in the central core regions .
we are most interested in these regions since most of the particle scattering depth accumulates near the center .
we explore the effects of decreasing the total mass of the system on our results in [ sec : lowmass ] .
columns 4 & 5 of table [ tab : obscond ] give the total projected mass within 150 kpc of the mass peak for the main cluster ( @xmath40kpc@xmath41 ) and the subcluster ( @xmath42kpc@xmath41 ) at the observed separation for each simulation run and from the lensing observations . some trial - and - error was necessary to determine what initial conditions gave the desired projected mass profiles at the desired separation . in general , larger @xmath1 values required a more concentrated subcluster initially . this effect occurs because the self - interaction of the dm particles causes them to be scattered away , particularly in high density regions .
consequently , the subcluster mass profile is spread out during core passage . for the purposes of comparing the simulations with observations , we take the simulation snapshot where the offset between the clusters is closest to the observed separation . the time resolution of the snapshots
was small enough to match this value to within a few kpc ( @xmath43 1 ) , which is within the observational error .
we require the simulated mass profiles at this moment to be consistent with the strong lensing mass map to within 10% in the inner regions ( @xmath44 500 kpc ) .
it is of interest to evaluate the stability of our simulated clusters , particularly in the presence of sidm . in the case of non - self - interacting dm ,
the phase - space distribution function can be computed and used to generate a gravitationally stable king model density profile . however , in the case of sidm , particle collisions will tend to transfer kinetic energy from one region of the cluster to another , consequently altering the density profile ( see , e.g. , burkert 2000 ) . in section
[ sec : mtol ] , we will draw conclusions based on the fraction of particles scattered away from the core of the subcluster due to the merger event .
it is therefore necessary to determine what fraction of particles might flow from the central region of the bullet due to the instability resulting from sidm collisions . to this end , we ran simulations of the subcluster dm halo , allowing it to evolve in isolation over the timescale of the merger simulations ( about 1 gyr ) .
we used the same cluster parameters as the subcluster that has the highest central density of all the clusters in table [ tab : ics ] .
the results are shown in figure [ fig : rhor ] , which gives the initial density profile ( solid line ) and the density profile after 1 gyr for @xmath45 @xmath3 g@xmath0 ( dotted line ) and @xmath46 @xmath3 g@xmath0 ( dashed line ) . the density in the inner regions
is marginally enhanced in the case of sidm .
this result is similar to the core - collapse phase seen by burkert ( 2000 ) , where weak interactions between the kinematically hot core and the cooler outer regions result in an outward transport of kinetic energy ( though this effect is expected to be somewhat curtailed here due to the near isothermality of the king profile at small radii ) . for the purposes of the test described in [ sec : mtol ] , we are only concerned with the total mass within _ projected _ radii .
this quantity is plotted for each run in figure [ fig : projr ] .
the above effect of sidm on the projected mass profile is negligible , particularly for projected radii @xmath47 kpc , which is the minimum radius considered for the test described in
[ sec : mtol ] .
thus , if we find that a large fraction of sidm particles scattered outside this radius , it can be assumed to be caused by the merger event as opposed to any halo instability .
furthermore , the collisionless galaxies are expected to adjust to any change in the overall potential ( which is dominated by the dm ) , thereby acting to further stabilize the mass - to - light ratio .
indeed , in these isolated subcluster runs , the mass - to - light ratios within a projected radius of 150 kpc from the cluster centers stay within 2% of their initial values , regardless of the dm self - interaction .
for non - self - interacting dm , the centroids of the subclump dm and galaxy distributions are expected to be coincident throughout the simulation , since gravity is the only operating force . however , when @xmath48 , the subcluster dm halo experiences a drag force as it passes through the main cluster , and subsequently lags the collisionless galaxies , just as the fluid - like subcluster gas core is observed to lag the dm halo ( see fig .
[ fig : image ] ) .
we ran simulations with a range of values for @xmath1 and calculated the centroids for each particle type by taking the average projected position of the particles in some large region , centering on this position with a smaller region , and repeating with smaller and smaller regions ( down to a region with a radius of 200 kpc ) .
column 6 of table [ tab : obscond ] gives @xmath49 , the offset between the subcluster galaxy and dm centroids , for each run , for the moment when the subcluster is close to the observed separation of 720 kpc from the main cluster .
the dependence of @xmath49 on @xmath1 is also plotted in figure [ fig : sigma ] ( solid line ) .
results from the run with @xmath50 indicate that the offsets from the simulations are accurate to about @xmath51 kpc ( @xmath52 ) .
it is clear from table [ tab : obscond ] that the centroid offset is a strong function of @xmath1 .
an x - ray image close - up of the bullet region with error contours for the subcluster total mass and galaxy centroids overlain is shown in figure [ fig : bullet ] .
details of the derivation of the total mass centroids are given in c06 .
the centroid of the galaxy distribution was calculated from the acs photometry , using all galaxies for which the f814w - f606w color is within 0.15 mag of the red sequence .
we used an epanechnikov kernel with @xmath53 ( merritt & tremblay , 1994 ; gonzalez et al .
2002 ) to determine the centroid , and a bootstrap technique to quantify the uncertainty .
the centroid of the subcluster galaxies is found to be @xmath54 ( @xmath55 kpc ) west of the corresponding weak lensing mass peak .
given the observational errors on the centroid positions ( roughly 5@xmath56 , or 22 kpc , on the subcluster mass peak and galaxy centroid ) , the absence of a larger offset means that @xmath57 1.25 @xmath3 g@xmath0 .
we note that , although this upper limit is greater than the best constraint of @xmath57 1 @xmath3 g@xmath0 found by m04 , it is more robust , since it does not rely on the assumption that the subcluster and the main cluster had equal @xmath7 ratios prior to the merger , as is the case with the limit from m04 ( see [ sec : mtol ] ) .
this distinction is relevant since , although there is evidence for a universal @xmath7 ratio for clusters , the level scatter for individual clusters is not negligible ( see dahle 2000 ) . in a merger scenario ,
sidm is expected to give a lower @xmath7 ratio for the subcluster that has just passed through a dense core as compared to collisionless dm .
this is because during the merger , dm particles are scattered away due to collisions , while the collisionless galaxies are relatively unaffected . to estimate the change in the @xmath7 ratio in the simulations due to the merger
, we simply take the ratio of the total mass to galaxy mass within 150 kpc ( projected ) of the bullet dm centroid and compare the values at the start of the simulations and at the observed separation .
the results are tabulated in column 7 of table [ tab : obscond ] , which gives @xmath58 , the fractional decrease in the bullet @xmath7 ratio within 150 kpc , and also plotted in figure [ fig : sigma ] ( dashed line ) .
we note that for @xmath59 @xmath3 g@xmath0 , the subcluster loses about 38% of its mass within 150 kpc , which is in agreement with a conservative estimate of 20 - 30% given by m04 .
as expected , the numerical results yield somewhat tighter constraints on @xmath1 as compared to the analytic estimates when using the same method and observational constraints . using the latest lensing mass map from b06 ,
we rederived @xmath7 ratios for each of the two subcluster within a projected 150 kpc of the total mass peaks ( for previous results see clowe et al . , 2004 ) .
for the subcluster , the mass contribution from the outskirts of the main cluster has been approximately subtracted , whereas for the main cluster , the total mass is used , since the contribution from the subcluster is negligible .
the projected mass contribution from the main cluster to the subcluster is estimated by taking the average mass in an annulus at the distance of the subcluster ( excluding the region of the subcluster itself ) .
this gives a conservative estimate for the upper limit on @xmath1 , since scattering due to putative dm collisions is expected to result in an anomalously low @xmath7 value for the subcluster as compared to the main cluster , and by reducing the observed mass of the subcluster we minimize the effect of the collisions that we want to constrain .
we find @xmath60 and @xmath61 for the subcluster and the main cluster , respectively ( for a discussion of the errors on the mass measurements , see b06 ) .
the ratios agree with one another to within about the 68% confidence intervals . from the @xmath62 band data , we find that the ratio of @xmath7 ratios of the subcluster and main cluster is @xmath63 .
we conservatively choose to use the @xmath62 band data only , since we want to put a firm lower limit on this ratio , and @xmath64 is larger for the subcluster than for the main cluster . assuming each cluster started out with similar @xmath7 values , which appears to be a reasonable assumption for clusters in general ( e.g. , mellier 1999 ; dahle 2000 ) , we conclude that the subcluster could not have lost more than @xmath65 of its initial mass .
a comparison with the results from simulations plotted in figure [ fig : sigma ] ( dashed line ) shows that this implies @xmath66 @xmath3 g@xmath0 , which is a slight improvement over the previous best limit of @xmath67 @xmath3 g@xmath0 from the conservative estimates of m04 .
m04 suggested that scattered dm particles , which would account for about 1/5 of the total subcluster mass , might form tail features in the dm distribution , similar to the tails seen in the x - ray image of the gas bullet ( see figure [ fig : image ] ) .
the simulations allow us to determine whether the non - observation of such tails in the mass map could be used to constrain @xmath1 .
we find that , rather than forming a tail , the scattered particles are mostly deposited in the core of the main cluster , and do not form any features at a level that is interesting for constraining @xmath1 .
as m04 argue , the morphology of the x - ray image , in particular , the symmetry of the north - south x - ray bar ( most likely an oblate spheroid viewed edge - on ) between the main cluster and subcluster mass peaks around the axis of symmetry set by the shape of the x - ray bullet ( which gives its present velocity direction ) , combined with the line - of - sight velocity and x - ray derived mach number , indicate a merger axis that is @xmath68 from the plane of the sky , and that the cluster cores must have passed close to one another , certainly within the @xmath69 kpc core radius of the main cluster . in all simulations
previously discussed , it was assumed that the bullet subcluster passed directly through the center of the main cluster core , i.e. , that the impact parameter of the merger , @xmath70 , is zero . for @xmath71
, we expect that the effects of self - interacting dm will be reduced , since the density is at a maximum when the core centers pass directly through one another , and the scattering probability is proportional to the density ( eqn .
[ eqn : prob ] ) . to test the strength of this effect , we re - ran the simulation r4 ( see table [ tab : ics ] ) with an impact parameter of @xmath72 kpc . aside from the impact parameter , the initial mass and velocity distributions were identical to those for run r4 , so that the relative effects of @xmath73 could be investigated ( specifically , no adjustments were made to the initial conditions to more closely match the current observed mass profiles ) . the resulting projected total mass profiles for the subcluster within 150 kpc and 250 kpc at the observed separation agreed with those from the @xmath74 run to within 4% . for the main cluster ,
the match was better than 1% .
the resulting offset between the galaxy and dm centroids during the post core passage phase was systematically smaller than the offset seen in the @xmath74 run . at the observed separation ,
the difference in the centroid offsets , as compared to the @xmath75 run , was about 4 kpc , which is on the order of both the observational error and the accuracy of our numerical technique .
the fractional change in the @xmath7 ratio of the subcluster was similarly affected ; for the @xmath74 case , the @xmath7 ratio within 150 kpc drops by about 27% , whereas for the run with @xmath72 kpc , it drops by 22% .
assuming , as we did in
[ sec : mtol ] , that the subcluster could not have lost more than @xmath76% of its initial mass , we find the constraint that @xmath77 @xmath3 g@xmath0 .
we therefore conclude that , although a non - zero impact parameter reduces the effects of self - interacting dm as expected , the level of the effect is relatively small .
this is likely due to the assumed king mass profile of the main cluster .
the radial density gradient is relatively small within the core radius of the main cluster ( which in this case is 151 kpc ) , so it is not surprising that the effects of self - interacting dm are not significantly reduced by increasing the impact parameter , so long as it is comparable to the core radius of the main cluster .
naturally , the effects of a non - zero @xmath70 would be increased if the main cluster had a strongly peaked mass profile , though the current lensing data suggest otherwise .
we conclude that any value of impact parameter that is consistent with the observations will only slightly alter our results . as noted in
[ sec : sims ] , the choice of a king mass profile for the subcluster is expected to give conservative estimates on the effects of collisional dm , based on the @xmath7 ratio , since the central density is low as compared to models with cuspy cores such as nfw and hernquist models . however , in the case of the galaxy / dm centroid offset test , one might argue that since the subcluster galaxies are more tightly bound in the center for more highly concentrated mass profiles , it will be more difficult to displace them , which could lead to a smaller offset between the centroids despite the increased action of dm collisions .
we therefore ran a test simulation , using a king model for the main cluster and a hernquist model for the subcluster , with @xmath78 @xmath3 g@xmath0 .
the hernquist profile is given by @xmath79 where @xmath80 is the total cluster mass and @xmath81 is the scale length ( hernquist 1990 ) . as before ,
initial parameters were chosen such that the bullet mass profile roughly matches the observed profile at the current separation ( we used @xmath82 m@xmath83 , @xmath84 kpc ) .
this is expected to be the most conservative model combination for this test , since the main cluster king model minimizes the effects of dm self - interaction while the subcluster hernquist model maximizes the binding energy of the subcluster galaxies .
the results show that , when comparing to run r4 in table [ tab : obscond ] , the galaxy / dm centroid offset was only slightly less than that found with the king model subcluster , on the order of the accuracy of the simulation offset values ( less than 1 kpc ) .
the change in the subcluster s @xmath7 was similarly only weakly affected ( @xmath85 for the king model bullet subcluster , whereas for the hernquist model we find @xmath86 , consistent with the king profile being the conservative case ) .
the agreement is likely due to the fact that the centroids become offset from one another after core passage , and it is during core passage that the central density peak of the bullet is mostly `` smoothed away '' due to dm collisions ( recall that dm scattering is more frequent in high density regions , so high density structures are more efficiently destroyed by dm self - interactions ) .
although strongly peaked density profiles have been found to be unstable to sidm ( e.g. , burkert 2000 ; yoshida et al .
2000b ) , in our simulation a significant change in density only occurred at small radii , such that the total projected mass of the subcluster within 50 kpc remained stable up until the merger event .
therefore the subcluster mass distribution remained significantly more peaked than a king profile cluster with the same projected mass within 150 kpc .
we conclude that our results are only weakly dependent on the mass profile chosen for the bullet , so long as we require that the observed mass profile is reproduced . for the initially more centrally concentrated profiles ,
the effects of the increased binding energy in the core are balanced by the increased scattering frequency in this region . as mentioned in
[ sec : sims ] , weak lensing is expected to underestimate the mass of the lens by 10 - 20% .
there are two separate effects that contribute to this underestimation .
first , near the core of a cluster there is a large region without weak lensing galaxies , and this region is effectively smoothed over when computing the mass map . additionally , galaxies near the regions where strong lensing dominates are measured in the weak lensing approximation , which also leads to an underestimate of the mass in the core .
second , the total cluster mass can be underestimated due to mass - sheet degeneracy , where the mass - map is affected by the non - detection of mass at the edge of the field of view .
although projection of foreground and/or background structures unassociated with the clusters will artificially increase the mass , it is highly unlikely that such projected structures significantly contributed to the detected lensing signal ( c06 ) .
results from strong lensing , which is not susceptible to the same systematic underestimation as weak lensing , do indeed give systematically higher projected masses for this system , by about a factor of 2 within the inner few hundred kpc , which is the region we are most interested in for this analysis ( compare b06 and c06 ; see c06 for further discussion of this discrepancy ) .
though we chose to use the mass estimates from strong lensing , since it should give a more reliable estimate of the projected mass near the cluster cores , it is interesting to explore the dependence of our results on the lensing mass estimates . to this end
, we conducted a simulation run similar to run r4 , but with the initial cluster central densities chosen such that the projected mass profiles at the observed separation were about 2 times lower than the masses derived from strong lensing observations , roughly in agreement with the weak lensing results given by c06 ( the initial core radii were the same as for run r4 ) .
since the scattering probability depends on the density , we expect these less massive halos to be more weakly affected by sidm .
results obtained from the simulations with the lower mass normalization show that the effects of sidm are diminished , as expected for a linear dependence of the scattering probabilities on the projected mass . in run
r4 , the @xmath7 ratio dropped by 27% , whereas for the run with 1/2 the total mass it dropped by 14% ( roughly a factor of 2 less ) .
similarly , the galaxy / dm centroid offset was 11.1 kpc , again , about a factor of 2 down from the 24.1 kpc offset seen in run r4 .
if we assume that this factor of two effect can be applied to all of values given in columns 6 & 7 of table [ tab : obscond ] and consider the more sensitive @xmath7 test , we find that requiring @xmath87 would correspond to @xmath88 @xmath3 g@xmath0 .
this is done as a test of the method only , since these low halo mass values are not realistic , as they are insufficient to produce the system of strong arcs observed in the _ hst _ images ( c06 ) .
all of the simulations discussed so far have assumed a merger velocity that is consistent with that derived from x - ray observations ( markevitch , 2005 ) , which give a mach number for the shock front of @xmath10 , and it is assumed that the subcluster has the same velocity as the shock ( though see springel & farrar , 2007 ) . since the subcluster could have slowed down , or the shock front accelerated , it is interesting to ask what effect a lower velocity would have on the inferred upper limit on @xmath1 , particularly since the observed velocity is larger than would be expected from free fall of the subcluster onto the main cluster ( farrar & rosen 2007 ) . in order to test the dependence of our upper limit on merger velocity
, we ran a simulation with @xmath78 @xmath3 g@xmath0 such that the relative velocity of the cluster dm halos at observed separation was 1.5 times lower , about 3100 km s@xmath0 ( @xmath89 ) .
this is close to the expected free - fall velocity of the subcluster , and to the relative velocity of 2860 km s@xmath0 found by springel & farrar ( 2007 ) from hydrodynamical simulations of this system .
the results showed little difference from the higher velocity run ( compare to run r4 in table [ tab : obscond ] ) : @xmath49 was 30.2 kpc ( vs. 24.1 kpc ) and @xmath58 was 0.25 ( vs. 0.27 ) .
we therefore conclude that our results are relatively insensitive to merger velocities that are not in large disagreement with the observations .
a weak dependence of the @xmath7 ratio on the subcluster velocity @xmath90 is easy to understand : the particle scatters out of the subcluster as long as @xmath91 is much greater than the escape velocity from the subcluster , which it is by a large margin ( m04 ) . as mentioned in
[ sec : sims ] , the intracluster gas observed in the x - ray band was not included in the simulations ( doing so would greatly increase the computing time and the complexity involved with matching the observations in detail ) . the only way for the gas to affect the results is via gravitational interaction ( we ignore the possibility of non - gravitational baryon - dm interactions , the cross - section of which has been shown to be extremely small , e.g. , chen et al .
2002 ) . in general
, the gas is expected to contribute about 10% of the total mass of the system , a figure which appears to be consistent with the lensing and x - ray observations ( b06 ) .
one might worry that , when matching the observed mass profiles , some `` extra '' dm is needed to account for the missing gas . as mentioned in
[ sec : sims ] , gas masses have been subtracted from the lensing masses using a detailed model of the gas distribution derived from fitting the x - ray observations . in terms of the test involving the decrease in the subcluster @xmath7 ratio ( see [ sec : mtol ] ) , we need nt worry about the subcluster gas for the simple reason that the gas bullet is far from the subcluster mass peak ( roughly 23@xmath56 , or 102 kpc ) . therefore , the gas in the region of the bullet mass peak is not centrally concentrated and will not significantly add to the binding energy of potentially scattered dm particles . for the galaxy and total mass centroid offset test ( see [ sec : noffset ] )
, the exclusion of the gas is expected to give a conservative result : the gas bullet and bar feature seen in figure [ fig : image ] will act to decelerate the subcluster dm halo and galaxies . however , if , as is the case with sidm , the dm halo starts to lag behind the galaxies and gets closer to the gas cores than the main concentration of the galaxies , it will experience a larger deceleration , thereby increasing the offset between the two . due to the relatively low mass of the gas components , and the large distance between the gas peaks and the subcluster dm halo and galaxies ( as compared to the offset of the latter ) , the strength of this effect will be quite small .
we therefore conclude that including gas in the simulations would not significantly affect our results .
we have combined results from new x - ray , optical and lensing observations and our n - body simulations of the merging galaxy cluster 1e 0657 - 56 in order to derive an upper limit on the self - interaction cross - section of dark matter particles , @xmath1 .
we give constraints on @xmath1 based on two independent methods : from the lack of offset between the total mass peak and galaxy centroid of the subcluster that would arise during the merger due to drag on the subcluster halo from dm particle collisions , and from the lack of a decreased mass - to - light ratio of the subcluster due to scattering of dm particles . from the former ,
we find @xmath92 @xmath3 g@xmath0 , and from the latter , @xmath4 @xmath3 g@xmath0 , which includes the uncertainty in the impact parameter of the merger ( upper limits are from 68% confidence intervals ) .
our best constraint is a modest improvement of the previous best constraint from conservative analytic estimates of @xmath8 @xmath3 g@xmath0 ( m04 ) .
furthermore , our limit of @xmath92 @xmath3 g@xmath0 is more robust than the best analytic limit , since this method does not depend on the assumption that the subcluster and main cluster @xmath7 ratios were equal prior to the merger .
previous studies have found that @xmath93 @xmath3 g@xmath0 is needed produce the observational effects that self - interacting dark matter has been invoked to explain ( e.g. , non - peaked galaxy mass profiles and the underabundance of small halos within larger systems ) .
our results rule out almost this full range of values , at least under the assumption that @xmath94 is velocity - independent .
we would like to thank volker springel , naoki yoshida , yago ascasibar , and alexey vikhlinin for useful discussions and for providing access to various private codes .
simulations were performed on a beowulf cluster at the itc in the harvard - smithsonian center for astrophysics .
support for this work was partially provided for by the nasa _
grants g04 - 5152x and tm6 - 7010x , and nasa contract nas8 - 39073 .
ccccccc r1 & 10@xmath23&0 & 3.27 & 213 & 4.59 & 149 + r2 & 10@xmath23&0.24 & 3.27 & 213 & 4.59 & 149 + r3 & 10@xmath23&0.48 & 4.42 & 183 & 6.57 & 129 + r4 & 10@xmath23&0.72 & 7.03 & 151 & 11.75 & 108 + r5 & 10@xmath23&0.96 & 6.26 & 167 & 9.76 & 124 + r6 & 10@xmath23&1.25 & 6.26 & 167 & 9.76 & 124 + ccccccc r1 & @xmath95&0 & 12.0 & 11.1 & 1.8 & 0.0 + r2 & @xmath95&0.24&11.5 & 10.4 & 5.4&0.08 + r3 & @xmath95&0.48&11.8 & 10.4 & 15.0&0.16 + r4 & @xmath95&0.72&12.6 & 11.0&24.1&0.27 + r5 & @xmath95&0.96&12.4 & 10.9&37.9&0.32 + r6 & @xmath95&1.25&11.4 & 9.8&53.9&0.38 + obs . & & & [email protected] & [email protected]&@xmath97 & @xmath98 + | we compare recent results from x - ray , strong lensing , weak lensing , and optical observations with numerical simulations of the merging galaxy cluster 1e 0657 - 56 .
x - ray observations reveal a bullet - like subcluster with a prominent bow shock , which gives an estimate for the merger velocity of 4700 km s@xmath0 , while lensing results show that the positions of the total mass peaks are consistent with the centroids of the collisionless galaxies ( and inconsistent with the x - ray brightness peaks ) .
previous studies , based on older observational datasets , have placed upper limits on the self - interaction cross - section of dark matter per unit mass , @xmath1 , using simplified analytic techniques . in this work , we take advantage of new , higher - quality observational datasets by running full n - body simulations of 1e 0657 - 56 that include the effects of self - interacting dark matter , and comparing the results with observations . furthermore , the recent data allow for a new independent method of constraining @xmath1 , based on the non - observation of an offset between the bullet subcluster mass peak and galaxy centroid .
this new method places an upper limit ( 68% confidence ) of @xmath2 @xmath3 g@xmath0 .
if we make the assumption that the subcluster and the main cluster had equal mass - to - light ratios prior to the merger , we derive our most stringent constraint of @xmath4 @xmath3 g@xmath0 , which comes from the consistency of the subcluster s observed mass - to - light ratio with the main cluster s , and with the universal cluster value , ruling out the possibility of a large fraction of dark matter particles being scattered away due to collisions .
our limit is a slight improvement over the previous result from analytic estimates , and rules out most of the @xmath5 @xmath3 g@xmath0 range invoked to explain inconsistencies between the standard collisionless cold dark matter model and observations . |
after the discovery of graphene@xcite , interest in atomically thin materials@xcite has grown rapidly due to their extraordinary physical properties.@xcite in the last decade , several 2d materials have been synthesized and theoretically predicted , such as silicene,@xcite germanene,@xcite stanene,@xcite transition metal dichalcogenides ( tmds such as mos@xmath0 , ws@xmath0),@xcite and iii - v binary compounds ( e.g. @xmath1-bn , @xmath1-aln).@xcite beside the single crystal of 2d materials , recently emerging field is their vertically stacked heterostructures.@xcite because of the van der waals type weak interlayer interaction , the synthesis of heterostructures is not restricted with the lattice matching of the each layers .
this provides a wide variety of combinations of layers which exhibit different electronic and optical properties.@xcite stable hexagonal crystalline structure of aln was first theoretically predicted by sahin _
et al_.@xcite and experimentally synthesized by tsipas _ _ et al.__@xcite they found that , differing from its bulk structure which is an insulator , monolayer @xmath1-aln is a semiconductor with the indirect band gap where the valance band maximum and conduction band minimum at the k and @xmath3 points , respectively .
_ et al.__@xcite investigated properties of defects , such as vacancies , antisites , and impurities , in @xmath1-aln . it was reported that n vacancies and si impurities lead to the breaking of the planar symmetry and cause significant changes in the electronic properties .
_ et al.__@xcite calculated the magnetic properties of bare and transition - metal ( tm ) doped aln nanosheets by using first - principles calculations .
they reported that nonmagnetic @xmath1-aln can be magnetized upon a single tm atom .
moreover , the electronic structures of nanoribbon of aln were investigated by zheng _ _ et al.__@xcite and it
is predicted that zig - zag edge nanoribbons have an indirect band gap while armchair edge nanoribbons have a direct band gap , and band gaps of both zig - zag and armchair decrease monotonically with increasing ribbon width .
furthermore , bacaksiz _ _ et al.__@xcite reported that the electronic band structure changes significantly when the number of layer increases and the structure having 10 or more layers exhibits direct bandgap as bulk form .
more recently , kecik _ _ et al.__@xcite investigated the optical properties of mono- and few - layer @xmath1-aln under strain .
they reported that the absorption peaks stand outside the visible - light regime , on the other hand , the applied tensile strain gradually redshifts the optical spectra . as a constituent of hetero - bilayer , physical properties of mg(oh)@xmath0 were investigated previously in several studies.@xcite it is a layered metal hydroxide with a wide band gap@xcite . as a member of the alkaline - earth - metal hydroxides ( aemhs ) family , ca(oh)@xmath0 was predicted to be stable in the bilayer and monoayer forms by aierken _ _ et al.__@xcite in addition , torun _ et al .
_ investigated the optical properties of gas - ca(oh)@xmath0 bilayer heterostructure by using gw+bse and reported that in spite of the similarities of electronic structures , the different stacking types have different optical spectra.@xcite recently , tsukanov _ et al .
_ investigated the interaction of organic anions with layered double hydroxide nanosheets consist of mg and al by using molecular dynamics simulations.@xcite in addition , very recently , the monolayer mg(oh)@xmath0 was synthesized and reported that in spite of the optically inactive nature of mg(oh)@xmath0 , the photoluminescence intensity of monolayer mos@xmath0 was assisted by mg(oh)@xmath0 and enhanced@xcite .
most recently , yagmurcukardes _
_ investigated the hetero - bilayer of the mg(oh)@xmath0 and ws@xmath0.@xcite they reported that the lower energy optical spectrum of the mg(oh)@xmath0-ws@xmath0 hetero - bilayer is dominated by the excitons originates ws@xmath0 layers .
these two studies indicate that the mg(oh)@xmath0 is a candidate for tuning the optical property of other monolayer materials . in the present work ,
the heterostructure of two perfectly - matching monolayers of @xmath1-aln , a member of iii - v binary compounds , and mg(oh)@xmath0 , a member of alkaline - earth - metal hydroxide ( aemh ) , are considered .
the similar lattice constants of the layers provide us to use smaller supercell which is important especially for the calculating the optical properties .
we found that the vertically stacked heterostructure possesses excitonic peaks which appear only after the construction of the hetero - bilayer .
more significantly , the wave function calculation shows that the first peak of the imaginary part of the dielectric function for the hetero - bilayer originates from spatially indirect exciton where the electron and hole localized at @xmath1-aln and mg(oh)@xmath0 , respectively .
the paper is organized as follows : in sec .
[ comp ] we give details of our computational methodology . in sec .
[ monolayers ] we present a brief overview of the structural and electronic properties of monolayer @xmath1-aln and mg(oh)@xmath0 . in sec .
[ bilayer_hetero ] different stacking orders of bilayer heterostructure of @xmath1-aln and mg(oh)@xmath0 and also optical property of the energetically favorable stacking order are investigated in detail .
finally , we present our conclusion in sec .
employing ( dft)-based methods , we investigated the the structural , electronic and optical properties of monolayer @xmath1-aln and mg(oh)@xmath0 and their hetero - bilayers .
we used the vienna ab - initio simulation package vasp@xcite which solves the kohn - sham equations iteratively using a plane - wave basis set . to describe electron exchange and correlation , the perdew - burke - ernzerhof ( pbe ) form of the generalized gradient approximation
( gga)@xcite was adopted .
the van der waals ( vdw ) forces which is important for layered materials was taken into account by using the dft - d2 method of grimme.@xcite to obtain partial charge on the atoms , a bader charge analysis was used.@xcite structural optimizations were performed with the following parameters .
the kinetic energy cut - off of the plane - wave basis set was 500 ev in all calculations . the total energy difference between the sequential steps in the iterations
was taken 10@xmath4 units as convergence criterion . the convergence for the hellmann - feynman forces per unit cell atom was taken to be 10@xmath5 ev / .
gaussian smearing of 0.05 ev was used and the pressures on the unit cell were decreased to a value less then 1.0 kb in all three directions . for the determination of accurate charge densities ,
brillouin zone integration was performed using a @xmath6 @xmath3-centered mesh for the primitive unit cell . to avoid interactions between adjacent monolayer and hetero - bilayers ,
our calculations were performed with a large unit cell including @xmath718 vacuum space .
we also calculated the cohesive energy ( @xmath8 ) , which was formulated as @xmath9 / \sum\limits_{i } n_{i } $ ] where @xmath10 stands for the atoms which contract the structure ; @xmath11 , @xmath12 , and @xmath13 are the number of @xmath10 atom in the unit cell , the energy of free @xmath10 atom , and the total energy per unit cell .
( color online ) upper panel illustrates the structure of monolayer @xmath1-aln ( left ) and mg(oh)@xmath0 ( right ) .
( a ) and ( b ) are the structures ( top and side view ) , the solid black parallelograms show the unitcell of the structural , ( c ) and ( d ) are the band structures of @xmath1-aln and mg(oh)@xmath0 , respectively .
the blue and light - red curves with square are for gga+soc and gga+gw@xmath14 , respectively .
the dashed vertical line is shows the fermi energy . in the band diagram , the charge densities of the vbm and cbm and the spin - orbit splitting at the @xmath3 point .
, width=8 ] [ cols="<,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] first of all , to determine energetically favorable structure , six different stacking orders are considered . as shown in fig .
[ hetero_fig](a ) , possible stacking orders are : @xmath15 ( al on mg , n on h ) , @xmath16 ( n on mg , al on h ) , @xmath17 ( al on h , n at the midpoint of three h ) , @xmath2 ( n on h , al at the midpoint of three h ) , @xmath18 ( al and n on top of midpoint of h triangles and n on mg ) , and @xmath19 ( switched of al and n form of @xmath18 ) . the energy difference between different stackings are given in table [ hetero_table ] .
in general , the hetero - bilayer has three type of properties with respect to localization of al or
n on trigonal h surface of mg(oh)@xmath0 ; when n atom is on top of h atom , which means the @xmath2 and @xmath15 stacking types , the hetero - bilayers posses the lowest energies with highest binding energies .
they have also similar band gap of 2.30 and 2.31 ev , respectively .
the work functions are found to be 4.31 and 4.32 ev which are very close to each other . instead of these energetically favorable forms ,
the minimum interlayer distances are obtained when the al and n atoms coincide to the midpoint h triangles ( @xmath19 and @xmath19 ) .
the band gap and work function values of the @xmath19 and @xmath19 stacking are found to be similar as well .
when al is on top of h , the structures have the highest interlayer distances highest energy and lowest binding energy .
the @xmath15 with @xmath2 , @xmath16 with @xmath17 , and @xmath18 with @xmath19 have similar parameters .
the minimum energy of the hetero - bilayer is obtained when the layer are stacked in the form of @xmath2 stacking which exhibits the maximum interlayer interaction of 237 mev .
interlayer distance , which is defined as the perpendicular distance from surface h atom to the @xmath1-aln plane , is 2.13 .
the work function of mg(oh)@xmath0 side is calculated to be 4.31 ev . the vbm and cbm are at @xmath3 as in monolayer mg(oh)@xmath0 , however the band gap values of 2.30 ev within gga which are significantly low as compared to that of monolayer of @xmath1-aln and mg(oh)@xmath0 .
in addition , the @xmath15 stacking order is energetically very close to @xmath2 . for both @xmath2 and @xmath15 , the layer - layer interaction , interlayer distances , and also band gap values are almost same .
in addition , considering vacuum level positions , formation of a staggered gap , with a mismatch of 1.02 ev at the intimate contact point of monolayers is predicted . moreover , since the minimum energy difference between the vb and cb is at @xmath3 point ( see fig .
[ hetero_fig ] ( b ) ) , the lower energy optical activity takes place at this high symmetry point of brillouin zone .
such a contact between two atomically - flat surfaces forms an ultra - clean type - ii heterojunction . in this part ,
the origin of the prominent excitonic peaks in the aln / mg(oh)@xmath0 hetero - bilayer structure , which corresponds to the ab@xmath20 stacking , are investigated in details .
there are two effects that determine characteristic properties of exciton in two - dimensional materials : the dielectric screening and the structural confinement . in general ,
when the crystal structure of a usual semiconductor is reduced from 3d to 2d , the dielectric screening between electron and hole takes place only inside layer .
therefore , as a consequence of the dimensional reduction , the screening effect on electron - hole pair is lower for 2d crystals as compared to the bulk . in fig .
[ dis_osci_fig ] , we show interlayer - spacing - dependent excitonic properties to investigate the origin and evolution of the excitons by tracking their peak positions and corresponding oscillator strengths in the energy spectrum . as shown in fig .
[ dis_osci_fig ] ( a ) , starting from the interlayer spacing of 6.0 ( equilibrium distance is set to 0 ) , the first peak of the isolated mg(oh)@xmath0 splits into two different peaks due to the broken symmetry of the top and bottom surface charge densities .
it is seen that the first peak of the bilayer structure , * i@xmath21 * , appears at the lower energy level than the first peak of both individual @xmath1-aln and mg(oh)@xmath0 .
as calculated in the previous section , the vbm and cbm electronic states are mainly formed by the o and n , respectively .
therefore , the first peak of the bilayer heterostructure , * i@xmath21 * , corresponds to a spatially indirect exciton in which the electron localizes at the n of @xmath1-aln and the hole is at o of mg(oh)@xmath0 .
as shown in fig .
[ dis_osci_fig ] ( b ) , the reduction in the oscillator strength of * i@xmath21 * ( minimum of all ) at around 3.0 , reveals that electronic state of mg(oh)@xmath0 at the intimate contact surface overlaps with @xmath1-aln layer as the distance decreases .
however , at the optimum interlayer distance , other mg(oh)@xmath0-originated exciton state evolves into the second peak of the heterostructure labeled as * ii@xmath21*. apparently , the electronic state of * ii@xmath21 * exciton peak originates from the outer surface state of mg(oh)@xmath0 . on the other hand , the origin and interlayer - spacing - dependence of * iii@xmath21 * , differs from those of * i@xmath21 * and * ii@xmath21*. firstly , in fig .
[ dis_osci_fig ] , it is evident that by decreasing interlayer spacing the first excitonic peak of @xmath1-aln crystal gradually evolves into third state of hetero - bilayer , labeled as * iii@xmath21*. it is seen from figs . [ dis_osci_fig ] ( a ) and ( b ) that the shape and the oscillator strength of the * iii@xmath21 * do not change significantly .
therefore , one can conclude that the characteristic properties of the * iii@xmath21 * are almost same with the first excitonic state of @xmath1-aln monolayer crystal .
when the distance decrease under 4.0 , a significant charge distribution appears on the o atom .
it appears that the screening on the @xmath1-aln increases when the mg(oh)@xmath0 layers become closer .
therefore as a consequence of decrease in exciton binding energy corresponding peak shows a significant blue shift .
( color online ) ( a ) the imaginary part of the dielectric functions of the @xmath2 stacked hetero - bilayer are shown for different interlayer distances starting from ground state distance of 2.13 .
@xmath22 refers to interlayer distance while the ground state distance is set to 0 and labeled as 0.0 .
the lowest three exciton peaks are labeled as * i@xmath21 * , * ii@xmath21 * , and * iii@xmath21*. ( b ) the distance dependent oscillator strengths of lowest three excitons are shown .
color code is given .
the position of the prominent peaks are illustrated by shaded curves .
the length of the out - of - plane lattice vector is fixed for different interlayer distance .
, width=8 ] for a deeper understanding of nature of excitonic states of the vdw mg(oh)@xmath0/@xmath1-aln heterostructure we also investigate the band decomposed charge densities and exctionic wave functions . in fig .
[ chg_dis ] ( a ) , the interlayer - spacing - dependent band decomposed charge densities of the first and second lowest levels of the cb at @xmath3 point are shown .
the lowest and second lowest levels are labeled as cb-@xmath23 and cb-@xmath24 , respectively .
the changes in the charge densities of these two levels contain remarkable informations about the evolution of the electrons involved in the states * i@xmath21 * , * ii@xmath21 * , and * iii@xmath21*. firstly , when the distance is 6.0 , the charge density shown in cb-@xmath24 part of the fig . [ chg_dis ] ( a ) corresponds to the charge density at the cbm of individual mg(oh)@xmath0 .
it is seen that this density appears mostly on o atoms and also at the surfaces of the mg(oh)@xmath0 .
as the interlayer distance decreases , the charge density vanishes at the inner surface and increases at the outer surface . on the other hand ,
the density shown in the cb-@xmath23 for the interlayer distance of 6.0 corresponds to the charge density at the cbm of individual @xmath1-aln .
as the distance decreases , charge changes negligibly at the vicinity of @xmath1-aln but an additional charge density appears on inner o atom when the distance is less than 4.0 .
it appears that when the layers approach each other , the inner surface electron of the mg(oh)@xmath0 is transfered to the @xmath1-aln and the outer surface electron states of the mg(oh)@xmath0 remains at its original position .
furthermore , we also calculate the real space exciton localizations by using the methodology given in appendix [ wavefunc ] .
the wave functions of the * i@xmath21 * , * ii@xmath21 * , and * iii@xmath21 * are shown in fig . [ chg_dis ]
( b ) where the yellow regions correspond to electrons and the black points stand for holes .
as shown in the lower panel , it is evident that the * i@xmath21 * is the spatially indirect exciton in which the electron and hole localize on the @xmath1-aln and mg(oh)@xmath0 , respectively .
the * ii@xmath21 * localizes at the outer side of the mg(oh)@xmath0 as shown in middle panel of the fig .
[ chg_dis ] ( b ) .
it seems that the * ii@xmath21 * is not affected by the @xmath1-aln .
the * iii@xmath21 * shown in upper panel of the fig .
[ chg_dis ] ( b ) stays mostly on the @xmath1-aln .
a small portion of the * iii@xmath21 * appears on the hydroxide regions of the mg(oh)@xmath0 .
in this study , structural , electronic , and optical properties of the recently synthesized novel 2d materials of @xmath1-aln and mg(oh)@xmath0 and their bilayer heterostructure are investigated by performing dft calculation .
the excitonic states are also calculated by solving bse over the @xmath25 approximation on top of dft .
our investigation revealed that : ( i ) the individual monolayer of the @xmath1-aln and mg(oh)@xmath0 can form a vertically stacked hetero - bilayer , ( ii ) when the hetero - bilayer formed , although the interlayer interaction is weak novel exciton states appear , ( iii ) while @xmath1-aln - states have weak spacing dependence , mg(oh)@xmath0 states are strongly affected by the presence of a neighboring layer , ( iv ) exponential increase in screening of mg(oh)@xmath0 on @xmath1-aln - states was also predicted , ( v ) while the first exciton peak is a spatially indirect one , the second and third exciton states are spatially direct states .
although the constituents have limited optical activity , the heterostructure shows unexpected optical properties .
in particular , @xmath1-aln - mg(oh)@xmath0 hetero - bilayer exhibits spatially indirect excitons which is important for the optoelectronic application , especially based on photoexcited electron collecting .
the calculations were performed at tubitak ulakbim , high performance and grid computing center ( tr - grid e - infrastructure ) .
cb and rts acknowledge the support from tubitak project no 114f397 .
hs acknowledges support from bilim akademisi - the science academy , turkey under the bagep program .
hs acknowledges financial support from the scientific and technological research council of turkey ( tubitak ) under the project number 116c073 .
ar and ad acknowledge financial support from the european research council(erc-2015-adg-694097 ) , spanish grant ( fis2013 - 46159-c3 - 1-p ) , grupos consolidados ( it578 - 13 ) , and afosr grant no .
fa2386 - 15 - 1 - 0006 aoard 144088 , h2020-nmp-2014 project mostophos ( ga no . 646259 ) and cost action mp1306 ( euspec )
theoretical results on optical properties , especially many - body and excitonic effects , are strongly depend on computational parameters such as brillouin zone sampling and the vacuum spacing between the adjacent layers .
the calculation with fine parameters , on the other hand , requires large computational resource and time .
therefore , we performed calculations to examine the convergence of the frequency dependent imaginary dielectric function with respect to k - point sampling and the vacuum spacing .
( color online ) the results of the k - point sampling tests : ( a ) and ( b ) the imaginary part of the dielectric functions of the monolayer _ h_-aln and mg(oh)@xmath0 for different k - point samplings , respectively .
( c ) the first exciton peak positions are shown as the blue and red squares for the monolayer _ h_-aln and mg(oh)@xmath0 , receptively.,width=8 ] the imaginary part of the dielectric function for the monolayer _ h_-aln and mg(oh)@xmath0 with respect to k - point sampling are shown in fig .
[ k - point ] ( a ) and ( b ) , respectively .
when the number of k - points increases , the spectrum of both monolayers are blue - shifted and rapidly converges by the 12@xmath2612@xmath261 sampling ( see fig .
[ k - point ] ( c ) ) .
therefore , using 12@xmath2612@xmath261 for optical properties of monolayer _ h_-aln and mg(oh)@xmath0 , provides reliable results . as shown in fig .
[ vacuum ] ( a ) and ( b ) , the effect of the vacuum spacing is also evident that the peak positions of both monolayers are red - shifted . for the _
h_-aln , the curves of 30 and 35 vacuum spacing values are almost the same . for mg(oh)@xmath0 , the curve does not change significantly when the vacuum spacing increases from 25 to 30 .
moreover , as shown in fig .
[ vacuum ] ( c ) , the difference between the first exciton energies of 30 and 35 vacuum spacing values is 11 mev for the monolayer _ h_-aln .
the smaller energy difference of 5 mev is obtained between 25 and 30 vacuum spacing values for the monolayer mg(oh)@xmath0 .
it is clearly seen that , the vacuum spacing of 30 and 25 are sufficient for the monolayers _ h_-aln and mg(oh)@xmath0 , respectively .
therefore , when we consider the hetero - bilayer , using the vacuum spacing of @xmath728 for optical properties provides reliable results . after having convergence tests , the best parameter set of 12x12x1 kpoint and 28 vacuum spacing
are used in our calculations .
( color online ) the results of vacuum spacing tests : ( a ) and ( b ) the imaginary part of the dielectric functions of the monolayer _ h_-aln and mg(oh)@xmath0 for different vacuum spacing values , respectively .
( c ) the first exciton peak positions are shown .
the blue and red curves correspond to the monolayer _ h_-aln and mg(oh)@xmath0 , receptively .
, width=8 ]
for a deeper understanding of the origin of excitonic states and determination of their wave functions of the vdw hetero - bilayer , three lowest exciton wave functions were calculated by using berkeleygw package@xcite on top of quantumespresso code.@xcite the eigenvalues and the eigenfunctions in the independent particle picture were obtained by using pbe form of gga for exchange and correlation . the quasi - particle energies were calculated within the gw approximation .
we employed an energy cutoff of 160 ry for the pbe calculations . for the computation of the dielectric matrix , we used 1987 conduction bands and g - vectors with energy up to 17 ry , whereas the self - energy operator was computed using 1987 conduction bands and a g - vector cutoff of 17 and 160 ry for the screened and bare coulomb matrices , respectively . to sample the bz we employed a @xmath27 k - point grid for the pbe and gw calculations , whereas for the solution of the bse
, we used a @xmath28 k - point grid with linearly interpolated gw quasi - particles energies .
the 8 lowest conduction bands and 8 topmost valence bands were included to solve the bse .
after obtaining the excitonic states , the probability of localization of the electron is calculated for each excitonic level when the hole is fixed to a specific point .
the result are given in fig .
[ chg_dis ] ( b ) .
s. z. butler , s. m. hollen , l. cao , y. cui , j. a. gupta , h. r. guti rrez , t. f. heinz , s. s. hong , j. huang , a. f. ismach , e. johnston - halperin , m. kuno , v. v. plashnitsa , r. d. robinson , r. s. ruoff , s. salahuddin , j. shan , l. shi , m. g. spencer , m. terrones , w. windl , and j. e. goldberger , acs nano * 7 * , 2898 ( 2013 ) .
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 , 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 ) .
p. giannozzi , s. baroni , n. bonini , m. calandra , r. car , c. cavazzoni , d. ceresoli , g. l chiarotti , m. cococcioni , i. dabo , a. d. corso , s. de gironcoli , s. fabris , g. fratesi , r. gebauer , u. gerstmann1 , c. gougoussis , a. kokalj , m. lazzeri5 , l. martin - samos , n. marzari , f. mauri , r. mazzarello , s. paolini , a. pasquarello , l. paulatto , c. sbraccia1 , s. scandolo , g. sclauzero , a. p seitsonen , a. smogunov , p. umari , and r. m wentzcovitch , j. phys .
matter * 21 * , 395502 ( 2009 ) . | motivated by recent studies that reported the successful synthesis of monolayer mg(oh)@xmath0 [ suslu _ et al .
_ , sci . rep .
* 6 * , 20525 ( 2016 ) ] and hexagonal ( _ h_-)aln [ tsipas _ et al_. , appl .
. lett . * 103 * , 251605 ( 2013 ) ] , we investigate structural , electronic , and optical properties of vertically stacked @xmath1-aln and mg(oh)@xmath0 , through _ ab initio _ density - functional theory ( dft ) , many - body quasi - particle calculations within the gw approximation , and the bethe - salpeter equation ( bse ) . it is obtained that the bilayer heterostructure prefers the @xmath2 stacking having direct band gap at the @xmath3 with type - ii band alignment in which the valance band maximum and conduction band minimum originate from different layer . regarding the optical properties ,
the imaginary part of the dielectric function of the individual layers and hetero - bilayer are investigated .
the hetero - bilayer possesses excitonic peaks which appear only after the construction of the hetero - bilayer .
the lowest three exciton peaks are detailedly analyzed by means of band decomposed charge density and the oscillator strength .
furthermore , the wave function calculation shows that the first peak of the hetero - bilayer originates from spatially indirect exciton where the electron and hole localized at @xmath1-aln and mg(oh)@xmath0 , respectively , which is important for the light harvesting applications . |
motif search problems are of central importance for sequence analysis in computational molecular biology .
these problems have applications in fields such as genetic drug target identification or signal finding ( see @xcite and the references cited therein for more details and further applications ) .
two core problems in this context are closest substring @xcite and consensus patterns @xcite : * input : * @xmath1 strings @xmath7 over alphabet @xmath8 and non - negative integers @xmath9 and @xmath10 .
+ + * question in case of closest substring : * is there a string @xmath11 of length @xmath10 , and for @xmath12 , a substring @xmath13 of length @xmath10 such that , for all @xmath14 , @xmath15 ? ( here @xmath16 denotes the hamming distance between @xmath11 and @xmath17 . ) + + * question in case of consensus patterns : * is there a string @xmath11 of length @xmath10 , and for @xmath12 , a substring @xmath13 of length @xmath10 such that , @xmath18 ? + what is currently known about these two problems is summarized as follows . *
the closest substring problem . * 1 .
closest substring is @xmath6-complete , and remains so for the special case of the closest string problem , where the string @xmath11 that we search for is of same length as the input strings .
closest string is @xmath6-complete even for the further restriction to a binary alphabet @xcite .
2 . on the positive side , both closest substring and
closest string admit polynomial time approximation schemes ( ptas s ) , where the objective function is the maximum length of the string @xmath11 @xcite .
3 . in the ptas s for both closest string and closest substring , the exponent of the polynomial bounding the running time depends on the goodness of the approximation .
these are not efficient ptas s ( eptas s ) in the sense of @xcite and therefore are probably not useful for bioinformatics practice . whether eptas s are possible for these approximation problems , or whether they are @xmath19$]-hard ( for the parameter @xmath20 , where the approximation is to within a factor of @xmath21 of optimal ) , currently remains open .
closest string is _ fixed - parameter tractable _ with respect to the parameter @xmath9 , and can be solved in time @xmath22
closest string is also fixed - parameter tractable with respect to the parameter @xmath1 , but here the exponential parametric function is much faster growing , and the algorithm is probably of less practical use ( see , however , @xcite for some encouraging experimental results also in this case ) . *
the consensus patterns problem . * 1 .
consensus patterns is @xmath6-complete and remains so for the restriction to a binary alphabet @xcite .
2 . consensus patterns admits a ptas @xcite , where the objective function is the maximum length of the string @xmath11 .
3 . the known ptas s for consensus patterns
are not eptas s , and whether eptas s are possible , or whether ptas approximation for this objective function is @xmath19$]-hard , is an important issue that also currently remains open .
the key distinguishing point between closest substring and consensus patterns lies in the definition of the distance measure @xmath9 between the `` solution '' string @xmath11 and the substrings of the @xmath1 input strings .
whereas closest substring uses a maximum distance metric , consensus patterns uses the sum of distances metric .
this is of particular importance when discussing values of parameter @xmath9 occurring in practice . whereas it makes good sense for many applications to assume that @xmath9 is a fairly small number in case of closest substring , this is much less reasonable in the case of consensus patterns
this will be of some importance when discussing our result for consensus patterns .
many algorithms applied in practice try to solve motif search problems exactly , often using enumerative approaches in combination with heuristics @xcite . in this paper
, we explore the parameterized complexity of the basic motif problems in the framework of @xcite . * our main results . * unfortunately , our main results are negative ones : we show that closest substring and consensus patterns are w[1]-hard with respect to the parameter @xmath1 of the number of input strings , even in case of a binary alphabet . for unbounded alphabet size ,
we show that the problems are @xmath0$]-hard for the combined parameters
@xmath10 , @xmath9 , and @xmath1 . in the case of constant alphabet size
, the complexity of the problems remain open when parameterized by @xmath9 and @xmath1 together , or by @xmath9 alone .
note that in the case of consensus patterns our result gains particular importance , because here the distance parameter @xmath9 usually is not small , whereas assuming that @xmath1 is small is reasonable . until now
, it was known only that if one additionally considers the substring length @xmath10 as a parameter , then running times exponential in @xmath10 can be achieved @xcite . an overview on known parameterized complexity results for closest substring and consensus patterns is given in table @xmath23 .
.overview on the parameterized complexity of @xmath24 and consensus patterns with respect to different parameterizations , where @xmath1 is the number of given strings , @xmath10 is the length of the substrings we search for , and @xmath9 is the hamming distance allowed .
results from this paper are marked by @xmath25 .
the @xmath26 results for constant size alphabet can be achieved by enumerating all length @xmath10 strings over @xmath8 .
open questions are indicated by a question mark . [ cols="^,^,^",options="header " , ] as is obvious from these distance values , the indicated substrings in the choice strings all have hamming distance @xmath27 to the solution string and , therefore , are matches for @xmath11 . for the reverse direction , we assume that the @xmath24 instance has a solution .
we need the following statements : [ fronttag ] a solution @xmath11 and all its matches in the input instance start with the front tag .
since @xmath11 is of length @xmath28 , the only possible match in the template string is the template string itself .
therefore , @xmath29 can differ from @xmath30 in at most @xmath27 symbols .
we can show that the only substrings in a choice string @xmath31 that are possible matches for @xmath11 with hamming distance at most @xmath9 start with the front tag , as we argue in the following . since @xmath11 is a solution
, there is a match in @xmath31 and we denote it by @xmath32 .
denote the the first @xmath33 symbols of @xmath32 by @xmath34 .
since @xmath35 and @xmath36 , we necessarily ( triangle inequality for hamming metric ) have @xmath37 .
we show that this is only possible when @xmath34 coincides with a front tag of a block of @xmath31 . assuming that it does not
, we will show that @xmath38 , a contradiction .
firstly , assume that the starting position of @xmath34 and the starting position of a front tag in @xmath31 differ by @xmath39 positions , @xmath40 .
then , at least @xmath41 symbols `` 0 '' of @xmath30 are aligned with symbols `` 1 '' of the front tag in @xmath34 and at least @xmath41 symbols `` 1 '' of @xmath30 are aligned with symbols `` 0 '' of @xmath34 .
this implies @xmath42 .
secondly , assume that the starting position of @xmath34 and the starting position of its closest front tag in @xmath31 differ by @xmath43 positions .
then , a block of @xmath44 symbols `` 1 '' falls onto the encoding and/or the back tag part of @xmath34 .
since the encoding part and back tag contain together only @xmath45 ( under the assumption that @xmath46 ) many symbols `` 1 '' , we have more than @xmath47 mismatching symbols and @xmath48 . summarizing
, we conclude that @xmath34 coincides with a front tag in choice string @xmath49 , i.e. , @xmath50 .
[ exactlyk ] the encoding part of @xmath11 contains exactly @xmath1 symbols `` 1 '' .
assume that @xmath11 has less than @xmath1 symbols `` 1 '' in its encoding part , i.e. , @xmath51 contains less than @xmath1 symbols `` 1 '' .
then , because @xmath52 , @xmath53 , implying @xmath54 , a contradiction .
assume that @xmath11 has more than @xmath1 `` 1 '' symbols in its encoding part @xmath51 .
then , @xmath55 for the encoding part @xmath56 of a match in every choice string @xmath31 .
now consider the solution s back tag @xmath57 . to achieve @xmath58
, we need @xmath59 and @xmath57 must contain one or more symbols `` 1 '' .
every `` 1 '' symbol in @xmath57 will decrease the value @xmath60 for a block @xmath32 of _ one _ choice string @xmath31 by one , but will increase the solution s hamming distance to the selected blocks of _ all other _ choice strings .
no matter how many `` 1 '' symbols we have in the back tag of @xmath11 , there will always be a choice string @xmath31 with @xmath61 . in summary
, we will always have a choice string @xmath31 with @xmath62 , a contradiction .
[ onepersec ] every section of the encoding part of @xmath11 contains exactly one symbol `` 1 '' .
assume that not every section in the encoding part of @xmath11 contains exactly one `` 1 '' symbol .
then , there must be a section containing no symbol `` 1 '' , since , by lemma @xmath63 , the number of symbols `` 1 '' in the encoding part of @xmath11 adds up to @xmath1 .
let @xmath64 , @xmath65 , be the section containing no symbol `` 1 '' .
, consider a choice string @xmath66 , @xmath67 or , if @xmath68 , a choice string @xmath66 , @xmath69 . in
_ every _ block @xmath70 of @xmath66 , sections @xmath64 and @xmath71 of the encoding part are active and , therefore , contain exactly one symbol `` 1 '' each ; these are the only symbols `` 1 '' in @xmath72 . now consider the @xmath1 symbols `` 1 '' in the encoding part of @xmath11 : the `` 1 ' 's in all sections of @xmath73 except for section @xmath71 are all aligned with `` 0 ' 's in @xmath72 ; within section @xmath71 , only a single `` 1 '' of @xmath73 can be matched to a `` 1 '' of @xmath72 .
therefore , @xmath74 . as in the proof of lemma @xmath63
, we conclude that @xmath11 is no solution .
[ reversek ]
the @xmath1 symbols `` 1 '' in the solution string s
encoding part correspond to a @xmath1-clique in the graph .
let @xmath11 be a solution for the closest substring instance .
summarizing , we know by lemma @xmath75 that @xmath11 can have as a match only one of the choice string s blocks . by lemma
@xmath76 , every section of the encoding part @xmath51 contains exactly one `` 1 '' symbol ; therefore , we can read this as an encoding of @xmath1 vertices of the graph .
let @xmath77 be these vertices .
further , we know that the back tag @xmath57 consists only of `` 0 '' symbols : by lemma @xmath63 , the encoding part @xmath78 has only @xmath1 `` 1 ' 's ; would @xmath57 contain a `` 1 '' , then we would have @xmath79 .
we have @xmath80 for _ every _ choice string match @xmath32 and , since every @xmath56 contains only two `` 1 '' symbols , @xmath81 .
now consider some @xmath82 and the corresponding choice string @xmath31 . since @xmath11 is a solution , we know that there is a block @xmath32 with @xmath83 .
that means that the two `` 1 '' symbols in @xmath56 have to match two `` 1 '' symbols in @xmath51 ; this implies that the two vertices @xmath84 and @xmath85 are connected by an edge in the graph .
since this is true for all @xmath82 , vertices @xmath86 are pairwisely interconnected by edges and form a @xmath1-clique .
propositions @xmath87 and @xmath88 yield the following main theorem : @xmath24 is @xmath0$]-hard for parameter @xmath1 in the case of a binary alphabet .
our techniques for showing hardness of closest substring , parameterized by the number @xmath1 of input strings , also apply to consensus patterns .
because of the similarity to closest substring , we restrict ourselves to explaining the problem and pointing out new features in the hardness proof . given strings @xmath7 over alphabet @xmath8 and integers @xmath9 and @xmath10 , the consensus patterns problem asks whether there is a string @xmath11 of length @xmath10 such that @xmath89 where @xmath13 is a length @xmath10 substring of @xmath90 .
thus , consensus patterns aims for minimizing the _ sum _ of errors .
since errors are summed up over all strings , the value of @xmath9 will , usually , not be a small and , therefore , the most significant parameterization for this problem seems to be the one by @xmath1 .
the problem is @xmath6-complete and has a ptas @xcite . by reduction from clique
, we can show @xmath0$]-hardness results as for closest substring given unbounded alphabet size .
we omit the details here and focus on the case of binary input alphabet .
we can apply basically the same ideas as were used in section [ par_k ] ; however , some modifications are necessary . * choice strings . * as in subsection [ reduction_k ] , we generate a set of @xmath92 _ choice strings _
@xmath93 , @xmath94 with @xmath95 , encoding the @xmath96 edges of the input graph .
this time , however , every block consists only of a front tag and an encoding part .
no back tag is necessary .
therefore , we use @xmath97 , in which the encoding part @xmath98 is constructed as in subsection [ reduction_k ] . before we explain the front tags , we already fix the distance value @xmath9
. * distance value . *
we set the distance value @xmath99
. * front tags . * the front tag is now given by @xmath100 .
thus , the front tag has length @xmath101 .
the front tag here is more complex than the one used in subsection [ reduction_k ] .
the reason is as follows .
its purpose is to make sure that a substring which is not a block can not be a match . to achieve this
, the front tag lets such an unwanted substring necessarily have a distance value larger than @xmath9 to a possible solution ( as explained in the proof of lemma @xmath75 ) . since @xmath9 has a higher value here compared to section [ par_k ] , we need the more complex front tag
. * solution length .
* we set the substring length to the length of one block , i.e. , the sum of @xmath101 ( the length of the front tag ) and @xmath102 ( the length of the encoding part ) .
therefore , @xmath103 .
* template strings . * in contrast to subsection [ reduction_k ] , we produce not only one but @xmath104 many template strings . all template strings have length @xmath10 , i.e. , the length of one block .
the template strings are a concatenation of the front tag part ( as given above ) and an encoding part consisting of @xmath102 many symbols `` 1 '' . in summary
, the front tag ensures that only the block of a choice string can be selected as a substring matching a solution . regarding the distribution of mismatches ,
we note that a closest substring s front tag part will not cause any mismatches . in its encoding part ,
every of its @xmath102 positions causes at least @xmath104 mismatches .
it causes exactly @xmath104 mismatches for every position iff the input graph contains a @xmath1-clique .
[ onedircp ] for a graph with a @xmath1-clique , the construction in subsection @xmath105 produces an instance of consensus patterns which has a solution , i.e. , there is a string @xmath11 of length @xmath10 such that every @xmath31 , @xmath106 , has a substring @xmath32 with @xmath107 . given an undirected graph @xmath108 with @xmath109 vertices and @xmath96 edges , let @xmath110 be the indices of @xmath1-clique s vertices .
then , let string @xmath11 consist of the front tag described in the above construction , concatenated with the encoding part @xmath111 , which encodes all clique vertices . for every @xmath82 , we choose in choice string @xmath31 the block @xmath32 encoding the edge connecting vertices @xmath84 and @xmath85 .
we will show that these blocks have exactly total hamming distance @xmath112 to @xmath11 .
the front tags of @xmath11 and of each @xmath32 coincide , their hamming distance is @xmath113 .
recall from subsection [ reduction_k ] that the encoding parts consist of @xmath1 sections , each section of length @xmath109 .
we consider the encoding parts section by section and , within a section , columnwise . given a section @xmath64 , @xmath65 , there are @xmath114 choice strings in which this section is active , and this section in these blocks encodes vertex @xmath115
. consider the column at position @xmath116 in this section , over all selected substrings and all template strings .
we have @xmath104 `` 0 '' symbols from the choice strings in which this section is inactive ; in all other strings , there is a `` 1 '' at this position . in @xmath11 ,
this position is `` 1 , '' causing @xmath104 mismatches .
now consider the remaining columns of section @xmath64 .
in each of them , we have @xmath104 `` 1 '' symbols from the template strings ; all @xmath92 choice strings have `` 0 '' at the corresponding position . in @xmath11 , this position is `` 0 , '' causing @xmath104 mismatches .
thus , we have @xmath104 mismatches at every of the @xmath109 positions within a section , and this is true for all @xmath1 sections of the encoding part .
the sum of distances from @xmath11 to the matches in choice strings and the template strings is @xmath117 ; @xmath11 is a solution .
for the reverse direction , we use two lemmas to show important properties that a solution of the constructed instance has . the first lemma is proved in analogy to lemma @xmath75 .
[ fronttagcp ] a solution @xmath11 and all its matches in the input instance start with the front tag . the second property of a solution , although also valid for the solutions in subsection [ correctnessk ] , is established in a different way here .
it relies on the additional template strings that have been introduced in the construction of the consensus patterns instance .
[ onepositions ] a solution @xmath11 contains exactly one symbol `` 1 '' in every section of its encoding part .
let @xmath11 be a solution for the constructed consensus patterns instance .
by lemma @xmath118 , we know that @xmath11 and all its matches in the choice strings start with the front tag
. consequently , the matches in the choice strings must be blocks .
consider the encoding part of a solution @xmath11 together with the encoding parts of its matches in the input strings .
we note that we have at least @xmath104 mismatches for every column at positions @xmath39 , @xmath119 : on the one hand , all @xmath104 template strings have `` 1 '' symbols at position @xmath39 . on the other hand , all @xmath104 choice strings in which position
@xmath39 s section is inactive have `` 0 '' at this position , no matter which blocks we chose in these choice strings . since @xmath11 is a solution and only a total of @xmath112 mismatches
are allowed , we have _ exactly _ @xmath104 mismatches for every position of the encoding part of @xmath11 with the corresponding positions in the matches of @xmath11 .
now , consider an arbitrary section @xmath64 , @xmath120 , and consider all @xmath114 choice strings in which section @xmath64 is active . in these choice strings ,
section @xmath64 contains exactly one `` 1 '' symbol .
we will show that in these choice strings blocks that form the matches for @xmath11 , the `` 1 '' in section @xmath64 must be at the same position in all matches , because , otherwise , @xmath11 is no solution .
assume that we chose blocks in which the `` 1 '' symbols of section @xmath64 are at different positions .
we can easily check that this would cause more than @xmath104 mismatches for the columns corresponding to the positions of the `` 1 '' symbols ; this contradicts the assumption that @xmath11 is a solution .
we conclude that , for all matches in choice strings , the `` 1 '' symbols of section @xmath64 must be at the same position . for columns in which we have `` 1 '' symbols in choice strings
, there is a majority of `` 1 '' symbols , namely those in the @xmath121 choice strings in which section @xmath64 is active and those in the @xmath104 template strings .
therefore , the respective position in @xmath11 must be `` 1 . '' for all other columns , there is a majority of `` 0 '' symbols , namely those in all @xmath92 choice strings . therefore , the respective position in @xmath11 must be `` 0 . ''
these two lemmas allow us to show that also the reverse direction of the reduction is correct .
[ reversecp ]
the @xmath1 symbols `` 1 '' in the solution string s encoding part correspond to a @xmath1-clique in the graph .
let @xmath11 be a solution for the constructed consensus patterns instance .
by lemma @xmath122 , every section in the encoding part of @xmath11 encodes a vertex of the input graph . in the following ,
we show that all encoded vertices are interconnected by edges .
let @xmath123 be the vertices encoded in the solution s encoding part .
for every two sections @xmath82 , we select in choice string @xmath31 a substring in which the `` 1 '' symbols of sections @xmath124 and @xmath71 are at the same positions as the `` 1 '' symbols of sections @xmath124 and @xmath71 in the solution : selecting another substring would result in a hamming distance greater than @xmath125 in the @xmath126th and @xmath127th column and @xmath11 could not be a solution . hence , the selected block encodes the edge connecting @xmath84 and @xmath85 . since we find such a substring for every @xmath82 , every pair of vertices in @xmath128 is connected by an edge , @xmath128 is a @xmath1-clique .
propositions @xmath129 and @xmath130 yield the following main result .
[ mcpthm ] consensus patterns is @xmath0$]-hard for parameter @xmath1 in case of a binary alphabet .
we have proven that closest substring and consensus patterns , parameterized by the number @xmath1 of input strings and with alphabet size two , are @xmath0$]-hard .
this contrasts with related sequence analysis problems , such as longest common subsequence @xcite and shortest common supersequence @xcite , where , until now , parameterized hardness has only been established in the case of unbounded alphabet size .
now , it is also known that these problems , parameterized by the number of input strings , are w[1]-hard in case of bounded alphabet size @xcite . in our opinion , however , intuitively speaking , our w[1]-hardness result for consensus patterns is the most surprising one in this context , because consensus patterns seems to carry significantly less combinatorial structure than the other problems .
the parameterized complexity of closest substring and consensus patterns , parameterized by `` distance parameter '' @xmath9 , remains open for alphabets of constant size .
if these problems are also @xmath0$]-hard , then an efficient and practically useful ptas would appear to be impossible @xcite , unless further structure of natural input distributions is taken into account in a more complex aggregate parameterization of these basic computational problems of bioinformatics . | we show that closest substring , one of the most important problems in the field of biological sequence analysis , is @xmath0$]-hard when parameterized by the number @xmath1 of input strings ( and remains so , even over a binary alphabet ) .
this problem is therefore unlikely to be solvable in time @xmath2 for any function @xmath3 of @xmath1 and constant @xmath4 independent of @xmath1 .
the problem can therefore be expected to be intractable , in any practical sense , for @xmath5 .
our result supports the intuition that closest substring is computationally much harder than the special case of closest string , although both problems are @xmath6-complete .
we also prove @xmath0$]-hardness for other parameterizations in the case of unbounded alphabet size .
our @xmath0$]-hardness result for closest substring generalizes to consensus patterns , a problem of similar significance in computational biology . |
quantum phase transitions(qpt ) , as opposed to the classical transitions , driven by quantum rather than thermal fluctuations , occur at absolute zero and are induced by the change of some coupling constant or physical parameter .
these phenomena have captured the attention of many researchers during the last decade . in cavity qed networks , the trade off between the mott insulator ( mi ) and superfluid ( sf )
phases have been specially studied . in the present context ,
these are not truly `` phases '' in the thermodynamic sense , since in our model , we deal only with few cavities . however , it has been shown that even small systems , containing a reduced number of cavities display the superfluid - mott transition @xcite .
the importance of this transition is that it enables the system to go from a delocalized state ( mi ) , with the excitations equally distributed throughout the system , to the state where the excitations move freely ( sf ) , and more importantly , all the excitations could be localized in just one cavity . in the literature ,
e.g. , persists a common conviction that , there is a key parameter that controls this transition , which is the atom - cavity detuning , @xmath0 . in this work
we demonstrate the importance of other parameters like damping rate , hopping rate and quantum correlation quantifiers ( e.g. negativity ) in detecting and controlling the mentioned phases . in many problems of matter - light interactions ,
it is more convenient to model the atom and cavity field together as a quasi particle called polariton " .
these polaritons are the dressed states of the jaynes - cummings model . however , in the polariton basis , the variation of @xmath0 leads to a change in the nature of this basis , e.g. in the limit of very large @xmath0 , the polaritonic state goes to a purely photonic state , with the atoms in the ground state .
therefore , it is meaningless to talk about the sf polaritonic phase in that limit .
this problem was addressed by irish @xmath1 @xcite .
nevertheless , a qpt is still possible for small variations of @xmath0 .
recently , a similar qpt was found in @xcite , by controlling the atom - cavity coupling constant , @xmath2 .
for a fixed detuning , a variation of @xmath2 leads also to a change in the polaritonic basis , but not for @xmath3 , in which case , the states remain always maximally entangled between light and matter .
all these transitions require an external control parameter , e.g. an external laser acting on each cavity and splitting the atomic levels by stark shift , which increases @xmath0 .
on the other hand , quantum correlations ( qc ) and their role in quantum information is already well known @xcite and a great effort has been devoted to the understanding of the connection between qc and qpt , see @xcite and the references therein .
in particular , qc play an important role detecting quantum phase transitions @xcite . in our case , since we are dealing with qudits , our best candidate for computing the quantum correlations is the negativity @xcite . in this work , we show a critical phenomenon , which is closely related to a qpt , with the losses playing a crucial role .
we consider as initial a superfluid like state , and show that it evolves to a mott insulator like state during the time evolution of the system , only when the interaction with a reservoir is on . as in many models ,
we consider that the main source of dissipation originates from the leakage of the cavity photons due to imperfect reflectivity of the cavity mirrors . on the other hand ,
spontaneous emission will be neglected , assuming long atomic lifetimes , when dealing with the conditions imposed in our cavity qed system .
this paper is organized as follows , we first briefly discuss the mapping of the jaynes - cummings - hubbard model to the polaritons .
then we study the transition from superfluid like state to mott insulator like state or self - trapping effect triggered by dissipation using the quantum correlations as a witness .
finally , we analyze the critical damping and discuss the results .
in our model we consider a linear array of n coupled cavities , where each cavity supports a single field mode and contains a single two - level atom .
photons are allowed to hop between neighboring cavities .
the hamiltonian of such a system , found in @xcite and references therein , is given by eq.(@xmath4 ) @xmath5 \nonumber \\
& + & \sum_{j=1}^{n-1}j_j[a_j^{\dagger } a_{j+1 } + a_{j+1}^{\dagger}a_j],\end{aligned}\ ] ] where @xmath6 and @xmath7 are the ground and excited states of the two - level atom with transition frequency @xmath8 , @xmath9(@xmath10 ) is the creation(annihilation ) operator of the cavity mode @xmath11 , @xmath12 is the coupling strength between the atom and the cavity mode , @xmath13 is the coupling strength ( hopping ) between the neighboring cavities and @xmath14 represents the number of cavities .
for each cavity , the eigenstates of the first three terms of the hamiltonian ( [ fullh ] ) are the dressed states , which are known as polaritons @xcite .
these states are given by @xmath15 with @xmath16 , @xmath17 , and @xmath18 corresponds to the number of photons inside each cavity .
one would expect , since the first three terms of the hamiltonian ( [ fullh ] ) can be written in the polariton basis , that there is a map to rewrite the hopping term ( fourth term ) in the polaritonic basis .
we derive this mapping in a different way as compared to @xcite , but the results are similar , both being a generalization of the one proposed in @xcite .
we found that the creation operator , in the polaritonic basis reads @xmath19 the lowering operator @xmath20 , destroys the state @xmath21 . on the other hand ,
the raising operator @xmath22 creates the state @xmath21 .
it is easy to see that with these definitions , all the transitions relative to a state @xmath23 will correspond to the upwards and downwards unit jumps .
the first two terms in ( [ adagger ] ) correspond to the transitions between subspaces of the same sign , e.g. from @xmath24 to @xmath25 .
the last two terms are related to inter converting the subspace , i.e. from the positive subspace to the negative one , and vice versa .
we notice that the sum for these last terms start at @xmath26 , since there is no inter converting transition from one to zero
. it will be convenient for future notation , to define each term in the above equation as @xmath27,@xmath28,@xmath29 and @xmath30 , in the same order that they appear .
the coefficients are found to be @xmath31 next , we perform some approximations and discuss which of the terms in the above equation can be neglected and under what conditions .
the inter converting operators , @xmath32 and @xmath33 , are the first candidates to be neglected . for only one excitation ,
this operators vanish , since @xmath34 start from two excitations . also , for a higher subspace
, it is easy to see that the factor @xmath35 decreases considerably .
the worst setting is for the second subspace @xmath36 , if we chose @xmath37 , then @xmath38 . if comparing it with other coefficients , i.e. @xmath39 , we realize that the inter converting operators are not so important during the evolution , and in most of the cases they are completely negligible .
taking into account the previous discussion , the hopping hamiltonian can be written as @xmath40\ ] ] following the idea of reference @xcite and the conditions in @xcite , one can apply the rotating wave approximation in order to eliminate the terms of the type @xmath41 , since they are fast rotating as compared to @xmath42 .
finally , our simplified hopping term in the hamiltonian , becomes : @xmath43\ ] ] next , if one chooses the initial condition of a particular type , for example @xmath44 , then the state @xmath45 will never show up , and we are allowed to consider , throughout the paper , the lower branch only .
it is worth noting that for zero detuning ( @xmath46 ) and restricting the cavities to only the first excited state @xmath47 , we recover our previous results @xcite .
however , for finite detuning and allowing the cavities to have more than one excitation , the dynamics becomes more involved , getting new and interesting results .
to fix the ideas , we imagine starting our system in a superfluid like state , where all excitations are in a single cavity .
as the system evolves in time , there is a probability that these excitations will be uniformly distributed throughout the system , leading to a mott insulator like state .
this transition is strongly dependent on the detuning .
when the system reaches the mott insulator like state @xmath48 , the trapping takes place due to the blockade effect , that is maximized when @xmath3 . in fig.([fig1 ] ) we show the time evolution of the probabilities of finding the initial state @xmath49 and @xmath48 , when considering only two cavities without losses , for two different cases : with and without detuning . on one hand
, we notice that for @xmath3 , there is delay in the propagation of the initial state to the second cavity , result already discussed in some previous works @xcite .
although , in those papers the authors did not study the effect of the detuning to modify ( shorten ) this delay , as shown in fig.([fig1 ] ) , for @xmath50 .
we can explain this improvement in the propagation time based on the fact that when using the polaritons basis , it is easy to see that an effective hopping shows up , which not only depends on @xmath51 but also on the coefficient @xmath52 in eq.([coefficient ] ) , which increases with the detuning . on the other hand , for the case of zero detuning ,
the probability of finding the state @xmath48 is zero and for @xmath50 this probability oscillates with a very small amplitude , as seen from the fig.([fig1 ] ) . in what follows
, we will describe a more realistic situation , with each cavity connected to its individual reservoir and consider the same previous initial state , but this time for only the case of zero detuning .
rather than using the traditional master equation approach , we consider the system evolving with a non - hermitian hamiltonian , interrupted once in a while by instantaneous quantum jumps , process usually referred to as the quantum trajectory " or
monte - carlo wave function method " .
@xcite .
fig.([fig2 ] ) shows a self - trapping effect , where the sf like state @xmath49 goes to a mi like state during the time evolution .
it is interesting to notice that the presence of the interaction of the reservoir with the cavities triggers the effect , and there is no need of an external control since the system always goes to the same state . for a small loss rate ,
the polariton @xmath53 oscillates , going back and forth from the first to the second cavity and vice versa , for certain period of time until a probability associated to the state @xmath48 starts growing and eventually reaching a slowing decaying plateau .
nevertheless , if we increase the losses to a critical value , the oscillations vanish and the state @xmath48 appears faster .
we also show the time evolution of the negativity and observe that at the critical damping , the oscillations of the probability associated to the @xmath49 state die out and there is a single sharp peak in the negativity at the time region where the transition between the states @xmath49 and @xmath48 takes place .
we also observe that the critical damping rate depends on the coupling constant between the cavities @xmath51 , as well as the detuning @xmath0 .
the key feature in this work is the presence of losses . as we pointed out in the previous section , these lead to sudden transitions between two states . furthermore , we are able to find a critical loss rate , @xmath54 , for which there is no reflection or oscillation of the probability associated to the polariton @xmath53 .
this implies a break in the periodicity of the system , where the damping rate is large enough to eliminate the oscillation . in the following ,
we investigate the peculiarities of @xmath54 , and notice that it is a function of the coupling strength between the cavities , and also the detuning .
in this subsection , we focus on the case @xmath3 , with @xmath55 and postpone the discussion on the @xmath0 dependence to the conclusions .
as was mentioned before , the negativity can help us to throw some light on the search of criticality . in fig.([fig3 ] ) we present two cases , with @xmath56 and @xmath57 .
the case @xmath58 was discussed in fig.([fig2 ] ) .
on one hand , when we are below the critical value , there are oscillations of the initial state , leading to two or more peaks in the negativity . on the other hand , when going near or above @xmath54 , we find a single peak , but as we increase the damping constant , it displaces to the left in time , with a decreasing maximum . considering these two facts
, we numerically estimate the critical damping to satisfy the relation @xmath59 , that corresponds to a negativity with a single peak at its maximum value .
thus , we can clearly identify two regions as shown in fig.([fig4 ] ) , a lower area that corresponds to an oscillating probability associated to the initial state and the upper region with no oscillations and where the self trapping takes place with a single peaked negativity .
in this work , we demonstrated for a system of two coupled cavities with dissipation to individual reservoirs , the occurrence of self - trapping effect , implying a transition from a superfluid to a mott insulator phase .
further , we found a critical damping rate , above which , the initial superfluid states damps out in time while the mott insulator phase is being created .
we find the negativity to be a very good witness that shows , at the critical damping rate , a single peak at its maximum value in the same time region where the transition takes place .
we notice that in the lossless case , the initial state oscillates back and forth , that is , the two excitations oscillate between the two cavities and the mi phase @xmath48 will never be reached .
hence , we can assert that the true triggering mechanism for this self - trapping effect is the presence of the dissipation in the cavities . as was discussed above
, increasing the detuning also increases the effective hopping , and therefore one would expect a similar dependence of the critical losses with @xmath0 as with @xmath51 . even though , this is only true for small variation of @xmath0 ,
as we increase it , the slope of the straight line shown in fig.([fig4 ] ) becomes higher .
moreover , as we increase @xmath0 further , a rapid oscillation of the negativity shows up , which makes it impossible to identify one single peak . for even larger detuning ,
say @xmath60 , the state is almost photonic ( there is no entanglement between light and matter ) , implying that there will be no self - trapping at all , since the atoms are in the ground state and photons can hop freely .
furthermore , we studied the behavior of the transition when adding more cavities to the system , up to four . in that case
, the mi like state ( @xmath61 and @xmath62 for three and four cavities respectively ) shows up at a later time ( as compared to the two cavity case ) , which is to be expected since the system is bigger thus it takes more time to propagate through the array .
meanwhile , the decay of the initial state ( @xmath63 and @xmath64 ) is more abrupt , showing a sign that we are getting closer to a real phase transition in the thermodynamic sense .
this whole analysis was done at zero temperature . in a near future ,
we plan to investigate the thermal effects on these transitions . @xcite .
we acknowledge the financial support of the fondecyt projects no.1140994 , the project conicyt - pia anillo act-112 , `` red de analisis estocastico y aplicaciones '' , pontificia universidad catolica de chile , universidad diego portales and conicyt doctoral fellowships .
we also thank professor m.loewe for providing extra computer facilities . in a model with fiber coupled cavities
, we found that actually thermal effects can increase the correlations between atoms , in the cavities .
see v.eremeev , v.montenegro , m.orszag , phys.rev.a,*85 * , 032315 ( 2012 ) . | in a coupled cavity qed network model , we study the transition from a localized superfluid like state to a delocalized mott insulator like state , triggered by losses . without cavity losses ,
the transition never takes place .
further , if one measures the quantum correlations between the polaritons via the negativity , we find a critical cavity damping constant , above which the negativity displays a single peak in the same time region where the transition takes place .
additionally , we identify two regions in the parameter space , where below the critical damping , oscillations of the initial localized state are observed along with a multipeaked negativity , while above the critical value , the oscillations die out and the transition is witnessed by a neat single peaked negativity . |
carbon monoxide ( co ) , the most common molecule after @xmath4 , is routinely detected in external galaxies .
when exposed to photons of energy greater than 11.1 ev , co is readily photodissociated into atomic carbon and oxygen . as the c@xmath3 ionization potential is only 11.3 ev
, i.e. quite close to the co dissociation energy , neutral carbon may be quickly ionized .
because carbon monoxide , its isotopes , as well as neutral and ionized carbon respond differently to ambient conditions , observations of the relative emission strengths of @xmath2 , @xmath1 , c@xmath3 and c@xmath5 provide significant information on the physical condition of cloud complexes from which the emission arises .
even though far - infrared continuum and [ cii ] lines are much more efficient coolants overall , co and [ ci ] lines are especially important for the temperature balance of cool and dense molecular gas .
co may be observed from the ground in many transitions .
emission from c@xmath5 has been observed towards numerous galaxies from the kao and iso platforms .
emission from c@xmath3 can also be measured from the ground but only under excellent atmospheric conditions . to date , the total number of galaxies detected in [ ci ] is about 30 .
most of the extragalactic 492 ghz [ ci ] measurements are summarized in just two papers , by gerin @xmath7 phillips ( 2000 ) using the cso and by israel @xmath7 baas ( 2002 ) using the jcmt , both on mauna kea , hawaii .
these papers include references to more detailed discussions of individual galaxy results .
_ almost all galaxies mapped thus far show strong concentrations of both atomic carbon and molecular gas well contained within radii @xmath8 kpc . _
the area - integrated co and [ ci ] luminosities of the observed galaxy centers cover a large range .
quiescent centers ( ngc 7331 , ic 342 , maffei 2 , ngc 278 , ngc 5713 ) have modest [ ci ] luminosities @xmath9 } \leq 5 \kkms$ ] kpc@xmath10 ( with @xmath11kpc@xmath10 corresponding to @xmath12 w ) . starburst nuclei ( ngc 253 , ngc 660 , m 82 , ngc 3628 , ngc 6946 ) have luminosities @xmath13 } \leq 40 \kkms$ ] kpc@xmath10 , except m 83 which has only @xmath14 } = 3.6 \kkms$ ] kpc@xmath10 .
the highest luminosities , @xmath14 } \geq 50 \kkms$ ] kpc@xmath10 , are found around the active nuclei of ngc 1068 and ngc 3079 . in previously observed galactic photon - dominated regions ( pdrs ) , the intensities of @xmath15 and @xmath0=21 @xmath1 line emission
were generally found to be very similar ( cf .
keene et al . 1996 ; kaufman et al . 1999 ) .
such ratios of @xmath15 and @xmath0=21 @xmath1 close to or less than unity are considered to be characteristic of the effects of enhanced uv radiation on molecular gas in pdrs .
however , fig.[profiles ] shows that galaxy centers may have much stronger [ ci ] emission . in fig.[cilum ]
we present a full comparison of the intensities of the 492 ghz [ ci ] line , the 461 ghz @xmath16 line and the 220 ghz @xmath0=21 @xmath1 line . the @xmath15 line is stronger than @xmath0=21 @xmath1 in all except three galaxies . the highest [ ci]/@xmath1 ratios of about five belong to the active galaxies ngc 1068 and ngc 3079 . generally , the @xmath15 line is weaker than the @xmath0=43 @xmath2 line , but not by much . _
thus , galaxy centers have much stronger [ ci ] emission than the galactic pdr results would lead us to expect . _ in the galaxy sample independently observed by gerin @xmath7 phillips ( 2000 ) more than two thirds also have a ratio [ ci]/@xmath1 @xmath17 .
low ratios are expected in high - uv environments , and in environments with high gas ( column ) densities where virtually all carbon will be locked up in co. in the galaxy , high ratios are found in environments with low gas column densities and mild uv radiation fields , such as exemplified by translucent clouds and at cloud edges .
there , @xmath2 and especially @xmath1 will be mostly dissociated , and gas - phase atomic carbon may remain neutral
. in dense clouds , [ ci ] may be relatively strong if the dominant heating mechanism is some other than uv photons .
the data presented here and in gerin @xmath7 phillips ( 2000 ) suggest that most of _ the emission from galaxy centers does not come from very dense , starforming molecular cloud cores or pdr zones_.
very approximate column densities may be calculated assuming optically thin [ ci ] and @xmath1 emission in the high - temperature lte limit , but accurate results are obtained only by more detailed radiative transfer calculations .
curves illustrating the possible physical condition of gas clouds , shown in fig.[ratios ] , were calculated with the leiden radiative transfer models ( see http://www.strw.leidenuniv.nl/radtrans/ ) .
the four panels correspond to molecular gas densities ranging from @xmath18 to @xmath19 . in each panel , mostly horizontal tracks mark co gradients @xmath20(co)/d@xmath21 = 0.3 , 1.0 and 3.0 @xmath22 @xmath23 for [ c@xmath3]/co abundances of 0.1 , 0.3 , 1.0 and 3.0 respectively and kinetic temperatures ranging from 150 k ( left ) to 10 k ( right ) are marked .
lines of constant kinetic temperature would be mostly vertical in these panels .
the diagrams contain points representing the observed galaxy centers as well as star - forming regions ( white @xmath7 sandell 1995 ; israel @xmath7 baas , unpublished ; bolatto et al .
2000 ) and the milky way center ( fixsen , bennett @xmath7 mather 1999 ) .
the star - forming pdrs have relatively low neutral carbon versus co abundances ( c@xmath3)/co ) @xmath25 0.10.3 ) . _ most galaxy centers ( especially the active ones ) have significantly higher abundances exceeding unity _ , independent of the assumed gas parameters .
the diagonal distribution of galaxy centers roughly follows lines of constant kinetic temperature .
the actual temperatures @xmath26 and molecular gas densities @xmath27 can not be determined indepently .
for instance , if we assume @xmath28 , we find @xmath29 , whereas by assuming @xmath30 we find more modest values @xmath31 k similar to the dust temperatures 33 k @xmath32 52 k found for these galaxy centers by smith @xmath7 harvey ( 1996 ) . the only available direct determination , for m 82 by stutzki et al .
( 1997 ) , yields @xmath33 and @xmath26 = 50 100 k , in very good agreement with the above . where the molecular line emission in most galaxy centers appears to arise from warm , dense gas ( as opposed to either hot and tenuous or cold and very dense gas ) , the centers of ngc 7331 , m 51 and ngc 4826 seem to be cooler independent of assumed gas density .
fig.[ciifir ] combines [ ci ] line intensities with those of [ cii ] and the far - infrared continuum ( for references , see israel @xmath7 baas 2002 ) .
the results resemble those obtained by gerin @xmath7 phillips ( 2000 ) .
there is no longer a clear distinction between the various types of objects .
the [ cii]/fir ratio increases with [ ci]/fir , but the [ cii]/[ci ] ratio _ decreases _ as [ ci]/fir increases . an upper limit for [ cii ]
places the merger galaxy ngc 660 in the same diagram positions as the ultraluminous mergers arp 220 and mrk 231 observed by gerin @xmath7 phillips .
the highest [ cii]/fir ratios are reached for pdr model gas densities @xmath34 ( gerin @xmath7 phillips 2000 ) , but fully half of the ratios in fig.[ciifir ] are much higher than predicted by the pdr models . in low - metallicity environments , this may be explained by longer mean free pathlengths of energetic uv photons ( israel et al .
1996 ) . however , this is not a credible explanation for high - metallicity ( see zaritsky , kennicutt @xmath7 huchra 1994 ) galaxy centers .
ideally , the observations should be modelled by physical parameters varying as a function of location in a complex geometry . practically , we may attempt to approach such a reality by assuming the presence of a limited number of distinct gas components .
the analysis of multitransition @xmath2 , @xmath1 and [ ci ] observations of half a dozen galaxy centers ( israel @xmath7 baas 2003 and references therein ) suggests that , within the observational errors , good fits to the data can be obtained by modelling with combinations of dense / cool and tenuous / warm gas components . at the same time , however , the energy requirements of keeping relatively large mass fractions at relatively high temperatures , and the peculiar [ ci]/[cii ] ratios again suggest that exploration of excitation models other than pdrs might be fruitful .
bolatto a.d .
, jackson j.m . , israel f.p . , zhang x. , @xmath7 kim s. 2000 545 , 234 fixsen d.j .
, bennet c.l . , @xmath7 mather j.c .
1999 526 , 207 gerin m. , phillips t.g .
2000 537 , 644 israel f.p . ,
maloney p.r . , geis n. , et al .
1996 465 , 738 x israel f.p . , @xmath7 baas f. 2002 383 , 82 israel f.p .
, @xmath7 baas f. 2003 404 , 495 kaufman m.j .
, wolfire m.g . ,
hollenbach d.j . ,
@xmath7 luhman m.l .
1999 527 , 795 keene j. , lis d.c . , phillips t.g .
, @xmath7 schilke p. 1996 in ` molecules in astrophysics ' , iau symp .
178 , ed . e.f .
van dishoeck , p. 129
smith b.j . , @xmath7 harvey p.m. , 1996 468 , 193 stutzki j. , graf u.u . , haas s. , et al .
1997 373 , 423 white g.j . , @xmath7 sandell , g. 1995 , 299 , 179 zaritsky d. , kennicutt r.c . , @xmath7 huchra j.p 1994 420 , 87 | measurements of [ ci ] , @xmath0=21 @xmath1 and @xmath0=43 @xmath2 emission from quiescent , starburst and active galaxy centers reveal a distinct pattern characterized by relatively strong [ ci ] emission .
the [ ci ] to @xmath1 emission ratio increases with central [ ci ] luminosity .
it is lowest in quiescent and mild starburst centers and highest for strong starburst centers and active nuclei .
c@xmath3 abundances are close to , or even exceed , co abundances .
the emission is characteristic of warm and dense gas rather than either hot tenuous or cold very dense gas .
the relative intensities of co , [ ci ] , [ cii ] and far - infrared emission suggest that the dominant excitation mechanism in galaxy centers may be different from that in photon - dominated regions ( pdrs ) . |
during their propagation to earth , ultra - high energy cosmic rays ( uhecrs ) can suffer different interaction processes and be deflected by the pervasive cosmic magnetic fields , which directly affects the spectrum and chemical composition observables of these particles .
therefore , to address the question concerning the origin and nature of uhecrs , one has to simulate their propagation taking into account all the relevant interaction and energy loss processes , as well as the effects of cosmic magnetic fields .
if uhecrs have an extragalactic origin , the intervening cosmic magnetic fields can play an important role on propagation , and increase the trajectory length of the particles in such a way that cosmological effects , such as the adiabatic expansion of the universe and the redshift evolution of the photon backgrounds , can have a relevant contribution to the energy losses .
weak magnetic fields should only slightly increase the trajectory length for nearby sources , but should have considerable effects for distant sources . depending on the model of extragalactic magnetic field ( egmf )
used this increase in path length causes larger deflections . by propagating uhecrs through the large scale structure of the universe ( lss ) obtained from magnetohydrodynamical ( mhd ) simulations , dolag _ et al . _
@xcite concluded that even for heavy nuclei the expected deflection due to the cosmic magnetic fields are of the order of a few degrees , whereas sigl _
et al . _
@xcite concluded that for protons the deflections can be small in some specific regions of the sky , but are overall larger for heavier nuclei .
these results are crucial to understand whether it is possible to do astronomy with uhecrs . aiming for an interpretation of the experimental data , it is possible to simulate the propagation of uhe particles and check whether the theoretical predictions match the observations .
the main observables that propitiate this comparison are the spectrum , composition and anisotropy .
the shape of the spectrum is not affected by magnetic fields if the sources are identical and uniformly distributed , with separation distance much less than the characteristic propagation lengths , as stated by the propagation theorem @xcite .
if these conditions are satisfied , then the spectrum has a universal form .
the chemical composition of uhecrs is still unknown , and the experimental data hint towards contradictory results .
the pierre auger collaboration reported a gradual increase in mass number towards higher energies @xcite , whereas results from hires are compatible with a proton composition at the highest energies @xcite .
therefore , when searching for realistic scenarios of propagation of uhecrs the mass composition also has to be explained , together with the spectrum .
the third important observable is the anisotropy . in this sense ,
a realistic scenario for uhecr propagation should explain , apart from the spectrum and composition , also the observed ( an)isotropy .
however , this observable is not robust , since it is directly affected by the galactic magnetic field ( gmf ) . in the recent gmf model proposed by jansson and farrar @xcite
the average deflection is strongly non - uniform across the sky , and can be @xmath05.2@xmath1 for 60 eev protons .
one should also add the deflections induced by the egmf , and hence the combination of galactic and extragalactic deflections can strongly affect the anisotropy observable . in the present work we use the same lss simulation used by sigl _
_ @xcite . to propagate the uhecrs through the universe we have used the crpropa 3 code @xcite . in this code
there are two different modes of propagation available , namely a unidimensional ( 1d ) and a tridimensional case ( 3d ) .
the redshift evolution of the backgrounds and energy losses due to the adiabatic expansion of the universe are taken into account in the 1d case , but not in the 3d case .
the reason for this is that if magnetic fields are considered in a 3d simulation , the information regarding the effective trajectory length of the particles , and therefore the redshift , is not known beforehand . in the present work
we address this problem by using an approximate method to correct the 3d simulations taking into account redshift losses .
this work is divided as follows : in section [ sec : propagation ] we describe the main energy loss processes that take place at uhe ; in section [ sec : method ] we present a method used to account for cosmological effects in 3d simulations ;
in section [ sec : applications ] we apply this method to simulated data ; finally , in section [ sec : conclusions ] we summarize the work and present the final remarks and futures perspectives .
uhecrs lose energy during the propagation to earth due to four main processes : pair production , pion production , and nuclear decay / photodisintegration ( in the case of nuclei ) and adiabatic expansion of the universe .
the pair production is a highly inelastic process and causes a continuous energy loss that can be analytically computed .
the photopion production interactions for nuclei can be numerically treated with the sophia code @xcite .
nuclear decay comprises @xmath2 and @xmath3 decays , as well as nucleon dripping .
photodisintegration is the dominant energy loss process that takes place at ultra - high energies for nuclei heavier than hydrogen .
the adiabatic expansion of the universe is also a source of energy loss and is relevant for particles that transversed cosmological distances .
the cosmological framework used here assumes a flat universe with matter and dark energy only , ignoring the radiation component .
this assumption is valid for @xmath4 .
the matter and dark energy densities are respectively @xmath5 and @xmath6 , and the hubble constant is @xmath7 km / s / mpc .
the two main photon fields that should be taken into account in the propagation of uhecrs in the universe are the cmb and the cib .
the treatment of the cmb is straightforward , since it can be done analytically and its redshift evolution is given by a factor @xmath8 with respect to the photon density of this background at @xmath9 .
the cib has to be treated numerically , but its redshift evolution is assumed to be the same as the cmb within the framework of crpropa .
as discussed before , the redshift dependence of the photon backgrounds , namely the cmb and cib , and the evolution of the sources can affect the spectrum and composition of the detected cosmic rays .
also , the expansion of the universe causes energy losses . from a 3d simulation one can obtain the following quantities : initial energy of the particle ( @xmath10 ) , final energy of the particle ( @xmath11 ) , initial @xmath12 and final @xmath13 particle type , initial particle position @xmath14 and final particle position @xmath15 , effective propagation time ( @xmath16 ) , among others . to introduce redshift effects in the 3d simulation ,
we resimulate each individual particle in 1d , using the following input from the 3d simulation : @xmath10 , @xmath12 and @xmath16 . for each resimulated particle
we obtain a set of subproducts from the photodisintegration of the mother - nucleus .
let @xmath17 be an index associated to each one of these subproducts .
then the output of the 1d simulation provides us the final energy @xmath18 and the final particle type @xmath19 . as a first approximation we can choose @xmath17 uniformly among all daughter - nuclei .
the next step of this method is to replace the final energy and type of the particle in the 3d simulation by the one obtained from the 1d simulation , by doing @xmath20 and @xmath21 . to check this method ,
we have applied it to the trivial case of a 3d simulation without magnetic fields .
the sources were assumed to be uniformly distributed up to 4000 mpc and to have a spectral index of -2.2 . in the first case we considered only protons , whereas in the second we have considered only iron .
the parameters of the 1d simulation are exactly the same , so that the only difference between the 1d and 3d cases are the cosmological effects .
the simulated spectra in 1d and 3d , as well as the corrected one , can be seen in figure [ fig : speccheck ] for a pure iron ( upper ) and pure proton ( lower ) composition .
the simulations presented here assumed the source distribution from ref .
@xcite according to a differential energy spectrum that follows a power law of spectral index @xmath22 , reweighted afterwards to @xmath23 . for the 3d case the observer , assumed to be
a sphere of radius 0.5 mpc , lies within a cubic box concatenated with several copies of itself up to 2 gpc , assuming periodic boundary conditions .
the propagation is performed by integrating the lorentz force using an integrator with adaptative step sizes in the range between 10 kpc and 1 mpc .
the relevant interactions , namely pion production , photodisintegration , pair production and redshift losses ( only for the 1d case ) were used .
the astrophysical backgrounds considered were the cmb and cib .
the model by kneiske _
et al . _
@xcite was used to treat the interactions with cib .
the redshift evolution of the photon fields was also done according to this reference . to understand the effects of cosmic magnetic fields on the spectrum
, we have performed simulations with and without the magnetic field from ref .
a side view of this magnetic field box is shown in figure [ fig : miniatifield ] .
all the simulations were performed for two mass compositions , namely pure iron and pure proton .
the results of these simulations are shown in figure [ fig : simspectra ] , together with the corresponding universal spectra and the experimental spectrum measured by the pierre auger observatory @xcite , for the sake of comparison .
if the source distribution is not uniform , then the previously described propagation theorem does not predict a universal spectrum , and therefore a difference between the spectrum in the cases with and without magnetic field is expected .
this can be seen in figure [ fig : spectracomparison ] .
the effects of magnetic fields on the spectrum are much larger in the iron than in the proton scenario , since the charge of the cosmic ray produced at the source in the former case is 26 times the proton charge , thus implying deflection angles between 1 and 26 times larger , due to the containment of these particles around the source . in figure
[ fig : spectracomparison ] it is also possible to compare the original 3d spectrum with the one corrected for cosmology .
there are visible effects of cosmology for both simulated mass compositions scenarios .
this happens because when the cosmology correction is applied , many events from the lower energy tail of the spectrum will suffer another energy loss process not present in the original 3d simulation , and will drop below the minimum energy threshold , set to 1 eev in our simulation .
we are also interested in the change in mass composition due to magnetic fields and cosmology effects .
this is analyzed only in the pure iron scenarios , since the proton case is trivial .
the mean value of the depth of the shower maximum ( @xmath24 ) using the parametrization from ref .
@xcite is used as an observable for the mass composition .
the value of @xmath24 for different energy ranges are shown in figure [ fig : composition ] , together with the experimental values measured by the pierre auger observatory @xcite and the theoretical predictions assuming the epos 1.99 hadronic interaction model . in this figure
it can be seen that even though cosmological effects are relevant , magnetic fields are responsible for the largest change in composition at lower energies . for the simulated data , compared to measurements of the pierre auger observatory ( black circles ) and the theoretical predictions according to the epos 1.99 model ( filled area ) for protons ( green filled area ) and iron ( pink filled area ) . the scenarios without magnetic field
are represented by empty markers , whereas the scenarios with magnetic fields are represented by filled markers .
squares correspond to the simulated data set with the cosmology correction , and the triangles without it . ] in figure [ fig : aniso ] the mean deflection as a function of the energy is shown .
the differences in the energy dependence of the deflection due to magnetic fields and cosmology are small . in case of isotropy ,
the mean value of the deflection would be 90@xmath1 .
we notice that for lower energies the simulated data sets are highly isotropic , becoming more anisotropic at higher energies .
in this work we have presented a new method to account for cosmological effects in 3d simulations of propagation of uhecrs .
this method allows us to perform 3d simulations of the propagation of uhecrs , including heavy nuclei , considering all relevant energy loss and interaction processes , as well as the expected magnetic field and baryon density obtained from mhd simulations .
we applied this method to the same mhd simulation used by sigl _
et al . _
@xcite for a scenario with a pure proton composition , and another one with pure iron .
we have shown the effects of cosmology and magnetic fields , separately , upon the spectrum and observed mass composition observables . from the plots previously presented it is clear that a 3d simulation is required , since magnetic fields affect the shape of the spectrum .
it is also clear that cosmological effects play an important role in the spectrum and mass composition observables .
our results have strong implications for cosmic ray astronomy . in our particular example
the deflections are so large that it is impossible to trace back the sources , especially if the fraction of heavier nuclei is large . however , the results of this work are bounded by the baryon density and magnetic fields obtained from a non - constrained mhd simulation , which was done assuming a given set of initial conditions .
the effects of these assumptions on the propagation of uhecrs are not totally known .
future studies using new and constrained mhd simulations are currently underway and may favor or disfavor the results here presented regarding the possibility of doing cosmic ray astronomy .
this work was supported by the deutsche forschungsgemeinschaft through the collaborative research centre sfb 676 , by bmbf under the grant 05a11gu1 , by the helmholtz alliance for astroparticle phyics ( hap ) funded by the initiative and networking fund of the helmholtz association , and by the forschungs- und wissenschaftsstiftung hamburg through the program `` astroparticle physics with multiple messengers '' . | the origin , propagation , and mechanisms of acceleration of the ultra - high energy cosmic rays ( uhecrs ) are not yet well understood . aiming for a better interpretation of the available experimental data ,
these data have to be confronted with theoretical models .
a realistic simulation of the propagation of uhecrs in the universe should take into account all the relevant energy loss processes due to the interaction with astrophysical backgrounds , as well as the intervening cosmic magnetic fields .
cosmological effects , such as the redshift dependence of the photon backgrounds and the adiabatic expansion of the universe can play an important role in the aforementioned processes . here
we present results of simulations of the propagation of uhecr through the large scale structure of the universe considering cosmological and magnetic field effects simultaneously .
[ firstpage ] ultra - high energy cosmic rays , large scale structure of the universe , propagation of uhecrs |
the memristor is an electronic device predicted by l. chua in 1971@xcite and experimentally obtained a few years ago.@xcite these devices are characterized by distinct resistance states which could be used to store information as they show very fast switching and high retention times.@xcite there are reports of a wide range of materials ( particularly metal oxides ) which show this memristive effect.@xcite although the switching mechanism , and the origin of the different resistance states has not been completely elucidated , it is generally believed that the presence of oxygen deficient phases plays a major role in the memristive effect.@xcite within the possible materials , tio@xmath2 is one of the most studied ones.@xcite inside memristive devices , ti@xmath0o@xmath1 channels immersed in the tio@xmath2-based matrix have been observed , and the formation of such structures is believed to be largely responsible for the change in resistivity.@xcite while tio@xmath2 is known to be a wide gap semiconductor ( experimental band gap @xmath3 ev@xcite ) ti@xmath0o@xmath1 presents three phases with different electronic transport properties .
the low temperature ( @xmath4 140 k ) and intermediate temperature phases ( 140 k @xmath5 150 k ) are semiconductors while the high temperature phase ( @xmath6 150 k ) is metallic.@xcite the differences between those phases is mainly due to small atomic displacements which do not lead to significant changes in the unit cell . despite the large amount of theoretical work on either tio@xmath2 rutile @xcite and ti@xmath0o@xmath1,@xcite the band alignment of these two materials
is not known up to now , mainly because of the difficulty in building a supercell containing the ti@xmath0o@xmath1-tio@xmath2 interface - the two materials present very different unit cells , making the match between them very difficult .
this information could be very useful to the understanding of the atomic level mechanism of the memristor , given that the ti@xmath0o@xmath1-tio@xmath2 interface spans a great part of the device length after the formation of the conducting channels .
as the total energy is given with respect to an arbitrary reference energy in periodic calculations , a common reference should be determined for all the structures involved in this work.@xcite in fact , the knowledge of this common reference is necessary to find the offset between the valence band of different materials .
the knowledge of this band offset is important , for instance , in determining whether there is charge transfer at the interface between the two systems , the materials doping limits@xcite as well as other properties .
for instance , scanlon _ _ et al.__@xcite determined the band offset between rutile and anatase tio@xmath2 structures via a cluster approach , providing an understanding about the separation of charge carriers in photocatalysis using the mixed - phase structure .
our objective in this paper is to study the interface of tio@xmath2 and one of its oxygen deficient counterparts , ti@xmath0o@xmath1 . for this purpose
, we built a supercell containing the ti@xmath0o@xmath1-tio@xmath2 interface which was then used for our calculations . in particular
we analyze the band alignment , which has been found to be of type ii ( the conduction band minimum ( cbm ) is located in one material and valence band maximum ( vbm ) in the other ) , and its consequences to the overall electronic structure of the combined system .
[ fig : struct - ti4o7 ] [ fig : struct - tio2 ] [ fig : struct - interface ]
all density functional theory ( dft@xcite ) calculations in this work were performed using the vasp package.@xcite wave functions are described within the projector augmented waves ( paw ) scheme@xcite with periodic boundary conditions .
the @xmath7 and @xmath8 electrons were considered as valence electrons for ti and o atoms respectively and the pbe functional@xcite was used as the exchange - correlation potential .
the cutoff energy for the plane wave expansion was 700 ev and the k - point meshes used throughout this work were generated using a @xmath9 @xmath10-centered grid .
spin polarization was taken into account for all calculations .
relaxation of the ions was performed while the unit cell was kept fixed , until forces were smaller than @xmath11 ev / , for the isolated materials and @xmath12 ev / , for the supercell .
we have performed test calculations with a smaller supercell for the interface , but with more strict parameters for the convergence of the forces .
the results were similar for quantities calculated in this work .
a hubbard @xmath13 parameter was introduced for a better description of the ti(d ) electrons .
our earlier work @xcite shows that a value of 5 ev for @xmath13 is enough to achieve a good description of the electronic structure of ti@xmath14o@xmath15 ( @xmath16 ) magnli phases .
even though , test calculations were performed for other values of @xmath13 ( 0 , 2 and 4 ev ) and the same qualitative results were obtained for the type of alignment as well as the relative position of the unnocupied levels of the same materials . in the rotationally invariant scheme as implemented in vasp@xcite only the value of @xmath17 ( @xmath18 is the spherically averaged matrix element of the screened coulomb interaction between electrons ) is of relevance , thus @xmath18 was set to zero for all calculations .
one of the key points in this work is that building the interface requires matching two of the three unit cell vectors of each compound at the interfacial region . for that to be accomplished , it is necessary to find the smallest supercell where the deformation imposed to each compound is such that no significant change in electronic structure with respect to the bulk material arises in regions away from the interface . for the materials presented in this work ,
this is difficult , since the tio@xmath2 and the ti@xmath0o@xmath1 presented very distinct unit cells . the stable form of tio@xmath2 at ambient conditions is the rutile structure ( tetragonal unit cell , @xmath19 , @xmath20@xcite ) while the ti@xmath0o@xmath1 presents a complicated magnli structure ( triclinic cell , see table [ tab : struct ] ) . *
1m0.1 * 4>m0.2 & ht & it & lt & tr + @xmath21 ( ) & 5.60 & 5.60 & 5.63 & 5.46 + @xmath22 ( ) & 7.12 & 7.13 & 7.20 & 7.14 + @xmath23 ( ) & 20.43 & 20.36 & 20.26 & 20.70 + @xmath24 ( @xmath25 ) & 67.70 & 67.76 & 67.90 & 65.52 + @xmath26 ( @xmath25 ) & 57.16 & 57.44 & 57.69 & 57.22 + @xmath27 ( @xmath25 ) &
108.76 & 109.02 & 109.68 & 108.46 + to overcome this difficulty , we followed steps that lead to the derivation of the magnli structure from the rutile one as follows .
the first step is to obtain the magnli phase structures from tio@xmath2 rutile via the unit cell transformation@xcite @xmath28 & = & \left [ \begin{array}{ccc } -1 & 0 & 1 \\ 1 & -1 & 1 \\ 0 & 0 & 2n-1 \end{array } \right ] & \left [ \begin{array}{c } a_r \\ b_r \\ c_r \end{array } \right ] , \end{array } \label{eq : transf}\ ] ] where @xmath29 and @xmath30 are the unit cell vectors of the magnli structure of index @xmath31 .
the result of using this operation for the case when @xmath32 is illustrated in fig .
[ fig : struct - tio2 ] and the comparison with the ti@xmath0o@xmath1 unit cell is presented in table [ tab : struct ] . notice that this still represents a tio@xmath2 rutile structure , but
the unit cell is now the one from ti@xmath0o@xmath1 . from this point
, the ti@xmath0o@xmath1 magnli phase , as well as all magnli phases ( ti@xmath14o@xmath15 , @xmath33 ) can be generated via the operation@xcite @xmath34 , \label{eq : displ}\ ] ] that consists of periodic displacements ; after every @xmath31-th ( 121 ) plane , the atoms are translated by @xmath35 $ ] crystal vector of the rutile unit cell .
this operation leads to the formation of extended defects , which are composed of oxygen vacancies ( @xmath36 s ) restricted to planes known as shear planes .
the resulting arrangement is presented in fig .
[ fig : struct - ti4o7 ] .
since the operation given by equation [ eq : displ ] affects only the atomic positions , both structures present the same triclinic unit cell .
in essence , the transformation presented above results in unit cells with equal lengths .
the two materials can thus be interfaced in the plane that is secant to the @xmath23 axis ( highlighted in fig .
[ fig : struct - interface ] ) .
this ti@xmath0o@xmath1-tio@xmath2 interface is , by construction , strain free , defect free , presents exactly the same cell parameters and consequently no dipoles would arise at the heterojunction . in our calculations we created a supercell along the @xmath23 axis by repeating the cell for each material four times ( @xmath37 ) in order to ensure convergence of the potential away from the interface .
vector for all structures presented in this work : isolated tio@xmath2 and ti@xmath0o@xmath1 , and the interface . ]
after obtaining the three structures - both isolated materials and their interface - it is possible to calculate the natural band offset between the two systems studied in this work .
the band offset will be given by @xcite @xmath38 where @xmath39 and @xmath40 are the differences between the valence band maximum energy ( @xmath41 ) and the electrostatic potential averaged over a certain volume of the unit cell of the isolated materials , while @xmath42 is the difference between the averaged electrostatic potential of the different materials calculated in bulk - like regions of the relaxed heterostructure supercell .
the determination of the quantities needed by equation [ eq : band - offset ] can be achieved by , first , obtaining the electronic structure of each isolated ( bulk ) system . from this calculation ,
the average electrostatic potential along the crystal axis which is secant to the interface ( the @xmath23 axis , in this case ) is determined , @xmath43 as well as the position in energy of the vbm , @xmath41 for each system . using the heterostructure after ionic relaxation we also calculated @xmath44 in a similar fashion . from the plot of the electrostatic potential for all three structures ( shown in figure [ fig : potential ] ) it is possible to define regions , labeled @xmath45 and @xmath46 , where both the bulk materials and the heterostructure present similar characteristics .
these regions of the heterostructure are considered as bulk - like and should have , in principle , the same potential .
the average electrostatic potential is then obtained by the integration of @xmath44 along @xmath45 and @xmath46 for ti@xmath0o@xmath1 and tio@xmath2 respectively , and is used to calculate the first and second terms of equation [ eq : band - offset ] . from the heterostructure calculation , the same quantities are used to obtain the final term .
the third step would comprise the determination of the band energy shift due to the hydrostatic pressure . in this particular case
our heterostructure is completely strain free , as it is built by equal unit cells , thus this last step presents no contribution . after those steps were carried out , the band offset obtained in our case resulted in @xmath47 ev .
numerical errors due to the choice of a particular @xmath13 were estimated to be @xmath48 0.3 ev , by comparison with other test calculations performed using other values of @xmath13 .
even though the value of the valence band offset presented a variation with respect to the value of @xmath13 used , being smaller for smaller @xmath13 , the overall electronic structure of the interface has not changed , and the final result , which is the relative position of electronic levels of tio@xmath2 and ti@xmath0o@xmath1 , remained practically unchanged .
for this reason , we have chosen to report only the results for @xmath13 = 5 ev .
the projected density of states ( pdos ) was calculated for the two isolated structures and the results were shifted in energy by the band offset .
these results are shown in fig .
[ fig : pdos ] . from that graph , it is possible to notice that the vbm of the interface is located on ti@xmath0o@xmath1 , while the cbm is found in the tio@xmath2 , resulting in a type ii heterostructure .
rutile and ti@xmath0o@xmath1 .
gray shaded areas denote the occupied levels ( below @xmath49 , which is highlighted by the dotted vertical lines ) and the two spin components are pictured by positive and negative values along the vertical axis .
the upper panel shows the band alignment between the two materials . ]
given that the cbm is 0.16 ev above the vbm ( upper panel , fig .
[ fig : pdos ] ) , it is possible that electrons would leave the ti@xmath0o@xmath1 structure and arrive at the unoccupied levels at the tio@xmath2 structure due to the presence of an electric field . in this case , the ti@xmath0o@xmath1 channels present inside of the tio@xmath2 matrix that forms the memristor could act as donors to the neighboring region of that matrix .
this specific characteristic does not depend on the choice of @xmath13 . while it is known that the isolated oxygen vacancy in the tio@xmath2 rutile acts as a shallow donor,@xcite as to our knowledge , this same behavior has not been reported for the ordered defects present into the ti@xmath0o@xmath1 structure .
o@xmath1-tio@xmath2 interface after ionic relaxation .
the blue plane points out the interface region of the structure , being composed of ti@xmath0o@xmath1 to the left and tio@xmath2 to the right of the line.,title="fig : " ] [ fig : cbm ] o@xmath1-tio@xmath2 interface after ionic relaxation .
the blue plane points out the interface region of the structure , being composed of ti@xmath0o@xmath1 to the left and tio@xmath2 to the right of the line.,title="fig : " ] [ fig : vbm ] further evidence of the real - space separation of the vbm and cbm is given by the projected charge density over the frontier orbitals .
the projection was performed for the last occupied and first unoccupied bands , at the @xmath10 point . as a result , presented in fig .
[ fig : parchg ] , it is possible to notice that while the vbm ( fig [ fig : vbm ] ) is located mainly on the ti atoms of the ti@xmath0o@xmath1 structure close to the interface ( with a contribution of some of the ti atoms from tio@xmath2 ) , the cbm ( fig .
[ fig : cbm ] ) is located exclusively on the tio@xmath2 segment .
some hybridization may be responsible for the contribution coming from the tio@xmath2 structure to the vbm , as both levels present ti(d ) character ( see fig [ fig : pdos ] ) .
another explanation for this contribution would be the length of the @xmath23 axis , which is not long enough to describe the separation of the frontier orbitals .
indeed , calculations performed with a smaller supercell ( the supercell @xmath23 vector length was half of the one used in the calculations presented ) resulted in larger charge delocalization . therefore , we expect that , in the limit of longer unit cells ( closer to realistic systems ) the charge would be fully separated on the two phases .
in conclusion , we have reported the band offset given by the average of the electrostatic potential for the ti@xmath0o@xmath1-tio@xmath2 heterostructure .
the supercell containing the junction was built using transformations that resulted in the same unit cell for both materials present on this junction .
the localization of the vbm and cbm was determined to be at the ti@xmath0o@xmath1 and tio@xmath2 structures respectively , resulting in a type ii heterostructure .
as the band gap is 0.16 ev , our results indicate that the ti@xmath0o@xmath1 channels inside of the memristor tio@xmath2 matrix would then act as donors for the tio@xmath2 structure .
d. s. jeong , r. thomas , r. s. katiyar , j. f. scott , h. kohlstedt , a. petraru , and c. s. hwang , emerging memories : resistive switching mechanisms and current status , rep .
75 , 076502 ( 2012 ) .
b. j. choi , d. s. jeong , s. k. kim , c. rohde , s. choi , j. h. oh , h. j. kim , c. s. hwang , k. szot , r. waser , b. reichenberg , and s. tiedke , resistive switching mechanism of tio2 thin films grown by atomic - layer deposition , j. appl
98 , 033715 ( 2005 ) .
x. l. wang , p. s. ku , q. shao , w. f. cheng , c. w. leung , and a. ruotolo , magnetism as a probe of the origin of memristive switching in p - type antiferromagnetic nio , appl .
phys . lett .
103 , 223508 ( 2013 ) .
kim , y - j .
baek , y. jin choi , c. jung kang , h. ho lee , h - m .
kim , k - b .
kim , and t - s .
yoon , investigation of analog memristive switching of iron oxide nanoparticle assembly between pt electrodes , j. appl .
114 , 224505 ( 2013 ) .
huang , j - s .
huang , c - c .
lai , h - w .
huang , s - j .
lin , and y - l .
chueh , manipulated transformation of filamentary and homogeneous resistive switching on zno thin film memristor with controllable multistate , acs appl .
& interf . 5 , 601723 ( 2013 ) .
x. he , y. yin , j. guo , h. yuan , y. peng , y. zhou , d. zhao , k. hai , w. zhou , and d. tang , memristive properties of hexagonal wo3 nanowires induced by oxygen vacancy migration , nanosc .
lett . 8 , 50 ( 2013 ) .
kwon , k. m. kim , j. h. jang , j. m. jeon , m. h. lee , g. h. kim , x .- s .
li , g .- s .
park , b. lee , s. han , m. kim , and c. s. hwang , atomic structure of conducting nanofilaments in tio2 resistive switching memory , nature nanotech . 5 , 148 153 ( 2010 ) . d. o. scanlon , c. w. dunnill , j. buckeridge , s. a. shevlin , a. j. logsdail , s. m. woodley , c. r. a. catlow , m. j. powell , r. g. palgrave , i. p. parkin , g. w. watson , t. w. keal , p. sherwood , a. walsh , and a. a. sokol , band alignment of rutile and anatase tio2 , nature mat . 12 , 798801 ( 2013 ) .
s. l. dudarev , g. a. botton , s. y. savrasov , c. j. humphreys , and a. p. sutton , electron - energy - loss spectra and the structural stability of nickel oxide : an lsda+u study , phys . rev .
b 57 , 15051509 ( 1998 ) .
li , a. walsh , s. chen , w - j .
yin , j - h .
yang , j. li , j. l. f. da silva , x. g. gong , and s - h .
wei , revised ab initio natural band offsets of all group iv , ii - vi , and iii - v semiconductors , appl .
94 , 212109 ( 2009 ) . | the formation of conducting channels of ti@xmath0o@xmath1 inside tio@xmath2-based memristors is believed to be the origin for the change in electric resistivity of these devices .
while the properties of the bulk materials are reasonably known , the interface between them has not been studied up to now mostly due to their different crystalline structures . in this work we present a way to match the interfaces between tio@xmath2 and ti@xmath0o@xmath1 and subsequently the band offset between these materials
is obtained from density functional theory based calculations .
the results show that while the valence band is located at the ti@xmath0o@xmath1 , the conduction band is found at the tio@xmath2 structure , resulting into a type ii interface . in this case
, the ti@xmath0o@xmath1 would act as a donor to the tio@xmath2 matrix . |
the majority of stars form in clusters of 100 or more members @xcite .
the dynamics and distribution of stars within clusters have important ramifications for planet formation , stellar multiplicity , mass segregation , cluster disruption , and any other property of stars and planets that can be influenced by interactions between stars in the dense environment of a young cluster .
for example , @xcite show that close encounters with neighboring stars can disrupt the outer parts of planetary systems ( however , see @xcite ) . at an earlier stage ,
protoplanetary disks in dense clusters are subject to dynamical disruption ( e.g. @xcite ) , photoevaporation ( e.g. @xcite ) , and contamination with short - lived radio - nuclides ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . on larger scales ,
radiation from neighboring stars in a cluster is an important feedback effect that can modify the stellar imf @xcite , while gravitational interactions among stars lead to mass segregation @xcite .
all of these phenomena depend critically on the kinematics of newly formed stars , since this determines the frequency of star - star encounters .
kinematics also influence the fraction of star clusters that remain bound rather than dissolving into the field ( e.g. * ? ? ?
. observations of the kinematics of young stars are challenging because the stars spend much of their youth enshrouded by gas , making spectroscopy difficult .
however , measurements have been possible in a few regions , most notably the orion nebula cluster ( onc ) , a particularly interesting and well - studied young star cluster containing thousands of optically revealed stars as well as traces of remnant molecular gas .
the onc is observed to be centrally concentrated and mass segregated , with the most massive stars residing in the center ( e.g. @xcite ) .
radial velocity measurements of stars within the onc indicate a substantial stellar velocity gradient and significant correlation between the local gas and star velocities @xcite . as a result ,
@xcite suggest that the onc is kinematically young and not dynamically relaxed , which they interpret as requiring that the parent cloud be in a state of dynamical collapse @xcite , rather than evolving in quasi - equilibrum @xcite .
@xcite use n - body simulations to demonstrate that collapsing clusters with subvirial initial conditions can indeed generate such observational signatures . however , it is far from clear how to extrapolate from observations of stellar kinematics to the conditions within the host giant molecular cloud ( gmc ) , particularly since not all components of a gmc share the same velocity dispersion . instead , the dispersion of the line - of - sight ( centroid ) gas velocity decreases with density ( e.g. @xcite ) .
gas velocity dispersions in prestellar cores are comparable to the sound speed and are much less than the global cloud virial velocity @xcite .
furthermore , the dispersion of the bulk velocities of the cores themselves is found to be subvirial @xcite , indicating the the relative motion between cores is small .
molecular gas in star forming regions such as perseus and ophiuchus also exhibits large - scale linear velocity gradients similar to those observed in the orion stellar population @xcite .
perseus , in particular , resembles an elongated filament with a gradient of @xmath3 km s@xmath2 pc@xmath2 along its length , similar to the onc .
these observations of dense gas suggest that one can not simply assume that , because stars are subvirial , the parent cloud must be too .
indeed , for gas the opposite appears to be the case .
all of these subvirial signatures for dense components coexist with large - scale velocity dispersions that are consistent with virial balance ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
simulations of turbulent , virialized clouds are able to reproduce the observed kinematics of dense gas @xcite , while simulations of clouds undergoing subvirial collapse have trouble producing quiescent cores like those we observe @xcite . nor do clouds appear to convert large fractions of their mass into stars in a single dynamical time , as expected for a subvirial collapse ( e.g. * ? ? ?
* ; * ? ? ?
instead , observations both of individual star - forming regions @xcite and entire galaxies @xcite show that , even for gas as dense as the cloud that formed the onc , @xmath4 of the mass per dynamical time can form stars .
the complicated relationship between the kinematics of dense and diffuse gas in molecular clouds suggests that it would be profitable to perform a systematic investigation of how the kinematics of newborn stars relate to the properties of the clouds from which they formed .
such an investigation is also of broader interest , since , as noted above , the kinematics of young clusters have a strong influence on planet formation , mass segregation , and other aspects of star and star cluster evolution .
thus it is important to develop first - principles predictions for the expected motion of newborn stars as a function of parent cloud properties , a subject that has not been investigated prior to this work . in this letter , we use turbulent , self - gravitating , adaptive mesh refinement ( amr ) hydrodynamics simulations of star cluster formation to study the properties of the star clusters that form in clouds in a variety of dynamical states . in particular , we examine the stellar velocity dispersion , projected velocities , and velocity gradients .
we compare the simulation results with recent observations of the onc and discuss the relevance to current modeling of star clusters . in [ parameters ] we describe our simulation methods and parameters . in [ results ] we present our results , and we summarize our findings in [ conclusions ] .
for our study , we perform three simulations of an isothermal , non - magnetized molecular cloud using the orion amr code ( e.g. @xcite ) . we reexamine the driven and decaying turbulent simulations described in @xcite , and we perform an additional simulation with higher resolution and different turbulent conditions .
we generate the initial state of the runs by driving turbulent motions in the gas for two crossing times without gravity . at @xmath5 ,
a turbulent steady state has been achieved and the gravitational energy is comparable to the turbulent energy in the box .
the cloud virial parameter is then @xmath6 @xcite . in the simulation
we refer to as u1 energy injection is halted after the initial driving phase and the turbulence decays on a dynamical timescale , initiating a global collapse of the type posited by @xcite . in the simulation d1
we continue turbulent driving so that the mach number and virial parameter are approximately constant ; this is a very rough representation of the behavior expected in a cloud where feedback drives turbulence and prevents global collapse ( e.g. * ? ? ?
* ; * ? ? ?
the third simulation , d2 , begins with the same physical initial conditions , but we adopt a driving pattern with @xmath7 and drive the turbulence at a constant energy injection rate . as gravitational motions become significant , the net kinetic energy in the box increases and the virial parameter rises . the parameters for the three runs are summarized in table [ table1 ] .
isothermal self - gravitating gas is scale free , so it is possible to scale the simulation values to observed cloud properties using the characteristic jeans length and mach number ( see scaling relations in @xcite ) .
however , the main quantity of interest , the ratio of the stellar dispersion to the gas dispersion , is independent of the physical values we adopt . to estimate physical stellar velocities
, we adopt a gas temperature of @xmath8 k and a gas density of @xmath9 g @xmath10 , which are similar to the conditions observed in orion @xcite .
this corresponds to a box length of 2 pc and a cloud mass of 1150 @xmath11 .
we run the simulations with gravity for one mean - density freefall time , @xmath12 .
we introduce a sink particle whenever the jeans conditions is exceeded on the finest amr level @xcite .
these particles represent stars , serving as an indicator of the stellar cluster velocity dispersion imparted by the initial turbulent conditions .
our simulations neglect radiative feedback , which likely has negligible influence on the initial stellar velocity .
the simulations also neglect magnetic fields , which may affect the details of the simulated turbulence and exact distribution of gas densities .
however , strong magnetic fields are more likely to further constrain the motion of the dense gas and forming stars rather than enhance it ( relative to non - magnetized gas ) .
ccccccc u1 & 128@xmath13 & 200 & no driving & 3 .. 4 & 3.9 & 1.1 + d1 & 128@xmath13 & 200 & constant @xmath14 & 3 .. 4 & 8.5 & 1.7 + d2 & 512@xmath13 & 100 & constant @xmath15 & 1 .. 2 & 12.0 & 3.4 +
in this section , we analyze the distribution and velocity information of the stars that form in each simulation .
the simulation with decaying turbulence , u1 , produces only 20 stars .
d2 with the highest amount of turbulence and highest resolution forms 74 stars . in our simulations
we do not resolve close binaries , and so in some cases the stellar properties represent the net mass and momentum of a stellar system . at the final time
, the runs have efficiencies of @xmath16 10 - 15 % . because we do not include protostellar winds , we neglect ejection of mass from prestellar cores , and thus likely overestimate both the stellar masses and the overall efficiency by a factor of @xmath17 @xcite . however , this shift in stars masses is unlikely to affect their kinematics .
figure [ sigmavst ] shows the 3d stellar velocity dispersion as a function of time .
the larger driving scale in d2 works to delay the onset of collapse so that stars form much later than in the d1 and u1 calculations . for u1 and d1 ,
the stellar velocity dispersion is sonic to transonic .
once a sufficient number of stars are formed , the velocity dispersion trend appears to be mostly uncorrelated with the simulation mach number . in the undriven case ,
the stellar velocity dispersion declines slightly , while remaining nearly fixed in the simulation with constant mach number driving . in d2 ,
the dispersion decreases slightly with increasing mach number .
however , in all cases the stellar dispersion is significantly less than the global gas velocity dispersion : @xmath18 .
this corresponds to final virial parameters of @xmath16 0.10 , 0.09 and 0.10 for u1 , d1 , and d2 , respectively , using the total mass in stars and gas .
if all the remaining gas were to disperse , leaving only stellar mass , then the virial parameter of the stars would be measured to be @xmath16 0.8 , 0.6 , and 1.1 for the runs u1 , d1 , and d2 , respectively .
this indicates that only if the initial gas cloud is significantly supervirial does the virial parameter of the stellar component become greater than unity , and that star clusters forming in turbulent virialized clouds naturally begin with subvirial velocities .
consequently , the subvirial nature of the stellar velocity dispersion is not necessarily an indication of collapse .
indeed , @xcite find that they are able to reproduce the kinematic signature of orion assuming an initial virial ratio of @xmath19 ( for stars plus gas ) , the value produced by simulation d1 , in which the gas cloud is turbulent and virialized .
similarly , @xcite assume an initial cluster virial parameter of 0.8 , much less subvirial than we find for our simulated clouds with virial parameters of @xmath20 .
if the ratio of the stellar dispersion to the gas dispersion influences how quickly the cluster undergoes mass segregation , then most clusters with substantially subvirial initial stellar velocities dispersions may segregate on short timescales . in this section
we investigate the stellar velocities projected along the line - of - sight to the cluster . for the most detailed examination
, we use the final time of the d2 simulation , which has the best statistics .
figure [ plotvelbin ] shows the velocities projected along the three cardinal sight - lines .
the velocities are somewhat clustered with only a couple instances of large deviations .
we have further binned the projected velocities into groups containing 9 stars as a function of position and we have overlaid error bars indicating the velocity dispersion in each bin .
these plots look similar to figure 9 in @xcite of the onc , which shows evidence of a velocity gradient across the cluster .
we derive the stellar velocity gradient by fitting the line - of - sight velocities with the function @xmath21 .
although the data may not be exactly linear in nature ( e.g. the onc exhibits a linear gradient mainly in the cluster center ) , we can characterize the change in the velocities across the cloud .
since we use periodic boundary conditions the velocity bins tend to wrap around the domain .
thus , we can better compare to the data by dividing each projection into smaller boxes and investigating the velocity trend in each box separately . in figure [ grad ]
, we show the velocity gradients , @xmath22 , along each projection for the entire domain and for boxes half and a quarter of the domain .
we include gradients at three different times and plot onc data from @xcite fitted in the same manner .
to limit statistical fluctuations , we only include gradients fitted to 4 or more stars .
the full and half box fits have median velocity gradient magnitudes of 1 km s@xmath2 pc@xmath2 and 2.1 km s@xmath2 pc@xmath2 , respectively .
the magnitude of the gradients increases in both the simulations and the onc for smaller boxes due to smaller statistics , where stars separated by several shock fronts may easily acquire significantly different velocities . for comparison , @xcite find that the onc has a radial velocity gradient of 1.1 km s@xmath2 pc over half of its length , which corresponds to an overall gradient of @xmath16 0.5 km s@xmath2 pc .
this value is similar to what we find for our simulated virialized clusters .
thus velocity gradients are not necessarily a hallmark of a subvirial turbulence or a strongly collapsing cloud . instead , large - scale velocity gradients appear to be a signature of the fact that both our simulations and real molecular clouds have most of their turbulent power in large - wavelength modes .
we find that the linear fits of the projected velocities fluctuate somewhat as a function of time , but they tend to fall within 0.2 - 2 km s@xmath2 pc@xmath2 on the largest scales .
these values are similar to gradients in the gas centroid velocity observed in regions like perseus and ophiuchus @xcite .
observations of clusters frequently find that the star and gas velocities are correlated ( e.g. , @xcite ) .
we can reproduce such observations by making simulated observations of the gas in particular molecular tracers . following the procedure of @xcite
, we generate a position - position - velocity cube of the gas intensity in the @xmath23co ( j = 1 @xmath24 0 ) molecular line along one line - of - sight in the d2 run .
we solve for the line emission , assuming that the gas is in statistical equilibrium and that radiative pumping is negligible .
we adopt a cloud distance of 400 pc and 26 beam size .
to model the telescope resolution , we smear each velocity channel with the gaussian beam .
figure [ gas_kin ] shows the line - of - sight @xmath23co gas and star velocities ; following furesz et al .
( 2008 , their figure 10 ) , to produce this figure from the 3-dimensional position - position - velocity cube , we have integrated along one direction on the plane of the sky , so the intensity shown in a given @xmath25 pixel is the sum over all @xmath26 values for that @xmath27 and @xmath28 .
our simulated image shows a strong correlation between gas and star velocities , similar to that seen in the star and gas velocity maps presented by @xcite . for the projection shown in figure [ gas_kin ] , the mean stellar velocity is @xmath29 km s@xmath2 , somewhat to the left of the mean @xmath23co gas velocity at 0.003 km s@xmath2 .
most of the disagreement in the histogram occurs because not all regions traced by @xmath23co have formed stars .
projections along other directions show similar correlations between gas and star velocities .
the strong correlation between gas and star velocities might seem surprising , given that the gas and stars have such different velocity dispersions .
the explanation is that the anti - correlation between density and velocity found in turbulent gas serves to artificially enhance the gas - star correlation .
figure [ gas_kin ] is constructed by summing over the @xmath26 direction , so each pixel represents a density - weighted average velocity .
the dispersion , however , is a density - weighted average of velocity squared , which , since velocity and density are anti - correlated , is much more heavily weighted toward low density , high velocity gas than what is shown in figure [ gas_kin ] .
this low density gas is also that which is least correlated with the stars .
the bias is exacerbated because @xmath23co is thermally excited only in relatively dense gas that is closely associated with forming stars .
it is subthermally excited in the lower density gas that dominates the dispersion .
thus , both averaging and excitation serve to improve the correlation between @xmath23co emission and stellar velocities , while at the same time obscuring the low density gas that carries the bulk of the kinetic energy .
in this letter , we have investigated the kinematic properties of stars forming in simulated turbulent self - gravitating clouds .
these calculations represent the initial conditions of stars in clusters prior to dynamic evolution and cloud dispersal .
we demonstrate that subvirial stellar velocity dispersions arise naturally from clouds in virial equilibrium .
thus , long lived clouds satisfying a quasi - steady state are not ruled out by observations of clusters with small stellar velocity dispersions .
we find that gas virial parameters of order unity produce a star to gas velocity dispersion ratio of approximately 0.2 .
we find that turbulent initial conditions easily generate stellar velocity gradients of the magnitude observed in star - forming regions .
using a planar fit , we demonstrate typical gradients of 1 km s@xmath2 pc@xmath2 , similar to the observed gradient in the onc .
we also show that for young clusters there is a strong similarity in the simulations between the dense gas traced by @xmath23co and the star velocities .
thus , trends in the stellar velocities as a function of position may be indicative of the dominance of large - scale turbulent modes , rather than evidence of global collapse or cloud rotation as suggested by some authors .
, n. j. , dunham , m. m. , jrgensen , j. k. , enoch , m. l. , mern , b. , van dishoeck , e. f. , alcal , j. m. , myers , p. c. , stapelfeldt , k. r. , huard , t. l. , allen , l. e. , harvey , p. m. , van kempen , t. , blake , g. a. , koerner , d. w. , mundy , l. g. , padgett , d. l. , & sargent , a. i. 2009 , , 181 , 321 , s. t. , gaidos , e. , hester , j. j. , adams , f. c. , bally , j. , lee , j .- e . , & wolk , s. 2008 , in astronomical society of the pacific conference series , vol .
384 , 14th cambridge workshop on cool stars , stellar systems , and the sun , ed .
g. van belle , 393+ | the kinematics of newly - formed star clusters are interesting both as a probe of the state of the gas clouds from which the stars form , and because they influence planet formation , stellar mass segregation , cluster disruption , and other processes controlled in part by dynamical interactions in young clusters . however , to date
there have been no attempts to use simulations of star cluster formation to investigate how the kinematics of young stars change in response to variations in the properties of their parent molecular clouds . in this letter
we report the results of turbulent self - gravitating simulations of cluster formation in which we consider both clouds in virial balance and those undergoing global collapse .
we find that stars in these simulations generally have velocity dispersions smaller than that of the gas by a factor of @xmath0 , independent of the dynamical state of the parent cloud , so that subvirial stellar velocity dispersions arise naturally even in virialized molecular clouds .
the simulated clusters also show large - scale stellar velocity gradients of @xmath1 km s@xmath2 pc@xmath2 and strong correlations between the centroid velocities of stars and gas , both of which are observed in young clusters .
we conclude that star clusters should display subvirial velocity dispersions , large - scale velocity gradients , and strong gas - star velocity correlations regardless of whether their parent clouds are in virial balance , and , conversely , that observations of these features can not be used to infer the dynamical state of the parent gas clouds . |
imaging cherenkov telescopes have proven to be the most successful tool developed so far to explore the cosmic gamma rays of energies above few hundred gev .
a pioneering generation of installations has been able to detect a handful of sources and start a whole program of very exciting physics studies . now
a second generation of more elaborated telescopes is under construction and will provide soon with new observations .
one of the main characteristics of these new telescopes is the ability to detect lower energy gamma rays , which could fill the observational gap between 10 and 300 gev .
one of the most interesting results in fundamental physics from the existing telescopes is a limit , given by the whipple collaboration @xcite , on the quantum gravity scale to be larger than @xmath1 gev ( about 1/250th of the planck mass ) at the 95 % confidence level . in that study ,
the data from a tev gamma - ray flare associated with the active galaxy markarian 421 observed on 15 may 1996 was used to place bounds on the possible energy - dependence of the speed of light in terms of an effective scale for quantum gravitational effects .
the basic idea behind the measurement is that gamma rays traversing cosmological distances should notice the quantum fluctuations in the gravitational vacuum which unavoidably should happen in any quantum theory of gravitation .
these fluctuations may occur on scale sizes as small as the planck length @xmath2 cm or time - scales of the order of @xmath3 ( @xmath4 gev ) .
these gammas will therefore experience a `` vacuum polarization '' correction which should be very small ( @xmath5 where @xmath6 is the gamma energy and @xmath7 is an effective scale for quantum gravity , which might be as large as @xmath8 ) but might become measurable after the gamma has traversed cosmological distances . in this quantum gravity scenario ,
the requirement of violation of the lorenz - invariance symmetry @xcite emerges naturally , providing an energy - dependent propagation speed for electromagnetic waves .
therefore , gammas of different energies being emitted simultaneously by a distant source should reach our observatories at different times .
in fact , for practical purposes , this delay with respect to the ordinary case of an energy - independent speed c for massless particles , for a source at a distance l can be expressed as @xcite @xmath9 this delay should be one of the cleanest signatures that cherenkov telescopes are able to study in order to spot quantum gravity effects . nevertheless ,
as it has already been pointed out in the literature , it might be mimicked by effects related to the source physics ( internal production delays for higher energy gammas ) or propagation ( time delay due to cascading in intergalactic magnetic fields @xcite ) .
therefore it will be mandatory to use the scaling of the effect with redshift to distinguish between quantum gravity effects and any source - dependent phenomena @xcite .
the new generation of telescopes , should have an improved sensitivity and lower threshold .
the first characteristic might help in resolving faster time structures while the second one should allow to observe sources at much longer distances because of a smaller absorption in the intergalactic medium .
therefore , some clear improvement in this kind of measurements should be expected .
the goal of this work is trying to estimate how these characteristics would actually help in improving the quest for quantum gravity effects .
for that , this writeup is organized as follows : first , a very brief overview of the theoretical framework used in this study is presented .
second , the results of the expected delay as a function of the gamma ray energy and the source redshift location for different quantum gravity scenarios is discussed .
third , the effect of the existence of a gamma ray horizon , including consistently the effect of the lorenz invariance violation , is shown .
after that , a discussion about the observational picture when all these effects are put together is presented and finally , in the conclusions we give our prospects for these observations with the coming generation of cherenkov telescopes .
when developing any model for quantum gravity , it appears naturally the necessity to modify some of the most basic continuous symmetries of space - time , such as lorentz invariance @xcite . in this analysis
we shall focus in the violation of the lorentz invariance symmetry ( liv ) caused by quantum gravity because it is the responsible of the main changes in the kinematics in which is based this study .
the violation of the lorentz invariance symmetry ( liv ) modifies the dispersion relation giving the propagation speed for gamma rays in a theory - dependent manner .
however , since in all plausible approaches the actual effects are expected to be small , they can be studied from a phenomenological point of view using an expansion in terms of the gamma energy divided by the effective quantum gravity scale .
therefore , the actual gamma dispersion relation for a massless particle can be expressed in leading order as @xcite @xmath10 where @xmath6 and @xmath11 denotes the energy and the momentum of gammas , @xmath12 and @xmath13 are the liv free parameters and @xmath7 is the energy scale for quantum gravity . in this scenario ,
gamma rays travelling in vacuum can be seen like traversing a material medium @xcite .
hence , the propagation speed of gamma rays should be computed as @xmath14\ ] ] where @xmath6 is actually the gamma comoving energy . in terms of the measured gamma energy at the earth @xmath15 ,
the velocity when the gamma is at a redshift @xmath16 is @xmath17\ ] ] therefore , the actual time - of - flight for a gamma ray coming from a source at redshift @xmath16 , is given by @xmath18 where @xmath19^{1/2}}\ ] ] being @xmath20 the hubble constant and @xmath21 , @xmath22 and @xmath23 the standard cosmological parameters . for practical purposes , we need the difference in the time - of - flight for gamma rays of different energies ( @xmath15 and @xmath24 ) produced at the same redshift @xmath16 because this is the signature that gamma ray telescopes are able to study .
this difference is given by @xmath25 if in the above expressions we take @xmath26 , @xmath27 and the limit @xmath28 , we recover the simple expression used by the whipple collaboration .
the free parameters used in our phenomenological approach are the ones used in references @xcite , namely @xmath12 , @xmath13 , and @xmath7 .
@xmath12 is only a sign ambiguity and hence its possible values are @xmath30 . out of the two possible values , in this work
we have chosen to use @xmath31 because it fits better the present experimental constraints @xcite . in what concerns @xmath13 , in quantum gravity theories corrections going like @xmath32 , where @xmath33 is the planck energy , typically appear as leading order of more complex analytic expressions @xcite .
this motivates us to study the cases @xmath34 and @xmath35 .
nevertheless , @xmath36 means an @xmath7 suppression factor in the difference of time of flight .
this would lead to a non measurable @xmath37 ( @xmath38 ) .
therefore we can not exclude the possibility that the leading term of liv goes as @xmath39 but , if it was the case , the gamma ray telescopes would not have any chance to test it with the gamma time - of - flights .
that forces us to study only the most favourable case ( potentially largest quantum gravity effects ) hence @xmath34 . in this scenario equation
[ delta - time - of - flight ] is reduced to @xmath40 figure [ evst_1_without ] shows the lines in the plane @xmath41 versus @xmath37 provided by the above equations for a set of different source redshifts . from this plot
it is clear that , for a given energy difference between the detected gammas , being capable of observing sources at larger redshifts should allow to check the @xmath7 with less demanding time resolutions .
or conversely it should allow , for a fixed time resolution , to use lower energy gammas to check the @xmath7 . this should help because , given the steep power - law energy spectrum of all the known high energy gamma astrophysical sources , lowering the gamma energy scale should allow to observe a larger number of gammas and , hence , to improve the experimental accuracy . nevertheless , in spite of the fact that the effective distance integral giving the @xmath16 dependence of the time delay has a correction of @xmath42 to the lookback time integrand ( which increases the effective gamma propagation time ) it is clear from figure [ evst_1_without ] that this dependence on the redshift saturates quite quickly .
actually the main target of the time - of - flight studies is to explore the @xmath7 scale . in figure [ evst_1_without ]
it is shown the effect of observing sources at different redshifts to test a given value of @xmath7 ( @xmath43 ) .
alternatively , fixing the observable delay time scale , this result can be shown as in figure [ eqgvsz_t10 ] , as the capability of exploring higher quantum gravity scales as the source redshift increases .
the above considerations about the benefit of observing higher redshift sources to explore higher quantum gravity effective scales were implicitly assuming that the gamma rays traverse the intergalactic space without being altered . in reality
, we expect high energy gamma rays traversing cosmological distances to be absorbed by their interaction with the diffuse background radiation fields , or `` extragalactic background light '' ( ebl ) , producing @xmath44 pairs .
the @xmath45 cross section is strongly picked to @xmath46 and therefore , there is a specific range in the ebl energy which is `` probed '' by each gamma ray energy @xcite .
this effect should lead to the existence of a `` gamma ray horizon '' , limiting the feasibility of observing very high energy gamma rays coming from very far distances .
the actual value of this horizon distance for gamma rays of a given energy , depends on the number density of the diffuse background radiation of the relevant energy range , which is traversed by the gamma rays . in the range of gamma ray energies which can be effectively studied by the next generation of gamma ray telescopes ( from , say , 10 gev to 50 tev )
, the most relevant ebl component is the infrared contribution .
quite precise predictions of the gamma ray horizon have been made but , unfortunately , so far no clear confirmation can be drawn from the observations of the present generation of gamma ray telescopes . in this work
we have used the approach and procedures described in @xcite which predict a gamma ray horizon as shown in figure [ hor_classic ] .
nevertheless , the above prediction has been obtained by assuming purely standard model physics .
recent claims of observation of very high energy gamma ray events coming from mkn 501 @xcite , a blazar at redshift @xmath47 which would somehow contradict the above prediction of the gamma ray horizon , have motivated some authors @xcite to revise that prediction .
it has been pointed out that the actual existence of liv should also affect the calculation of the gamma ray horizon , since the threshold energy of the ebl needed to produce @xmath44 pairs is modified @xcite .
this effect could explain the very high energy observed events . therefore , to be consistent with the framework of our analysis , we ve also undertaken the task of implementing consistently in the complete calculation of the gamma ray horizon the inclusion of the threshold - modifying effects coming from the effective quantum gravity inspired dispersion relations discussed in the previous sections . for
that we ve followed the approach of reference @xcite , where the modified threshold value for the gamma ray momentum to produce @xmath44 pairs is deduced , which is ] @xmath48 where @xmath49 is the energy of the ebl gammas and @xmath50 it the threshold momentum for the gamma ray . in the standard calculation of the gamma ray horizon @xcite , the optical depth for gamma rays of energy @xmath15 coming from a source at redshift @xmath16
is obtained from the integration @xmath51\ ] ] where @xmath52 being @xmath53 the gamma - gamma scattering angle , @xmath54 the spectral density of the ebl gammas @xmath55 at the given z and @xmath56 the threshold energy for the ebl gammas which is given by @xmath57 after we include liv , we arrive to a modified threshold condition of @xmath58^{1+\alpha}}{4 e_{qg}^{\alpha } } \left ( \frac{1}{2^{\alpha}}-1 \right)\ ] ] where @xmath15 and @xmath59 are the energies at the earth and , therefore , we had to add some @xmath60 factors respect to the equation [ p_threshold ] .
the dependence on the scattering angle has also been introduced as a global factor @xmath61 .
the results comparing the optical depths with and without this threshold correction ( once again in the @xmath62 and @xmath63 scenario ) are shown in figure [ optdepth ] .
for these parameter values the net effect of the correction is increasing the threshold energy in such a way that , for any given gamma energy , harder ebl photons are responsible for their absorption . in the optical depth calculation without liv , the gamma rays of @xmath64 gev `` probe '' the ebl spectrum at around @xmath65 where the density has a maximum ,
while when the liv correction is present , they probe the ebl spectrum at lower wavelengths , where the value of the density is smaller @xcite .
the main consequence of this is that in the calculation with liv corrections the universe becomes transparent again for large energy gammas whereas the results for the lower energy gammas of the plots remain basically unchanged .
the correction for large energy gammas is so severe that avoids the existence of a gamma ray horizon for moderate redshift sources , like mkn 421 , giving then a plausible explanation for the observed gamma events as has already being mentioned .
the obtained gamma ray horizon is shown in figure [ hor_lid ] for different @xmath7 and compared with the `` standard '' one .
if one assumes high @xmath7 , it is clear from that plot that the main difference due to liv should happen for low and moderate redshift sources whereas for large redshift sources no effect should be observable in the gamma ray horizon . on the other hand for not so high @xmath7 , the universe becomes transparent again for a gamma ray energy that can be reached with current agn models @xcite ( 15 tev for @xmath66 ) .
the fact that @xmath0 rays of this high energies have been detected for close objects ( @xmath67 ) @xcite but not for objects at larger redshift suggests that the @xmath7 must be in excess of @xmath68 .
actually , one of the most recent results about the effects of liv in the gamma ray horizon was presented in @xcite where tev gamma ray observations of the active galaxy markarian 501 were used to set a lower limit on the lorentz invariance breaking parameters .
we shall now put together the effects discussed in the previous sections to get a complete picture of the expectations .
as already mentioned , on the one hand the capability of observing more distant sources should allow to have more lever arm to explore higher effective quantum gravity scales .
unfortunately , on the other hand , for more distant sources , the gamma absorption in the ebl is stronger and the gamma ray horizon happens at smaller gamma energies .
this trade - off is clearly summarized in figures [ evst ] where , on top of the @xmath41 versus @xmath37 lines predicting the propagation delays , one can see the parameter region beyond the gamma ray horizon .
it must be stressed here that the gamma ray horizon does not mean a `` hard boundary '' since it just gives the condition for an e - fold reduction of the observed flux . because of that , the shaded area given by the gamma ray horizon has to be understood as the region in which the flux reduction due to the absorption starts being strong enough to make the source observation difficult . from these figures
it is clear that the time resolution needed depends strongly on the actual effective quantum gravity scale that one would like to test . to explore up to the planck mass , time resolutions on the scale of just few seconds
will be mandatory if gamma rays of up to few hundred gev are used ( or few tens of seconds if gammas of few tevs are used ) quite independently on the actual source redshift .
but resolution of several minutes could be enough for a scale of one hundredth of the planck mass . in view of these figures
it is not obvious how much having a reduced threshold would help the next cherenkov telescope generation to improve in this kind of analysis .
in fact , from the above figures the straight conclusion is that the observation of the nearby blazars ( mkn 421 , at @xmath69 is actually the closest known blazar ) in an energy range of @xmath70 gev seems to be the most comfortable scenario for exploring the highest effective quantum gravity scale .
in this work the potential of the next generation of gamma ray telescopes in improving the existing tests of an effective quantum gravity scale from the study of the propagation delay for gamma rays of different energies coming from a distant astrophysical source has been analyzed .
these new telescopes should have better sensitivity and lower threshold than the present ones , enabling a better time resolution and the observation of sources at higher redshift and therefore , a clear improvement in this kind of measurements should be expected .
detailed calculations on the actual expected delay for sources at any redshift have been presented .
they show that the effective delay time grows with the redshift faster than the lookback time . in spite of that , the redshift dependence of the relationship between the gamma energy scale and the observable arrival time delay has been shown to saturate rather quickly as the source redshift increases .
in addition , a calculation of the gamma ray horizon due to the gamma absorption in the extragalactic background light ( ebl ) taking consistently into account the lorenz invariance violation ( liv ) effects has been carried out .
this absorption should limit the observability of high energy gammas coming from very distant sources but , as it has already been realized in the past , the inclusion of liv effects opens a window for very high energy gammas . putting all these effects together
, the global picture is that quite stringent requirements on the time resolution ( in the scale of few seconds ) should be achieved to be able to explore an effective quantum gravity scale close to the planck mass .
in fact , given the effect of the gamma ray horizon , this demanding time resolution constraint can not be avoided by using large redshift sources unless very high energy gamma rays are used .
although , this resolution is less restrictive ( several minutes ) for a scale around one hundredth of the planck mass . in view of this situation , maybe the observation of the nearby blazars in an energy range of @xmath70 gev could be the most comfortable scenario for exploring the highest effective quantum gravity scale . given this result , the new generation of telescope arrays might be the most promising choice since they can provide a much larger collection area for higher energy gamma rays
. we would like to thank our colleagues of the magic collaboration for their comments and support .
we are indebted to s.m.bradbury from the veritas collaboration for triggering our interest in this subject .
we thank r.pascual for discussions and comments on this manuscript .
we d like also to thank j.cortina , o. de jagger , j.garriga , a.grifols , e.lorenz , k.mannheim and a.saggion for their encouraging comments . and
finally , we d like to thank to o. bertolami , g. amelino - camelia and f. stecker for their comments in the preprint stage of this article .
g. amelino - camelia , notes based on lectures given at the xxxv karpacz winter school of theoretical physics _ from cosmology to quantum gravity _ ,
polanica , poland , 2 - 12 february , 1999 l.m .
krauss , submitted to world scientific on 27 mar 2001 . | the actual potential of the next generation of gamma ray telescopes in improving the existing tests of an effective quantum gravity scale from the study of the propagation delay for gamma rays of different energies coming from a distant astrophysical source is discussed .
it is shown that the existence of a cosmological gamma ray horizon , will impose very demanding conditions on the observations of the telescopes to try to test a quantum gravity scale close to the planck mass .
, , lorentz invariance , @xmath0-rays , quantum gravity , cherenkov telescopes , gamma ray horizon 95.85.pw , 95.30.sf |
supersymmetric particles have been so far elusive . nor we have seen any indirect signal of them in flavour physics , giving rise to the so - called _ supersymmetric flavour problem_. assuming that s - particles at the fermi scale indeed exist , is there any message in this ?
on one side , the discovery potential of the lhc in its full strength appears large enough to render this question not so urgent : if s - particles are there , they will be seen . on the other side ,
any orientation towards a specifically motivated configuration of s - particle masses and , possibly , of corresponding signals would clearly be valuable .
this would be especially true in the case of particular configurations that appear to have received relatively little attention in the otherwise vast literature on phenomenological supersymmetry .
we argue in the following that there is one such case , based on naturalness and on the supersymmetric flavour problem .
as we shall see , it is also simple to characterize some of its features , useful to define a strategy for its discovery .
since naturalness is the main motivation for supersymmetry at the fermi scale , it makes sense to ask which are the parameters affecting the s - particle spectrum that are mostly constrained by naturalness itself .
we have in mind a generic minimal supersymmetric standard model , unconstrained by specific supersymmetry breaking mechanisms or specific unification conditions .
the answer is well known : * the @xmath2-parameter , which affects the electroweak scale already at tree level and gives , in absence of mixing with the gauginos , the higgsino masses ; * the masses @xmath3 and @xmath4 of the third generation of squarks , which are coupled through the large top yukawa coupling , @xmath5 , to the higgs system and define , together with the stop - mixing term , the physical masses , @xmath6 , of the two _ stops _ , @xmath7 , and of the left - handed _ sbottom _
, @xmath8 , in absence of mixing with the right - handed one .
these can be the lightest s - particle masses , whereas all the other s - partner masses , allowed to go to their naturalness limits , can be heavier or be phenomenologically ( almost ) irrelevant for lhc .
we find this possibility - to be made more precise in a while - motivated by two main considerations . if these are the s - particle masses , the supersymmetric flavour problem gets definitely alleviated , if not solved completely , depending on the assumptions one is willing to make on the ( approximate ) flavour symmetries .
we have in mind , at least as an example , an approximate @xmath9 symmetry which enforces ckm - like angles in the extra supersymmetric interactions and some amount of degeneracy between the s - fermions of the first two generations@xcite@xcite@xcite .
it is also true - without any bias from specific models - that both the @xmath2-parameter and the masses of the squarks coupled to the higgs system via @xmath5 are sufficiently singled out from the other mass parameters that one may conceive for them a somewhat special role .
here it suffices to mention the much discussed @xmath2-problem and , e.g. , the effect of @xmath5 on the running of @xmath3 and @xmath4 .
let us be more specific on our choice of the mass parameters . a relevant range for the @xmath2-parameter is between 100 and 200 gev , to evade the lep2 bound on charginos but otherwise not significantly above the fermi scale itself , @xmath10 gev .
the masses of @xmath7 and @xmath11 should just be sufficiently heavy to allow a higgs boson mass again consistent with the lep2 bounds : depending on the mixing term between the stops , their typical range is in the hundreds of gevs ( see below ) .
the gluino mass , @xmath12 , is also constrained by naturalness and flavour considerations . on the upper side ,
@xmath13 drives a one loop effect on @xmath3 and @xmath4 , that wants to make them too heavy , depending on the messenger scale . on the low side , gluino - sbottom exchanges in @xmath14 @xcite@xcite and @xmath15 @xcite@xcite amplitudes with
ckm - like angles point to a not too light gluino .
an interesting range for the gluino mass is between 500 gev and 1 tev .
naturalness allows the remaining s - particles to be heavy enough to play little role in the lhc phenomenology , at least in a first stage .
this is the case for all the s - fermions with small coupling to the higgs system , that can go in the @xmath16 tev range and @xmath17 but all the third generation s - fermions be relatively lighter than those ones of the first two generations .
this could still be consistent with current flavor constraints , although bringing the focus also on the @xmath18 transition , mediated by stau - gaugino exchanges , and on the ongoing experiment at psi@xcite . from the point of view of the lhc signals discussed below nothing would change as long as the right sbottom mass is comparable or heavier than @xmath13 and both @xmath19 s and the third generation s - neutrino(s ) are heavier than @xmath2 . in turn the heaviness of @xmath20
relative to @xmath17 could be due to the smallness of the bottom yukawa coupling relative to the top one in their running effects on the squark masses.]@xcite , and for the electroweak gauginos , that can safely be taken at about 500 gev or even higher .
as we are going to see , an advantage of considering this configuration of s - particle masses is that its main lhc phenomenology is fully determined by a few physical parameters , irrespective of the underlying theoretical model .
fig . [ spettro ] shows a particular representation of the generic s - particle spectrum that we are led to consider below 1 tev .
the lightest s - fermions are mostly higgsino - like , have an average mass @xmath2 , a small splitting among them , in the 10 gev range , and a definite order , @xmath21 .
the decay among them of the heavier into the lighter ones via virtual @xmath22 and @xmath23 produces also leptons , whose possible detection has been analyzed in ref .
the softness of these leptons , however , make us consider other signals of this configuration of s - particles . as a consequence ,
the splitting among the light higgsinos can be safely neglected .
furthermore , as anticipated , the heavier ( mostly ) electroweak gauginos play no relevant role .
although they can occur in two body decays of some squarks , the corresponding branching ratios are negligibly small , at the relative _ per cent _ level , because of the stronger coupling of the squarks to the higgsinos and because of phase space .
the most significant signals in the relatively earlier stages of the lhc come from gluino pair production , followed , when allowed by phase space , by two - body gluino decays : @xmath24 and the corresponding conjugate modes , each with equal branching ratios .
all these branching ratios , given in appendix a , are determined by @xmath25 and by the stop mixing angle @xmath26 , defined in terms of the left and right stops by @xmath27 the mass of the single relevant sbottom is . ]
@xmath28 by definition @xmath29 . in a synthetic notation ,
all the two - body decays of the three squarks , when allowed by phase space and charge conservation , are @xmath30 @xmath31 @xmath32 where @xmath33 other than @xmath6 and @xmath34 , the branching ratios for these decays involve @xmath2 and @xmath35 , as is the case for the mass of the lightest scalar , @xmath36 .
all of these quantities are given in appendix a. a relevant range for @xmath37 is between 5 and 10 . at @xmath38
the lep bound on @xmath39 is problematic . on the contrary
, the higgsino - stop contribution to @xmath40 grows like @xmath37 and , with mixing angles analogous to the ckm ones , becomes exceedingly large . as an example , this is illustrated in fig .
[ bsgamma ] , which is well representative of the generic situation . while this is uncertain , also in view of possible cancellations with other contributions to the @xmath40 amplitude coming from charged higgs or gluino exchanges , we find it better to stick to @xmath41 .
in fact the specific value taken by @xmath37 between 5 and 10 constrains @xmath6 and @xmath34 through @xmath42 gev but is otherwise practically irrelevant in the branching ratios for the modes ( [ sqchi ] , [ sqv ] , [ sqs ] )
. incidentally notice also from fig.[bsgamma ] that @xmath40 disfavours highly asymmetric stop masses because of the correspondingly growing stop - mixing parameter for fixed values of @xmath34 . in the same @xmath43-plane , on the contrary
, there is no significant constraint from the @xmath44-parameter .
[ bsgamma ] from its measured value @xmath45 , in the illustrative case that the susy amplitude is _ only _ given by the higgsino - stop exchange with _ exactly _ ckm mixing angles .
the lep bound on the higgs boson mass ( see eq .
( [ higgsmass ] ) in the appendix ) is satisfied above the red dashed line.,title="fig:",width=302 ] from its measured value @xmath45 , in the illustrative case that the susy amplitude is _ only _ given by the higgsino - stop exchange with _ exactly _ ckm mixing angles .
the lep bound on the higgs boson mass ( see eq .
( [ higgsmass ] ) in the appendix ) is satisfied above the red dashed line.,title="fig:",width=302 ] largely dominant among the squark decays are the higgsino decays ( [ sqchi ] ) .
this is a cumulative effect of phase space and of suppression factors of the decay amplitudes into the vector bosons and into scalars .
furthermore , among the same higgsino decays , one of them , @xmath46 also has a suppressed amplitude , relative to the other ones , by a factor @xmath47 .
this is because the coupling between @xmath48 and @xmath49 only occurs in the down yukawa coupling . a good approximation for a general discussion , therefore , is provided by sticking to the following relevant decay modes : @xmath50 and the corresponding conjugate ones .
as already mentioned , the decays @xmath51 are meant to be summed over the two neutral almost degenerate higgsinos . in turn , the possible final states in the gluino chain decay are limited to : @xmath52 with the last two conjugate decays having an identical branching ratio , @xmath53 we find it a remarkable simplification that a single inclusive branching ratio may end up characterizing the main features of the phenomenology of this configuration of s - particle masses , as we are going to discuss .
this remains true even if one wants to include in the gluino decay chain the subdominant modes @xmath54 , @xmath55 . to this end
one needs to generalize @xmath56 to @xmath57 where @xmath58 or @xmath59 or @xmath60 . in any given point of the parameter space
the difference between @xmath56 and @xmath61 is small .
furthermore @xmath61 is equally useful to characterize the features of the events we study in the next section .
as it is clear from the previous considerations , the most characteristic feature of the lhc signals from this configuration of s - particle masses is an overproduction of top quarks from gluino decays .
starting from gluino pair production , there will in fact be four types of semi - inclusive final states : @xmath62 @xmath63 @xmath64 @xmath65 where , as before , @xmath66 stands for a charged or any one of the two neutral higgsinos .
their rates will depend in good approximation , up to the overall gluino - pair production rate , only on @xmath56 , @xmath67 @xmath68 @xmath69 @xmath70 @xmath56 is therefore the crucial quantity to study .
whenever @xmath71 is negligible in phase space factors , it is @xmath72 .
this follows from the form of the top yukawa coupling , which , as already mentioned , dominates over the bottom one and dictates the couplings of the left and right squarks to the higgsinos @xmath73 where both the charged @xmath74 and the neutral @xmath66 , in the limit of degenerate @xmath75 , can be considered as dirac spinors of the same mass @xmath2 .
neglect first @xmath76 mixing . then from ( [ yukawa - stops ] ) and @xmath77 ( up to @xmath71 corrections ) @xmath78 and @xmath79 with the same branching ratios for the charge conjugate modes .
hence indeed @xmath80 , either from ( [ gluino_tl ] ) or from ( [ gluino_tr ] ) , which does not change by @xmath76 mixing since @xmath56 is inclusive . on the other hand ,
whenever @xmath71 becomes important in phase space factors , @xmath56 increases since the @xmath81-reach modes are favoured over the @xmath82-reach ones .
[ btb ] as defined in the text , for @xmath83 , @xmath84 , @xmath85 and @xmath86 .
the higgs boson mass satisfies @xmath87 ( see eq .
( [ higgsmass ] ) in the appendix ) above the red dashed line.,title="fig:",width=302 ] as defined in the text , for @xmath83 , @xmath84 , @xmath85 and @xmath86 .
the higgs boson mass satisfies @xmath87 ( see eq .
( [ higgsmass ] ) in the appendix ) above the red dashed line.,title="fig:",width=302 ] this is confirmed and illustrated in fig.[btb ] for some sample values of @xmath88 , where the bound on the higgs mass from lep is also shown .
these figures all exhibit the same pattern , which is in particular almost symmetric in @xmath34 above or below @xmath89 . from @xmath90 @xmath56 increases and eventually reaches its limiting value @xmath91 for asymmetric physical stop masses , where @xmath92 gets inhibited by phase space , unlike @xmath93 , or when both @xmath94 and @xmath95 are heavy enough that @xmath96 gets kinematically constrained or even forbidden , unlike @xmath97 .
[ rates ] . the red ( blue )
dashed line indicates the rate to have _ at least _ two same - sign leptons ( three leptons).,title="fig:",width=340 ] high @xmath98 same - sign di - lepton and tri - lepton events both from supersymmetry as from other beyond - the - standard - model physics@xcite@xcite@xcite , have been extensively discussed in the literature .
they look particularly relevant in the present case , coming from the semi - leptonic decays of the abundant top quarks . adding up the same - sign di - leptons , either electrons or muons , from ( [ 4t],[3t],[2 t ] ) , one finds a total rate , for any gluino pair produced , @xmath99 where @xmath100 and @xmath101 are the branching ratios of the @xmath22 into @xmath102 or into hadrons respectively .
each of these events contains four @xmath81-jets and up to four extra partonic jets from @xmath22 decays , together with the missing energy associated with the two @xmath66 s and two neutrinos from the top quarks .
similarly the semi - leptonic decay of three top quarks from either ( [ 4 t ] ) or ( [ 3 t ] ) give a rate of tri - lepton events @xmath103 where this time , other than the four @xmath81-jets , there are up to two additional jets .
these relative rates are shown in fig.[rates ] versus @xmath56 .
[ tabella ] .leading - order cross sections for the main backgrounds to the signal .
the same - sign di - lepton cross section is estimated including the possibility to misidentify the lepton charge with a 1% probability in the @xmath104 and in the @xmath105 cases .
[ cols="^,^,^,^",options="header " , ] these signals should allow the discovery of the configuration of s - particles that we are discussing already in the relatively early stages of lhc . taking , e.g. , @xmath106 gev and @xmath107 , the cross section at @xmath108 tev for the same sign di - leptons ( tri - leptons ) from top decays is @xmath109 .
this should be compared with the background raw rates estimated in table 1 , without the imposition of any cut .
( right ) for the signal events where at least two same - sign leptons with @xmath110 are found .
also shown are the same distributions from the dominant @xmath104 background ( see text).,title="fig:",width=340 ] ( right ) for the signal events where at least two same - sign leptons with @xmath110 are found .
also shown are the same distributions from the dominant @xmath104 background ( see text).,title="fig:",width=340 ] the two dominant backgrounds exceeding the signal arise from @xmath104 or @xmath111 with one of the leptons mis - identified to have the wrong charge with a @xmath112 probability .
[ dist ] shows some distributions for the same sign di - lepton signal that should allow to discriminate it relatively easily from these backgrounds .
particularly effective would be a cut on the scalar sum , @xmath113 , of all the transverse momenta and of the transverse missing energy , that would only marginally affect the signal . in fig .
[ dist ] we have taken @xmath106 gev , @xmath114 gev , @xmath115 gev , @xmath116 and @xmath117 gev , corresponding to @xmath118 . to see a possible effect of the leptons in the cascade decays among the higgsinos we have also introduced a splitting among them by mixing with the electroweak gauginos at @xmath119 gev and @xmath120 gev . before any cuts are applied , the cross section at @xmath108 tev for the same sign di - leptons ( tri - leptons ) from top decays is @xmath121 . with an integrated luminosity of @xmath122 , cutting at @xmath123 tev leaves about 14 events with at least two same - sign leptons with @xmath110 against less than one event of our simulated total background .
no attempt is made to optimize the cuts , since a more realistic description of the background may be required .
a problem one encounters in studying the phenomenology of the mssm is the huge number of parameters that plagues the most general version of it .
even specific theoretically motivated versions of the mssm , based on particular mechanisms of supersymmetry breaking and transmission , involve quite a number of parameters , making difficult to give a compact description of the relevant phenomenology .
this is at least one good reason for looking at a particular configuration of s - particle masses that appears theoretically motivated and has the advantage of being describable in terms of relatively few physical parameters .
the typical spectrum of one such configuration of s - particles is shown in fig .
[ spettro ] , with the s - particle masses allowed to vary in the range given in sect .
1 . no doubt this is a particular case of the generic mssm .
yet it is useful to concentrate one s attention on it - we believe - as a particularly interesting case , for the reasons also given in sect . 1 .
the most characteristic signal to be expected at the lhc is due to the relative abundance of multi - top events .
there may be several ways to attack the detection of this signal .
most significantly , however , the top decays will produce a number of hard @xmath98 same - sign di - leptons or tri - lepton events with rates , for any given gluino mass , that can be easily read from fig.s [ btb ] and [ rates ] in any point of the parameter space . what counts most is the inclusive branching ratio @xmath56 defined in eq .
( [ defbtb ] ) .
although the study of the precise discovery potential of the lhc has not been our main motivation , we believe that it should be possible to explore the entire range of masses described in sect .
1 already in a relatively early stage of the lhc , say at @xmath124 tev and an integrated luminosity of 1 @xmath125
. we would not be too surprised if this turned out to be the first signal of manifest supersymmetry .
how generic are these considerations from the point of view of the lhc signals ?
while they are certainly not completely generic , they can apply with little or even no modifications at all also to some other configurations of s - particle masses , more or less theoretically motivated but certainly possible .
the important thing is that no squarks other than @xmath7 and @xmath11 are introduced below 1 tev .
sleptons , on the contrary , can be anywhere above @xmath2 . these modifications alone would introduce essentially no change of the picture described above .
some change would on the contrary occur by bringing down the electroweak gaugino masses relative to fig .
[ spettro ] .
this would make the full chargino - neutralino sector enter into the gluino decay chain , with the net result , however , of increasing the number of high @xmath98 multi - lepton events .
other than those ones coming from the semi - leptonic top decays , one would also have the leptons from the cascade decays inside the chargino - neutralino sector .
this is partly true even in the case of the spectrum of fig .
[ spettro ] , depending on the cut of the @xmath98 of the leptons . a motivation for looking at some modifications of the s - particle spectrum of fig .
[ spettro ] along the lines mentioned above might come from dark matter considerations . if thermally produced in the early universe , the cosmological abundance of the lightest higssino in fig .
[ spettro ] would in fact be insufficient to make the observed amount of dark matter .
finally one can ask what happens on the low side of the gluino mass range that we have considered , or even below it , for the same configuration of the other s - particles .
the answer is clear .
as long as @xmath126 is kinematically allowed to go , there would again be no significant modification of the picture described above with @xmath56 close to 0.5 . at extreme values of the parameters ,
- say @xmath127 gev and @xmath128 gev - , this might even be of interest for the current searches at the tevatron@xcite .
this work is supported by the eu under rtn contract mrtn - ct-2004 - 503369 , by the miur under contract 2006022501 and by the swiss national science foundation under contract no .
200021 - 116372 .
we thank gino isidori for many useful discussions and comments ; d.p .
thanks roberto franceschini and jan mrazek for valuable hints and discussions .
for ease of the reader we collect here the expressions for the relevant decay widths used in the text . in each partial amplitude
the same phase - space factor appears , so we define @xmath129 the two - body gluino decay amplitudes are given by @xmath130 finally we list the partial widths for the _ stops _ and the _ sbottom _ s - quarks @xmath131 @xmath132 we quote the formula used for the calculation of the higgs boson mass .
this accounts for the one - loop and the leading - log two loops contribution from the top and the stops @xcite .
@xmath133 @xmath134 with @xmath135 and @xmath136 where @xmath137 is the running top quark mass evaluated at the pole .
99 a. pomarol and d. tommasini , nucl .
b * 466 * , 3 ( 1996 ) [ arxiv : hep - ph/9507462 ] .
r. barbieri , g. r. dvali and l. j. hall , phys .
b * 377 * , 76 ( 1996 ) [ arxiv : hep - ph/9512388 ] .
r. barbieri , l. j. hall and a. romanino , phys .
b * 401 * ( 1997 ) 47 [ arxiv : hep - ph/9702315 ] .
for a recent review see g. f. giudice , m. nardecchia and a. romanino , nucl .
b * 813 * ( 2009 ) 156 [ arxiv:0812.3610 [ hep - ph ] ] .
r. barbieri and g. f. giudice , phys .
b * 309 * ( 1993 ) 86 [ arxiv : hep - ph/9303270 ] .
for a recent review see u. haisch , arxiv:0805.2141 [ hep - ph ] .
s. dimopoulos and g. f. giudice , phys .
b * 357 * , 573 ( 1995 ) [ arxiv : hep - ph/9507282 ] . c. bemporad
, http://www.slac.stanford.edu/spires/find/hep/www?irn=8121486[spires entry ] _ prepared for 4th international workshop on neutrino oscillations in venice : ten years after the neutrino oscillations , venice , italy , 15 - 18 apr 2008 _ r. kitano and y. nomura , phys .
d * 73 * ( 2006 ) 095004 [ arxiv : hep - ph/0602096 ] .
g. gamberini , z. phys .
c * 30 * ( 1986 ) 605 .
h. baer , v. d. barger , d. karatas and x. tata , phys .
d * 36 * ( 1987 ) 96 .
r. barbieri , f. caravaglios , m. frigeni and m. l. mangano , nucl .
b * 367 * ( 1991 ) 28 .
r. m. barnett , j. f. gunion and h. e. haber , phys .
b * 315 * ( 1993 ) 349 [ arxiv : hep - ph/9306204 ] .
m. guchait and d. p. roy , phys .
d * 52 * , 133 ( 1995 ) [ arxiv : hep - ph/9412329 ] .
j. hisano , k. kawagoe , r. kitano and m. m. nojiri , phys .
d * 66 * , 115004 ( 2002 ) [ arxiv : hep - ph/0204078 ] .
s. kraml and a. r. raklev , phys .
d * 73 * ( 2006 ) 075002 [ arxiv : hep - ph/0512284 ] .
s. p. martin , phys .
d * 78 * , 055019 ( 2008 ) [ arxiv:0807.2820 [ hep - ph ] ] .
m. toharia and j. d. wells , jhep * 0602 * ( 2006 ) 015 [ arxiv : hep - ph/0503175 ] .
g. mercadante , j. k. mizukoshi and x. tata , phys .
d * 72 * ( 2005 ) 035009 [ arxiv : hep - ph/0506142 ] .
h. baer , v. barger , g. shaughnessy , h. summy and l. t. wang , phys .
d * 75 * ( 2007 ) 095010 [ arxiv : hep - ph/0703289 ] .
b. s. acharya , p. grajek , g. l. kane , e. kuflik , k. suruliz and l. t. wang , arxiv:0901.3367 [ hep - ph ] . b. lillie , j. shu and t. m. p. tait , jhep * 0804 * , 087 ( 2008 ) [ arxiv:0712.3057 [ hep - ph ] ] .
r. contino and g. servant , jhep * 0806 * , 026 ( 2008 ) [ arxiv:0801.1679 [ hep - ph ] ] .
a. pomarol and j. serra , phys .
d * 78 * , 074026 ( 2008 ) [ arxiv:0806.3247 [ hep - ph ] ] .
torbjorn / pythia.html | we draw attention on a particular configuration of supersymmetric particle masses , motivated by naturalness and flavour considerations .
all its relevant phenomenological properties for the lhc are described in terms of a few physical parameters , irrespective of the underlying theoretical model .
this allows a simple characterization of its main features , useful to define a strategy for its discovery .
= 1 1.0 cm * s - particles at their naturalness limits * 1.0 cm riccardo barbieri@xmath0 and duccio pappadopulo@xmath1 + _
@xmath0 scuola normale superiore and infn , piazza dei cavalieri 7 , 56126 pisa , italy _
+ _ @xmath1 institut de thorie des phnomnes physiques , epfl , ch1015 lausanne , switzerland_1.0 cm |
recent observations of unevolved halo stars have confirmed the surprising linear relationship of the be and b abundances with [ fe / h ] ( @xcite , @xcite ) .
this primary behaviour of be and b can not be understood on the basis of the standard model ( sm ) of libeb production in the interstellar medium ( ism ) from gcr spallation .
this is because in the sm the libeb production rates are proportional to the global metallicity of the ism , so that a secondary behaviour with [ fe / h ] is expected .
however , the sm assumes a _ gcr flux constant in time _ and equal to that currently oberved at the earth .
there are many reasons to believe that the gcr flux was different in early epochs . in this contribution we solve the time - dependent propagation equation of crs in the galaxy in a simple and parametrized way to test whether the sm can be compatible with observations when the hypothesis of the cr flux constancy in time is relaxed .
we consider a general time - dependent equation for the gcr propagation : @xmath0 where @xmath1 is the number density of gcr particles per energy interval as a function of energy per nucleon and time , @xmath2 is the loss time - scale including escape , @xmath3 , and inelastic nuclear interaction , @xmath4 .
@xmath5 represents the ionization energy losses in the ism and finally , @xmath6 is the source spectrum of gcrs .
we make the following assumptions : \1 ) the energy and time dependences of @xmath7 , @xmath8 and @xmath9 are separable : @xmath10 2 ) we assume that gcrs are produced by supernovae ( sn ) so that @xmath11 , the sn rate , taken from a chemical evolution model which reproduces the current sn rates in the galaxy and the main characteristics of the solar neighbourhood .
the source energy spectrum is : @xmath12^{1.5}}\ ] ] with @xmath13 mev .
the constant @xmath14 is chosen in such a way that we can reproduce the observed current gcr flux at the earth .
\3 ) @xmath2 can be split into two parts @xmath15 where @xmath16 represents the escape time - scale from the galaxy , and @xmath17 is the loss time - scale against inelastic nuclear interaction .
the energy dependence of @xmath16 is taken from @xcite and that of @xmath17 from @xcite assuming a gas density of 1 @xmath18 .
\4 ) the energy losses due to ionization in the ism have been considered according to @xcite .
\5 ) the time - dependence of the ionization losses and of the inelastic nuclear interactions can be plausibly caused by temporal changes in the gas density .
the time - dependence of the escape time - scale is more difficult to understand . here
we treat it as a free parameter . for both we assume , for the sake of simplicity , exponential functions
: @xmath19 where @xmath20 is the characteristic e - folding time of gas exhaustion in the galaxy and @xmath21 is the characteristic time - scale for variations of @xmath16 .
we assume that @xmath2 was higher in the past , during the halo collapse ( t@xmath22 gyr ) .
eq . ( 1 ) can be solved using the method of characteristics .
the solution can be simplified considerably if @xmath23 yr .
plausible values for @xmath20 and @xmath21 are longer than @xmath24 yr , so in that case the solution is @xmath25 where @xmath26 is the ionization energy loss time - scale calculated as : @xmath27 then , the rate of accumulation of libeb in the ism is calculated in the usual way ( see @xcite ) .
we have considered the following cases : * case 1 : * the energy dependence of the gcr source spectrum is given by eq .
( 3 ) with @xmath28 equal to the sn rate evolution .
the remaining parameters are constant in time
. * case 2 : * case 1 plus @xmath29 from eq .
( 5 ) with @xmath30 gyr during the first 2 gyr of galactic evolution and @xmath31 afterwards ( i.e no time variation of @xmath32 ) .
this means that the cr escape time - scale at @xmath33 was a factor @xmath34 higher than now .
figures 1 and 2 show the calculated time - dependent gcr fluxes as well as the corresponding libeb evolution with [ fe / h ] .
the results are : * case 1 : * the flux is _ higher in the past _ and decreases slowly with time in the whole energy range .
this higher flux implies a larger libeb production in the past than the sm model , although , this is still not enough to reproduce the observations . * case 2 : * the flux in the past is _ also higher and the spectrum is flatter at energies e@xmath35 _ mev
/ n . in consequence
the libeb production is even higher at early epochs than in _ case 1_. note that the predicted be evolution is nearly in agreement with the observations .
on the other hand , the effect of considering the temporal evolution of the gas content in the galaxy is not very important for the libeb production . in this case
the flux and spectrum are modified in a complex way depending on the value of @xmath20 and the interplay between @xmath36 and @xmath17 .
our results show that variations of the gcr flux in time have to be considered to study the early production of libeb by spallation in the ism .
a higher flux and a flatter cr spectrum in the past are able to improve the agreement between the predictions of the ( modified ) sm with observations .
this is evident for be and b. for li , however , another important source of production is still required .
a higher flux in the past can be achieved by taking into account the temporal evolution of the gcr sources , which is naturally explained by the temporal variation of the sn rate , or by assuming that the escape time - scale @xmath37 was higher in the past .
the later assumption can improve the agreement between data and model considerably .
however , it still requires a more solid theoretical justification ( @xcite ) .
99 duncan d.k . , primas f. , rebull l.m . , boesgaard a.m. , deliyannis c.p . , hobs l.m . , king j.r . ,
1997 , 488 , 338 fields b. , olive k. , schramm d. 1994 , , 435 , 185 garca - lpez r. , lambert d.l , gustafsson b. , rebolo r. 1998 , ( in press ) garca - muoz m. , simpson j.a .
, guzik t.g .
, wefel j.p . , margolis s.h . , 1987 , s 64 , 269 longair m.s .
1992 , _ high energy astrophysics _ , cambridge university press , cambridge , p. 49
meneguzzi m. , audouze j. , reeves h. , 1971 , 15 , 337 prantzos n. , cass m. , vangioni - flam e. , 1993 , 403 , 630 | it is well known that no time - independent galactic cosmic ray ( gcr ) flux model can account for the observed linear relationship of be and b abundances with metallicity . in this contribution
we solve the time - dependent cr propagation equation using simple analytic time dependences for the cr spectrum at the source , the galactic cr escape - length and gas density evolution .
the cr fluxes so obtained are used to study the early evolution of the libeb abundances in the galaxy . |
in its most generic form , the persistence problem is concerned with the fraction of space which persists in its initial @xmath0 state up to some later time @xmath1 .
it is a classic problem which falls into the general class of so - called first passage problems [ 1 ] and has been extensively studied over the past decade or so for model spin systems by physicists [ 2 - 6 ] .
persistence has been investigated at both zero [ 2 - 5 ] and non - zero [ 6 ] temperatures .
typically , in the non - equilibrium dynamics of spin systems at zero - temperature [ 2 - 5 ] , the system is prepared initially in a random state and the fraction of spins , @xmath2 , that persists in the same state as at @xmath3 up to some later time @xmath1 is monitored . at a finite temperature , on the other hand , one is interested in the global persistence behaviour and one monitors the change in the sign of the magnetization in a collection of non - interacting systems [ 6 ] .
it is now well established that the persistence probability decays algebraically [ 2 - 6 ] @xmath4 where @xmath5 is the non - trivial persistence exponent .
note that the value of @xmath6 depends not only on the spatial [ 4 ] ( @xmath7 ) and the spin [ 7 ] ( @xmath8 ) dimensionalities , but also on whether the temperature , @xmath9 , is zero or finite .
it is only for @xmath10 and @xmath11 that @xmath12 is known exactly [ 5 ] ; see ray [ 8 ] for a recent review .
we merely mention here that at criticality , @xmath13 , [ 6 ] , @xmath14 for the pure two - dimensional ising model .
more recently it has been discovered that disorder [ 9 - 11 ] also alters the persistence behaviour .
a key finding [ 9 - 10 , 12 ] is the appearance of blocking in systems containing disorder .
blocked spins are effectively isolated from the behaviour of the rest of the system in the sense that they _
never _ flip . as a result ,
@xmath15 .
the persistence exponent has also been obtained from a wide range of experimental systems and the values range from 0.19 - 1.02 , depending on the system [ 13 - 15 ] .
a considerable amount of time and effort has been taken up in trying to obtain estimates of @xmath5 for different models and systems . in this work
, we present one of the first estimates of the persistence exponent from financial data from the japanese market .
the main motivation behind the present study is to compare and contrast the behaviour of the japanese and uk markets .
it was recently [ 16 ] found that the persistence behaviour of the london market displays an interesting double power law behaviour . as we shall see
, we find very similar features in the japanese market . in the next section
we give a brief outline of the methodology used .
section iii presents our results and we finish with a brief conclusion in section iv .
financial markets contain many of the features found in model systems that have been studied over the past 50 years or so in statistical physics .
for example , share values of most banks would tend to move in the same direction if the central bank rate is raised or reduced .
this behaviour is analagous to the tendency of spins connected by a ferromagnetic coupling to align themselves at low temperature .
market behaviour analagous to antiferromagnetic coupling is also seen in real markets .
for example ,
the above news regarding the central bank rate would have the opposite effect on the values of shares in manufacturing companies .
furthermore , interactions between different companies exist and are highly non - trivial to model . in our data analysis
we make no attempt to model these interactions .
we simply work with historical data from the nikkei index from 2002 .
the pre - processed financial datatset was kindly supplied to us by t. kaizoji and l. pichl ( international christian university , tokyo , japan ) . for the purposes of this study we extracted end - of - day ( eod ) data from the database to enable a direct comparison with the earlier study made with the data from the london ftse100 market [ 16 ] .
as mentioned , we work with eod share prices quoted in japanese yens . we follow the earlier work on the london market and map the share values onto ising spins . the dataset for 2002
was first partitioned into quarterly data ( jan - mar , apr - jun , jul - sep , oct - dec ) . in figure 1
we display the behaviour of the share values of 3 typical companies over a randomly chosen quarter .
the share values are in japanese yen .
the dashed lines indicate the respective share values at @xmath3 . in this work
we restrict ourselves to the first passage time , that is the first time a given share value crosses its value at @xmath3 . .
the top plot ( a ) indicates that the share value remains below the base price for the duration of the quarter considered . in the middle plot ( b )
the time at which the share value returns to the base price for the first time is indicated by the first arrow .
similarly , the arrows in the bottom plot ( c ) indicate the first 4 times at which the share value returns to the price at @xmath3 . ]
the start of each quarter is designated as day 0 and the share values take on their base values .
the share values at the end of trading on the next day ( day 1 ) are compared with the corresponding base values .
a spin @xmath16 is associated with each share value @xmath17 .
we allocate the values as follows @xmath18 table 1 shows a typical mapping .
.a typical example of the mapping used . note that the value of @xmath16 is determined with reference to the price at @xmath3 . [ cols="^,^,^,^,^",options="header " , ] it should be emphasised that the value of the spin is always with reference to the share value at @xmath3 and _ not _ the eod price at the previous day .
consequently , in this work we disregard all fluctuations which take place during the day . in our example in table 1 , the spin has flipped at @xmath19 .
the values of @xmath20 form the initial configuration for our spin system .
each 3-monthly dataset is converted into possible values of ising spins , @xmath16 with the share values at @xmath3 as the reference prices .
this allows us to track the share values relative to their prices at @xmath3 . in figure 1 the first passage time
is indicated by the first arrow for any given share .
subsequent arrows indicate the second ( and later ) passage times .
to complete the analogy with the persistence problem as studied in statistical physics , we look for the _ first time _
@xmath21 as this corresponds to the underlying share value either going above ( if @xmath22 ) or below ( if @xmath23 ) the price at @xmath3 , also for the first time . for an anlysis of the first passage properties ,
all subsequent values of @xmath16 are disregarded . at each time step
, we count the number of spins that still remain in their initial @xmath0 state and the total number of spins , @xmath24 , which have never flipped until time @xmath1 is given by [ 12 ] @xmath25 the density of non - flipping spins , @xmath26 , is the key measure of interest here , namely @xmath27 where @xmath28 , the number of companies appearing in the nikkei index . [ in the actual analysis discussed in this paper , we had to work with @xmath29 as the share values of one of the constituent companies of the nikkei index were not quoted due to some financial irregularity . ]
we now discuss our main results . figure 2 shows a log - log plot of the density of non - flipping spins against time .
the data has been averaged over all of the samples . against @xmath1 for the japanese data .
we can clearly see the appearance of a double power law with the presence of a distinctive shoulder .
the inset shows the corresponding plot for the data from the london market [ 16 ] .
the line with slope @xmath30 is for reference purposes only and is the exact result in @xmath31 dimension for a symmetric random walk . ]
we can clearly see the presence of a double power law with a distinctive shoulder . whereas for short times we estimate the slope to be @xmath32 , for longer times we find @xmath33 .
for information we also display the corresponding plot [ 16 ] from the london ftse100 data as an inset in figure 2 . in the earlier analysis of the london data
[ 16 ] it was found that the short - time ( typically up to about a week ) persistence exponent is @xmath34 and for longer times its @xmath35 . note that as here we are averaging over only 4 samples with the nikkei data , the error - bars are much bigger than those obtained for the london data which was averaged over 25 samples .
nevertheless , the plot confirms that the behaviour of the japanese data is consistent with that seen on the london ftse100 market .
a financial market contains different types of traders .
there are usually both short - term ( speculative ) and long - term traders . clearly , those traders speculating on the market are likely to react over a much shorter time scale than those trading for the long - haul .
the double power law discovered in our analysis could be a signature of the presence of both types of traders .
of course , in addition , there are intraday traders hoping to make a profit by trading at even higher frequency ( eg hourly ) .
as we work with eod data , we do not expect to see the behaviour of such traders in our data .
both the japanese nikkei and the london ftse100 markets are well developed and yield very similar features .
we note that the corresponding global persistence exponent for a one - dimensional symmetric random walk is given from [ 1 ] @xmath36 for reference , the value of the exponent for the @xmath37 case is shown in figure 2 .
clearly , from our data we can exclude such a model .
to conclude we have used a novel mapping to map eod share values on the nikkei index onto ising spins .
the mapping can be re - interpreted in term of the first passage crossing time .
we find clear evidence for a double power law decay with different exponents for short and long time behaviour .
our observation is consistent with the recent findings on the london ftse100 market .
we suggest that the double - power is the signature of the presence of different types of traders in the market .
namely , traders speculating on a daily basis and those taking a long - term view .
it would be interesting to see whether the presence of really high frequency intraday traders can be deduced from minute - by - minute tick data .
we would like to thank the daiwa anglo - japanese foundation for funding this research via a small grant ( ref : 6124/6356 ) .
t. kaizoji ( icu , tokyo , japan ) is acknowledged for providing access to tick data on the nikkei index and l. pichl ( icu , tokyo , japan ) for pre - processing the data .
[ 1 ] s. redner , _ a guide to first - passage processes _ , cambridge university press ( 2001 ) . | the persistence phenomenon is studied in the japanese financial market by using a novel mapping of the time evolution of the values of shares quoted on the nikkei index onto ising spins .
the method is applied to historical end of day data from the japanese stock market during 2002 . by studying the time dependence of the spins , we find clear evidence for a double - power law decay of the proportion of shares that remain either above or below starting values chosen at random
. the results are consistent with a recent analysis of the data from the london ftse100 market .
the slopes of the power - laws are also in agreement .
we estimate a long time persistence exponent for the underlying japanese financial market to be 0.5 . |
generalization of some well - known stochastic processes indexed by a single parameter to processes indexed by two parameters has attracted considerable interest recently .
for example , houdre and villa @xcite generalized fractional brownian motion parametrized by a single hurst index to the bifractional brownian motion characterized by two indices
. another example is provided by the multidimensional stationary gaussian fields with generalized cauchy covariance indexed by two parameters introduced by gneiting and schlather @xcite .
these processes can be regarded as extension of the gaussian processes with cauchy covariance used in geostatistics . for simplicity , we call such processes the gaussian field with generalized cauchy covariance ( gfgcc ) . here
we would like to point out that one should not confuse such a process with a stable process with cauchy marginals . in general
, processes parametrized by two indices can provide more flexibility in their applications in modeling physical phenomena .
in particular , the gfgcc model has an additional nice and useful property as it allows separate characterization of fractal dimension and long range dependence ( lrd ) by two different indices .
this is in contrast to models based on fractional brownian motion ( or fractional brownian noise ) which use a single index to characterize these two properties .
models based on a stochastic process or field parametrized by a single index seem inadequate .
a detailed analysis on network traffic carried by park et al @xcite shows that fractional brownian motion / fractional gaussian noise model is inadequate for description of network traffic for all scales since at very small time scales the traffic fluctuations are no longer statistically self - similar .
the need to replace global scaling by local scaling is essential for processes such as multifractional brownian motion introduced independently by peltier and vehel @xcite and benassi , jaffard and roux @xcite .
the main aim of this paper is to study gfgcc and its anisotropic counterpart , which we call gaussian sheet with generalized cauchy covariance ( gsgcc ) . in view of the fact that gfgcc is widely used in geostatistics and other applications @xcite
, it will be useful to consider its properties in more detail . in these existing applications , usually only the covariance structure of gfgcc
are used , and the sample path properties of gfgcc are rarely mentioned . however , a better understanding of the sample properties of gfgcc and gsgcc will render more versatility and flexibility to their applications .
our approach to this subject is mainly from a physical viewpoint .
basic sample properties such as the long range dependence and the local self - similarity properties of gfgcc and gsgcc are investigated .
a simpler method is used to derive the asymptotic properties of the spectral densities of gfgcc and gsgcc . by generalizing the lamperti transformation to @xmath2-dimensional processes ,
new types of random field and random sheet with global self - similar property associated with gfgcc and gsgcc are obtained .
properties of these random field and random sheet are also studied .
in this section we consider gfgcc , which is a multidimensional isotropic gaussian random field in @xmath2-dimensional euclidean space .
we first introduce some notations and state some basic definitions and properties of gfgcc .
denote by @xmath3 , @xmath4 , @xmath5 and @xmath6 the sets of positive integers , integers , real numbers and positive real numbers respectively .
let @xmath7 and @xmath8 be two vectors in @xmath9 , and @xmath10 , @xmath11 be its euclidean norm . by @xmath12 and @xmath13 , we mean @xmath14 and @xmath15 respectively for all @xmath16 . a random field @xmath17 on @xmath9
is called a gaussian field with generalized cauchy covariance ( or gfgcc ) if it is a stationary gaussian field with mean zero and covariance given by @xmath18 where @xmath19 $ ] and @xmath20 .
note that has the same functional form as the characteristic function of the generalized multivariate linnik distribution first studied by anderson @xcite .
@xmath21 is positive - definite for the above ranges of @xmath22 and @xmath23 , and it is completely monotone for @xmath24 , @xmath25 .
@xmath17 becomes the gaussian field with usual cauchy covariance when @xmath26 .
recall that a random field @xmath27 is @xmath28-self - similar ( @xmath28ss ) if @xmath29 , where @xmath30 denotes equality in the sense of finite - dimensional distributions of @xmath31 .
self - similar property requires scale invariance to hold for all scales .
this is rather too restrictive for many applications .
we also know from samorodnitsky and taqqu @xcite that a stationary gaussian random field such as @xmath17 can not be a self - similar field .
however , @xmath17 satisfies a weaker self - similar property known as local self - similarity considered by kent and wood @xcite .
[ def2]let @xmath32 $ ] .
a centered stationary gaussian field is locally self - similar ( lss ) of order @xmath33 if for @xmath34 , its covariance @xmath35 satisfies @xmath36\end{aligned}\]]for some positive constants @xmath37 , @xmath38 and @xmath39 . since as @xmath34 , @xmath40,\end{aligned}\ ] ] the gfgcc @xmath17 is @xmath33 lss , with @xmath41 , @xmath42 and @xmath43 .
adler @xcite called the class of gaussian fields which satisfy the indexed-@xmath22 fields .
these processes are also known as the adler processes according to some authors , for example lang and roueff @xcite .
they form a very rich class of gaussian random fields , which include the centered gaussian field @xmath44 with powered exponential covariance @xmath45which have the same functional form as the characteristic function of the multivariate symmetric stable distribution as given in kotz , kozubowski and podgorski @xcite , and garoni and frankel @xcite . instead of using to characterize local self - similarity , one can also use the definition of locally asymptotically self - similar ( lass ) property first introduced by benassi , jaffard and roux @xcite for multifractional brownian motion @xmath46 , which is a generalization of fractional brownian motion with the hurst index replaced by the hurst function @xmath47 , @xmath48 .
it can be shown that under some regularity conditions on @xmath47 , the multifractional brownian motion @xmath46 is lass .
this property can be adapted to gfgcc if we take @xmath47 as constant with its value in ( 0,1 ) .
[ def1 ] a stochastic process @xmath27 is lass at a point @xmath49 with order @xmath50 if @xmath51and @xmath52 is nontrivial . here
the convergence and equality are in the sense of finite dimensional distributions , and @xmath52 is called the tangent field of @xmath27 at the point @xmath49 .
[ p1 ] gfgcc is a lass random field of order @xmath33 ; and its tangent field is l@xmath53vy fractional brownian field of index @xmath33 . by using ,
the covariance of the increment field @xmath54 , for @xmath55 , @xmath56 is given by @xmath57 let @xmath58 and @xmath59 , then @xmath60 where @xmath61 is the index@xmath33 l@xmath53vy fractional brownian field with zero mean and covariance given by @xmath62 thus up to a multiplicative constant @xmath63 , the tangent field of gfgcc at any point @xmath64 is the l@xmath53vy fractional brownian field indexed by @xmath33 .
the tangent field of @xmath17 at the point @xmath49 reflects the local structure of the random field at @xmath49 .
in other words , gfgcc behaves locally like a l@xmath53vy fractional brownian field .
this provides an example to the general results on tangent fields considered by falconer @xcite .
the fractal dimension of the graph of a random field @xmath27 depends on the local property of the random field .
the local irregularities of the graph are measured by the parameter @xmath22 , which can be regarded as the fractal index of the random field .
thus the behavior of the covariance function at the origin to a great extent determines the roughness of the random field .
the results on the fractal dimension of an lss field are treated by adler @xcite , kent and wood @xcite , davies and hall @xcite .
a nice property of @xmath33lss fields is that their fractal dimension is determined by @xmath22 .
let @xmath27 be a stationary gaussian field and let @xmath65 be the variance of the increment process @xmath66 .
if there exists @xmath67 $ ] satisfying @xmath68 then @xmath33 is called the fractal index of the random field @xmath27 .
equivalently , @xmath33 is the local hlder index of the random field .
clearly , condition is fulfilled by random field which satisfies such as the gfgcc @xmath17 .
adler had shown that @xmath33 is the upper bound of the indices for which , with probability one , the graph of @xmath27 satisfies a global regularity of the same order .
thus , @xmath22 characterizes the roughness of the sample path .
the fractal dimension of gfgcc can be obtained by using the following result for the fractal ( hausdorff ) dimension of an lss field as given in adler @xcite , chapter 8 .
[ p1026_4 ] the fractal dimension @xmath69 of the graph of a locally self - similar field @xmath70 , of fractal index @xmath33 , over a hyperrectangle @xmath71 $ ] , is given by @xmath72 the estimation of @xmath22 for lss field has been studied extensively , see for example wood and chan @xcite , istas and lang @xcite , kent and wood @xcite , lang and roueff @xcite .
the parameter @xmath23 is also known as topothesy in the studies of roughness of surfaces by wood and chan @xcite , and davies and hall @xcite .
the topothesy of a cross section provides a measure of the roughness which is scale dependent in contrast to fractal dimension which is scale invariant .
the gaussian random field @xmath17 can have srd ( short range dependence ) or lrd ( long range dependence ) , depending on the values of the parameters @xmath22 and @xmath23 .
for this purpose we make use of the following definition which is a generalization of the one - dimensional case considered by flandrin et al @xcite , lim and muniandy @xcite : a stationary centered gaussian field with covariance @xmath35 is said to be a long range dependent process if @xmath73 otherwise it is short range dependent . [ p2]the gfgcc @xmath17 is a long range dependent random field if and only if @xmath74 . in order to obtain the condition for the gaussian random field with covariance to be lrd , we make use of the following integral identity given in gradshteyn and ryzhik @xcite , 3.251 , no.11 : @xmath75 where @xmath76 , @xmath77 and @xmath78 is the beta function . using polar coordinates
, one gets @xmath79 for large @xmath80 , @xmath81 therefore , the integral is divergent for all @xmath20 , @xmath74 . for @xmath20 and @xmath82
, we have @xmath83 therefore the condition for @xmath17 to be a gaussian field with lrd is @xmath74 .
the discussion above shows that it is possible to characterize the fractal dimension @xmath69 and the lrd property separately .
if the covariance is re - expressed as @xmath84 , which behaves like @xmath85 in the large@xmath86 limit , then @xmath87 is lrd if and only if @xmath88 . thus @xmath22
( @xmath89 ) and @xmath90 ( @xmath91 ) respectively provide separate characterization of fractal dimension and lrd / srd .
the separate characterization of the fractal dimension ( local property ) and lrd ( global property ) for gfgcc appears to offer a more natural and flexible model than that based on a single parameter such as in l@xmath53vy fractional brownian field .
we note that this feature of separate characterization of local self - similarity ( hence fractal dimension ) and long range dependence is present in any stationary gaussian field with covariance @xmath35 satisfying the asymptotic behaviors @xmath92 as @xmath93 , and @xmath94 as @xmath95 , with @xmath19 $ ] , @xmath91 .
similarly , one can also have a gaussian stationary process which has separate parametrization of fractal dimension and short range dependence @xcite .
the ability to have separate characterization of fractal dimension and hurst effect is a desirable property in the modeling of physical and geological phenomena .
in this section , we consider the spectral density of @xmath17 and its asymptotic properties .
though the covariance of gfgcc is given by a relatively simple expression , the analytic simplicity of the covariance function is not inherited by the corresponding spectral density .
this is similar to the case of the stationary gaussian field @xmath44 with powered exponential covariance which have simple form , but its spectral density in general does not have closed analytic expression . in the case of gfgcc in @xmath5 , a detailed study of spectral densities ( in terms of probability distributions correspond to the characteristic functions of generalized linnik distributions ) have been carried out by kotz et al .
@xcite for @xmath96 , @xmath97 , @xmath98 ; by ostrovskii @xcite for @xmath96 , @xmath97 , @xmath99 ; and by erdogan and ostrovskii @xcite for @xmath96 , @xmath20 , @xmath98 .
they employed the contour integration representations and series expansions of the generalized linnik distributions .
however the techniques used in these works are less accessible to practitioners . in this section ,
we derive the asymptotic properties of the spectral densities of gfgcc for @xmath89 and @xmath20 , which can be regarded as an extension to the results on generalized multivariate linnik distributions .
the techniques used in our derivations are mathematically more tractable .
recall that the spectral density @xmath100 of a stationary field @xmath27 is defined as the fourier transform of its covariance function @xmath101 : @xmath102if the integral is convergent .
if the integral does not converge , we consider @xmath103 as a generalized function and define @xmath100 as the fourier transform of @xmath103 in the schwartz space of test functions @xcite .
namely , for any test function @xmath104 in the schwartz class of @xmath9 , we require @xmath105where @xmath106 alternatively , the spectral density can also be defined to be the function satisfying @xmath107 for gfgcc , since @xmath108as @xmath109 ( @xcite , page 209 ) , we find that when @xmath110 , it s spectral density is @xmath111where @xmath112 is the bessel function .
when @xmath113 , using no.4 of 6.565 in @xcite , we have the explicit formula @xmath114if @xmath115 . here
@xmath116 is the modified bessel function .
the formula no.7 of 6.576 in @xcite shows that if @xmath117 , then @xmath118therefore is still the spectral density when @xmath119 $ ] . for general @xmath120 , no explicit formula such as can be found for @xmath121 . when @xmath98 , the formula gives the spectral density of @xmath17 for all values of @xmath19 $ ] and @xmath20 . for @xmath122
, we would also like to find a formula for the spectral density that is valid for all @xmath123 and @xmath20 .
for this purpose , it would be beneficial to investigate the case @xmath98 first .
when @xmath98 , we can rewrite as @xmath124let @xmath125 and consider the region @xmath126 in the complex plane defined by @xmath127when @xmath128 , the function @xmath129 is an analytic function on the domain @xmath126 .
therefore , by cauchy integral formula , @xmath130notice that the boundary of @xmath126 , @xmath131 , consists of three components : the line segment @xmath132 along the real axis from @xmath133 to @xmath80 , the arc @xmath134 of the circle @xmath135 from @xmath80 to @xmath136 , and the line segment @xmath137 along the imaginary axis from @xmath136 to @xmath133 . on the arc @xmath134 ,
if @xmath138 , then @xmath139therefore @xmath140and implies that @xmath141this gives us @xmath142 for @xmath122 , we can derive a formula similar to .
recall that the hankel s function of the first kind @xmath143 is defined as @xmath144 where @xmath145 is the modified bessel function of the second kind or called the neumann function . using hankel s function ,
we can rewrite as @xmath146for @xmath109 , we have ( @xcite , no.3 of 8.451 ) @xmath147therefore , we can show as in the @xmath98 case that @xmath148using the formula ( @xcite , no.1 of 8.407 ) @xmath149 we have finally @xmath150this formula agrees with the formula for multivariate linnik distribution proved in @xcite for @xmath151 and @xmath99 .
notice that the right hand side of is well - defined for all @xmath152 . using the formula ( no.2 of 6.521 in @xcite ) ,
@xmath153we have @xmath154 @xmath155 when @xmath123 , residue calculus implies that this last integral is equal to @xmath156this shows that is indeed the spectral density of gfgcc for all @xmath123 and @xmath20 .
we would also like to remark that although the formula is derived under the assumption @xmath122 , but since @xmath157 therefore when @xmath98 , the formula reduces to the formula .
we summarize the result as follows .
if @xmath123 and @xmath20 , the spectral density of the gfgcc @xmath17 is given by @xmath158if @xmath113 and @xmath20 , the spectral density of the gfgcc @xmath159 is given by @xmath160 to find the high frequency behavior of the spectral density , we first consider the case where @xmath113 .
using the fact that ( @xcite , no.6 of 8.451 ) as @xmath109 , @xmath161this implies that as @xmath162 , @xmath163with leading term @xmath164 for general @xmath123 , to find the high frequency behavior of @xmath121 , we make use of eq . . making a change of variable and using @xmath165 as @xmath166 , we find that @xmath167using the formula ( @xcite , no.16 of 6.561 ) @xmath168and the definition of asymptotic expansion ( @xcite , page 611 ) , we conclude that as @xmath169 , @xmath121 behaves asymptotically as @xmath170 notice that there is a drastic change of high frequency limit of @xmath121 when @xmath120 and @xmath113 . in fact , naively putting @xmath113 in give identically zero terms .
this is a hint that as @xmath162 , @xmath171 does not have polynomial decay , instead it decays exponentially as is verified by .
we summarize the results as follows : [ p1113_1 ] if @xmath123 and @xmath20 , the high frequency limit of the spectral density @xmath121 is given by the following asymptotic series @xmath172if @xmath113 and @xmath20 , the high frequency limit of the spectral density @xmath171 is given by the following asymptotic series @xmath173 when @xmath174 and @xmath175 , agrees with the results given in @xcite and @xcite respectively . in particular
, we observe that when @xmath123 , the high frequency behavior of the spectral density of gfgcc is @xmath176which is independent of @xmath23 .
kent and wood @xcite have shown that if a random field has spectral density satisfying , then its covariance satisfies with locally self similar property . however , the converse is not true .
in this connection we remark that for the gaussian stationary field @xmath44 with powered exponential covariance which is lss with @xmath177 , \hspace{1cm}\vert\tau\vert\rightarrow 0^+,\end{aligned}\]]its small @xmath86 behavior has a similar form as that of gfgcc .
thus it is not surprising that this two random fields have the same tail behavior for their spectral densities at high frequencies as given by .
the detailed calculation carried out by garoni and frankel @xcite for the probability distribution of the multivariate l@xmath53vy stable distribution with characteristic function given by confirms this
. we also note that can be used to verify that the tangent field at any point @xmath49 has spectral density which varies as @xmath178 for @xmath179 .
if we let @xmath180 , then the tangent field is just the l@xmath53vy fractional brownian field in @xmath9 .
such a relationship can be viewed as a consequence of the tauberian
abelian theorem ( see e.g. @xcite ) . for the low frequency behavior of the spectral density @xmath121
, we first consider the case @xmath113 .
using 8.485 , 8.445 , 8.446 of @xcite , we find that if @xmath181 , @xmath182whereas when @xmath183 , where @xmath184 is a nonnegative integer , @xmath185here @xmath186 is the logarithm derivative of the gamma function .
therefore from , we find that : @xmath187 if @xmath188 , then as @xmath189 , @xmath190 @xmath187 if @xmath191 , then as @xmath189 , @xmath192where @xmath90 is the euler constant .
@xmath187 if @xmath193 , then as @xmath189 , @xmath194 in fact , by considering the cases @xmath195 and @xmath196 separately and substituting the series and into , we can express the spectral density @xmath171 in terms of convergent power series in @xmath197 .
= .8 = .8 = .8 = .8 = .8 = .8 for general @xmath123 , the low frequency behavior of @xmath121 depends on the arithmetic nature of @xmath22 and @xmath23 .
the method we are going to employ does not allow the derivation of the whole asymptotic series as obtained by kotz et al @xcite , erdogan and ostrovskii @xcite .
we will only derive the leading behavior of the spectral density @xmath121 , which only depends on the algebraic conditions @xmath82 , @xmath198 or @xmath199 .
these conditions are less stringent than the arithmetic conditions considered in @xcite .
however , the simpler method employed here provides the necessary @xmath200 , @xmath189 asymptotic behaviors which are sufficient for most practical purposes . when @xmath82 , since ( @xcite , eq .
8.402 ) @xmath201we find from that as @xmath202 , @xmath203(@xcite , no.11 of 3.251 ) .
when @xmath198 , we make a change of variable on to get @xmath204 in view of the leading behavior of the bessel function @xmath112 as @xmath205 , we write @xmath121 as the sum of two terms @xmath206 and @xmath207 where @xmath208and @xmath209as @xmath202 , and show that @xmath206 has a finite limit . by lebesgue
s dominated convergence theorem , the limit is given by @xmath210 .
namely @xmath211this expression can be evaluated using regularization method .
more precisely , using lebesgue s dominated convergence theorem again , we find that @xmath212the formula no.14 of 6.561 in @xcite then gives @xmath213for the term @xmath207 , we make a change of variable @xmath214 or equivalently @xmath215 , to get @xmath216we split @xmath207 again into a sum of two terms @xmath217 and @xmath218 , where@xmath219and@xmath220when @xmath189 , @xmath221the integral is a convergent integral with value given by ( @xcite , 3.219 page 316 ) @xmath222 putting everything together , we find that @xmath223therefore , as @xmath224 , @xmath225 when @xmath199 , care has to be taken since is defined only for @xmath110 . making a change of variable , one finds that @xmath226it is easy to verify that for any @xmath227 and @xmath20 , when @xmath189 , @xmath228now using no.1 of 6.621 in @xcite , we have @xmath229where @xmath230 is the hypergeometric function @xmath231using the formula no.2 of 9.131 in @xcite , we find that @xmath232since @xmath233therefore , @xmath234where we have used the formula @xmath235 .
this gives for @xmath236 , @xmath237it is easy to verify that by putting @xmath113 in , and , we get back , and respectively . the low frequency behaviors of @xmath121 are summarized below .
[ p1113_2]for all @xmath19 $ ] and @xmath20 , the low frequency limit of the spectral density @xmath121 is given by @xmath238 we see that if @xmath74 , @xmath121 is divergent at the origin and if @xmath236 , @xmath239 as @xmath189 .
note that the condition @xmath74 agrees with the lrd condition .
in fact , it is a basic fact that for a stationary field with positive covariance , it is lrd if and only if its spectral density diverges at the origin .
we would also like to point out that the low frequency limit of the spectral density of the gaussian field with powered exponent covariance @xmath44 is finite as shown by garoni and frankel @xcite , in agreement with the fact that @xmath44 is srd .
we remark that by considering @xmath240 we can find the next order term in the asymptotic expansion of @xmath121 when @xmath189 using the same methods we employed above .
the results depend on more complicated conditions on @xmath22 and @xmath23 .
we also briefly remark that the asymptotic behavior of the spectral density at low frequency is connected to the large time behavior of the covariance function . for the covariance function @xmath21 which satisfies @xmath241where
@xmath242 is a slowly varying function for large @xmath86 , i.e. @xmath243 as @xmath244 for all positive constant @xmath245 , hardy littlewood karamata tauberian theorem ( see e.g. @xcite ) implies that the spectral density @xmath121 has the following asymptotic behavior @xmath246where @xmath247
we can also introduce another type of @xmath2-dimensional process @xmath248 , called gaussian sheet with generalized cauchy covariance ( gsgcc ) .
the gsgcc is a centered gaussian random field @xmath248 indexed by two multidimensional parameters @xmath249 and @xmath250 , with @xmath251 $ ] , @xmath252 for all @xmath16 , and with covariance given by @xmath253where @xmath254 is the covariance of one - dimensional gfgcc indexed by @xmath255 and @xmath256 .
heuristically , one can regard @xmath248 as the product of @xmath2 independent one - dimensional gfgcc processes @xmath257 .
however , since the product of independent normal random variables is not a normal random variable , we shall not write @xmath258 .
nevertheless , it is true that for any integer @xmath259 , @xmath260^m\right)=\prod_{i=1}^n e\left(\left[x_{\alpha_i,\beta_i}(t_i)\right]^m\right).\ ] ] the gsgcc @xmath248 is an anisotropic gaussian random field and it does not satisfy the definition of local self similarity given in definition [ def2 ] . however , we can show that it is lass and has tangent field according to definition [ def1 ] .
assume that @xmath261 .
namely , for some @xmath262 , @xmath263 , and for all @xmath264 , @xmath265 . here @xmath266 [ p1026_2]gsgcc is lass with tangent field @xmath267 , which is a stationary centered gaussian field with covariance @xmath268 from definition , it is easy to see that@xmath269\right\ } \hspace{0.5cm}\text{as}\hspace{0.5cm}|\tau_i|\rightarrow 0^+ , \;i=1,\ldots , n .
\end{aligned}\ ] ] therefore , @xmath270 from this proposition we see that in general , the tangent field @xmath271 of the gsgcc @xmath248 defined by definition [ def1 ] are uncorrelated in some of the directions of @xmath9 .
it does nt fully capture the locally self similar property of @xmath248 .
there is another measure of local self similarity , called local multiple self - similarity that better describe this kind of process .
recall that @xcite a stochastic process @xmath27 is called multi - self similar ( mss ) of index @xmath272 if and only if for any @xmath273 , @xmath274 it is obvious that a @xmath28-mss field is a @xmath275$]-ss field , but in general a self similar field is not multi self similar .
the notion of lss can be generalized accordingly .
a centered stationary gaussian field is @xmath2-ple locally self similar of order @xmath33 if its covariance function @xmath35 satisfies for @xmath276 , @xmath277\right),\end{aligned}\]]for some @xmath278 , @xmath279 . in view of ,
it is easy to see that @xmath248 is @xmath2-ple locally self similar with @xmath280 , @xmath281 and @xmath282 .
we also note that the local multi - self similarity is equivalent to @xmath283\;\;\text{as}\;\ ; \tau_i\rightarrow 0^+,\end{aligned}\ ] ] for any @xmath284 . here
@xmath285 is the unit vector in the @xmath286 direction .
this can also be rephrased as @xmath2-ple lass .
@xmath248 is @xmath2-ple lass .
more precisely , for every @xmath16 , @xmath287where @xmath61 is the one dimensional fractional brownian motion of index @xmath33 .
the proof is the same as proposition [ p1 ] .
one should observe that the @xmath2-ple lass here is different from the lass defined in definition [ def1 ] .
we take the limit process in each of the @xmath286direction separately and the limit process can be regarded as the partial derivative process of @xmath248 in the @xmath286-direction . a disadvantage of this definition
is that the limit process of an @xmath2-dimensional field is a one - dimensional process .
therefore , we will define another limit process that will better reflect local multi - self similarity , as were considered in @xcite and @xcite . given a stochastic process @xmath288
, we define its total increment by @xmath289 at the point @xmath290 as @xmath291when @xmath98 , this is the same as the increment , i.e. @xmath292 .
when @xmath293 , we have @xmath294which is also known as the rectangular increment . for general @xmath2 ,
given @xmath295 , the set of points @xmath296 are the vertices of the hyperrectangle with @xmath297 and @xmath298 as one of the main diagonals . by giving
the vertex @xmath298 a weight @xmath299 , the other vertices of the hyperrectangle can be given a weight @xmath299 and @xmath300 alternatingly so that adjacent vertices has different weights .
@xmath301 is then the weighted sum of the field @xmath27 at the vertices of the hyperrectangle .
recall that the fractional brownian sheet @xmath302 of index @xmath303 is a centered gaussian process with covariance @xmath304it is well known that @xmath305 is a self similar field of order @xmath306 , and multi self similar of order @xmath33 .
however , unlike the l@xmath53vy fractional brownian field , the fractional brownian sheet is not a process with stationary increment .
nevertheless , it is a process with stationary total increment , i.e. , @xmath307 now returning to the local asymptotic multi self
similarity of gsgcc , we can show the following : [ p1026_5]@xmath308 b_{\alpha/2}^{\#}(u).\end{aligned}\]]here @xmath309 .
if we heuristically write @xmath248 as @xmath310 , then heuristically the total increment @xmath311 can be written as the product of increments @xmath312 , and the result follows from proposition [ p1 ] . for a more rigorous proof ,
notice that @xmath313e_i\right)\\ = & \sum_{\delta\in \{0,1\}^n}\sum_{\eta\in\{0,1\}^n}(-1)^ { \sum_{i=1}^n \delta_i+\sum_{i=1}^n \eta_i } \prod_{i=1}^n c_{\alpha_i,\beta_i}\left({\varepsilon}_i \left[\delta_i u_i-\eta_i v_i\right ] \right)\\ = & \prod_{i=1}^n \bigl\{c_{\alpha_i,\beta_i}(0)-c_{\alpha_i,\beta_i}({\varepsilon}_i u_i)-c_{\alpha_i,\beta_i}({\varepsilon}_i v_i)+c_{\alpha_i,\beta_i}({\varepsilon}_i[u_i - v_i])\bigr\}\\ = & \prod_{i=1}^n \left\langle \delta_{{\varepsilon}_i u_i}x_{\alpha_i,\beta_i}(t_i)\delta_{{\varepsilon}_i v_i}x_{\alpha_i,\beta_i}(t_i)\right\rangle.\end{aligned}\]]the result follows . from this proposition
, we see that the total increment process can capture the locally multi - self similar property of the gsgcc better .
it also shows that locally , gsgcc behaves similarly as the fractional brownian sheet .
one would tend to use proposition [ p1026_2 ] and proposition [ p1026_4 ] to conclude that the hausdorff dimension of the graph of gsgcc over a hyperrectangle is @xmath314 however , proposition [ p1026_4 ] can not be applied here since proposition [ p1026_2 ] does not imply that the fractal index of gsgcc is @xmath315 . for the fractional brownian sheet @xmath316 , kamont @xcite showed that the hausdorff dimension of its graph is bounded above by @xmath317 . in @xcite , ayache used wavelet method to show that the hausdorff dimension of the graph of the fractional brownian sheet @xmath316 is indeed equal to @xmath317 .
this fact was proved again by ayache and xiao @xcite as a special case of a more general result on hausdorff dimension of fractional brownian sheets from @xmath9 to @xmath318 , using a different method .
therefore , it is natural for us to conjecture that the hausdorff dimension of the graph of gsgcc is also @xmath317 .
in fact , implies that @xmath319 ^ 2\right\rangle = o\left(\vert \tau\vert^{\min\alpha}\right)\hspace{1cm}\text{as}\;\ ; \vert\tau\vert\rightarrow 0^+.\end{aligned}\ ] ] by a well - known theorem ( see e.g. @xcite ) , this implies that the hausdorff dimension of the graph of gsgcc is bounded above by @xmath317 . on the other hand ,
it is easy to show that [ lemma1122_1 ] let @xmath320 $ ] be a hyperrectangle in @xmath9 .
there exist constants @xmath321 and @xmath322 such that @xmath323 ^ 2\right\rangle \leq c_2\sum_{i=1}^n |t_i - s_i|^{\alpha_i},\end{aligned}\ ] ] for all @xmath324 . then by adapting the proof in @xcite for the lower bound on the hausdorff dimension of general anisotropic gaussian fields with stationary increments
, it can be verified that the graph of gsgcc over a hyperrectangle has hausdorff dimension equal to @xmath317 .
the condition for lrd can be generalized to random sheet and it becomes @xmath325 since @xmath326 and proposition [ p2 ] for @xmath98 gives @xmath327we have gsgcc is lrd if and only if for some @xmath279 , @xmath328 ; it is srd if and only if @xmath329 for all @xmath279 . in order to determine the asymptotic behavior for the spectral density of the generalized cauchy sheet , we observe that the spectral density of gsgcc , which we denote by @xmath330 can be expressed as products of spectral densities of one dimensional isotropic gfgcc considered in section [ sec3 ] .
namely , @xmath331 therefore , the high frequency and low frequency behaviors of the spectral density @xmath330 can be obtained from proposition [ p1113_1 ] and proposition [ p1113_2 ] respectively .
the high frequency limit of the spectral density @xmath330 is given by @xmath332where @xmath333and@xmath334 the low frequency limit of the spectral density @xmath330 is given by @xmath335where@xmath336 here we have used the fact that @xmath337 ( @xcite , no.1 of 8.335 ) , @xmath338 and @xmath339 ( @xcite , no.2 of 8.366 ) .
in his seminal paper @xcite , lamperti introduced a transformation which provides a one to one correspondence between a self similar process and a stationary process . for a stationary process @xmath27 , @xmath340 ,
we let @xmath341 for @xmath342 , @xmath343 , and @xmath344 be its @xmath28-lamperti transfrom .
then @xmath345 is an @xmath28-self similar ( @xmath28-ss ) process .
conversely , if @xmath346 is @xmath28-ss with zero mean , then the inverse lamperti transformation of @xmath345 defined by @xmath347 is a stationary process .
there are two ways to define extensions of lamperti transformation to @xmath9 linking stationary random field to a self - similar random field . given a stationary random field @xmath288 , the first way to define its lamperti transformation
is , given @xmath348 , defined by @xmath349for @xmath350 .
it is easy to show that [ p3 ] let @xmath351 be the @xmath28-lamperti transform of a stationary field @xmath70 defined by .
then @xmath345 is a @xmath28-ss field . for any @xmath284 , we have @xmath352,\ldots , \ln [ ct_n]\right)\\ & = _ dc^h\vert t\vert^h x(\ln t_1 + \ln c,\ldots , \ln t_n+\ln c)\\ & = _ d c^h \vert t\vert^h x(\ln t_1,\ldots , \ln t_n)\\ & = _ d c^h y(t),\end{aligned}\]]i.e
. @xmath345 is @xmath28-ss .
the inverse to this lamperti transformation transforms a random field @xmath345 to @xmath353for @xmath354 . for a field @xmath355 to be stationary , it must satisfy @xmath356 for all @xmath357 .
this is an @xmath2-parameter family of conditions .
however , in proving that @xmath345 is @xmath28-ss , we only use one parameter family of these conditions , i.e. , we only look at those @xmath358 of the form @xmath359
. therefore we can not expect that the inverse lamperti transform of a @xmath28-ss field is a stationary field . as an example , consider the l@xmath53vy fractional brownian field with index @xmath28 , @xmath360 , which is a @xmath28-ss field . define its inverse @xmath28lamperti transform @xmath361 by .
we find that its covariance is @xmath362 ^ 2+\ldots+[e^{t_n+\tau_n}-e^{t_n}]^2\right)^{h}\biggr\}.\end{aligned}\]]it is easy to verify that for @xmath122 , this expression is not independent of @xmath354 .
therefore , @xmath361 is not a stationary process . to make full use of the @xmath2-parameter families of symmetry in a stationary process
, there is another definition of lamperti transformation introduced by @xcite . given a stationary process @xmath70 , and a multi - index @xmath363 , the second @xmath28-lamperti transform of @xmath27 , denoted by @xmath364 ,
is defined as @xmath365for @xmath350 .
it was shown in @xcite that [ p4 ] let @xmath366 be the @xmath28-lamperti transform of a stationary field @xmath70 defined by .
then @xmath364 is a @xmath28mss field .
the inverse of this second lamperti transformation transform a field @xmath367 to @xmath70 , where @xmath368 it has a nice property , namely , as was shown in @xcite : let @xmath288 be the inverse second @xmath28lamperti transform of a @xmath28-mss field @xmath367 defined by .
then @xmath27 is a stationary field . as a side remark ,
it can be shown that is essentially the unique ( up to some multiplicative constants ) transformation that takes a stationary process to a @xmath28-mss . however ,
if for @xmath369 , @xmath370 is a function of the form @xmath371then if @xmath372,\ldots , c[m]\in { \mathbb{r}}_+^n$ ] , @xmath373 and @xmath374 are such that @xmath375 , it is easy to verify that the transform @xmath376,\beta_i}(t)^{\gamma_i}x(\ln t_1,\ldots , \ln t_n)\ ] ] takes a stationary process to a @xmath28ss process , but its inverse in general does not take a @xmath28ss process to a stationary process .
the first lamperti transformation corresponds to @xmath377 , @xmath372_1=\ldots = c[1]_n=1 $ ] , @xmath378 and @xmath379 .
returning to the gfgcc @xmath17 and the gsgcc @xmath248 , proposition [ p3 ] shows that their first lamperti transforms , @xmath380 and @xmath381 , are @xmath28ss fields with covariances @xmath382^{-\beta},\end{aligned}\]]and @xmath383^{-\beta_i}\end{aligned}\]]respectively .
on the other hand , proposition [ p4 ] shows that their second lamperti transforms , @xmath384 and @xmath385 , are @xmath28mss fields with covariances @xmath386^{-\beta},\end{aligned}\]]and @xmath387^{-\beta_i}\end{aligned}\]]respectively .
next we consider whether the lrd property of gfgcc and gsgcc is preserved under the lamperti transformations .
first we note that the lrd condition can be extended to non - stationary field in the following way .
let @xmath388 be the correlation function of a non - stationary gaussian field @xmath27 .
then the condition of lrd for the non - stationary gaussian field @xmath27 is given by @xmath389 + a. the @xmath28-ss gaussian field @xmath380 is lrd for all @xmath19 $ ] and @xmath20 .
+ b. the @xmath28-ss gaussian field @xmath381 is lrd for all @xmath390 satisfying @xmath251 $ ] and @xmath252 , @xmath279 .
+ c. the @xmath28-mss gaussian field @xmath384 is lrd for all @xmath19 $ ] and @xmath20 .
+ d. the @xmath28-mss gaussian field @xmath385 is lrd for all @xmath390 satisfying @xmath391 $ ] and @xmath252 , @xmath279 .
+ it is easy to verify that both the processes @xmath380 and @xmath384 have the same correlation function @xmath392\right)^2\right)^{\alpha/2}\right]^{-\beta } , \hspace{0.5cm}\tau\in { \mathbb{r}}_+^n,\nonumber\end{aligned}\]]which is independent of @xmath28 . in order to show that these processes are lrd , we have by condition , @xmath393\right)^2\right)^{\alpha/2}\right]^{-\beta}d^n\tau\\ = & \left[\prod_{i=1}^n
t_i\right]\int\limits_{{\mathbb{r}}_+^n}\left[1+\left(\sum_{i=1}^n\left(\ln \left[1+u_i\right]\right)^2\right)^{\alpha/2}\right]^{-\beta}d^nu\\ = & \left[\prod_{i=1}^n t_i\right]\int\limits_{{\mathbb{r}}_+^n}\left[1+\left(\sum_{i=1}^nv_i^2\right)^{\alpha/2}\right]^{-\beta}\left[\prod_{i=1}^n e^{v_i}\right]d^nv,\end{aligned}\ ] ] where we have used the substitutions @xmath394 and @xmath395 . clearly , when @xmath396 , the integrand also approaches @xmath397 .
therefore , the last integral diverges for all @xmath398 , which is the condition for @xmath380 and @xmath384 to be lrd . the statements for @xmath399 and @xmath385 are proved analogously .
note that the lrd property of gfgcc and gsgcc is preserved under the lamperti transformations , but the srd property is not preserved .
here we have an example that the application of lamperti transformation to a lrd stationary process ( in this case the gfgcc and gsgcc ) gives a ( multi)self similar process with lrd .
examples of lamperti transformation encountered so far relate either two short memory processes ( for example , ornstein - uhlenbeck process and brownian motion ) , or between a srd process and a lrd process ( in the case of fbm and its inverse lamperti transformed process ) . for examples
, one can show that the inverse lamperti transformation of fractional brownian sheet with lrd property gives rise to a stationary random sheet with short range dependence , and the stationary field associated with the l@xmath53vy fractional brownian field is a stationary field with srd .
recall that l@xmath53vy fractional brownian field is essentially the only self similar gaussian field with stationary increments @xcite .
similarly , one can show that fractional brownian sheet is essentially the only multi
similar gaussian random field that has stationary total increments .
hence @xmath380 , @xmath399 , @xmath384 and @xmath385 , the self - similar and multi self similar fields associated with gfgcc and gsgcc , do not have stationary increments or total increments .
there exists a weaker stationary property known as asymptotically locally stationarity , which requires the field to be stationary in the limit @xmath34 .
a centered gaussian random field @xmath27 is said to have asymptotically locally stationary increment if and only if as @xmath34 , the variance of its increment @xmath400 ^ 2\right\rangle$ ] is independent of @xmath297 .
more precisely , @xmath401where @xmath402 is independent of @xmath297 , and for any fixed @xmath297 , @xmath403 as functions of @xmath404 .
similarly , a centered gaussian random field @xmath27 is said to have asymptotically locally stationary total increment if and only if as @xmath34 , the variance of its total increment @xmath405 ^ 2\right\rangle$ ] is independent of @xmath297 . for the random field @xmath380
, we have @xmath406 ^ 2\right\rangle = \left\langle \left[y_{\alpha,\beta}(t+\tau)\right]^2\right\rangle+ \left\langle \left[y_{\alpha,\beta}(t)\right]^2\right\rangle-2\bigl\langle y_{\alpha,\beta}(t+\tau ) y_{\alpha,\beta}(t)\bigr\rangle\\ & = \vert t+\tau\vert^{2h } + \vert t\vert^{2h}-2\vert t+\tau\vert^{h}\vert t\vert^h \left[1+\left(\sum_{i=1}^n \left(\ln \left[1+\frac{\tau_i}{t_i}\right]\right)^2\right)^{\alpha/2}\right]^{-\beta}.\end{aligned}\]]using @xmath407and @xmath408\right)^2\right)^{\alpha/2}\right]^{-\beta}= & \left[1+\left(\sum_{i=1}^n \left ( \frac{\tau_i}{t_i}+o(\tau_i^2)\right)^2\right)^{\alpha/2}\right]^{-\beta}\\ = & \left[1+\left(\sum_{i=1}^n \left [ \frac{\tau_i}{t_i}\right]^2+o(\vert \tau\vert^3 ) \right)^{\alpha/2}\right]^{-\beta}\\ = & \left[1+\left(\sum_{i=1}^n \left [ \frac{\tau_i}{t_i}\right]^2\right)^{\alpha/2}+o(\vert \tau\vert^{\alpha+1 } ) \right]^{-\beta}\\ = & 1-\beta\left(\sum_{i=1}^n \left [ \frac{\tau_i}{t_i}\right]^2\right)^{\alpha/2}+o\left(\vert\tau\vert^{\min\{2\alpha , \alpha+1\}}\right),\end{aligned}\]]we find that as @xmath409 , @xmath410 ^ 2\right\rangle=&\vert t\vert^{2h}\left(1 + 2h\frac{\sum_{i=1}^n t_i\tau_i}{\vert t\vert^2}\right)+\vert t\vert^{2h } \\\nonumber & - 2\vert t\vert^{2h}\left(1+h\frac{\sum_{i=1}^n t_i\tau_i}{\vert t\vert^2}\right)\left(1-\beta\left(\sum_{i=1}^n \left [ \frac{\tau_i}{t_i}\right]^2\right)^{\alpha/2}\right)\\&+o\left(\vert\tau\vert^{\min\{2,2\alpha , \alpha+1\}}\right)\nonumber\\ \nonumber= & 2\beta \vert t\vert^{2h}\left(\sum_{i=1}^n \left [ \frac{\tau_i}{t_i}\right]^2\right)^{\alpha/2}+o\left(\vert\tau\vert^{\min\{2,2\alpha , \alpha+1\}}\right).\end{aligned}\]]it is obvious that the leading term is not independent of @xmath297 for any @xmath28 and @xmath22 , unless when @xmath98 and @xmath180 .
therefore for @xmath122 , @xmath380 does not have asymptotically locally stationary increments . similarly , for the fields @xmath384 , @xmath399 and @xmath385 , one can show similarly that as @xmath409 , @xmath411 ^ 2\right\rangle= & 2\beta t_1^{2h_1}\ldots t_n^{2h_n}\left(\sum_{i=1}^n \left [ \frac{\tau_i}{t_i}\right]^2\right)^{\alpha/2}+o\left(\vert\tau\vert^{\min\{2,2\alpha , \alpha+1\}}\right),\end{aligned}\ ] ] @xmath412 ^ 2\right\rangle= & 2\vert t\vert^{2h}\sum_{i=1}^n \beta_i \left|\frac{\tau_i}{t_i}\right|^{\alpha_i}+o\left(\vert\tau\vert^{\min\{2,2\alpha_i , \alpha_i+1\}}\right),\end{aligned}\]]@xmath413 ^ 2\right\rangle= & 2t_1^{2h_1}\ldots t_n^{2h_n}\sum_{i=1}^n
\beta_i \left|\frac{\tau_i}{t_i}\right|^{\alpha_i}+o\left(\vert\tau\vert^{\min\{2,2\alpha_i , \alpha_i+1\}}\right).\end{aligned}\]]therefore for @xmath414 , none of the fields @xmath384 , @xmath399 and @xmath385 have asymptotically locally stationary increments .
when @xmath98 , all the fields @xmath380 , @xmath384 , @xmath399 and @xmath385 are actually the same , and they have asymptotically locally stationary increment if and only if @xmath180 , in which case the variance of the increment @xmath415 behaves like @xmath416 next we consider the total increments . since increment is the same as total increment when @xmath98 , we only need to consider @xmath122 . using similar computations as given above , one can verify that for @xmath122 , the fields @xmath380 , @xmath384 and @xmath399 do not have asymptotically locally stationary increments , but @xmath385 have if @xmath180 .
we show the computation of the latter case here . by definition , @xmath417 ^ 2\right\rangle=\sum_{\delta\in\{0,1\}^n } \sum_{\eta\in\{0,1\}^n } ( -1)^{\sum_{i=1}^n \delta_i+\sum_{i=1}^n \eta_i}\\&\hspace{4cm}\left\langle \mathbb{y}^{\#}_{\alpha,\beta}\left(t+\sum_{i=1}^n\delta_i\tau_i e_i\right)\mathbb{y}^{\#}_{\alpha,\beta}\left(t+\sum_{i=1}^n\eta_i\tau_i e_i\right ) \right\rangle\\ = & \prod_{i=1}^n \sum_{(\delta_i,\eta_i)\in\{0,1\}^2}(-1)^{\delta_i+\eta_i}(t_i+\delta_i\tau_i)^{h_i } ( t_i+\eta_i\tau_i)^{h_i}c_{\alpha_i,\beta_i}\bigl(\ln[t_i+\delta_i\tau_i ] , \ln[t_i+\eta_i\tau_i]\bigr)\\ = & \prod_{i=1}^n \biggl\{t_i^{2h_i}-2(t_i+\tau_i)^{h_i}t_i^{h_i } \left(1+\left|\ln\left[1+\frac{\tau_i}{t_i}\right]\right|^{\alpha_i}\right)^{-\beta_i}+(t_i+\tau_i)^{2h_i}\biggr\}\\ = & \prod_{i=1}^n\left\{2\beta_i t_i^{2h_i}\left|\frac{\tau_i}{t_i}\right|^{\alpha_i}+o\left(|\tau_i|^{\min\{2,2\alpha_i,\alpha_i+1\}}\right)\right\}\\ = & 2^n\prod_{i=1}^n\left\{\beta_i t_i^{2h_i}\left|\frac{\tau_i}{t_i}\right|^{\alpha_i}\right\}+o\left(\vert \tau\vert^{\sum_{i=1}^n\alpha_i+\delta}\right),\end{aligned}\]]where @xmath418 . if @xmath180 , then as @xmath34 , @xmath419 ^ 2\right\rangle\sim 2^n\prod_{i=1}^n [ \beta_i|\tau_i|^{\alpha_i}].\end{aligned}\ ] ] therefore , we have verified the following : the total increments of @xmath385 are asymptotically locally stationary if @xmath180 .
next we consider the tangent fields ( definition [ def1 ] ) of the lamperti transforms of gfgcc @xmath380 and @xmath384 .
we have [ p1026_1 ] + a. the field @xmath380 is lass of order @xmath33 , with tangent field at @xmath420 being @xmath421 + b. the field @xmath384 is lass of order @xmath33 , with tangent field at @xmath420 being @xmath422 b_{\alpha/2}\left(\frac{u_1}{t_1},\ldots,\frac{u_n}{t_n}\right).\ ] ] + using the formula @xmath423 ^ 2\right\rangle + \left\langle\left[\delta_{{\varepsilon}v}y_{\alpha,\beta}(t)\right]^2\right\rangle-\left\langle\left[\delta_{{\varepsilon}(u - v)}y_{\alpha,\beta}(t+{\varepsilon}v)\right]^2\right\rangle\bigr\},\end{aligned}\]]we obtain immediately from that as @xmath424 , @xmath425 ^ 2\right)^{\alpha/2}+ \left(\sum_{i=1}^n\left[\frac{v_i}{t_i}\right]^2\right)^{\alpha/2}- \left(\sum_{i=1}^n\left[\frac{u_i - v_i}{t_i}\right]^2\right)^{\alpha/2}\right\}.\end{aligned}\]]notice that this is up to the factor @xmath426 , the covariance of the scaled l@xmath53vy brownian field @xmath427the statement for @xmath384 is proved similarly .
in fact , one can also deduce from and that the local fractal index of the fields @xmath380 and @xmath384 are both equal to @xmath33 , and the hausdorff dimension of their graphs over a hyperrectangle is @xmath428 .
for the tangent fields of the lamperti transforms of gsgcc @xmath399 and @xmath385 , we recall that given @xmath19^n$ ] , @xmath429 , and we assume wlog that @xmath430 for some @xmath431 . then as in proposition [ p1026_1 ]
, we can use and to show that [ p1026_3 ] + a. the field @xmath399 is lass of order @xmath432 , with tangent field at @xmath420 being @xmath433where the field @xmath434 is defined in proposition [ p1026_2 ] .
+ b. the field @xmath385 is lass of order @xmath435 , with tangent field at @xmath420 being @xmath436 t_{\alpha,\beta}\left(\frac{u_1}{t_1},\ldots,\frac{u_n}{t_n}\right).\ ] ] + from propositions [ p1 ] and [ p1026_1 ] and propositions [ p1026_2 ] and [ p1026_3 ] , we find that the tangent fields of gfgcc and gsgcc are related to the tangent fields of their lamperti transforms by some change of variable formulas .
namely , if @xmath27 has tangent field @xmath437 at @xmath354 , then its first lamperti transform @xmath438 has tangent field @xmath439 at @xmath440 , and its second lamperti transform @xmath441 has tangent field @xmath442 t\left(\frac{u_1}{s_1},\ldots,\frac{u_n}{s_n}\right)\ ] ] at @xmath440 .
we also notice that the order of self - similarity of the fields @xmath380 and @xmath399 , @xmath28 , is in general different from their order of local asymptotic self similarity , being @xmath33 and @xmath435 respectively .
since the stochastic process @xmath385 resembles the gsgcc @xmath248 in the sense that both their covariances can be written as products of covariance of one - dimensional processes , we can consider the limit of the total increments of @xmath385 .
it is easy to obtain as in proposition [ p1026_5 ] , using the @xmath98 case of proposition [ p1026_1 ] that @xmath443b_{\alpha/2}^{\#}(u).\end{aligned}\ ] ] here @xmath309 .
we have studied some of the basic properties of gfgcc and gsgcc , and their associated lamperti transforms .
the asymptotic properties of the spectral densities of gfgcc and gsgcc are considered .
one expects the separate characterization of fractal dimension and long range dependence for gfgcc and gsgcc will provide more flexibility in their applications to modeling various surfaces and images . in the one - dimensional case ,
gfgcc has been applied to model the havriliak - negami relaxation law @xcite .
the estimations of parameters for stationary gaussian processes have been widely studied .
some of these estimations can be adapted for gfgcc and gsgcc , and they are crucial to the applications .
further applications in spatial - temporal processes are possible if gfgcc is extended and modified to a space - time field to include non - stationarity and anisotropy @xcite .
we hope to apply results obtained in this paper to model various physical systems such as thin film surfaces in semiconductors @xcite , surface ocean waves @xcite , geological morphology @xcite , etc , in a future work .
* acknowledgements * the authors would like to thank malaysian academy of sciences , ministry of science , technology and innovation for funding this project under the scientific advancement fund allocation ( saga ) ref .
no p96c .
anderson , _ a multivariate linnik distribution _ , stat .
* 14 * , 333 336 ( 1992 ) .
george e. andrews , richard askey , and ranjan roy , _ special functions _ , encyclopedia of mathematics and its applications , vol .
71 , cambridge university press , cambridge , 1999 .
p. flandrin , p. borgnat and p.o .
amblard , _ from stationarity to self - similarity , and back : variations on the lamperti transformation _ , in _ processes with long - range correlations , theory and applications _ , ( lecture notes in physics vol .
g. rangarajan and m. ding , springer , new york , 2003 ) .
kent and a.t.a .
wood , _ estimating fractal dimension of a locally self - similar gaussian process by using increments _ , statistics research report srr 03495 , centre for mathematics and its applications , austalia national university , canberra ( 1995 ) ; j. r. statist .
b * 59 * , 679699 ( 1997 ) . s. c. lim and l.p .
teo , _ fractional oscillator process with two indices _ , preprint arxiv : 0804.3906 ( 2008 ) .
j. mateu , e. porcu , g. christakos and m. bevilacqua , _ fitting negative spatial covariances to geothermal field temperatures in nea kessani ( greece ) _ , envoronmetrics * 18 * , 759773 ( 2007 )
. n.k .
sahoo , s. thakur and r.b .
tokas , _ fractals and superstructures in gadolinia thin film morphology : influence of process variables on their characteristic parameters _ , thin solid films * 503 * , 8595 ( 2006 ) . | two types of gaussian processes , namely the gaussian field with generalized cauchy covariance ( gfgcc ) and the gaussian sheet with generalized cauchy covariance ( gsgcc ) are considered .
some of the basic properties and the asymptotic properties of the spectral densities of these random fields are studied .
the associated self - similar random fields obtained by applying the lamperti transformation to gfgcc and gsgcc are studied .
@xmath0faculty of engineering , multimedia university , jalan multimedia , cyberjaya , 63100 , selangor darul ehsan , malaysia .
@xmath1faculty of information technology , multimedia university , jalan multimedia , cyberjaya , 63100 , selangor darul ehsan , malaysia . |
-input - multiple - output ( mimo ) communications @xcite have continued to be one of the key technologies of the next generation wireless systems due to their potential of providing higher data rate and better reliability compared to the conventional single - input - single - output ( siso ) systems .
when the channel state information ( csi ) is available at both the transmitter and receiver , it is well known that the closed - loop gain can be further acquired by jointly designing the precoder and the equalizer . among these closed - loop transceiver design schemes , singular - value - decomposition
( svd)-based linear transceiver decomposes the mimo channel into multiple parallel subchannels and is known to achieve the channel capacity if proper power allocation @xcite is applied .
however , due to the variation of the signal - to - noise - ratio ( snr ) in each subchannel , the bit error rate ( ber ) performance is dominated by the subchannel with the worst snr .
consequently , without sophisticated bit - allocation schemes , fundamental trade - off between the ber and capacity can not be avoided in this type of design @xcite .
in addition to the svd - based linear design , geometric - mean - decomposition ( gmd)-based nonlinear transceiver design has also been proposed @xcite . with the help of gmd @xcite
, the mimo channel is decomposed into multiple subchannels with identical snr , and hence the simple identical bit allocation can be used for all subchannels .
it has also been shown that the gmd - based transceiver under the zero - forcing ( zf ) constraint asymptotically achieves both the optimal ber and capacity simultaneously at sufficiently high snr . due to these nice properties , various extensions and generalizations of gmd - based transceivers
have been reported in the literature @xcite .
as the gmd is the core of many advanced mimo transceiver designs , the associated implementation issues started to draw researchers attention @xcite . in @xcite ,
a scaled geometric mean decomposition algorithm was proposed in order to simplify the detection logic . in @xcite
, the authors presented a constant throughput of gmd implementation which also supports hardware sharing between precoding and signal detection modules . in this correspondence ,
a new implementation issue of the gmd algorithm is addressed .
we noticed that all the existing gmd algorithm requires the computation of the geometric mean @xmath0 of all the positive singular values of the matrix to be decomposed in the initialization step .
this requires the capability of computing the @xmath1th root @xmath2{a}$ ] of some positive real number @xmath3 , in which @xmath4 for all @xmath5 . for special cases where @xmath6 with @xmath7 being some positive integer , it is possible to decompose the computation of @xmath0 into successive geometric mean computations of two numbers , @xmath8 where the square root operation can be carried out efficiently using cordic - based computing @xcite .
for @xmath9 , while finding @xmath2{a}$ ] can generally be achieved by using newton s type of @xmath1th - root algorithm @xcite , the difficulty lies in the fact that a good initial guess is often required for the algorithm to converge .
another possible way of computing @xmath2{a}$ ] is to first transform @xmath10 into logarithmic domain and then convert it back after divided by @xmath1 .
cordic - based algorithms can be used to implement algorithmic and exponential functions , but the inherent bounded input range often limits their applications unless extra pre- and post - processing steps are applied .
an alternative method is to use lookup tables and/or a piecewise polynomial ( including linear ) algorithms to realize both the logarithmic and exponential functions .
a mass of memory and extra computations are therefore required to ensure the accuracy for such high dynamic - range computations . in this correspondence , we propose an iterative gmd ( igmd ) algorithm based on the technique of successive approximation .
the proposed algorithm has a regular structure and is applicable to matrices of any dimension .
it does not require explicit computation of the @xmath1th root , and hence eliminates the hardware difficulties mentioned above .
the convergence proof of proposed iterative gmd algorithm is provided , and numerical results show the performance of the proposed algorithm converges to that of the gmd very quickly within a few iterations . _
notations : _ throughout this paper , matrices and vectors are set in boldface , with uppercase letters for matrices and lower case letters for vectors .
the superscripts @xmath11 , @xmath12 denote the transpose and conjugate transpose of a matrix , respectively .
@xmath13 denotes the diagonal matrix with diagonal elements @xmath14 .
@xmath15_{p , q}$ ] and @xmath15_{m : n , p : q}$ ] denote the @xmath16th component and the submatrix formed by the consecutive @xmath17th to @xmath18th rows and @xmath19th to @xmath20th columns of @xmath21 , respectively .
the proposed iterative gmd algorithm can be described as follows .
* initialization : * given the matrix @xmath22 of rank @xmath23 , the algorithm starts with some general orthogonal decomposition of @xmath24 @xmath25 where @xmath26 and @xmath27 are both semi - unitary , and @xmath28 is upper - triangular . for general @xmath29 , @xmath30 , and @xmath1
, one can always choose the singular value decomposition for initialization .
for special cases where @xmath24 is full column rank with @xmath31 , other orthogonal decompositions such as the qr decomposition @xcite can also be used .
the algorithm then initializes by setting @xmath32 , @xmath33 , @xmath34 , and starts with iteration index @xmath35 .
* iteration : * in each iteration , the algorithm performs @xmath36 stages of operations as the stage index @xmath37 ranges from @xmath38 to @xmath36 . at stage @xmath37
, the algorithm first computes the singular value decomposition for the @xmath39 submatrix of @xmath40 @xmath41=\mathbf{u}^{(k)}_\gamma \boldsymbol{\sigma}^{(k)}_\gamma \mathbf{v}_\gamma^{(k)\mathrm{h } } , \label{r_tilde_22}\end{aligned}\ ] ] where the singular matrices @xmath42 and @xmath43 are both unitary , and @xmath44 is a diagonal matrix with singular values @xmath45 and @xmath46 . without loss of generality , we assume @xmath47 . after the singular values are obtained , carefully designed planar rotations are then applied to obtain an upper triangular matrix with positive diagonal elements @xmath48 and @xmath49 , where @xmath50 is a continuous mapping from @xmath51 to @xmath52 with some desired property to be discussed in details shortly . in matrix notations
, we then have @xmath53 .
\label{eq : planar_rotation_v1}\end{aligned}\ ] ] note that the planar rotations @xmath54 and @xmath55 applied in ( [ eq : planar_rotation_v1 ] ) always exist as long as @xmath56^\mathrm{t}$ ] multiplicatively majorizes @xmath57^\mathrm{t}$ ] , or equivalently when @xmath58 holds @xcite .
it is easy to verify that the matrices @xmath54 and @xmath55 can be constructed as @xmath59,\label{eq : phi_l}\\ \boldsymbol{\phi}^{(k)}_{\mathrm{r}}&=\left[\begin{array}{cc}c & -s \\ s & c\end{array}\right],\end{aligned}\ ] ] where @xmath60 combining the relations in ( [ r_tilde_22 ] ) and ( [ eq : planar_rotation_v1 ] ) , we then obtain @xmath61 , \label{eq : transformation}\end{aligned}\ ] ] where @xmath62 , and @xmath63 . since @xmath64 and @xmath65 are both products of unitary matrices , they are unitary matrices as well .
after @xmath66 and @xmath67 are obtained , @xmath68 , and @xmath69 are then constructed from the identity matrix @xmath70 with the submatrix @xmath71_{k : k+1,k : k+1}$ ] and @xmath72_{k : k+1,k : k+1}$ ] replaced by @xmath73 and @xmath74 , respectively .
the matrices @xmath40 , @xmath75 , and @xmath76 are then updated as @xmath77 it is clear that @xmath40 remains upper - triangular , and @xmath75 and @xmath76 both remain unitary after ( [ eq : update_r])-([eq : update_s ] ) are performed at the end of each stage . if the stage index @xmath37 is smaller than @xmath36 , the algorithm set @xmath78 and performs the procedure ( [ r_tilde_22])-([eq : update_s ] ) .
otherwise , the algorithm set the iteration index @xmath79 and start a new iteration unless the prescribed number of iterations is attained .
for the convenience of the subsequent discussion , we denote @xmath80}$ ] , @xmath81}$ ] , @xmath82}$ ] as the updated @xmath75 , @xmath40 , @xmath76 respectively at the end of @xmath83th stage in the @xmath84th iteration .
then the following relations hold for the proposed iterative geometric mean decomposition algorithm : @xmath85}=&\mathbf{q}^{[\ell]}\mathbf{g}^{(1)\mathrm{t}}_\mathrm{l}\ldots\mathbf{g}^{(k-1)\mathrm{t}}_\mathrm{l},\label{eq : q_recursive}\\ \mathbf{s}^{[\ell+1]}=&\mathbf{s}^{[\ell]}\mathbf{g}^{(1)}_\mathrm{r}\ldots\mathbf{g}^{(k-1)}_\mathrm{r},\\ \mathbf{r}^{[\ell+1]}=&\mathbf{g}^{(k-1)}_\mathrm{l}\cdots\mathbf{g}^{(1)}_\mathrm{l}\mathbf{r}^{[\ell]}\mathbf{g}^{(1)}_\mathrm{r}\ldots\mathbf{g}^{(k-1)}_\mathrm{r}. \label{eq : r_recursive}\end{aligned}\ ] ] for all @xmath86 . here
@xmath87}$ ] , @xmath88}$ ] , and @xmath89}$ ] are defined as @xmath90 , @xmath91 , and @xmath92 , respectively .
the planary rotation matrices @xmath93 and @xmath94 clearly also depend on the iteration index @xmath84 , but the dependency is not denoted explicitly in ( [ eq : q_recursive])-([eq : r_recursive ] ) for simplicity as long as no confusion results . in the following section ,
we show that one can design the mapping @xmath50 such that @xmath95}\right]_{k , k}=\bar{\sigma}=\left(\prod_{k=1}^k \left[\mathbf{\tilde{r}}\right]_{k , k}\right)^{1/k } , \end{aligned}\ ] ] for all @xmath96 .
the geometric mean decomposition is therefore obtained when the algorithm converges .
for the ease of following discussions , we introduce several new notations . for given @xmath97 ,
we define a subset @xmath98 : @xmath99 where @xmath100^\mathrm{t}$ ] .
we also define continuous mappings @xmath101 , given by @xmath102 \right)&= \left[\begin{array}{c } \mathbf{x}_{1:j-1 } \\ \omega\left(x_j , x_{j+1}\right)\\
\frac{x_j x_{j+1}}{\omega\left(x_j , x_{j+1}\right ) } \\ \mathbf{x}_{j+2:k } \end{array}\right ] .
\label{eq : tj}\end{aligned}\ ] ] if we denote the vector on the main diagonal of @xmath81}$ ] as @xmath103}$ ] , then the diagonal vectors of @xmath104}$ ] and @xmath81}$ ] can be related from ( [ eq : update_r ] ) and ( [ eq : r_recursive ] ) using the new notations @xmath105 } & = t^{(k-1)}\left ( \ldots\left(t^{(2)}\left(t^{(1)}\left(\mathbf{r}^{[\ell]}\right)\right)\right)\ldots\right)\label{eq : definition_t}\\ & = t\left(\mathbf{r}^{[\ell]}\right)=t^{\ell+1}\left(\mathbf{r}^{[0]}\right),\end{aligned}\ ] ] where @xmath106 , @xmath107 is the @xmath108-fold repeated composition of the mapping @xmath109 , and @xmath110}=\mathrm{diag}\left\{\mathbf{\tilde{r}}\right\}$ ] . with the aforementioned notations , we now present the main results for the @xmath50 design .
@xmath111}=\bar{\sigma}\mathbf{1}$ ] if the mapping @xmath112 satisfies the following property @xmath113 for all @xmath114 , and the equality holds if and only if @xmath115 . given @xmath110}=\left[r^{[0]}_1,\ldots , r^{[0]}_k\right]=\mathrm{diag}\{\mathbf{\tilde{r}}\}$ ]
, we let @xmath116}_k$ ] , and consider the function @xmath117 , @xmath118 . from the arithmetic - geometric inequality , @xmath119 it is clear that @xmath120 attains its absolute minimum in @xmath121 at @xmath122 . from the definition of @xmath109 and ( [ eq : condition_omega ] ) ,
it is clear that @xmath123 for all @xmath124 , and the equality holds if and only if @xmath125 .
it follows readily that @xmath109 , which is a composite mapping of @xmath126 , satisfies @xmath127 with the equality holds if and only if @xmath128 .
consequently , @xmath129}=f(t(\mathbf{r}^{[\ell-1]}))$ ] is a monotonically decreasing sequence in @xmath52 , and hence is guaranteed to converge to the greatest lower bound @xmath130 @xcite .
furthermore , since @xmath131 is continuous , we have @xmath132}\right)\right)=f\left(\lim_{\ell\rightarrow \infty}t\left(\mathbf{r}^{[\ell-1]}\right)\right)=k\bar{\sigma},\end{aligned}\ ] ] which is attained when @xmath133}\right)=\bar{\sigma}\mathbf{1}$ ] . as a result , we have @xmath134}=\bar{\sigma}\mathbf{1}$ ] , which completes the proof . there exists potentially many functions that satisfy condition ( [ eq : condition_omega ] ) .
the geometric mean @xmath135 clearly satisfies ( [ eq : condition_omega ] ) as @xmath136 always holds due to the am - gm inequality , and equality holds if and only if @xmath115 .
in addition to @xmath137 , the arithmetic mean @xmath138 is another choice that also satisfies ( [ eq : condition_omega ] ) .
this can be observed by squaring both sides of the am - gm inequality @xmath139 as a result , @xmath140 , and the equality holds if and only if @xmath115 . note that @xmath141 is simply the harmonic mean function @xmath142 while @xmath143 , it is clear that @xmath144 also satisfies ( [ eq : condition_omega ] ) from the same relation we obtained in ( [ eq : am_gm_square ] ) .
based on @xmath145 , @xmath146 , and @xmath147 , we can then construct our iterative gmd algorithms : igmd - am , igmd - gm , and igmd - hm , respectively .
since these mappings are highly nonlinear , it is very challenging to compare the convergence speed of the proposed algorithm in these three constructions .
in fact , the convergence behaviour not only depends on the topological property of the mapping but also depends on how the algorithms are initialized . from the implementation point of view , igmd - am and igmd - hm may have some advantages over the igmd - gm as only basic arithmetic operations are required in computing @xmath148 and @xmath144 rather than the square root operations required in computing @xmath137 .
in this section , we present some simulation results of the proposed igmd algorithm under three different constructions : igmd - am , igmd - gm , and igmd - hm . throughout the simulation , we assume standard @xmath149 i.i.d .
rayleigh fading channel in which every element in the channel matrix @xmath24 is modelled as zero - mean circularly symmetric complex gaussian random variable with unit variance . to highlight the applicability of the proposed algorithm in the challenging large @xmath9 case , we choose @xmath150 in the simulation . each simulation point in the figure
is averaged over @xmath151 channel realizations .
[ fig : convergence_mse_svd ] and fig .
[ fig : convergence_mse_qr ] show the mean - square - error ( mse ) convergence behaviour of the diagonal elements of @xmath40 using svd and qr as initializations , respectively . for svd initialization
, we propose an alternative interleaved - svd ( intrlv - svd ) , defined as @xmath152 where @xmath153 \right\}$ ] and @xmath154 to enable more efficient averaging in each stage . for qr initialization ,
the qr factorization with vblast ordering ( vbqr ) @xcite is also proposed to speed up convergence .
as the diagonal elements of @xmath92 in vbqr generally has less spread than those in qr , vbqr initialization provides a smaller mse when used in the initialization as shown in fig .
[ fig : convergence_mse_qr ] . due to the highly nonlinear nature of the corresponding mappings @xmath155 , @xmath156 , and @xmath157 , it is very difficult to analytically compare the convergence behaviour of the three igmd constructions .
hence we resort to numerical simulations and leave the more challenging theoretical analysis to our future work . the simulation results in fig .
[ fig : convergence_mse_svd ] and fig .
[ fig : convergence_mse_qr ] show that the igmd - hm has the fastest convergence rate , followed by the igmd - gm , and the igmd - am when qr , vbqr , and svd are used as initialization . on the other hand , when intrlv - svd are used , the igmd - am has the fastest convergence rate , followed by the igmd - gm , and finally the igmd - hm . in the second simulation setting ,
we investigate the convergence behaviour of the proposed igmd on the error rate performance of a gmd - based mimo system .
a @xmath158 gmd - based zero - forcing tomlinson - harashima precoded ( zfthp ) mimo system @xcite using @xmath159-quadrature amplitude modulation is simulated .
[ fig : convergence_ber_qr ] and fig .
[ fig : convergence_ber_vbqr ] show the error rate of the proposed igmd using qr and vbqr , respectively . by comparing the ber of zfthp - qr in fig .
[ fig : convergence_ber_qr ] and zfthp - vbqr in fig .
[ fig : convergence_ber_vbqr ] , it is clear that the vbqr provides a better initialization for the proposed igmd , and results in faster convergence . at the @xmath38st iteration
, the error rate performance of igmd - am and igmd - hm appears to be similar .
however , for iteration number greater than @xmath38 , the igmd - gm and igmd - hm both outperform the igmd - am and perform very close to the optimal gmd after @xmath160 iterations .
[ fig : convergence_ber_svd ] and fig .
[ fig : convergence_ber_intrlvsvd ] show the error rate of the proposed igmd using regular svd and interleaved svd , respectively . in medium to low snr region , the igmd - hm shows comparable or even better performance than the igmd - gm , while for sufficiently high snr , the igmd - gm performs the best , followed by the igmd - hm , and the igmd - am . on the contrary , the proposed igmd using interleaved svd as initialization shows very different convergence behaviour .
the igmd - hm with interleaved svd performs much worse compared to the igmd - am and igmd - gm . for most snr region of practical interests in this
setting , the igmd - am is comparable to the igmd - gm for iteration number greater than @xmath38 . from fig .
[ fig : convergence_ber_vbqr ] and fig [ fig : convergence_ber_intrlvsvd ] , it is observed that the proposed igmd - intrlv - svd - gm and igmd - intrlv - svd - am achieve even better error rate than the igmd - vbqr - gm and igmd - vbqr - hm after @xmath160 iterations .
an iterative geometric mean decomposition algorithm for mimo communications is proposed .
the proposed algorithm has a regular structure and can be easily adapted to accommodate problems of different dimensions . through iteratively updating the constituents matrices ,
the algorithm is able to converge to the true gmd without performing the @xmath1th root computation .
the convergence of the algorithm under certain sufficient condition is proved analytically and verified through computer simulations .
the authors would like to thank prof .
chiu - chu melissa liu for the helpful discussion .
zhang , a. kavi , and k. m. wong , `` equal - diagonal qr decomposition and its application to precoder design for successive - cancellation detection , '' _ ieee trans .
inf . theory _
51 , no . 1 ,
154172 , jan .
2005 .
f. liu , l. jiang , and c. he , `` advanced joint transceiver design for block diagonal geometric mean decomposition based multiuser mimo system , '' _ ieee transactions vehicular technology _ , vol .
692703 , feb .
2010 .
liu and p. p. vaidyanathan , `` generalized geometric mean decomposition and dfe transceiver design
part i : design and complexity , '' _ ieee trans . signal process .
_ , vol . 60 , no . 6 , pp . 31123123 , jun
. 2012 .
k. meher , j. vallas , t .- b .
juang , k. sridharan , and k. maharatna , `` @xmath161 years of cordic : algorithms , architectures , and applications , '' _ ieee trans .
circuits syst .
i , reg . papers _ , vol .
56 , pp . 18931907 , sep . 2009 .
w. wolniansky , g. j. foschini , g. d. golden , and r. a. valenzuela , `` v - blast : an architecture for realizing very high data rates over the rich - scattering wireless channel , '' in _ proc .
ursi international symposium on signals systems and electronics _ , 1998 , pp .
295300 .
g. d. golden , g. j. foschini , r. a. valenzuela , and p. w. wolniansky , `` detection algorithm and initial laboratory results using v - blast space time communications architecture , '' _ electronic letters _ , vol .
1416 , jan . | this correspondence presents an iterative geometric mean decomposition ( igmd ) algorithm for multiple - input - multiple - output ( mimo ) wireless communications .
in contrast to the existing gmd algorithms , the proposed igmd does not require the explicit computation of the geometric mean of positive singular values of the channel matrix , and hence substantially reduces the required hardware complexity .
the proposed igmd has a regular structure and can be easily adapted to solve problems with different dimensions .
we show that the proposed igmd is guaranteed to converge to the perfect gmd under certain sufficient condition .
three different constructions of the proposed algorithm are proposed and compared through computer simulations .
numerical results show that the proposed algorithm quickly attains comparable performance to that of the true gmd within only a few iterations . geometric mean decomposition ( gmd ) , mimo , qr , vblast , tomlinson - harashima precoding ( thp ) |
quantum communication ( qc ) has recently emerged as a subject of much interest , due to its applications in distributed quantum computation and quantum cryptography @xcite . in qc ,
non - orthogonal quantum signals are typically transmitted through telecom fibers .
reducing noise in telecom fibers is crucial for qc applications , because the very weak signals carried by polarization states are usually employed .
ideally , a single photon in a superposition of two pulses separated in time with a controlled phase difference ( i.e. , @xmath0 ) may be used . however ,
on - demand single - photon sources remain an important technological challenge . currently , weak coherent states are often employed as approximate single photon signals .
it is well known in quantum cryptography @xcite ) a branch of qc that weak coherent states may open up loopholes in security because of the probabilistic existence of multi - photon signals @xcite . indeed , a weak coherent state , when phase randomized via decoherence , gives a poisson distribution in photon numbers .
an eavesdropper , eve , may , in principle , measure the photon number in such a signal by a non - demolition measurement . afterward
, she can stop all single photon signals from reaching the receiver , bob . for each multi - photon signal , she can steal one photon and keep it in her quantum memory and send the rest of the signals to bob by using , for example , an ideal channel instead . since eve now has an exact copy of the quantum state transmitted to bob , this creates a significant challenge in the security of quantum cryptography @xcite . thus , attenuation losses and decoherence in qc are a major issue and methods for reducing such quantum noise are therefore important . unfortunately , the classically successful amplifiers ( such as edfa @xcite ) can not be used in qc because of the no - cloning theorem @xcite , and new methods must be explored . here
, we propose a method to reduce noise in the transmission of quantum optical signals in a telecom fiber .
our method is inspired by the theory of quantum dynamical bang - bang ( bb ) decoupling @xcite .
however , a key novelty of our work is the following : we propose to implement bb control in space , rather than time , through the insertion at regular intervals of a sequence of simple linear optical elements ( phase - shifters and/or beam - splitters ) in sections of a telecom fiber .
we also discuss the experimental feasibility of our scheme , and propose a few simple experimental tests .
we do not expect our method to improve the fidelity of classical light transmission compared to , e.g. , edfa amplifiers , since our method turns out to be quite sensitive to reflection from optical elements and deviations from average fiber homogeneity , which is not the case for classical amplifiers .
an optical fiber provides boundary conditions that guide light along a ( locally ) straight trajectory .
an ideal fiber allows modes of traveling photons to propagate through unchanged .
a real fiber induces noise ( dispersion , loss , decoherence ) compared to the ideal case . the method we introduce in this paper
is designed to cancel ( in principle ) all quantum noise .
the dominant classical - light loss mechanisms in an optical fiber are uv absorption , rayleigh backscattering , oh absorption , and raman scattering , and lead to typical attenuation rates , for state - of - the - art commercial silica telecom fiber , of about 0.25db / km @xcite .
these mechanisms are active also in the quantum regime @xcite .
all noise processes affecting quantum light in optical fibers are derivable from a microscopic hamiltonian describing ( i ) the direct interaction between photons and the optical ( dielectric ) material of a fiber , and ( ii ) the indirect interaction between photons and quasi - particle excitations of the fiber material , such as polaritons and photon - phonon coupling .
these indirect interactions are , of course , in turn derivable from a microscopic hamiltonian that takes into account matter - matter interactions in the fiber , and couples them to photons .
the derivation of the resulting effective interactions ( e.g. , a non - linear schrdinger equation ) from such microscopic hamiltonians has been covered in detail , e.g. , in @xcite .
the starting point of our analysis is the observation that all interactions involving photons can be written in terms of polynomials in the bosonic raising and lowering operators @xmath1 ( where @xmath2 is the mode of the traveling photons @xcite ) .
a polynomial of order @xmath3 describes an @xmath3-photon process , and typically the cross - section of interactions decreases with increasing @xmath3 . in the case of the non - relativistic quantum electrodynamics of charged particles
one can decompose the photon - matter interaction hamiltonian into linear and quadratic terms with respect to the photon field , @xmath4 , where the linear part is @xmath5where the bath operators @xmath6 depend only on the variables of charged particles and/or quasi - particles , and the quadratic part @xmath7 is a function of the bilinear operators , @xmath8 , @xmath9 and @xmath10 .
higher - order interactions may originate from relativistic effects .
in general @xmath7 , which makes no contribution to one - photon processes , is much smaller than @xmath11 @xcite .
therefore , the quadratic term can usually be neglected .
let us substantiate these arguments by briefly reviewing the corresponding non - relativistic electrodynamics .
consider particles @xmath12 with charge @xmath13 and mass @xmath14 , which constitute the optical material of a fiber .
let @xmath15 and @xmath16 be the position and momentum of particle @xmath12 , and @xmath17 be the vector potential of the photon field .
the system - bath hamiltonian that describes the dynamics can be written , in the coulomb gauge , as @xmath18here @xmath19 depends only on the variables of the charged particles .
@xmath20 is the free photon hamiltonian , where @xmath21 are the photon annihilation and creation operators in the normal vibrational mode @xmath2 of the field identified by the wave vector @xmath22 , the polarization @xmath23 and the frequency @xmath24 , where @xmath25 is the speed of light in vacuum .
then the linear part with respect to the photon field @xcite is @xmath26where for a cubic box with dimension @xmath27 the operator @xmath6 can be expressed as @xmath28which only depends on the variables of charged particles . here
@xmath29 is the @xmath30 factor , @xmath31 is the permittivity of free space , and @xmath32 is the spin of particle @xmath12 .
note that the interaction is linear in the operators @xmath33 and @xmath34 .
the quadratic part of the interaction hamiltonian is found to be @xmath35and is a function of the bilinear operators , @xmath36 , @xmath37 and @xmath10 . under the long - wavelength approximation , where the spatial variations of the electromagnetic field over the size of the particles is negligible , first order perturbation theory of @xmath11 results in the widely applied dipole interaction ( e.g. , @xcite and references therein ) .
some effective interactions , such as atom - mediated photon - photon interactions and nonlinear photon - photon interactions ( kerr effect ) have been derived without consideration of @xmath7 @xcite .
we provide more details on these effective interaction in section [ leading ] . for simplicity of presentation we will first design an anti-_linear_-decoherence fiber by considering @xmath38 only . later on we show how to treat higher order interaction terms .
it is important to stress that in essence our method hardly depends on the details of the interaction , but _
depends on the statistics of photons as bosons_. for this reason our method is very general and is in principle applicable to the entire phenomenology of quantum noise processes affecting photons in fibers , though its practical applicability is a matter of being able to satisfy certain constraints that will be discussed in detail below .
we first consider quantum data transmission through a telecom fiber with noise induced by @xmath11 . since @xmath11 describes absorption and creation of photons , it generates photon loss , among other processes . to simplify
, we suppose that a _ polarization _ photon is transmitted from end @xmath39 to end @xmath40 .
one can define a logical qubit supported by @xmath41 and @xmath42 where the mode indices refer to the two polarization states .
the initial state at end @xmath39 is @xmath43 , where @xmath44 is the state of the bath ( dielectric material and quasi - particle excitations in the fiber ) . at the time @xmath45 ( where @xmath46 is the distance between @xmath39 and @xmath40 , and @xmath47 is the average speed of light in the telecom fiber ) the wave function is @xmath48 @xmath49 , where the evolution operator is ( in units where @xmath50 ) @xmath51 , where @xmath52dx$ ] is the average hamiltonian over the @xmath53th segment , where @xmath54 is a sum of the matter ( and/or excitations ) and photon self - hamiltonians , @xmath55 , and we have assumed that @xmath56 is large in order to expand the normal - ordered exact propagator @xmath57dx]:$ ] .
i.e. , we have neglected deviations from average fiber homogeneity , @xmath58)^{2}\rangle $ ] [ @xmath59 can easily be expressed including such second and higher order moments using a magnus expansion , and it is known how to generalize bb decoupling to treat such higher moments , at the expense of more bb pulses @xcite ] .
the interaction @xmath60 entangles the output wave function at end @xmath40 with the material / excitations in the fiber . by standard arguments it follows that , therefore , the quantum information encoded into the photon state will decohere @xcite . in order to solve this problem of decoherence , we draw inspiration from the idea of bb decoupling via time - dependent pulses @xcite ( we note that a method for finding such pulses directly from empirical data was proposed in @xcite ) .
we first recall the action of a phase - shifter .
it is simple to show [ using the baker - campbell - hausdorff ( bch ) formula @xcite ] for a boson that @xmath61where @xmath62 is a boson number operator .
physically , the operation @xmath63 is a @xmath64 phase - shifter ( it puts a phase of @xmath65 between the number states @xmath66 and @xmath67 , not to be confused with our logical qubit states ) . defining the @xmath64-phase - shifter operator @xmath68we therefore have @xmath69because the photons term of @xmath54 is @xmath70 , so that @xmath71=0 $ ] .
the crucial point is that _ the sign of the linear term of the interaction hamiltonian has been negated by the action of two phase - shifters , i.e. , effectively time - reversed_. now , if we install thin phase - shifters inside the fiber at positions @xmath72 , from @xmath39 to @xmath40 , the evolution will be modified to @xmath73,\end{aligned}\]]where in the second line we have introduced a self - explanatory notation that will be used repeatedly below .
note that in writing this expression we have neglected the variation of @xmath74 inside the phase - shifter ; this will hold provided that the phase - shifter width is much smaller than the distance over which deviations @xmath75 from average fiber homogeneity become significant .
further note that we are applying the parity - kick version of bb decoupling viola:98,vitali:99 , but are implementing it in space , rather than time .
now assume that the _ average _ hamiltonians over two successive segments are equal : @xmath76the better this approximation , the better our method will perform ; we address deviations in appendix [ app ] .
in this case , to first order in @xmath77 , and using eq .
( [ eq : flip ] ) , we have exact cancellation of @xmath38 between successive segments : @xmath78this yields the overall evolution operator @xmath79i.e . , the evolution is completely decoherence - free , in analogy to the ideal bb limit of infinitely fast and strong pulses @xcite .
because of the in - principle equivalence between the bb method and the quantum zeno effect @xcite , the proposed method can only work if the phase - shifters are inserted at small intervals @xmath80 over which coherence loss is quadratic ( zeno - like ) , rather than exponential ( markovian ) .
a reliable estimate of @xmath80 requires a first principles calculation which is beyond the scope of the present work ; we present a phenomenological model for a detailed estimate of @xmath80 in appendix [ app ] . here
we give a rough _ upper bound _ estimate of this distance .
we assume that the linear term of the interaction hamiltonian gives rise to the @xmath81db / km ( @xmath82 ) _ classical _ loss figure in a telecom fiber .
our main approximation now consists in further assuming that the insertion of phase - shifters into the fiber causes a reduction of loss from first to second order , and we use this to estimate the @xmath80 required in the _ quantum _ case . thus , imagine a distributed quantum computing scenario where small - scale quantum computers are connected by optical fibers of length about 1 km .
our goal is to have reliable quantum computation within the fault - tolerance threshold value of @xmath83 error rate for each elementary quantum logical operation .
[ we remark that for reliable quantum _ communication of entangled photon pairs _ , the current error rate of about @xmath82 is already acceptable provided one allows the application of entanglement purification @xcite ; our scheme is significantly simpler . ] therefore , we need to cut down the loss figure from @xmath84 to say @xmath83 .
suppose we need to insert @xmath3 phase - shifter within 1 km of a telecom fiber .
denote the attenuation between a pair of phase - shifter by @xmath85 .
then , without the @xmath3 phase - shifters , we have @xmath86 . for a sufficiently large @xmath3
, we can expand the expression binomially and obtain the approximation @xmath87 .
now , with the insertion of phase - shifters , we simply assume that the attenuation between two phase - shifters is due to a second order contribution of the form @xmath88 .
we further assume that those contributions sum up in usual addition .
therefore , we have @xmath89 .
this yields @xmath90 and @xmath91 .
recalling that two phase - shifters are needed per cancellation step , we see that about @xmath92 phase - shifters have to be inserted in a distance of 1 km which translates to one phase - shifter every @xmath93 m .
this figure is merely a rough upper bound estimate on the distance @xmath80 between two phase - shifters for our scheme to be useful ; one can also determine @xmath80 via the experiment we propose below . also note that we have assumed here that the fiber is straight as is typically done in theoretical models . in order to regain the straight fiber approximation , in the case of a curved fiber @xmath80
is upper - bounded by the local radius of curvature .
while in spirit our method is similar to bb decoupling @xcite , a major advantage here is that we do not need to apply any time - dependent pulses , which may result in significant uncertainties such as gate errors and off - resonance transitions . instead
, the phase - shifters may be incorporated into the fiber directly during the manufacturing process .
alternatively , time - independent ( say , electronic or pressure ) controls may be applied at various points of a telecom fiber to achieve the action of pulse shifters .
we now consider higher order processes .
although they are generally weak , the bilinear interactions appearing in @xmath7 may still cause decoherence . a direct harmful consequence is to change the polarization direction , through a term such as @xmath94 . in the classical case , the fiber structure
can be designed so that a _ _ known _ _ polarization direction can be preserved @xcite . in the quantum case
the polarization direction is _ not known _ prior to the transmission and the classical method is not applicable . in this case
one must in general consider a system - bath hamiltonian that is a linear combination of all @xmath95 possible independent bilinear terms : @xmath96 , @xmath97 , @xmath98 ( the grouping will be clarified momentarily ) .
it can be shown that all @xmath95 of these terms can be eliminated by installing @xmath99 linear optical devices that include beam - splitters in addition to phase - shifters , i.e. , in @xmath100 elementary steps ( we combine beam - splitting and phase - shifting into one step ) .
this result is based on eq .
( [ eq : n - b ] ) and the following identities [ that follow directly from eq .
( [ eq : n - b ] ) ] @xmath101the role of the beam - splitter is to eliminate the set of operators @xmath102 ; the beam - splitter is inserted after the first eight steps .
the @xmath100-step result can be considerably simplified in the realistic situation wherein the two polarizations used to represent our qubit are degenerate . in this case
@xmath102 becomes @xmath103 , which generates an _ overall _ phase and hence will not cause decoherence . in this degenerate case , as we now show , we need only phase - shifters to eliminate all contributions to decoherence .
let @xmath104i.e . , a pair of phase - shifters .
it follows immediately from eqs .
( eq : n - b),([eq : n - b2 ] ) that@xmath105while @xmath106 [ where @xmath107 was used above ] . from these and the results for the anti - linear - decoherence fiber ,
the sequence @xmath108 $ ] does not contain any linear terms , but still contains all bilinear terms .
then , the sequence @xmath109
\notag \\ & = & [ 4,\pi , 3,\pi \gamma ^{\dagger } , 2,\pi , 1,\pi \gamma ] \end{aligned}\]]has , in four elementary phase - shifter steps , eliminated @xmath11 as well as @xmath110 , and in particular the polarization - direction - changing terms @xmath94 and @xmath111 : at this point we have a _ polarization - preserving fiber_. note that the composite terms can be combined into a single phase - shifter , i.e. , @xmath112 the only remaining bilinear terms at this point are the counter - rotating terms @xmath113 , that are typically neglected in the rotating wave approximation @xcite . to eliminate them nevertheless , we note that @xmath114therefore the sequence that eliminates _ all _ linear and bilinear terms for degenerate qubit states is @xmath115= [ 8,\pi , 7,\pi \gamma ^{\dagger } , 6,\pi , 5,\pi \gamma \pi _ { 1},4,\pi , 3,\pi \gamma ^{\dagger } , 2,\pi , 1,\pi \gamma \pi _
{ 1}],\end{aligned}\]]which involves @xmath116 elementary phase - shifter steps ( note that @xmath117 ) . at this point
we have a fiber that is completely free of both linear and bilinear decoherence - causing terms for degenerate polarization qubits .
we can repeat the mixed - classical - quantum rough distance estimate above , by simply assuming that now contributions to decoherence come only due to third order in @xmath85 : @xmath118 .
this leads to @xmath119 , and recalling that @xmath116 phase - shifters are needed per cancellation step , we arrive at an upper - bound estimate of about @xmath95 phase - shifters per km , or one phase - shifter every 100 m .
these phase - shifters must be introduced in addition to the ones used above for cancellation of first order effects .
we have again assumed here that the fiber is straight ; local curvature may impose a lower upper bound .
so far we have considered linear and bilinear photon terms in the interaction hamiltonian . the most general two - mode photon - related term in a hamiltonian is @xmath120 .
provided @xmath121 and @xmath122 the identity@xmath123shows that such a term can be eliminated using only phase shifters .
for example , when @xmath124 is an odd number , our considerations in the linear case show that the term can be eliminated using the phase shifter @xmath125 , while @xmath126 can be eliminated using @xmath127 .
high - order terms with @xmath128 arise if one considers the relativistic contribution , and they appear also in most of the effective photon scattering theories .
it should be clear that if such terms arise , they can be reduced using additional phase - shifters , or beam - splitters in the case @xmath129 and/or @xmath130 , which arise due to terms containing photon number operators .
as mentioned in section [ noise ] the leading loss mechanisms in optical fibers are well characterized : uv absorption , rayleigh backscattering , oh absorption , and infrared absorption .
it is useful to quickly review how these processes arise , and then are treated by our method .
consider , for example , the case of rayleigh backscattering .
we base our discussion on the standard reference @xcite ( for a general description of absorption see p.168 ; the cross section of raleigh scattering is given on pp .
371 - 373 ) .
the discussion starts @xcite[eq .
( 4.9.9 ) ] from the _ dipole approximation _ to our general photon - matter interaction hamiltonian , eq . ( [ eq : hl ] ) : @xmath131 where @xmath132 is the interacting charged particle state , or the eigenstate of @xmath19 , @xmath133 is the atom position , @xmath134 is the volume , @xmath135 are the matrix elements of atomic dipole moment , and @xmath136 is the polarization .
a general scattering transition rate @xmath137 is @xcite[eq .
( 7.7.2 ) ] : @xmath138where @xmath139 and @xmath140 are the atomic ground state and final state . initially , there are @xmath141 photons with with frequency @xmath142 and wave vector @xmath143 . at the end
there are @xmath144 incident photons and a single scattered photon with frequency @xmath145 and wave vector @xmath146 .
then the cross - section follows from the relation @xmath147 , and the differential light - scattering cross - section is @xmath148 .
the differential cross - section of rayleigh scattering is the special case when the atom returns to its ground state , which is @xcite[eq .
( 8.8.1 ) ] : @xmath149 where the parameters are obtained from matrix elements of @xmath150 the important equation is ( [ eq : tau ] ) above : it shows that rayleigh scattering originates from the interaction @xmath151 .
clearly , the differential cross - section of rayleigh scattering vanishes when @xmath152 is zero . _
our spatial bb method does just that : it effectively eliminates the interaction _ @xmath151 .
of course , this is not unique to rayleigh scattering , which is just one of the processes derived from considering various cases involving @xmath151 .
for example , photon absorption and emission is mainly related to transitions involving two atomic or molecular levels .
the corresponding matrix element for absorption is @xcite[eq .
( 4.10.1 ) ] : @xmath153n_{\mathbf{k}\lambda } ^{1/2 } , \label{eq : absor}\end{aligned}\ ] ] where @xmath154 . the radiative lifetime is @xmath155and , of course , it follows from eq .
( [ eq : absor ] ) that this absorption is prevented when @xmath151 is zero .
note how @xmath151 , which is effectively eliminated by our method , involves the bosonic raising and lowering operators @xmath156 .
the reason that our method is so general is that it acts directly on these operators , and time - reverses @xmath151 by flipping their sign .
as mentioned above , a crucial requirement for the success of our proposed method is to insert the optical elements at intervals over which the coherence - loss is still quadratic , rather than exponential .
an experiment to test for this regime is thus useful .
this could be done by monitoring the coherence ( in particular , loss ) locally , by focusing onto the edge of the fiber and collecting light into a photon - counting device ( since the absolute intensity would be very small ) . by moving the focus along the fiber
, one should be able to track the decay as a function of distance from the fiber entry point , and observe the required quadratic - to - exponential transition , yielding an estimate of @xmath80 . to actually test the method in the presence of phase - shifters
, one could repeat the above experiment with a single fiber and write some phase - shift segments into it ( as in the manufacturing of fiber bragg gratings ) , at intervals bounded above by those determined from the first experiment .
we note that a point of some potential concern is the impedance mismatch between air and the phase - shifter material , that will lead to reflection .
let @xmath157 ( @xmath158 ) denote the indices of refraction : the reflected amplitude is @xmath159 , which leads , at normal incidence , to 4% loss per air - glass interface .
however , a standard anti - reflection coating can solve the problem : a quarter - wave layer of material at @xmath160 between the two materials ( two equal reflections out of phase cancel out ) . in fibers
the index changes will be smaller and reflection is typically neglected .
moreover , by writing a smooth phase profile as in the experiment proposed above , the reflection problem essentially disappears .
once @xmath80 has been estimated , one can proceed to directly test our method , as follows .
take two fiber segments and write a @xmath64 phase - shifter ( ps ) into each .
attach them co - linearly ( i ) in the order ps - fiber - ps - fiber , ( ii ) in the order fiber - ps - ps - fiber , and perform a photon counting measurement .
our method should reduce attenuation in ( i ) by comparison to ( ii ) .
we have proposed a method to reduce quantum noise in optical fibers , via the insertion of phase - shifters at appropriately spaced intervals .
we have shown that , in principle , this method can eliminate all quantum noise processes that do not involve photon number operators in the system - bath hamiltonian ; when such terms do arise , the phase shifters need to be supplemented with beam - splitters , and our conclusions remain .
thus , with simple linear - optical devices , quantum noise in optical fibers can be drastically reduced .
this conclusion has potentially important implications for quantum communication ( and its variants , quantum cryptography and distributed quantum computing ) via optical fibers .
the practical feasibility of our method hinges on the required distance between phase - shifters .
we have given a rough upper - bound estimate of several meters based on known attenuation rates .
we have also presented a more detailed calculation that predicts a range of distances , depending on the bath spectral density appropriate for a fiber .
ultimately we believe that the best way to test our proposal is to perform the relatively straightforward experiment that it implies .
support from nserc , the darpa - quist program ( managed by afosr under agreement no .
f49620 - 01 - 1 - 0468 ) , and the sloan foundation is gratefully acknowledged ( to d.a.l . ) .
we thank prof .
lo , prof .
a.m. steinberg , prof . t. sargent , and dr .
sun for very helpful discussions . 10 , _ the physics of quantum information _ ( springer , berlin , 2000 ) . ,
physics today , nov .
2000 , p.22 .
phys . * 74 * , 145 ( 2002 ) . ,
a * 59 * , 3301 ( 1999 ) . , _ nonlinear optics _ ( westview press , 2003 ) . ,
physics today * 44 * , 36 ( 1991 ) .
l. viola , s. lloyd , phys .
a * 58 * , 2733 ( 1998 ) . for a geometric interpretation see m.s .
byrd , d.a .
lidar , q. inf . process .
* 1 * , 19 ( 2002 ) . ,
rev . a * 59 * , 91 ( 1999 ) . , eprint quant - ph/991209 .
a * 62 * , 043819 ( 2000 ) .
, _ atom - photon interactions _ ( john wiley & sons , new york , 1992 ) .
lett . * 82 * , 2417 ( 1999 ) . ,
rev . a * 67 * , 012324 ( 2003 ) . , j. math .
* 41 * , 2434 ( 2000 ) .
d. vitali , p. tombesi , phys .
a * 59 * , 4178 ( 1999 ) . ,
a * 69 * , 032314 ( 2004 ) .
a * 59 * , 169 ( 1999 ) . , _ fiber - optic communication systems _
( wiley - interscience , new york , 2002 ) . , _ the quantum theory of light _ ( oxford university press , oxford , 2000 ) . , chem .
phys . * 268 * , 35 ( 2001 ) . , rev .
phys . * 59 * , 1 ( 1987 ) . , proc .
london ser .
a * 452 * , 567 ( 1996 ) . ,
phys . rev .
a * 65 * , 012305 ( 2002 ) .
b * 34 * , 5684 ( 1986 ) .
recall that our main approximation was the assumption of average fiber homogeneity , eq .
( [ eq : approxh ] ) . in this appendix
we relax this assumption in order to estimate an upper bound on the distance @xmath80 between phase shifters .
we do this by considering corrections to order @xmath161 and the non - ideal case @xmath162where @xmath163 and we take @xmath164 to be independent , identically distributed ( iid ) gaussian , local , and time - dependent operator - valued corrections .
this phenomenological model of fiber inhomogeneity may be the result of material non - uniformity along the fiber ( such as local defects ) , slow time - dependent fluctuations in fiber properties , or even the quadratic interaction ( [ eq : q ] ) . by virtue of the central limit theorem
it will be accurate in the case of a _
large number _ of defects .
we assume that the effective bb time - interval @xmath137 is chosen to be on the order of the small parameter @xmath165 ( though we make no attempt to estimate @xmath165 ) . in this case , using the bch formula @xmath166/2+ ... }$ ] to second order ( i.e. , keeping only terms of order @xmath167 ) , we find instead of the ideal eq .
( [ eq : cancel ] ) : @xmath168\tau } e^{-i[h_{0}(k\delta ) -h_{i}^{l}(k\delta ) ] \tau } \notag \\ & \approx & \exp \{-i\tau \lbrack 2h_{0}(k\delta ) + \varepsilon ( p_{k}+q_{k})]-\tau ^{2}[h_{i}^{l}(k\delta ) , h_{0}(k\delta ) ] \},\end{aligned}\ ] ] where in the second line the effect of the phase shifters was to flip the sign ( and thus cancel ) the @xmath169 term .
to the same order of accuracy the overall evolution operator becomes @xmath170\}\exp \{-i\varepsilon \tau \sum_{k=1}^{n/2}(p_{2k}+q_{2k})\}. \label{eq : u'}\]]let us evaluate the first exponential .
using eqs .
( [ eq : h0]),([eq2 ] ) : @xmath171 & = & -i\sum_{j , j^{\prime } } [ ( \hat{b}% _ { j}^{\dagger } ( 2k\delta ) b_{j}+\hat{b}_{j}(2k\delta ) b_{j}^{\dagger } ) , \hbar \omega _ { j^{\prime } } ( 2k\delta ) ( \hat{n}_{j^{\prime } } + 1/2)+h_{m}(2k\delta ) ] \notag \\ & = & -i\sum_{j=0,1}\{\hbar \omega _ { j}(2k\delta ) \hat{b}_{j}(2k\delta ) + [ \hat{b% } _ { j}(2k\delta ) , h_{m}(2k\delta ) ] \}b_{j}^{\dagger } \notag \\ & & -\{\hbar \omega _ { j}(2k\delta ) \hat{b}_{j}^{\dagger } ( 2k\delta ) -[\hat{b}_{j}^{\dagger } ( 2k\delta ) , h_{m}(2k\delta ) ] \}b_{j } \notag \\ & \equiv & h^{\prime } , \end{aligned}\]]where @xmath172 is an effective hamiltonian ( it is hermitian ) , which plays the role of a lamb shift @xcite . we thus have for the first exponential in eq .
( [ eq : u ] ) : @xmath173\}=\exp ( -i\tau ^{2}h^{\prime } ) , \]]whose effect is an energy renormalization ( i.e. , a phase shift ) , and does not contribute to decoherence .
next , consider the second exponential in eq .
( [ eq : u ] ) .
the operator @xmath174 defined through @xmath175dt\equiv \int_{0}^{t}g(t)dt$ ] is gaussian distributed by our assumption that @xmath176 are gaussian , iid random variables .
we would like to estimate the average deviation in @xmath177 that results from its presence .
since @xmath178 is gaussian distributed the average can be computed as follows @xcite : @xmath179\rangle \notag \\ & = & \exp [ -i\varepsilon \tau \int_{0}^{t}\int_{0}^{t}\langle g(t)g(t^{\prime } ) \rangle dtdt^{\prime } ] \notag \\ & \equiv & \exp [ -\varepsilon \tau \gamma ( t)].\end{aligned}\ ] ] expressed in terms of fourier components @xmath180 of @xmath178 we have for the decoherence factor : @xmath181where @xmath182but in the gaussian case we have ( as in the spin - boson model leggett:87)@xmath183where @xmath184 is the spectral density ( of matter in the fiber ) and @xmath185 is the inverse temperature .
hence our result is that the correction is the attenuation is thus strongly dependent upon the form of @xmath184 , but also depends sensitively on temperature . in particular ,
the thermal time - scale @xmath187 is important in separating thermal effects from effects due purely to vacuum fluctuations @xcite . in order to formally separate the two
it is convenient to write @xmath188where @xmath189is the average number of field excitations at inverse temperature @xmath190 . in the limit of very low temperatures ( @xmath191 )
we have @xmath192and we can analytically evaluate the integral in eq .
( [ eq : sbint ] ) , e.g. , for the class of ohmic - type spectral densities , i.e. , for the case @xmath193where @xmath12 is the coupling strength and @xmath194 is the high - frequency cutoff ( note that @xmath12 is not dimensionless ) .
the result in the zero - temperature case is @xmath195\right ) , & \quad n\neq 1 \end{array } \right . .\end{aligned}\ ] ] to obtain the non - zero temperature correction in the approximation ( eq : n - t - finite ) take these results , multiply by @xmath196 , replace @xmath194 by @xmath197 everywhere , and add to the zero temperature case .
we tabulate a few cases of interest in the zero temperature limit , letting @xmath198 : @xmath199=\left\ { \begin{array}{lr } ( 1+x^{2})^{-\alpha \varepsilon \tau /4 } , & \quad n=1\text { ( ohmic ) } \\ \exp [ -\frac{1}{2}\alpha \varepsilon \tau \omega _ { c}\frac{x^{2}}{1+x^{2 } } ] , & \quad n=2\text { ( super - ohmic ) } \\ \exp [ -\frac{1}{2}\alpha \varepsilon \tau \omega _ { c}^{2}\frac{x^{2}(3+x^{2 } ) } { ( 1+x^{2})^{2 } } ] , & \quad n=3\text { ( debye ) } \end{array}% \right .
.\]]let @xmath200 be the desired coherence value after time @xmath201 ( or distance @xmath46 ) ; then we need to solve for the phase shifter spacing @xmath80 from @xmath199>1-\delta ( t).\]]we find ( assuming @xmath202 ) : @xmath203/\ln [ ( 1+x^{2 } ) ] , & \quad n=1 \\ \delta
^{2}<-\frac{2v^{2}}{\alpha \omega _ { c}}\frac{1+x^{2}}{x^{2}}\ln [ 1-\delta ( t ) ] , & \quad n=2 \\
\delta ^{2}<-\frac{2v^{2}}{\alpha \omega _ { c}^{2}}\frac{(1+x^{2})^{2}}{% x^{2}(3+x^{2})}\ln [ 1-\delta ( t ) ] , & \quad n=3% \end{array } .
\label{eq : delta}\ ] ] the present model is , unfortunately , too phenomenological to make a reliable estimate of @xmath80 .
nevertheless , it is of some interest to see its prediction .
e.g. , we could wish to improve upon the current figure of merit of @xmath81 db / km to the threshold value of @xmath204 . recall that @xmath205 , @xmath206 and we assumed @xmath207 .
the coupling strength @xmath12 is typically of order unity viola:98,vitali:01 ; we shall set @xmath208 .
we take @xmath209 , the speed of light in a typical fiber , and @xmath204 .
the results in the three cases , with @xmath210 , are displayed in fig .
[ fig ] , as a function of the high - frequency cutoff @xmath194 . as a rough reference ,
the debye temperature of amorphous silica is @xmath211 @xcite , yielding a debye frequency estimate of @xmath212 .
the corresponding value of @xmath80 is @xmath213 ( @xmath214 ) , @xmath215 ( @xmath216 ) , @xmath217 ( @xmath218 ) .
this strong sensitivity to the decoherence model underscores the need for the proposed experiment in order to settle the question of the actual required distance between phase shifters . between phase shifters ( in meters ) , as a function of high - frequency cutoff @xmath219 ( in hz ) .
note the double logarithmic scale .
upper , middle , bottom curves correspond to @xmath220 respectively in eq . _ _ ( [ eq : delta]).,height=302 ] | noise in optical telecom fibers is an important limitation on optical quantum data transmission .
unfortunately , the classically successful amplifiers ( such as edfa ) can not be used in quantum communication because of the no - cloning theorem .
we propose a simple method to reduce quantum noise : the insertion of phase - shifters and/or beam - splitters at regular distance intervals into a fiber .
we analyze in detail the case of qubits encoded into polarization states of low - intensity light , which is of central importance to various quantum information tasks , such as quantum cryptography and communication .
we discuss the experimental feasibility of our scheme and propose a simple experiment to test our method . |
detection of the mott insulating transition is a desirable function for the first - principles calculations , which has been demanded for years .
a recently developed multi - reference density functional theory with the fluctuation reference method defines a self - consistent first - principles calculation , in which a short - range correlation effect is explicitly included.@xcite this technique is a generalization of the kohn - sham scheme of the electronic structure calculation.@xcite incorporation of the effective many - body system to determine the total energy and the single - particle charge density of the electronic state became possible through 1 ) introduction of a fluctuation - counting term and 2 ) the reformulation of the exchange correlation energy functional as a residual exchange correlation energy functional .
the effective hamiltonian appearing in the theory is a kind of the hubbard model@xcite or the anderson model@xcite .
the lda+u hamiltonian@xcite is derived as an approximation for the residual exchange correlation functional.@xcite in this short paper , i show a technique to determine the mott insulating phase in this first - principles calculation .
the method is an application of the momentum boost technique known in the literature.@xcite to introduce the momentum boost method in the density functional theory , one needs to show the n representability .
this is done by extending the harriman construction@xcite for the twisted boundary condition .
a test calculation will be shown using a simple artificial system , which may be represented by a one - dimensional hubbard model . in the last part
, i will summarize the present work .
i consider a born - von - karman boundary condition with a twist .
consider an array of atoms in one dimension and let @xmath0 the number of atoms in the direction , which is called the @xmath1 direction below . formally using a phase @xmath2 with @xmath3 and a unit flux @xmath4
, i introduce a twisted boundary condition in this @xmath1 direction .
@xmath5 here , i used a combined coordinate @xmath6 with the space and spin coordinates .
@xmath7 is a unit vector and @xmath8 is the lattice constant in the @xmath1 direction .
@xmath9 is a many - body wavefunction and @xmath10 is a creation operator defined with a properly determined wannier basis @xmath11 .
indeed , determining a wannier transformation fixing a gauge in the unitary transformation,@xcite the set of @xmath11 is uniquely determined in each self - consistent step of the extended kohn - sham scheme.@xcite using another gauge transformation , the twisted boundary condition is represented by shifted @xmath12 vectors , @xmath13 , for the single - particle bloch orbitals . here
@xmath14 is an integer within a range of @xmath15 . from the single - particle energy @xmath16
, one may construct a tight - binding model written in @xmath10 and @xmath17 .
determining the eigen energy of the effective fermion system , one has an energy - flow diagram as a function of @xmath18 .
if the system is a metal , the flow given by adiabatic connection of the ground state should show a long extended ab period , while the period has to be @xmath4 , when there is a charge gap due to formation of a mott gap .
although the density - functional theory utilized in the present work is the ground - state formulation , one can have a signal of the change in the period as disappearance of a cusp in the lowest - energy flow . in case of the mott insulator ,
the flux line becomes a smooth curve without any cusp .
the appearance of the isolated lowest flow lie implies uniqueness of the ground state separated by a gap from charge excitations in the whole range of twist and thus indicates the formation of the mott gap .
if one has a difficulty in finding the change of the period with a three - dimensional sample , one may consider a one - dimensional setup using a needle like sample . before discussing the result ,
i show that the @xmath19 representability of the present problem is guaranteed . in the present setup ,
only the boundary condition in the @xmath1 direction is a twisted one .
thus i consider a slice of the charge density @xmath20 fixing @xmath21 and @xmath22 coordinate for simplicity .
we are seeking for @xmath23 satisfying @xmath24 where @xmath12 is an integer , @xmath25 and @xmath26 . in the twisted boundary condition
, however , @xmath27 obeys the periodic boundary condition as @xmath28 here , integration with respect to @xmath29 should be interpreted as a summation .
thus , we can readily prepare the orbital @xmath23 obeying the twisted boundary condition as , @xmath30 with @xmath31 we can immediately show that a set of @xmath23 forms complete orthonormal and that @xmath32 using the orbital wavefunction @xmath23 obeying the twisted boundary condition , we can show existence of a many - body state @xmath33 whose coordinate expression is the single slater determinant made on @xmath23 .
as a test calculation , i consider a one - dimensional hydrogen array .
the system is denoted by an outer unit cell with ten atoms for the many - body calculation and an inner unit cell with a single atom for the single - particle problem .
the lattice parameters for the inner cell are @xmath34 and @xmath35
. the determination of @xmath36 may be achieved by setting a reference calculation.@xcite here , to show the change in the period explicitly , i consider @xmath37 as a parameter and do only the self - consistent calculation for the extended - kohn - sham system . here
, @xmath38 is the value of the nearest - neighbor transfer parameter .
note that the present system is represented by a tight - binding model with long - range hopping terms .
for the exchange - correlation energy functional , i utilized the perdew - zunger parameterization of the ceperley - alder diffusion monte - carlo data.@xcite the troullier - martins soft pseudopotential is used with the cut - off energy of 20 [ ry].@xcite this setup is confirmed to be enough accurate for the discussion below by increasing the parameters .
the numerical diagonalization with the lanczos algorithm is used to obtain the many - body state of the first - principles hubbard model .
the result of the momentum boost is depicted in fig .
[ mom - boost ] . for @xmath39 ,
the lda calculation shows a crossing in energy flow lines . in this case ,
the lowest energy flow can be traced , since the constrained lda calculation fixing the filling of each @xmath12 point is available .
once a finite @xmath36 is introduced , the many - body calculation automatically concludes the lowest branch of the energy flows . in this system , we see a continuous change of the energy flow which has an energy gap structure at @xmath40 .
this result is qualitatively the same as the single - band hubbard model with only the nearest neighbor hopping.@xcite if the fluctuation reference method is precisely applied , a finite value of @xmath36 is expected , since the inter - atomic distance of @xmath41 is in a strong - correlation regime for the hydrogen molecule.@xcite thus the present result gives a concrete test for the method of determination of the mott insulator from the first - principles .
[ mom - boost ] in the boundary condition .
the system has 10 atoms .
the value of @xmath37 is set to be @xmath42 with @xmath38 is the nearest - neighbor transfer parameter . for the case of @xmath39 ( lda calculation ) ,
crossing of the flow lines occur at @xmath40 , while the crossing becomes anti - crossing due to the charge - gap formation for finite @xmath36 .
, width=264 ]
we have shown that the momentum boost method is formulated rigorously in the density functional theory .
the n representability is shown for the twisted boundary condition . using the multi - reference density functional theory ( mr - dft ) , we are able to construct an effective interacting fermion system , which may undergo the mott insulating transition .
the momentum boost technique is applicable for this problem to detect the transition . in a realistic system
, the formation of the mott gap by the applied pressure or by the effective carrier doping would be seen as a change in the period of the lowest energy flow .
this work is partly supported by the grand - in - aid for the scientific research ( no .
17064006 , no .
15gs0213 and no .
19310094 ) and the 21st coe program by the japan society for promotion of science . | it is shown that a momentum - boost technique applied to the extended kohn - sham scheme enables the computational determination of the mott insulating transition .
self - consistent solutions are given for correlated electron systems by the first - principles calculation defined by the multi - reference density functional theory , in which the effective short - range interaction can be determined by the fluctuation reference method .
an extension of the harriman construction is made for the twisted boundary condition in order to define the momentum - boost technique in the first - principles manner .
for an effectively half - filled - band system , the momentum - boost method tells that the period of a metallic ground state by the lda calculation is shortened to the least period of the insulating phase , indicating occurrence of the mott insulating transition . |
currently the most popular activation function for neural networks is the rectified linear unit ( relu ) , which was first proposed for restricted boltzmann machines @xcite and then successfully used for neural networks @xcite .
the relu activation function is the identity for positive arguments and zero otherwise .
besides producing sparse codes , the main advantage of relus is that they alleviate the vanishing gradient problem @xcite since the derivative of 1 for positive values is not contractive @xcite .
however relus are non - negative and , therefore , have a mean activation larger than zero .
units that have a non - zero mean activation act as bias for the next layer .
if such units do not cancel each other out , learning causes a _ bias shift _ for units in next layer .
the more the units are correlated , the higher their bias shift .
we will see that fisher optimal learning , i.e. , the natural gradient @xcite , would correct for the bias shift by adjusting the weight updates . thus
, less bias shift brings the standard gradient closer to the natural gradient and speeds up learning .
we aim at activation functions that push activation means closer to zero to decrease the bias shift effect
. centering the activations at zero has been proposed in order to keep the off - diagonal entries of the fisher information matrix small @xcite . for neural network
it is known that centering the activations speeds up learning @xcite .
`` batch normalization '' also centers activations with the goal to counter the internal covariate shift @xcite .
also the projected natural gradient descent algorithm ( prong ) centers the activations by implicitly whitening them @xcite .
an alternative to centering is to push the mean activation toward zero by an appropriate activation function .
therefore @xmath0 has been preferred over logistic functions @xcite .
recently `` leaky relus '' ( lrelus ) that replace the negative part of the relu with a linear function have been shown to be superior to relus @xcite .
parametric rectified linear units ( prelus ) generalize lrelus by learning the slope of the negative part which yielded improved learning behavior on large image benchmark data sets @xcite .
another variant are randomized leaky rectified linear units ( rrelus ) which randomly sample the slope of the negative part which raised the performance on image benchmark datasets and convolutional networks @xcite .
in contrast to relus , activation functions like lrelus , prelus , and rrelus do not ensure a noise - robust deactivation state .
we propose an activation function that has negative values to allow for mean activations close to zero , but which saturates to a negative value with smaller arguments .
the saturation decreases the variation of the units if deactivated , so the precise deactivation argument is less relevant .
such an activation function can code the degree of presence of particular phenomena in the input , but does not quantitatively model the degree of their absence .
therefore , such an activation function is more robust to noise .
consequently , dependencies between coding units are much easier to model and much easier to interpret since only activated code units carry much information .
furthermore , distinct concepts are much less likely to interfere with such activation functions since the deactivation state is non - informative , i.e. variance decreasing .
to derive and analyze the bias shift effect mentioned in the introduction , we utilize the natural gradient .
the natural gradient corrects the gradient direction with the inverse fisher information matrix and , thereby , enables fisher optimal learning , which ensures the steepest descent in the riemannian parameter manifold and fisher efficiency for online learning @xcite . the recently introduced hessian - free optimization technique @xcite and the krylov subspace descent methods @xcite
use an extended gauss - newton approximation of the hessian , therefore they can be interpreted as versions of natural gradient descent @xcite . since for neural networks the fisher information matrix is typically too expensive to compute , different approximations of the natural gradient have been proposed .
topmoumoute online natural gradient algorithm ( tonga ) @xcite uses a low - rank approximation of natural gradient descent .
factorized natural gradient ( fang ) @xcite estimates the natural gradient via an approximation of the fisher information matrix by a gaussian graphical model .
the fisher information matrix can be approximated by a block - diagonal matrix , where unit or quasi - diagonal natural gradients are used @xcite .
unit natural gradients or `` unitwise fisher s scoring '' @xcite are based on natural gradients for perceptrons @xcite .
we will base our analysis on the unit natural gradient .
we assume a parameterized probabilistic model @xmath1 with parameter vector @xmath2 and data @xmath3 .
the training data are @xmath4 with @xmath5 , where @xmath6 is the input for example @xmath7 and @xmath8 is its label .
@xmath9 is the loss of example @xmath10 using model @xmath11 .
the average loss on the training data @xmath12 is the empirical risk @xmath13 .
gradient descent updates the weight vector @xmath2 by @xmath14 where @xmath15 is the learning rate .
the _ natural gradient _ is the inverse fisher information matrix @xmath16 multiplied by the gradient of the empirical risk : @xmath17 . for a multi - layer perceptron @xmath18 is the unit activation vector and @xmath19 is the bias unit activation .
we consider the ingoing weights to unit @xmath20 , therefore we drop the index @xmath20 : @xmath21 for the weight from unit @xmath22 to unit @xmath20 , @xmath23 for the activation , and @xmath24 for the bias weight of unit @xmath20 .
the activation function @xmath25 maps the net input @xmath26 of unit @xmath20 to its activation @xmath27 . for computing the fisher information matrix , the derivative of the log - output probability @xmath28 is required .
therefore we define the @xmath29 at unit @xmath20 as @xmath30 , which can be computed via backpropagation , but using the log - output probability instead of the conventional loss function .
the derivative is @xmath31 .
we restrict the fisher information matrix to weights leading to unit @xmath20 which is the _ unit fisher information matrix _ @xmath32 .
@xmath32 captures only the interactions of weights to unit @xmath20 .
consequently , the unit natural gradient only corrects the interactions of weights to unit @xmath20 , i.e. considers the riemannian parameter manifold only in a subspace .
the unit fisher information matrix is @xmath33_{kj } \ = \ { \mathbf{\mathrm{e}}}_{p({\bm{z}};{\bm{w } } ) } \ !
\left ( \frac{\partial \ln p({\bm{z}};{\bm{w } } ) } { \partial w_{k } } \ \frac{\partial \ln p({\bm{z}};{\bm{w } } ) } { \partial w_{j } } \right ) \ = \ { \mathbf{\mathrm{e}}}_{p({\bm{z}};{\bm{w } } ) } ( \delta^2 \ a_k \ a_j ) \ .\end{aligned}\ ] ] weighting the activations by @xmath34 is equivalent to adjusting the probability of drawing inputs @xmath35 .
inputs @xmath35 with large @xmath34 are drawn with higher probability . since @xmath36 , we can define a distribution @xmath37 : @xmath38 using @xmath37 , the entries of @xmath32 can be expressed as second moments : @xmath39_{kj } \ & = \ { \mathbf{\mathrm{e}}}_{p({\bm{z } } ) } ( \delta^2 \ a_k \ a_j ) \ = \ \int \delta^2 \ a_k \ a_j \ p({\bm{z } } ) \
d{\bm{z}}\ = \ { \mathbf{\mathrm{e}}}_{p({\bm{z } } ) } ( \delta^2 ) \ { \mathbf{\mathrm{e}}}_{q({\bm{z } } ) } ( a_k \ a_j ) \ .\end{aligned}\ ] ] if the bias unit is @xmath19 with weight @xmath40 then the weight vector can be divided into a bias part @xmath40 and the rest @xmath2 : @xmath41 . for the row @xmath42_{0}$ ] that corresponds to the bias weight , we have : @xmath43 the next theorem [ th : th1 ] gives the correction of the standard gradient by the unit natural gradient where the bias weight is treated separately ( see also @xcite ) .
[ th : th1 ] the unit natural gradient corrects the weight update @xmath44 to a unit @xmath20 by following affine transformation of the gradient @xmath45 : @xmath46 where @xmath47_{\neg 0,\neg 0 } = { \mathbf{\mathrm{e}}}_{p({\bm{z } } ) } ( \delta^2 ) { \mathbf{\mathrm{e}}}_{q({\bm{z}})}({\bm{a}}{\bm{a}}^t)$ ] is the unit fisher information matrix without row 0 and column 0 corresponding to the bias weight .
the vector @xmath48_{0}$ ] is the zeroth column of @xmath32 corresponding to the bias weight , and the positive scalar @xmath49 is @xmath50 where @xmath18 is the vector of activations of units with weights to unit @xmath20 and @xmath51 .
multiplying the inverse fisher matrix @xmath52 with the separated gradient @xmath53 gives the weight update @xmath44 : @xmath54 where @xmath55_{0 } \quad , \quad c \
\left[{\bm{f}}({\bm{w}})\right]_{00 } \quad , \quad { \bm{u}}\ = \ - \ s \ { \bm{a}}^{-1 } \
{ \bm{b}}\quad , \quad s \ = \ \left ( c \ - \ { \bm{b}}^t{\bm{a}}^{-1 } { \bm{b}}\right)^{-1 } \ .\end{aligned}\ ] ] the previous formula is derived in lemma [ th : lemma1 ] in the appendix .
using @xmath56 in the update gives @xmath57 the right hand side is obtained by inserting @xmath58 in the left hand side update . since @xmath59 , @xmath60 , and @xmath61 , we obtain @xmath62 applying lemma [ th : lemma2 ] in the appendix gives the formula for @xmath49 . the _ bias shift _ ( mean shift ) of unit @xmath20
is the change of unit @xmath20 s mean value due to the weight update .
bias shifts of unit @xmath20 lead to oscillations and impede learning . see section 4.4 in @xcite for demonstrating this effect at the inputs and in @xcite for explaining this effect using the input covariance matrix . such bias shifts are mitigated or even prevented by the unit natural gradient .
the _ bias shift correction _ of the unit natural gradient is the effect on the bias shift due to @xmath63 which captures the interaction between the bias unit and the incoming units . without bias shift correction ,
i.e. , @xmath64 and @xmath65 , the weight updates are @xmath66 and @xmath67 .
as only the activations depend on the input , the bias shift can be computed by multiplying the weight update by the mean of the activation vector @xmath18 .
thus we obtain the bias shift @xmath68 .
the bias shift strongly depends on the correlation of the incoming units which is captured by @xmath69 .
next , theorem [ th : th2 ] states that the bias shift correction by the unit natural gradient can be considered to correct the incoming mean @xmath70 proportional to @xmath71 toward zero .
[ th : th2 ] the bias shift correction by the unit natural gradient is equivalent to an additive correction of the incoming mean by @xmath72 and a multiplicative correction of the bias unit by @xmath73 , where @xmath74 using @xmath75 , the bias shift is : @xmath76 the mean correction term , indicated by an underbrace in previous formula , is @xmath77 the expression eq
. for @xmath73 follows from lemma [ th : lemma2 ] in the appendix .
the bias unit correction term is @xmath78 . in theorem
[ th : th2 ] we can reformulate @xmath79 .
therefore @xmath73 increases with the length of @xmath80 for given variances and covariances .
consequently the bias shift correction through the unit natural gradient is governed by the length of @xmath80 .
the bias shift correction is zero for @xmath81 since @xmath82 does not correct the bias unit multiplicatively . using eq . , @xmath80 is split into an offset and an information containing term : @xmath83 in general , _ smaller positive @xmath70 lead to smaller positive @xmath80 , therefore to smaller corrections . _
the reason is that in general the largest absolute components of @xmath84 are positive , since activated inputs will activate the unit @xmath20 which in turn will have large impact on the output . to summarize ,
the unit natural gradient corrects the bias shift of unit @xmath20 via the interactions of incoming units with the bias unit to ensure efficient learning .
this correction is equivalent to shifting the mean activations of the incoming units toward zero and scaling up the bias unit . to reduce the undesired bias shift effect without the natural gradient , either the ( i ) activation of incoming units
can be centered at zero or ( ii ) activation functions with negative values can be used .
we introduce a new activation function with negative values while keeping the identity for positive arguments where it is not contradicting .
the _ exponential linear unit _ ( elu ) with @xmath85 is @xmath86 the elu hyperparameter @xmath87 controls the value to which an elu saturates for negative net inputs ( see fig . [ fig : activationfunction ] ) .
elus diminish the vanishing gradient effect as rectified linear units ( relus ) and leaky relus ( lrelus ) do .
the vanishing gradient problem is alleviated because the positive part of these functions is the identity , therefore their derivative is one and not contractive .
in contrast , @xmath88 and sigmoid activation functions are contractive almost everywhere . r0.45 in contrast to relus , elus have negative values which pushes the mean of the activations closer to zero .
mean activations that are closer to zero enable faster learning as they bring the gradient closer to the natural gradient ( see theorem [ th : th2 ] and text thereafter ) .
elus saturate to a negative value when the argument gets smaller .
saturation means a small derivative which decreases the variation and the information that is propagated to the next layer .
therefore the representation is both noise - robust and low - complex @xcite .
elus code the degree of presence of input concepts , while they neither quantify the degree of their absence nor distinguish the causes of their absence .
this property of non - informative deactivation states is also present at relus and allowed to detect biclusters corresponding to biological modules in gene expression datasets @xcite and to identify toxicophores in toxicity prediction @xcite .
the enabling features for these interpretations is that activation can be clearly distinguished from deactivation and that only active units carry relevant information and can crosstalk .
in this section , we assess the performance of exponential linear units ( elus ) if used for unsupervised and supervised learning of deep autoencoders and deep convolutional networks .
elus with @xmath89 are compared to ( i ) rectified linear units ( relus ) with activation @xmath90 , ( ii ) leaky relus ( lrelus ) with activation @xmath91 ( @xmath92 ) , and ( iii ) shifted relus ( srelus ) with activation @xmath93 .
comparisons are done with and without batch normalization .
the following benchmark datasets are used : ( i ) _ mnist _ ( gray images in 10 classes , 60k train and 10k test ) , ( ii ) _ cifar-10 _
( color images in 10 classes , 50k train and 10k test ) , ( iii ) _ cifar-100 _
( color images in 100 classes , 50k train and 10k test ) , and ( iv ) _ imagenet _ ( color images in 1,000 classes , 1.3 m train and 100k tests ) . we first want to verify that elus keep the mean activations closer to zero than other units . fully connected deep neural networks with elus ( @xmath89 ) , relus , and lrelus ( @xmath94 )
were trained on the mnist digit classification dataset while each hidden unit s activation was tracked .
each network had eight hidden layers of 128 units each , and was trained for 300 epochs by stochastic gradient descent with learning rate @xmath95 and mini - batches of size 64 .
the weights have been initialized according to @xcite . after each epoch we calculated the units average activations on a fixed subset of the training data .
[ fig : mnistplots ] shows the median over all units along learning .
elus stay have smaller median throughout the training process .
the training error of elu networks decreases much more rapidly than for the other networks .
section [ sec : variancemean ] in the appendix compares the variance of median activation in relu and elu networks .
the median varies much more in relu networks .
this indicates that relu networks continuously try to correct the bias shift introduced by previous weight updates while this effect is much less prominent in elu networks . to evaluate elu networks at unsupervised settings
, we followed @xcite and @xcite and trained a deep autoencoder on the mnist dataset .
the encoder part consisted of four fully connected hidden layers with sizes 1000 , 500 , 250 and 30 , respectively .
the decoder part was symmetrical to the encoder . for learning we applied stochastic gradient descent with mini - batches of 64 samples for 500 epochs using the fixed learning rates ( @xmath96 ) .
[ fig : mnistauto ] shows , that elus outperform the competing activation functions in terms of training / test set reconstruction error for all learning rates . as already noted by @xcite , higher learning rates seem to perform better . in this subsection we show that elus indeed possess a superior learning behavior compared to other activation functions as postulated in section [ sec : elu ] .
furthermore we show that elu networks perform better than relu networks with batch normalization .
we use as benchmark dataset cifar-100 and use a relatively simple convolutional neural network ( cnn ) architecture to keep the computational complexity reasonable for comparisons .
+ + + the cnn for these cifar-100 experiments consists of 11 convolutional layers arranged in stacks of ( @xmath97 , [ 1\times192\times1 , 1\times240\times3 ] , [ 1\times240\times1 , 1\times260\times2 ] , [ 1\times260\times1,1\times280\times2 ] , [ 1\times280\times1,1\times300\times2 ] , [ 1\times300\times1 ] , [ 1\times100\times1]$ ] ) layers @xmath98 units @xmath98 receptive fields .
2@xmath982 max - pooling with a stride of 2 was applied after each stack . for network regularization we used the following drop - out rate for the last layer of each stack ( @xmath99 ) .
the @xmath100-weight decay regularization term was set to @xmath101 .
the following learning rate schedule was applied ( @xmath102 , 35k-85k [ 0.005 ] , 85k-135k [ 0.0005 ] , 135k-165k [ 0.00005]$ ] ) ( iterations [ learning rate ] ) . for fair comparisons
, we used this learning rate schedule for all networks . during previous experiments ,
this schedule was optimized for relu networks , however as elus converge faster they would benefit from an adjusted schedule .
the momentum term learning rate was fixed to 0.9 .
the dataset was preprocessed as described in @xcite with global contrast normalization and zca whitening .
additionally , the images were padded with four zero pixels at all borders .
the model was trained on @xmath103 random crops with random horizontal flipping . besides that , we no further augmented the dataset during training .
each network was run 10 times with different weight initialization . across networks with different activation functions the same run number had the same initial weights .
mean test error results of networks with different activation functions are compared in fig .
[ fig : rescifar100 ] , which also shows the standard deviation .
elus yield on average a test error of 28.75(@xmath1040.24)% , while srelus , relus and lrelus yield 29.35(@xmath1040.29)% , 31.56(@xmath1040.37)% and 30.59(@xmath1040.29)% , respectively .
elus achieve both lower training loss and lower test error than relus , lrelus , and srelus .
both the elu training and test performance is significantly better than for other activation functions ( wilcoxon signed - rank test with @[email protected] ) .
batch normalization improved relu and lrelu networks , but did not improve elu and srelu networks ( see fig .
[ fig : rescifar100_bn ] ) .
elu networks significantly outperform relu networks with batch normalization ( wilcoxon signed - rank test with @[email protected] ) .
the following experiments should highlight the generalization capabilities of elu networks .
the cnn architecture is more sophisticated than in the previous subsection and consists of 18 convolutional layers arranged in stacks of ( @xmath107 , [ 1\times384\times1 , 1\times384\times2,2\times640\times2 ] , [ 1\times640\times1,3\times768\times2 ] , [ 1\times768\times1,2\times896\times2 ] , [ 1\times896\times3,2\times1024\times2 ] , [ 1\times1024\times1,1\times1152\times2 ] , [ 1\times1152\times1 ] , [ 1\times100\times1]$ ] ) .
initial drop - out rate , max - pooling after each stack , @xmath100-weight decay , momentum term , data preprocessing , padding , and cropping were as in previous section .
the initial learning rate was set to 0.01 and decreased by a factor of 10 after 35k iterations .
the mini - batch size was 100 .
for the final 50k iterations fine - tuning we increased the drop - out rate for _ all _ layers in a stack to ( @xmath108 ) , thereafter increased the drop - out rate by a factor of 1.5 for 40k additional iterations .
* 1>p8em*2>p11.5em*1>p4.25em * network & * cifar-10 ( test error % ) & * cifar-100 ( test error % ) & * augmented + alexnet & 18.04 & 45.80 & + dsn & 7.97 & 34.57 & @xmath109 + nin & 8.81 & 35.68 & @xmath109 + maxout & 9.38 & 38.57 & @xmath109 + all - cnn & 7.25 & 33.71 & @xmath109 + highway network & 7.60 & 32.24 & @xmath109 + fract .
max - pooling & * 4.50 & 27.62 & @xmath109 + elu - network & 6.55 & * 24.28 & + * * * * * * elu networks are compared to following recent successful cnn architectures : alexnet @xcite , dsn @xcite , nin @xcite , maxout @xcite , all - cnn @xcite , highway network @xcite and fractional max - pooling @xcite .
the test error in percent misclassification are given in tab .
[ tab : tab_res1 ] .
elu - networks are the second best on cifar-10 with a test error of 6.55% but still they are among the top 10 best results reported for cifar-10 .
elu networks performed best on cifar-100 with a test error of 24.28% .
this is the best published result on cifar-100 , without even resorting to multi - view evaluation or model averaging . finally , we evaluated elu - networks on the 1000-class imagenet dataset .
it contains about 1.3 m training color images as well as additional 50k images and 100k images for validation and testing , respectively . for this task
, we designed a 15 layer cnn , which was arranged in stacks of ( @xmath110 ) layers @xmath98 units @xmath98 receptive fields or fully - connected ( fc ) .
2@xmath982 max - pooling with a stride of 2 was applied after each stack and spatial pyramid pooling ( spp ) with 3 levels before the first fc layer @xcite . for network regularization we set the @xmath100-weight decay term to @xmath101 and used 50% drop - out in the two penultimate fc layers .
images were re - sized to 256@xmath98256 pixels and per - pixel mean subtracted . trained was on @xmath111 random crops with random horizontal flipping .
besides that , we did not augment the dataset during training .
[ fig : resimagenet ] shows the learning behavior of elu vs. relu networks .
panel ( b ) shows that elus start reducing the error earlier .
the elu - network already reaches the 20% top-5 error after 160k iterations , while the relu network needs 200k iterations to reach the same error rate .
the single - model performance was evaluated on the single center crop with no further augmentation and yielded a top-5 validation error below 10% .
currently elu nets are 5% slower on imagenet than relu nets .
the difference is small because activation functions generally have only minor influence on the overall training time @xcite . in terms of wall clock time ,
elus require 12.15h vs. relus with 11.48h for 10k iterations .
we expect that elu implementations can be improved , e.g. by faster exponential functions @xcite .
we have introduced the _ exponential linear units _ ( elus ) for faster and more precise learning in deep neural networks .
elus have negative values , which allows the network to push the mean activations closer to zero .
therefore elus decrease the gap between the normal gradient and the unit natural gradient and , thereby speed up learning .
we believe that this property is also the reason for the success of activation functions like lrelus and prelus and of batch normalization .
in contrast to lrelus and prelus , elus have a clear saturation plateau in its negative regime , allowing them to learn a more robust and stable representation .
experimental results show that elus significantly outperform other activation functions on different vision datasets .
further elu networks perform significantly better than relu networks trained with batch normalization .
elu networks achieved one of the top 10 best reported results on cifar-10 and set a new state of the art in cifar-100 without the need for multi - view test evaluation or model averaging .
furthermore , elu networks produced competitive results on the imagenet in much fewer epochs than a corresponding relu network .
given their outstanding performance , we expect elu networks to become a real time saver in convolutional networks , which are notably time - intensive to train from scratch otherwise . [ [ acknowledgment . ] ] acknowledgment .
+ + + + + + + + + + + + + + + we thank the nvidia corporation for supporting this research with several titan x gpus and roland vollgraf and martin heusel for helpful discussions and comments on this work .
[ th : lemma1 ] the positive definite matrix @xmath112 is in block format with matrix @xmath113 , vector @xmath63 , and scalar @xmath114 .
the inverse of @xmath112 is @xmath115 where @xmath116 for block matrices the inverse is @xmath117 where the matrices on the right hand side are : @xmath118 further if follows that @xmath119 we now use this formula for @xmath120 being a vector and @xmath121 a scalar .
we obtain @xmath122 where the right hand side matrices , vectors , and the scalar @xmath49 are : @xmath123 again it follows that @xmath124 a reformulation using @xmath125 gives @xmath126
[ th : lemma2 ] for a random variable @xmath18 holds @xmath127 and @xmath128 furthermore holds @xmath129 the sherman - morrison theorem states @xmath130 therefore we have @xmath131 using the identity @xmath132 for the second moment and eq . , we get @xmath133 the last inequality follows from the fact that @xmath134 is positive definite . from last equation , we obtain further @xmath135 for the mixed quadratic form we get from eq .
@xmath136 from this equation follows @xmath137 therefore we get @xmath129
to compare the variance of median activation in relu and elu networks , we trained a neural network with 5 hidden layers of 256 hidden units for 200 epochs using a learning rate of 0.01 , once using relu and once using elu activation functions on the mnist dataset .
after each epoch , we calculated the median activation of each hidden unit on the whole training set .
we then calculated the variance of these changes , which is depicted in figure [ fig : varianceofchanges ] .
the median varies much more in relu networks .
this indicates that relu networks continuously try to correct the bias shift introduced by previous weight updates while this effect is much less prominent in elu networks . | we introduce the `` exponential linear unit '' ( elu ) which speeds up learning in deep neural networks and leads to higher classification accuracies . like rectified linear units
( relus ) , leaky relus ( lrelus ) and parametrized relus ( prelus ) , elus alleviate the vanishing gradient problem via the identity for positive values .
however elus have improved learning characteristics compared to the units with other activation functions .
in contrast to relus , elus have negative values which allows them to push mean unit activations closer to zero like batch normalization but with lower computational complexity .
mean shifts toward zero speed up learning by bringing the normal gradient closer to the unit natural gradient because of a reduced bias shift effect .
while lrelus and prelus have negative values , too , they do not ensure a noise - robust deactivation state .
elus saturate to a negative value with smaller inputs and thereby decrease the forward propagated variation and information .
therefore elus code the degree of presence of particular phenomena in the input , while they do not quantitatively model the degree of their absence . in experiments ,
elus lead not only to faster learning , but also to significantly better generalization performance than relus and lrelus on networks with more than 5 layers . on cifar-100 elus networks
significantly outperform relu networks with batch normalization while batch normalization does not improve elu networks .
elu networks are among the top 10 reported cifar-10 results and yield the best published result on cifar-100 , without resorting to multi - view evaluation or model averaging . on imagenet , elu networks considerably speed up learning compared to a relu network with the same architecture , obtaining less than 10% classification error for a single crop , single model network . |
agn classification is still somewhat heuristic , but in the unified model , blazars are believed to be sources viewed close to the axis of a powerful relativistic jet ( * ? ? ? * and references therein ) .
they thus have compact flat spectrum radio counterparts , with apparent superluminal motion at vlbi scales .
optically , the sources are often variable , exhibit significant polarization and show either broad emission lines ( flat spectrum radio quasar ) or continuum - dominated ( bl lac ) spectra .
perhaps the most interesting consequence of the jet / line - of - sight alignment is the observability of a compton scattered component at x - ray to gev or ( for nearby sources ) tev energies .
the interaction of the high energy jet particles and radiation with surrounding photon fields allows unique probes of the extragalactic background light ( ebl ) .
one of the principal discoveries of the _ egret _ experiment on the _ compton gamma ray observatory _ was the large population of blazars emitting at gev energies .
these represented the largest identified population in the third _ egret _
( 3eg ) catalog ; still many high latitude sources remained unidentified .
we have developed a method to extend identifications to fainter radio flux levels @xcite , finding counterparts for 70% of the @xmath3 sources .
follow - on observations show that most of these are previously unidentified blazars .
upcoming gev @xmath0-ray missions , especially _ glast _ , will have much higher sensitivity and are expected to detect several thousand blazars .
this substantially exceeds the total number of blazars cataloged to date @xcite . thus in preparation for this mission
, we have selected radio ( & x - ray ) sources with properties like the _ egret _ blazars and are obtaining optical identifications and spectroscopy ( * ? ? ?
* srm04 ) . to prioritize the optical observations ,
we have also developed a method for estimating the probability that a given position has an excess of @xmath0-ray photons in the _ egret _ survey observations , including the effect of strong variability .
the radio blazar candidates selected above show , as a set , a clear excess of @xmath0-ray flux over random sky positions ( srm04 ) ; individually , they are well below the 4@xmath4 criterion for inclusion in the 3eg catalog . for example , radio blazar candidates with @xmath575% probability of _ egret _ @xmath0-ray excess are found at @xmath6 , and thus may represent 10 - 20% of the new sources detectable to _ glast_. these are prime candidates for optical follow - up ; we find that most are indeed flat spectrum radio quasars with redshifts 1 - 2.5 .
q0609 + 6930 , however , has a very high redshift of 5.47 .
observations described below support a blazar identification for this source ; this would be the highest redshift blazar discovered to date .
in selecting @xmath0-ray blazar candidates we start from compact 8.4ghz sources ( from class , meyers et al . 2003 or from our own vla snapshots ) and identify flat / inverted spectrum sources using nvss counterparts .
our selection algorithm also applies a weighting toward x - ray sources sources detected in the _ rosat _ all - sky survey ( rass ) ; this is weak as many of the _ egret _ blazars are below the rass survey threshold .
this identification scheme differs from other blazar surveys ( eg
. dxrbs , * ? ? ?
* ) and gives a much higher correlation with the 3eg sources .
we follow up with optical spectroscopy obtained with the marcario low resolution spectrograph ( lrs ; * ? ? ?
* ) at the hobby - eberly telescope , obtaining source classifications and redshifts .
q0906 + 6930 has a strongly inverted spectrum @xmath7 @xmath8 with a class - epoch flux of @xmath9mjy .
the radio position is 09:06:30.75 + 69:30:30.8 ( j2000.0 ) .
this source was cataloged as compact at the @xmath10 scale , but the short class snapshots provide only limited @xmath11mjy sensitivity to flux on arcsecond scales .
this source is just included in our ` sub - threshold ' @xmath0-ray target list , with a likelihood analysis of the @xmath0-ray counts giving a 75% probability for emission in excess of background at this position .
we obtained 2@xmath12300s exposure with the het lrs on 1/18/04 , under poor @xmath13 conditions , which nonetheless showed a classic ly@xmath14 forest - absorbed qso spectrum with high @xmath15 redshift .
follow - on exposures of 2@xmath12600s on 1/27/04 with a 300l / mm grating , @xmath16 slit and 5150 long - pass filter ( to preclude 2nd order contamination ) and improved @xmath17 seeing produced the spectrum shown in figure 1 .
these data were subject to optimal extraction , calibration and telluric correction using standard iraf routines .
uncorrectable fringing and imperfect sky subtraction introduce large noise beyond 9500 .
the resulting spectrum has a dispersion of 4 / pixel and an effective resolution of 16 , covering @xmath185200 - 10,000 . in the spectrum ly@xmath14
is strong , but absorbed ; we estimate a pre - absorption rest equivalent width of 50 - 60 .
the redshift , z=5.47@xmath19 , is best constrained by the nv and oiv lines ; the oiv rest ew is 22 and the kinematic width is @xmath20km / s .
the uv continuum luminosity at 1350 is @xmath21erg / s ( we use @xmath22km / s / mpc , @xmath23 , @xmath24 ) .
civ is not clearly detected , so if we use the oiv kinematic width in the @xcite uv m - l relationship we obtain @xmath25 \approx 10.2 $ ] .
this is near the upper end of masses inferred for high z sources and formation of such a high @xmath26 black hole after @xmath11gyr is difficult to understand .
there are several candidate civ absorption line systems redward of ly@xmath14 .
two candidate systems with apparent damped ly lines are marked .
the presence of intervening absorption raises the possibility that q0906 + 6930 is lensed , which could inflate our estimate of the bh mass .
however , class finds no radio lens , there is no extended emission visible on the poss and the source appears point - like in a short pre - spectrum acquisition image .
improved optical / near ir spectroscopy is needed to tighten up the mass estimate , and to improve the identification of the absorption systems as strong civ doublets . to search for evidence of compact jet structure that would support the blazar designation of q0906 + 6930 , we obtained snapshots with the vlba at 2 cm and 7 mm wavelengths ( figure 2 ) . with 10 stations recording two 8-mhz bandpasses at 15.36ghz ,
each in two polarizations , we obtained an average of 60 minutes of on - source per baseline under nominal observing conditions on 2004 february 27 ( mjd 53062 ) .
we used 0716 + 714 to calibrate phases and delays prior to self - calibration of the blazar , using aips .
we constructed deconvolved images with natural weighting of the data ( half - power beam width [email protected] at -2@xmath27 ) .
the rms background noise in the image was @xmath28 above the thermal noise limit of 0.18 mjy .
the 7 mm observations recorded @xmath29 minutes on - source per baseline on 2004 march 3 ( mjd 53067 ) , tuned to 43.21 ghz .
about 30% of the data could not be self - calibrated because of rapid ( @xmath30minute ) tropospheric fluctuations .
in addition , we obtained no usable data from the los alamos antenna , due to weather , and from one polarization - baseband combination recorded at the saint croix antenna .
the half - power beamwidth was @xmath31 mas at @xmath32 for natural weighting .
the compact core was unresolved at both wavelengths .
in the 2 cm image a clear jet component is seen at pa 225@xmath33 .
for a two component gaussian model we fit a core flux of 115@xmath34mjy and a jet flux of 6.3@xmath35mjy . the jet is marginally detected in our 7 mm image with a maximum @xmath36 beams from the position of the 2 cm peak , although it is only comparable to the largest peaks in the image background .
the two component fit gives a core flux 42@xmath37mjy and jet flux @xmath38mjy , for a 3.7@xmath4 detection .
our combined _ egret _ likelihood analysis finds a nominal @xmath39 excess of @xmath0-ray photons at the radio position . in the viewing period with the most significant detection , vp0220 , the source produced 1.12@xmath40(e@xmath1100mev ) .
however , if other sources beyond the standard 4@xmath4 3eg sources are present in this region the fit flux is even lower . at this point
it is best to employ the mission - averaged upper limit @xmath41(e@xmath1100mev ) .
we can collect the existing data into a crude spectral energy distribution ( sed , figure 3 ) . comparing to the well known blazar 3c279 placed at z=5.47 , q0906 + 6930
has a brighter uv flux , fainter cm wavelength emission and ( potentially ) a gev peak power nearly 10@xmath12 greater .
the source has by far the highest radio loudness @xmath42 of any @xmath43 qso .
we also guide the eye with a simple synchrotron - compton spectrum @xcite for a broken power - law approximating a cooled electron spectrum .
the synchrotron component is broadly similar to that of 3c279 .
the compact core spectrum ( inset ) appears to turn over at cm wavelengths and does not connect to the falling optical spectrum .
this is not however unexpected , as most blazars show extra radio components at cm wavelengths .
if the 7 mm jet detection is trusted , the @xmath44 of the jet component is somewhat flatter than that of the core @xmath45 .
this is unusual , although not unique for pc - scale jets ( a. marscher , private comm . ) ; additional vlbi observations will be needed to see if this is temporary due to jet component brightening at 7 mm . with only a marginal gev detection and an x - ray upper limit ,
the amplitude of the compton component is poorly constrained .
however , it is interesting to note that the broad line region flux is much larger than that of 3c279 with with a @xmath46 brighter optical continuum and rest equivalent line widths @xmath47 larger .
since comptonized blr flux is generally believed to dominate the gev emission , a bright @xmath0-ray peak may be motivated , as shown by the dotted curves .
we expect the compton scattered emission of blazar jets to be highly beamed .
thus , few blazar x- and @xmath0-ray sources are visible , but these may appear quite bright . the glast mission will have a large duty cycle and a survey sensitivity @xmath48(e@xmath1100mev ) , so even if the source is a flaring 3c279 analog , the prospects for detection are good .
if the gev spectrum of this source can be measured , then glast can observe the effect of pair attenuation of the @xmath1gev flux by the optical - ir ebl @xcite . at its high redshift @xmath4910gev observations of q0906 + 6930
will be probing @xmath50 photons from high @xmath51 star formation .
this ebl is presently poorly understood , but an approximate attenuation curve based on the lower @xmath51 ebl is applied for the combined spectrum ( full line ) .
the cut - off is well within the glast range .
it may be measured if , for example , a stable absorption cutoff is seen on a gev spectrum whose slope and amplitude varies with time . in the x - ray regime ,
core emission probes the low energy compton component , with the best connection to the radio .
however , most interesting is the possibility of detecting a resolved @xmath17 ( @xmath52kpc ) scale x - ray jet produced by compton up - scatter of cmb photons . such extended
( @xmath13 ) jet structure has been seen in _ cxo _
observations of the qso gb 1508 + 5714 @xcite at z=4.3 .
this feature is also seen as a faint radio jet @xcite and the flux ratio @xmath53 supports the cmb up - scatter hypothesis .
extended emission x - ray observed @xmath5420@xmath55 from the @xmath56 qso j1306 + 0356 @xcite may represent a second high - z jet , although radio emission is not seen . with a @xmath57 increase in the energy density of the cmb seed photons
, such jets may be detectable to very high redshifts@xcite .
q0906 + 6930 at z=5.47 experiences @xmath58 the local cmb energy density .
as we already have evidence for an energetic jet , prospects for detection of a compton up - scatter component are good and we are pursuing the required x - ray and radio observations to test this hypothesis .
if seen , the link to the external seed field gives x - ray / radio measurements unique power to probe the particle and field populations in this high - z jet .
the hobby - eberly telescope ( het ) is a joint project of the university of texas at austin , the pennsylvania state university , stanford university , ludwig - maximilians - universitt mnchen , and georg - august - universitt gttingen .
the het is named in honor of its principal benefactors , william p. hobby and robert e. eberly .
the marcario low resolution spectrograph is named for mike marcario of high lonesome optics who fabricated several optics for the instrument but died before its completion .
the lrs is a joint project of the hobby - eberly telescope partnership and the instituto de astronomia de la universidad nacional autonoma de mexico .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . | we report the discovery of a radio - loud flat - spectrum qso at z=5.47 with properties similar to those of the _ egret _ @xmath0-ray blazars .
this source is the brightest radio qso at z@xmath15 , with a pc - scale radio jet and a black hole mass estimate @xmath2 .
it appears to be the most distant blazar discovered to date .
high energy observations of this source can provide powerful probes of the background radiation in the early universe . |
in iron pnictides , the superconductivity @xcite emerges near an antiferromagnetically ordered state @xcite in the phase diagram .
this implies a strong interplay between the superconductivity and magnetism in these materials .
it has been observed that the parent compounds of iron pnictides have a @xmath7 antiferromagnetic order , which could arise either within a weak - coupling approach invoking a fermi surface nesting,@xcite or from a strong - coupling approach whose starting point is a local moment @xmath8 model.@xcite in the strong - coupling approach , the @xmath7 antiferromagnetic order can arise from a @xmath9 model for sufficiently large @xmath10 .
but the strong anisotropic magnetic excitations in the paramagnetic phase observed from inelastic neutron scattering measurements prompts the necessity and importance of including a biquadratic exchange interaction @xmath11 between the nearest - neighbor ( n.n . )
pair of spins.@xcite from the theoretical perspective , the emergence of the biquadratic interaction is natural when the system contains multiple orbitals and when the correlations are strong in the metallic ground state : the low - energy effective hamiltonian in the spin sector should consist of not only two - spin heisenberg interactions , such as @xmath12 and @xmath10 heisenberg exchange between n.n .
and next - nearest - neighbor ( n.n.n . ) spins on a square lattice , but also interactions involving larger number of spins , particularly , the interaction for spin size @xmath13@xcite . in a previous work
, we have shown that the biquadratic interaction is crucial in understanding the spin dynamics in the paramagnetic phase of the parent iron pnictides.@xcite given the close correlations between magnetic excitations and the nearby superconductivity in the phase diagram , it would be equally interesting to ask how the biquadratic interaction may affect superconductivity in these materials .
the pairing symmetry of the iron - based superconductors has been studied via various experimental techniques .
angle resolved photoemission spectroscopy ( arpes ) measurements find the superconducting gap to be nodeless and isotropic at both the hole and electron pockets in a number of materials @xcite .
neutron scattering measurements on these compounds observe a clear spin resonance mode in the superconducting state @xcite .
these are consistent with an @xmath5 pairing channel , with the pairing order parameter @xmath14 changing sign between the hole pockets near the brillouin zone ( bz ) center and the electron pockets near the zone corner , which arises within both weak - coupling and strong - coupling approaches @xcite .
recent experiments find evidences of anisotropic or even nodal superconducting gaps on several iron pnictide compounds in either underdoped or heavily doped regimes .
for example , for the heavily hole doped compound kfe@xmath15as@xmath15 , thermal conductivity and penetration - depth measurements indicate the presence of nodes,@xcite suggesting a different pairing symmetry to its optimally doped counterpart.@xcite experimentally , the issue is not settled , with thermal conductivity measurements interpreted as evidence for the @xmath16-wave symmetry of the order parameter@xcite , consistent with predictions of the functional renormalization group @xcite , whereas the arpes measurements are indicative of an extended @xmath17-wave state with accidental nodes.@xcite a recent arpes study of heavily hole - doped ba@xmath18k@xmath19fe@xmath15as@xmath15 found nodes in the small @xmath20 hole pockets off - centered around the @xmath21 point at the edge of the brillouin zone @xcite .
also in the underdoped nafeas and in the undoped lifeas , anisotropic but nodeless superconducting gaps along the fermi pockets have been identified.@xcite all these results question the validity of isotropic @xmath5 pairing in the underdoped and/or overdoped regimes . in this paper , we study the superconducting states of a strong - coupling two - orbital @xmath22 model with electron doping using a slave - boson mean - field theory . by comparing to the results of a @xmath23 model ,
we show that a moderate biquadratic coupling @xmath24 changes the superconducting pairing phase diagram dramatically .
it lowers the energy of both the @xmath1 @xmath2 and @xmath3 @xmath4 pairing channels , and thus favors an @xmath6 pairing .
we are particularly interested in the doping evolution of superconductivity .
our results indicate that the biquadratic interaction is a key ingredient in tuning the relative stability of various superconducting states as a function of doping .
this is in dramatic difference to the case of absent biquadratic interaction ( @xmath25 ) , where the @xmath5 pairing is found to be robust in the entire range of studied dopings . by contrast
, a moderate biquadratic interaction leads to a serious change of the pairing symmetry , pairing strength , and gap anisotropy with doping , favouring @xmath6 and even pure @xmath16-wave pairing in the heavily electron- and hole - doped regimes .
we show that this is important in understanding the different pairing states observed in recent experiments .
the remainder of the paper is organized as follows . in section
[ sec : model ] , we introduce the two - orbital @xmath22 model and describe the slave - boson mean - field theory for superconductivity used in this work . in section [ sec : phd ]
we show how the biquadratic coupling @xmath24 affects the superconducting phase diagram . in section [ sec : doping ]
we discuss how the superconducting states evolve with doping .
further discussion in connection with existing theories and experiments is included in section [ sec : disc ] .
section [ sec : conclusion ] contains a few concluding remarks .
we consider a two - orbital @xmath22 model on a two - dimensional ( 2d ) square lattice
. it can be obtained from a two - orbital hubbard model via a perturbative expansion in the strong coupling limit @xcite .
the reason behind studying the two - orbital ( as opposed to the full five - orbital ) model is the universally accepted fact that the major contribution to the fermi surface comes from the iron @xmath26 orbitals ( @xmath27 ) , whereas the @xmath28 orbital weight is very small@xcite .
additionally , the @xmath29 and @xmath30 orbitals carry most of the spectral weight@xcite and the @xmath31 orbital can thus be neglected in the first approximation .
it is true that in order to obtain _ all _ the fermi pockets observed in arpes , one needs to consider all 5 fe orbitals@xcite , and one of us has studied superconductivity in such a 5-orbital @xmath23 model @xcite
. however , inclusion of the biquadratic @xmath24 term would be highly non - trivial for such a model and for reasons of transparency of the analysis , we chose to focus on the two - orbital model first , with the hope that our central results should remain valid upon inclusion of other orbitals .
the detailed comparison with the model will be made in section [ sec : disc ] .
the two - orbital @xmath22 hamiltonian reads as follows : @xmath32 ^ 2,\end{aligned}\ ] ] where @xmath33 creates an electron at site @xmath34 , with orbital index @xmath35 and spin projection @xmath17 ; @xmath36 is the chemical potential and @xmath37 the hopping matrix .
the orbital index @xmath38 correspond to the iron @xmath39 and @xmath40 orbitals , respectively .
the nearest - neighbor ( n.n . , @xmath41 ) and next - nearest - neighbor ( n.n.n . , @xmath42 )
exchange interactions are respectively denoted by @xmath43 and @xmath44 .
@xmath24 is the coupling for the n.n .
biquadratic interaction .
the spin operator @xmath45 and the density operator @xmath46 , with @xmath47 representing the pauli matrices .
for the hole - doped case , the constraint prohibiting the double - occupancy of the fermion is @xmath48 for the electron - doped case , we first apply a particle - hole transformation to the @xmath49-fermions , then enforce the above double - occupancy constraint . the constraint in eq .
[ eq : constraint1 ] can be treated in the standard way by a slave - boson mean - field theory@xcite .
we introduce a slave boson operator @xmath50 and a fermionic spinon operator @xmath51 in each site and for each orbital , and rewrite the electron operator as @xmath52 .
the spin operator is rewritten to @xmath53 . in the slave - boson representation , the constraint in eq .
[ eq : constraint1 ] becomes : @xmath54 in the slave - boson mean - field approach , we introduce a lagrange multiplier @xmath55 to impose the constraint on average @xcite , and perform a mean - field decomposition between the slave boson and the fermionic spinon operators .
we further assume that the slave bosons are bose condensed with @xmath56 , where @xmath57 , is the doping concentration ( with @xmath58 for hole doping and @xmath59 for electron doping ) . here
we into account the fact that in the parent compound , the @xmath29 and @xmath30 orbitals are half - filled , with @xmath60 . with these simplifications , we obtain the following effective hamiltonian for the spinons : @xmath61 \nonumber \\ & + & \sum_{\langle ij\rangle,\alpha,\beta } j_{1}^{\alpha \beta } \left(\vec{s}_{i\alpha}\cdot \vec{s}_{j\beta}-\frac{1}{4}n_{i\alpha } n_{j\beta}\right ) \nonumber \\ & + & \sum_{\langle \langle ij\rangle \rangle,\alpha,\beta } j_{2}^{\alpha \beta } \left(\vec{s}_{i\alpha}\cdot \vec{s}_{j\beta}-\frac{1}{4}n_{i\alpha } n_{j\beta}\right)\nonumber\\ & -&\sum_{\langle ij\rangle } k \left[(\sum_{\alpha}\vec{s}_{i\alpha})\cdot ( \sum_{\beta}\vec{s}_{j\beta})\right]^2,\end{aligned}\ ] ] where @xmath62 , and @xmath63 . in eq .
[ eq : ham : tj1j2k_ren ] , the constraint is taken implicitly by renormalizing the hopping matrix with the hole doping concentration . in the following ,
we work with this effective hamiltonian , and consider the superconductivity via a bcs mean - field theory .
we define the spin singlet ( @xmath64 ) and triplet ( @xmath65 , @xmath66 ) operators in the pairing channel to be @xmath67 we then rewrite the exchange interactions in eq .
[ eq : ham : tj1j2k_ren ] in terms of @xmath64 and @xmath65 . for iron pnictides
, experiments suggest that the dominant superconducting pairing is in the spin singlet channel .
hence we project the exchange interactions onto the spin singlet sector .
we thus obtain @xmath68 ^ 2 = \frac{9}{32 } \sum_{\alpha\beta } \mathcal{s}^\dagger_{i\alpha , j\beta } \mathcal{s}_{i\alpha , j\beta } \nonumber\\ & & + \frac{9}{64 } \sum_{\alpha\beta\alpha^\prime\beta^\prime } \mathcal{s}^\dagger_{i\alpha , j\beta } \mathcal{s}^\dagger_{i\alpha^\prime , j\beta^\prime } \mathcal{s}_{i\alpha , j\beta } \mathcal{s}_{i\alpha^\prime , j\beta^\prime}.\end{aligned}\ ] ] in the bcs mean - field theory , the superconductivity corresponds to the bose condensation of the singlet @xmath64 , with the pairing function @xmath69 , where @xmath70 .
we make a further simplification by considering only intra - orbital pairing , namely , @xmath71 . including the inter - orbital pairing only changes the results quantitatively .
after some algebra , we find the mean - field hamiltonian in the momentum space : @xmath72 where @xmath73 is the total number of iron sites in the system , and we denote @xmath74 , @xmath75 and @xmath76\delta_{\hat{\epsilon}\alpha } \cos(\vec{k}\cdot\hat{\epsilon}).\end{aligned}\ ] ] the constant term depends on the pairing functions @xmath77 as follows : @xmath78 ^ 2 + \frac{1}{n } \sum_{k\alpha } \xi^{\alpha\alpha}_k - \tilde{\mu } x.\end{aligned}\ ] ] the pairing functions @xmath77 can be determined by minimizing the free energy associated with the mean - field hamiltonian in eq .
[ eq : ham : tj1j2k_mf ] . in the two - orbital model
, they can be combined as follows according to the way they transform under the @xmath79 group symmetry operations : @xmath80 each of the above eight pairing channels contains an amplitude and a phase . , @xmath81 , @xmath82 , from bottom to top .
( b ) and ( c ) : fermi surface in the 1-fe brilluion zone at @xmath81 [ in ( b ) ] and @xmath82 [ in ( c ) ] . in each case , the fermi surface contains two electron pockets ( @xmath83 and @xmath84 ) and two hole pockets ( @xmath85 and @xmath86 ) .
@xmath87 is defined as the winding angle along each fermi pocket .
, width=302 ]
for the fermiology , we take the two - orbital tight - binding model in ref . .
as shown in fig .
[ fig:1 ] , within a certain range of electron doping , the fermi surface contains two hole pockets ( @xmath85 and @xmath86 ) and two electron pockets ( @xmath83 and @xmath84 ) .
but the bandstructure is highly asymmetric about half - filling ( @xmath88 ) , and for large hole doping concentration @xmath89 , the fermi surface contains only hole pockets .
this allows us to study the effects of heavy hole doping .
figure [ fig:2 ] compares the superconducting phase diagrams for the @xmath23 ( with @xmath25 ) and @xmath22 models at doping concentrations @xmath90 .
the corresponding pairing amplitudes at @xmath91 are shown in fig .
[ fig:3 ] .
here we have scaled the exchange coupling by the renormalized bandwidth of spinons , @xmath92 . in the @xmath23 model [ fig .
[ fig:2](a ) and ( c ) ] , for both the hole and electron doping , the phase diagram contains four phases with different dominant pairing symmetries . as clearly shown in fig .
[ fig:2 ] , when @xmath93 , the pairing symmetry is @xmath4 , and the leading pairing channel is @xmath94 .
the pairing function changes sign between the hole and electron pockets , and is thus denoted as @xmath5 . with increasing @xmath12 ,
the @xmath95 channel emerges and coexists with @xmath96 ( see fig .
[ fig:3 ] ) . at zero temperature ,
the phase difference of these two pairing functions is fixed to be @xmath97 , and this leads to a @xmath98 ( denoted as @xmath99 ) which breaks the time reversal symmetry . further increasing @xmath12 ,
the pairing amplitudes of the @xmath4 channels vanish , and there is a regime where the pairing symmetry is pure @xmath2 , with the dominant pairing to be @xmath1 . the @xmath4 symmetry reappears at a higher @xmath12 value , but the dominant pairing channel becomes @xmath3 .
the phases between the @xmath4 and @xmath2 are again locked , resulting in an @xmath100 pairing ( denoted as @xmath101 ) .
the phase boundaries are sensitive to the bandstructure .
we see from fig .
[ fig:2 ] that the @xmath1 phase is narrower for the electron doping than for the hole doping .
this phase is completely suppressed when the doping concentration changes from @xmath82 to @xmath102 .
( see supplementary figure 3 of ref . . )
we discuss the doping effects in more detail in section [ sec : doping ] . along the hole ( @xmath85 ) and electron ( @xmath83 ) fermi pockets at @xmath81 , @xmath103 , and @xmath91 .
( a ) : @xmath104 , in the @xmath5 phase ; ( b ) : @xmath105 , in the @xmath1 phase . , width=302 ] and imaginary [ in ( b ) ] parts of @xmath106 along the hole ( @xmath85 ) and electron ( @xmath83 ) fermi pockets in the @xmath99 phase where @xmath81 , @xmath103 , @xmath107 , and @xmath91 . ,
width=302 ] having established the phase diagram of the @xmath23 model , we now discuss the effects of the biquadratic coupling @xmath24 . by comparing the phase diagrams in fig .
[ fig:2 ] , we see that when a moderate @xmath24 is turned on , despite the pairing symmetries remaining the same as in the @xmath23 model , the phase boundaries change a lot . to understand this , note that there are two terms in eq .
[ eq : ksinglet ] .
the first term is bilinear in @xmath64 , and can be absorbed into the @xmath12 term .
the second term gives nonlinear couplings between the gap functions .
if the exchange couplings @xmath12 , @xmath10 , and @xmath24 are small compared to the renormalized bandwidth of the spinons , @xmath92 , then @xmath108 is small . as a result ,
the main effect of the biquadratic term is to renormalize the n.n .
exchange coupling @xmath12 by promoting both the @xmath95 and the @xmath109 pairing channels , as illustrated in fig .
[ fig:3 ] .
this is the main reason for the change of the phase boundaries .
the nonlinearity effect of the gap function appears at sufficiently strong exchange couplings , where several pairing channels become quasi - degenerate .
note that the degeneracy of the pairing channels at infinitely strong @xmath12 ( or @xmath10 ) couplings are associated with gauge symmetries in the exchange interactions,@xcite which are still preserved in presence of the biquadratic coupling .
hence the nonlinearity effects associated with the finite @xmath24 term will not lead to a strong change of the phase boundaries in the moderate coupling regime , however they may affect the values of pairing amplitudes in each channel .
superconductivity opens up a gap in the bcs quasiparticle spectrum . in a single - band model ,
the momentum distribution of this gap directly reflects the superconducting pairing symmetry .
in the two - orbital model , we must transform the pairing function @xmath110 into the band basis : @xmath111 , where the matrix @xmath112 diagonalizes @xmath113 .
the @xmath114 dependence of @xmath106 along the hole ( @xmath85 ) and electron ( @xmath83 ) pockets in the @xmath5 and @xmath1 phases are shown in fig .
[ fig:4 ] , respectively . in the @xmath5 phase ,
the dominant pairing is @xmath96 .
the pairing function is nodeless on both electron and hole pockets , but has different signs along these two pockets .
we see from fig .
[ fig:4](a ) that @xmath106 is isotropic along both the electron and hole pockets , and this leads to an isotropic gap . on the other hand , in the @xmath1 phase , where the dominant pairing is @xmath115 ,
the pairing function has nodes along the hole pockets at @xmath116 ( corresponds to @xmath117 ) . as a result ,
the gap is highly anisotropic along the hole pockets .
interestingly , in the two @xmath6 phases , we find that the gap can be still anisotropic , but it is nodeless . in these phases , nodes appear in either the real or imaginary part of @xmath106 ( see fig .
[ fig:5 ] ) .
but due to the different pairing symmetry , the nodes in @xmath118 and @xmath119 necessarily locate at different momenta .
the quasiparticle gap is proportional to @xmath120 ^ 2+[\mathrm{im}({\tilde{v}}_{k\gamma})]^2}$ ] , and is always nodeless .
in the previous section , we have discussed the superconducting phase diagram of the @xmath22 model at fixed hole or electron doping . we find that several phases with different pairing symmetries can be stabilized depending on the interactions . an important question with more experimental relevance is that of the pairing symmetry for realistic model parameters , and its doping dependence . in this section
we investigate the evolution of superconducting states with doping . in the @xmath22 model ,
the doping concentration @xmath121 renormalizes the electron bandwidth , @xmath62 . to make a fair comparison at different doping concentrations , for bafe@xmath15as@xmath15 , the couplings are estimated to be @xmath122 mev , and the bare bandwidth projected to the @xmath123 orbitals from dft calculations is about @xmath124 ev . from these values , we find @xmath125 at optimal ( electron / hole ) doping @xmath90 .
to study the influence of the biquadratic interaction , we compare the results at three different values of @xmath24 : @xmath126 , @xmath127 , and @xmath128 . in each case
, the dependence of the pairing amplitudes with doping for the dominant pairing channels is shown in fig .
[ fig:6 ] .
the overall pairing amplitudes are maximal at half - filling , which is understood because we only consider the superconducting pairing , but have ignored the magnetic order in the model .
when @xmath129 is decreased toward half - filling , it is expected that the superconductivity is suppressed , and the ground state is eventually antiferromagnetic , as is found experimentally .
hence our results are more pertinent to the optimally and over - doped regimes ( @xmath130 ) .
the pairing amplitudes are strongly suppressed by heavy electron / hole doping . given that the pairing amplitudes are proportional to the superconducting transition temperature @xmath131 , this result is consistent with the very low superconducting @xmath132 k for the heavily hole doped compound kfe@xmath15as@xmath15 . from fig .
[ fig:6 ] we see that the biquadratic interaction may strongly affect the pairing symmetry . for @xmath25 ,
the pairing symmetry is @xmath4 for almost the entire doping regime studied . at any doping concentration ,
the dominant pairing channel is @xmath5 .
this picture changes at a moderate biquadratic coupling @xmath127 . depending on doping ,
the pairing symmetry is either a pure @xmath4 , or a pure @xmath2 , or a @xmath133 . very interestingly , we find that the dominant pairing is : 1 .
@xmath101 for @xmath134 , 2 .
@xmath5 near optimal doping ( @xmath135 ) , and 3 .
@xmath95 in the overdoped regime ( @xmath136 ) .
such a doping evolution of the pairing symmetry and amplitudes also leads to the change of superconducting gap anisotropy with doping . from the discussion in sec .
[ sec : phd ] we see that the gap is anisotropic when the @xmath95 is dominant .
hence our results suggest a change of pairing from the isotropic @xmath137 in the optimally doped regime to the anisotropic @xmath1 in the overdoped regime . and various @xmath138 values .
see text for the choice of @xmath12 , @xmath10 , and @xmath24 values .
, width=302 ]
as mentioned in section [ sec : model ] , besides the degenerate @xmath29 and @xmath30 orbitals , the other three fe @xmath139 orbitals also contribute to the low - energy bandstructure .
therefore generally speaking , a five - orbital model is more appropriate when describing the electronic properties of these materials . for superconducting pairing ,
several calculations based on five - orbital models have been done.@xcite to see how accurate is the superconducting phase diagram we obtained in sec .
[ sec : phd ] for the two - orbital model , we compare our results with the strong - coupling phase diagram of a five - orbital @xmath23 model in ref . .
we see that the pairing phase diagrams of the two- and five - orbital models are quite similar .
this suggests that the simplified two - orbital model already captures the correct pairing symmetry and the dominant pairing channel in each phase , as alluded to earlier in section [ sec : model ] .
it is a natural result of the strong - coupling theory : the dominant pairing symmetry is determined by @xmath140 and @xmath141 , and the details of the fermi surface structure are secondary .
nonetheless , the orbital character along the fermi surface is important to the anisotropy of the superconducting gap , especially when the superconducting pairing has a strong orbital selectivity.@xcite in this case , the inclusion of other fe @xmath139 orbitals ( such as the @xmath31 orbital ) may be crucial and will be the subject of future work . in sec . [
sec : doping ] we have shown that the dominant pairing symmetries at different dopings are sensitive to the biquadratic coupling @xmath24 . without the biquadratic coupling ,
the dominant pairing is @xmath5 at any doping . with a moderate @xmath24 , the @xmath5 pairing is dominant only near optimal doping . for overdoped and underdoped regimes ,
the dominant pairing channels are respectively @xmath1 and @xmath142 .
the change of pairing symmetry with doping is a consequence of the band renormalization . with increasing @xmath129 , the renormalized bandwidth @xmath92 increases , and @xmath140 and @xmath141 reduces .
the variation of @xmath140 and @xmath141 accounts for both the change of the pairing symmetry and the dominant pairing amplitude , as we have already seen in figs . [ fig:2 ] and [ fig:3 ] .
note that our theory on the doping evolution of superconducting pairing is very different from the weak - coupling theories , in which the doping evolution reflects the interplay between the intra- and inter - pocket electron interactions.@xcite
in conclusion , we have studied the effect of the biquadratic spin - spin interaction @xmath11 on the superconducting states of the iron pnictides via a two - orbital @xmath22 model within the slave - boson mean - field theory . unlike the @xmath23 model , where the dominant pairing channel at any doping concentration is always @xmath143 for realistic @xmath144 ratios , we find that a moderate biquadratic interaction @xmath24 favors both the @xmath95 and @xmath109 pairing channels in the @xmath22 model , resulting in a serious change of pairing symmetry with doping .
though the dominant pairing is still the isotropic @xmath143 ( referred to as @xmath96 ) near the optimal doping , it changes to @xmath95 in the overdoped regime and to @xmath100 in the underdoped regime , and in both cases the gap can be anisotropic , as figs .
[ fig:4 ] and [ fig:5 ] illustrate .
our results are in qualitative agreement with the gap anisotropy observed in recent arpes measurements for several iron - based superconductors , including nafeas , lifeas @xcite and the heavily hole - doped ba@xmath145k@xmath146fe@xmath15as@xmath15 .
@xcite theoretically , the interplay between @xmath17-wave and @xmath16-wave pairing channels has been previosuly studied , however to stabilize the @xmath16-wave state , one had to rely on varying the applied pressure @xcite , the pnictogen height @xcite , the @xmath147 orbital hybridization @xcite , or the strength of nel fluctuations@xcite . in this study , we show that the @xmath16-wave state is naturally stabilized in the over - doped regime , without any additional fine - tuning and for realistic values of the biquadratic interaction @xmath24 inferred @xcite from fitting the inelastic neutron spectra to the @xmath22 model .
one of the outstanding questions is the effect of electron nematicity on the pairing symmetry and resulting anisotropy of the superconducting order parameter .
recent theoretical work indicates that non - time - reversal symmetry breaking @xmath148 pairing can be realized due to coupling to the nematic order parameter @xcite .
study of this effect in the framework of the strong - coupling @xmath22 model will be the subject of future work , potentially relevant to superconductivity in the thin fese films @xcite and in the stoichiometric single - crystalline fese @xcite . | we study the symmetry and strength of the superconducting pairing in a two - orbital model for iron pnictides using the salve boson mean - field theory .
we show that the nearest - neighbor biquadratic interaction @xmath0 inflences the superconducting pairing phase diagram by promoting the @xmath1 @xmath2 and the @xmath3 @xmath4 channels .
the resulting phase diagram consists of several competing pairing channels , including an isotropic @xmath5 @xmath4 channel , an anisotropic @xmath1 @xmath2 channel , and two @xmath6 pairing channels .
we have investigated the evolution of superconducting states with electron doping , and find that with the biquadratic interaction various pairing channels may dominate at different doping concentrations .
we show that this is crucial in understanding the doping evolution of superconducting gap anisotropy observed in some angle resolved photoemission spectrum measurements . |
sexual reproduction is the observed mode of reproduction for nearly all multicellular organisms . as such
, the evolution of sex has been one of the central outstanding questions in evolutionary biology .
one of the biological explanations for the existence of sex is that it provides a natural mechanism for diploid organisms to eliminate deleterious mutations from a population @xcite .
the idea is that , by reproducing via a haploid intermediate , it is possible for haploids without defective genes to recombine with one another , thereby preventing the accumulation of deleterious mutations .
other explanations that have been advanced are that sex leads to greater variability in a population , making the population more adaptable in adverse conditions .
it has also been postulated that sex evolved as a mechanism for coping with parasites @xcite . in recent years
, there have been a number of numerical studies focusing on the evolutionary dynamics of sexual replication @xcite .
these studies have established that , depending on the choice of parameters , either sexual or asexual modes of reproduction are the advantageous replication strategy .
one study in particular argues that sexual reproduction is favored when the number of daughter genomes produced by the parents is high , since this reduces the amount of time required to find a reproductive partner @xcite . in this paper
, we present a relatively simple evolutionary dynamics model that allows us to compare sexual and asexual replication strategies .
the model is analytically solvable , and treatable within the quasispecies formalism @xcite .
the essential result is that sexual replication is favored in populations with low replication rates , and when the characteristic time associated with finding a reproductive partner is small compared with the time scale associated with replication .
these results suggest that increasing population density favors the sexual replication strategy , since it reduces the time scale associated with finding a mate .
this paper is organized as follows : in the following section ( section ii ) , we develop a simplified model for sexual replication , whose steady - state behavior we proceed to characterize in section iii . in section
iv we compare sexual and asexual replication , and establish regimes where each is the preferred mode of reproduction .
we conclude the paper in section v with a brief discussion and an outline of avenues for future research .
in a simplified model for sexual replication , we assume that we have a population of single - celled organisms , where each organism has a genome consisting of two chromosomes .
we assume that each chromosome may be denoted by a linear symbol sequence @xmath2 , where each letter , or base , @xmath3 , is chosen from an alphabet of size @xmath4 ( @xmath5 for terrestrial life ) .
we further assume that there exists a `` master '' sequence @xmath6 for which a given chromosome is functional .
it is assumed that a chromosome is nonfunctional whenever @xmath7 ( that is , the genes on such a chromosome are defective ) . within this approximation , there are three distinct types of genomes in the population : ( 1 ) @xmath8 genomes where both chromosomes are identical to the `` master '' sequence .
( 2 ) @xmath9 genomes where only one of the chromosomes is identical to the `` master '' sequence , while the other chromosome is defective . ( 3 ) @xmath10 genomes where both of the chromosomes are defective .
we are therefore dealing with a diploid population .
if we assign the gene sequence @xmath6 as _ viable _ , while all other gene sequences are _ unviable _ , then our three genome types may be classified as @xmath11 , @xmath12 , and @xmath13 , where @xmath14 stand for viable / unviable .
we assume that the organisms replicate with a first - order growth rate constant .
for the three distinct genome types , the first - order growth rate constants are taken to be @xmath15 , @xmath16 , and @xmath17 .
we have that @xmath18 .
the sexual replication of the population occurs as follows : the diploid organisms divide to form a population of haploid organisms .
it is assumed that those haploid organisms containing a genome of type u are incapable of participating further in the reproductive process , so that only viable haploids can recombine with each other .
the newly formed diploids then divide via the normal mitotic pathways to form two new daughter cells . to develop a set of ordinary differential equations governing the replication dynamics described above , we first let @xmath19 denote the total number of organisms with genome of type @xmath11 , @xmath20 denote the total number of organisms with genome of type @xmath12 , and @xmath21 denote the total number of organisms with genome of type @xmath13 .
we also let @xmath22 denote the population of viable haploids .
we then wish to obtain expressions for @xmath23 , @xmath24 , @xmath25 , and @xmath26 .
first note that , the diploid to haploid division leads to destruction of each of the diploid genomes at a rate given by @xmath27 for @xmath11 , and similarly for the other genomes , and a creation of viable haploid genomes at a rate given by @xmath28 .
if we let @xmath0 denote the average amount of time a viable haploid spends searching for a viable haploid mate , then in a given amount of time @xmath29 the total number of viable haploids who have recombined is given by @xmath30 ( the individual times are poisson distributed ) .
therefore , recombination leads to a destruction rate of haploids given by @xmath31 , and a creation rate of diploids given by @xmath32 .
if we let @xmath33 denote the probability of correctly replicating a chromosome , then neglecting backmutations we have that @xmath34 with probability @xmath33 , @xmath35 with probability @xmath36 , and @xmath37 with probability @xmath38
. using this , we can construct the various possible replication pathways and their associated probabilities , illustrated in figure 2 . from these pathways
, we can construct the contribution to @xmath19 , @xmath20 , and @xmath21 in turn . for @xmath19 , all three replication pathways in figure 1 give a contribution .
taking into account probabilities and degeneracies , we have a total contribution of @xmath39 from the first pathway , @xmath40 from the second pathway , and @xmath41 from the second pathway , giving a total rate of production of @xmath42 = ( 1/4 ) n_v/\tau_{seek } ( 1 + p)^2 $ ] . for @xmath20 , similar reasoning gives a rate of production of @xmath43 = ( 1/2 ) n_v\tau_{seek }
( 1 - p^2 ) $ ] . for @xmath21
, we obtain @xmath44 . putting everything together
, we obtain the system of differential equations , @xmath45 we can non - dimensionalize these equations by defining @xmath46 , and @xmath47 , @xmath48 , @xmath49 .
then we obtain , @xmath50 define @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 .
note that @xmath60 , @xmath61 , @xmath62 , and @xmath63 .
then defining @xmath64 , @xmath65 , we have , @xmath66 re - expressing the dynamical equations in terms of the @xmath67 , @xmath68 , @xmath69 , @xmath70 population fractions , we have , @xmath71
at steady - state , the above time derivatives may all be set to @xmath72 , giving , @xmath73 therefore , at steady - state , we have that , @xmath74 + \frac{x_v}{1 - x_v } \nonumber \\ & = & \frac{1}{1 - x_v}[x_v + \bar{\kappa}(\tau = \infty ) - x_v ( 1 + \bar{\kappa}(\tau = \infty ) ) ] \nonumber \\ & = & \bar{\kappa}(\tau = \infty)\end{aligned}\ ] ] and so it is equivalent to measure the mean fitness of the population at steady - state using either @xmath75 and @xmath76 .
this is important , because , in comparing mean fitness results with an asexually replicating population , the natural mean fitness result to use is @xmath76 .
the equivalence between @xmath75 and @xmath76 means that we can compute @xmath75 at steady - state and compare the results directly with the value of @xmath75 for the asexually replicating population . plugging the steady - state value of @xmath70 into the steady - state equations for @xmath67 and @xmath68 , we obtain , @xmath77 x_{vv } \nonumber \\ & & + \frac{1}{4 } ( 1 + p)^2 \frac{\tilde{\kappa}_{vu } x_{vu}}{1 + \bar{\kappa}(\tau = \infty ) } \nonumber \\ & & 0 = [ \tilde{\kappa}_{vu } ( \frac{1}{2 } ( 1 - p^2 ) \frac{1}{1 + \bar{\kappa}(\tau = \infty ) } - 1 ) - \bar{\kappa}(\tau = \infty ) ] x_{vu } \nonumber \\ & & + ( 1 - p^2 ) \frac{1}{1 + \bar{\kappa}(\tau ) } \tilde{\kappa}_{vv } x_{vv}\end{aligned}\ ] ] we can solve the first equation for @xmath68 in terms of @xmath67 . plugging the resulting expression into the second equation ,
we obtain , after some algebra , the quadratic , @xmath78 \bar{\kappa}(\tau = \infty ) \nonumber \\ & & - \tilde{\kappa}_{vv } ' \tilde{\kappa}_{vu } ' p\end{aligned}\ ] ] where @xmath79 , @xmath80 .
we can further simplify the notation by defining @xmath81 , and @xmath82 .
then @xmath83 is the solution to the quadratic , @xmath84 where , @xmath85 - \frac{1}{2 } \frac{\alpha}{1 + \alpha \kappa } ( 1 + p^2 ) \nonumber \\ & & b(p , \kappa , \alpha ) = \frac{1}{1 + \kappa } \frac{\alpha}{1 + \alpha \kappa } p\end{aligned}\ ] ] differentiating both sides of the quadratic , it is possible to show , after some manipulation , that @xmath86 , and hence that the mean fitness is an increasing function of @xmath33 .
if , for simplicity , we assume that @xmath87 , then for asexual replication the steady - state value for @xmath83 may be readily characterized @xcite : it is given by @xmath88 . therefore , if @xmath89 is defined by the equality @xmath90 , then , @xmath91 $ }
\\ \alpha p & \mbox{if $ p \in ( 0 , p_{crit } ] $ } \end{array } \right.\end{aligned}\ ] ] we can compare these values with that obtained for sexual replication .
we begin by considering the case where there is no time cost associated with sex , so that @xmath93 .
then @xmath94 - \frac{1}{2 } \alpha ( 1 + p^2 ) \nonumber \\ & & b(p , 0 , \alpha ) = \alpha p\end{aligned}\ ] ] we claim that for @xmath92 , @xmath95 , with equality occurring only when @xmath96 for arbitrary @xmath97 , and @xmath98 . to prove this claim , note first that for @xmath96 , we have @xmath99 , and that @xmath100 .
so since the claim is true for @xmath96 , we may now consider @xmath101 . if @xmath102 , then @xmath103 , while if @xmath104 , then @xmath105 , since @xmath106 for @xmath107 $ ] , with equality occuring only when @xmath96 .
so , we now consider the case where @xmath108 , and @xmath101 .
if we define @xmath109 , then we have two possibilities : either @xmath110 , or @xmath111 .
we will consider each of these two cases in turn .
so , first assume that @xmath112 .
we wish to show that , @xmath113 > k(p)\ ] ] after some manipulation , we obtain that this condition is equivalent to the condition that @xmath114 to establish this inequality for @xmath101 , note that @xmath115 .
since @xmath116 for @xmath101 , and since @xmath117 for @xmath101 , the inequality follows .
so now suppose that @xmath118 , so that @xmath119 . then for our calculations , we first re - write @xmath120 as @xmath121 .
then we wish to show that , @xmath122 \nonumber \\ & & > \alpha p\end{aligned}\ ] ] after some manipulation , this becomes equivalent to the condition that , @xmath123 which is certainly true , since @xmath124 for @xmath101 , and @xmath108 by assumption .
therefore , we have proven our claim , and hence , within this model , sexual replication leads to a greater mean fitness for a population than asexual replication , assuming that there is no cost associated with sex .
figure 2 shows a plot of @xmath83 for both sexual and asexual replication , assuming @xmath125 and @xmath92 .
note that for @xmath92 , if a sexually and asexually replicating population were placed in an identical flask , then under the circumstances dictated by our model the sexually replicating population would eventually dominate the population ( that is , the fraction of sexually replicating organisms would increase to @xmath38 , while the fraction of asexually replicating organisms would decrease to @xmath72 ) . for @xmath126 ,
we have , for sexual replication , that @xmath127 for @xmath96 . since for asexual replication
we get @xmath128 for @xmath96 , it follows by continuity that there exists a regime @xmath129 $ ] for which asexual reproduction leads to a greater mean fitness of the population than sexual reproduction .
presumably , as @xmath130 increases , @xmath131 should decrease .
we can determine , for a given @xmath130 , the mutation regime where asexual replication outcompetes sexual replication , and the mutation regime where sexual replication outcompetes asexual replication . to do this ,
we do not attempt to compute @xmath131 directly . rather , as a function of @xmath33 , we seek to determine @xmath132 , the value of @xmath130 for which asexual and sexual replication yield identical mean fitnesses . since by definition @xmath133
, the function @xmath132 may be inverted to obtain @xmath131 . in what follows
, we restrict our analysis to the case where @xmath108 (
if @xmath98 , then for @xmath126 asexual reproduction outcompetes sexual reproduction for all values of @xmath33 ) . because the mean fitness of an asexually replicating population falls into two distinct regimes defined by the cutoff @xmath89 ( at least , within the context of our model )
, we must determine @xmath132 separately for @xmath134 and @xmath135 .
for @xmath134 , we have for asexual replication that @xmath136 , and hence we must have , @xmath137 assuming that @xmath138 , this expression may be re - arranged and simplified to , @xmath139 so that @xmath140 , and hence @xmath141 , may be solved using the quadratic formula .
we claim that @xmath141 is an increasing function of @xmath33 on the interval @xmath142 $ ] .
we can prove this by showing that @xmath140 is an increasing function of @xmath33 on the interval @xmath142 $ ] .
defining @xmath143 , we have , @xmath144 differentiating with respect to @xmath33 , we obtain , @xmath145 so we wish to show that @xmath146 for @xmath147 . multiplying both sides of the inequality by @xmath33 , and noting that @xmath148 , we have that for @xmath149 we need to establish the inequality , @xmath150 to prove this inequality , note first that @xmath151 , and @xmath152 . by continuity , @xmath153 in a neighborhood of @xmath154 .
if @xmath155 for some @xmath149 , then by the intermediate value theorem there exists at least one @xmath149 for which @xmath156 . if @xmath157| x'(p ) = 0\ } $ ] , then by continuity it follows that @xmath158 , and hence @xmath159
therefore , @xmath153 for @xmath160 , otherwise by the intermediate value theorem there would exist a @xmath161 such that @xmath162 , contradicting the definition of @xmath163 .
but , since @xmath153 for @xmath160 , it follows that @xmath164 is increasing on @xmath165 $ ] , hence @xmath166 , which implies that @xmath167 .
therefore , @xmath153 for @xmath107 $ ] , hence on @xmath142 $ ] , @xmath132 increases from @xmath168 to @xmath169 .
now , for @xmath170 $ ] , we have for an asexually replicating population that @xmath171 , hence , in this regime , @xmath132 is defined by , @xmath172 which after some manipulation may be re - arranged to give , @xmath173 where @xmath174 , @xmath175 .
we then have that , @xmath176 we claim that @xmath132 is a decreasing function of @xmath33 for @xmath170 $ ] .
we will prove this by showing that @xmath177 and @xmath178 are both decreasing functions of @xmath33 for @xmath170 $ ] . to prove that @xmath177 is a decreasing function of @xmath33 for @xmath170 $ ] , we need
show that @xmath179 is a decreasing function of @xmath33 for @xmath170 $ ] .
differentiating , we obtain , @xmath180 so @xmath177 is certainly a decreasing function of @xmath33 for @xmath170 $ ] . now , after some manipulation , we can show that , @xmath181 note that @xmath182 is decreasing for @xmath170 $ ] , and that @xmath183 is increasing .
we also have that , @xmath184 so that @xmath185 is a decreasing function of @xmath33
. therefore , @xmath186 is an increasing function of @xmath33 , hence @xmath178 is decreasing for @xmath170 $ ] , as we wished to show .
figure 3 illustrates the behavior of @xmath132 for three values of @xmath97 . in what follows , we shall change our notation slightly to explicitly indicate that @xmath141 also depends on @xmath97 .
thus , we shall re - denote @xmath132 by @xmath187 .
this notation was not needed in the previous arguments , since we were considering the behavior of @xmath141 at a fixed @xmath97 .
we may now summarize the behavior of @xmath187 as a function of @xmath33 : from @xmath72 to @xmath89 , @xmath187 increases from @xmath168 to @xmath188 , while from @xmath89 to @xmath38 , @xmath187 decreases from @xmath189 to @xmath72 .
this behavior leads to three distinct regimes of @xmath130 . for @xmath190 $ ]
, there exists only one value of @xmath33 for which asexual and sexual replication yield identical mean fitness results .
this value of @xmath33 is contained in the interval @xmath191 $ ] .
as @xmath130 increases from @xmath72 to @xmath168 , this value of @xmath33 decreases . for these values of @xmath130 ,
asexual replication is advantageous over sexual replication at low mutation rates .
however , there is a crossover replication fidelity where sexual replication becomes advantageous .
as @xmath130 increases , this crossover replication fidelity gets pushed to lower values .
this makes sense , since a higher value of @xmath130 corresponds to a greater penalty associated with sex . for @xmath192 $ ]
, there exist exactly two values of @xmath33 for which asexual and sexual replication yield identical mean fitness results .
one value of @xmath33 is contained in the interval @xmath142 $ ] , while the other value is contained in the interval @xmath191 $ ] .
as @xmath130 increases from @xmath168 to @xmath189 , the value of @xmath33 in @xmath142 $ ] increases from @xmath72 to @xmath193 , while the value of @xmath33 in @xmath191 $ ] decreases to @xmath193 . for these values of @xmath130 ,
asexual replication is also advantageous over sexual replication at low mutation rates . as with the previous regime
, there is a crossover replication fidelity where sexual replication becomes advantageous .
however , in contrast to the first @xmath130 regime , there is a second crossover replication fidelity where asexual replication again becomes advantageous .
for these values of @xmath130 , the cost associated with sex is still sufficiently low that sexual replication can become the advantageous strategy at higher mutation rates .
however , the cost of sex is sufficiently high that , at even higher mutation rates , sexual recombination no longer offsets the production of unviable chromosomes from viable ones to an extent that makes the strategy advantageous .
finally , for @xmath194 , the cost associated with sex is so high that sexual replication is never the advantageous strategy . as a final note for this subsection
, we can show that @xmath189 is a decreasing function of @xmath97 .
differentiating the quadratic equation given by eq .
( 19 ) with respect to @xmath97 ( where @xmath195 ) , we have , @xmath196 differentiating both sides of the equality defining @xmath89 , it is possible to show that @xmath197 .
we also have , from a previous analysis , that @xmath198 , so that , to show @xmath199 , we need to show that @xmath200 multiplying by @xmath201 and simplifying , this is equivalent to the inequality , @xmath202\ ] ] since we showed that the expression for @xmath164 in eq .
( 19 ) , valid over @xmath203 $ ] , is an increasing function of @xmath33 for @xmath107 $ ] , then solving the quadratic in eq .
( 19 ) for @xmath96 gives @xmath204 .
therefore , @xmath205 , hence the inequality holds .
we have therefore shown that @xmath206 is a decreasing function of @xmath97 , hence @xmath189 is a decreasing function of @xmath97 . when @xmath104 , @xmath207 , so @xmath208 .
as @xmath209 , @xmath210 , so @xmath211 \rightarrow \kappa_{= } \rightarrow \infty $ ] .
when @xmath212 , the results for sexual replication remain unchanged .
however , the results for asexual replication change somewhat , since now an additional localization to delocalization transition can occur , once @xmath213 .
it is therefore possible to have a situation where sexual replication only becomes advantageous once the mean fitness of an asexually replicating population is @xmath17 . with an appropriate choice of parameters
, it is then possible that the mean fitness of the sexually replicating population is less than @xmath17 , so that sexual replication will never be the preferred replication strategy .
we leave the investigation of this phenomenon for future work . in any event , for @xmath212 , asexual replication will become the advantageous mode of replication at sufficiently high mutation rates , since the mean fitness of the sexually replicating population decreases to zero , while after complete delocalization over the genome space has occurred , the mean fitness of the asexually replicating population becomes @xmath17 .
this result , however , is likely due to a mating strategy that essentially `` throws away '' the `` u '' chromosomes .
other mating strategies , where all haploids are capable of mating , may not exhibit the same effect .
this paper developed a simplified model for sexual replication , and showed , within the context of the model , that a sexually replicating population will outcompete an asexually replicating one when there there is no cost associated with sex .
we further showed that if the cost associated with sex is sufficiently low , then sexual replication becomes advantageous at higher mutation rates , because recombination prevents the accumulation of defective mutations in the diploid genomes ( assuming that there is a fitness penalty associated with the defective mutations ) .
the cost for sex was measured by the dimensionless parameter @xmath130 , defined to be the product of the first - order growth rate constant of the mutation - free genomes ( @xmath15 ) , and the characteristic time associated with finding a recombination partner ( @xmath0 ) . since @xmath214 , where @xmath215 denotes a characteristic replication time
, it follows that @xmath216 .
therefore , the cost associated with sex is measured by the ratio of the time a haploid spends finding a recombination partner with the time scale for replication
. the smaller this ratio , the smaller the fitness penalty incurred by reproducing via a haploid intermediate , and the greater the selective advantage for sex .
the implications of this model are that sexual replication is favored in environments where organisms replicate relatively slowly , and in environments where the time spent finding a recombination partner is small compared with the time scale for replication .
thus , sexual replication is favored in environments with high population density .
these results are therefore consistent with the observation that sexual replication is the preferred ( and generally the only ) mode of reproduction for nearly all multicellular organisms . that sexual replication only becomes the preferred mode of reproduction at low @xmath130 suggests why sexual replication occurs as a stress response in some organisms , such as _ saccharomyces cerevisiae _ ( baker s yeast ) . when conditions are favorable , @xmath15 , and hence @xmath130 , are relatively high , so asexual replication is the advantageous strategy . under sufficiently adverse conditions
, @xmath130 can drop to levels where the sexual strategy becomes advantageous .
the replicative strategy that can adopt the optimal replication strategy for the given environment will have a selective advantage ( assuming that resource costs for maintaining this switching behavior are not prohibitive ) , and so organisms carrying this strategy in their genomes will dominate the population . however , as one moves toward more complex life forms , the replication rate drops to values such that asexual replication is almost never the preferred mode of reproduction , so that the ability for an organism to switch between the two modes of reproduction disappears . at this point
, we postulate that the division of haploid cells into two distinct types of gametes , and then later the division of the organisms themselves into male and female , are the result of selection for evolutionary pathways leading to the division of labor and specialization of tasks associated with sexual replication . when replication rates are low , and when the time cost associated with sex is low , then it is likely more efficient ( in terms of resource utilization ) to divide the reproductive tasks associated with sexual replication among two types of organisms ( `` male '' and `` female '' )
. the relative fitness advantage as a result of such savings in resource costs likely increases with the complexity of the organism , leading to a stronger selection pressure for a male - female split as organismal complexity grows . in this paper , we assumed that only haploid cells with viable chromosomes are capable of engaging in sexual recombination .
this allowed a simplified analysis within the standard quasispecies formalism .
while we obtained a selective advantage for sexual replication using this mating strategy , a fuller analysis will require the consideration of various mating strategies on the selective advantage for sex .
an important such mating strategy , which is the opposite of the one considered in this paper , is the random mating strategy , whereby all haploids are capable of engaging in sexual recombination , and do so with a pairwise distribution given by the hardy - weinberg equilibrium . in this vein
, one interesting question is to determine , for a given fitness landscape , whether there always exists a mating strategy for which sexual replication will outcompete asexual replication .
additionally , while this paper implicitly assumed that the strategy for sexual or asexual replication is inherited , future studies should consider genomes where genes for sex are explicitly included .
this leads to the ability for sexual organisms to mutate into asexual ones .
as the selective advantage for sexual replication disappears ( as a function of mutation rate , for instance ) , the models may exhibit localization to delocalization transitions over the portions of the genome controlling sex . | this paper develops a simplified model for sexual replication within the quasispecies formalism .
we assume that the genomes of the replicating organisms are two - chromosomed and diploid , and that the fitness is determined by the number of chromosomes that are identical to a given master sequence .
we also assume that there is a cost to sexual replication , given by a characteristic time @xmath0 during which haploid cells seek out a mate with which to recombine . if the mating strategy is such that only viable haploids can mate , then when @xmath1 , it is possible to show that sexual replication will always outcompete asexual replication . however , as @xmath0 increases , sexual replication only becomes advantageous at progressively higher mutation rates .
once the time cost for sex reaches a critical threshold , the selective advantage for sexual replication disappears entirely .
the results of this paper suggest that sexual replication is not advantageous in small populations per se , but rather in populations with low replication rates . in this regime ,
the cost for sex is sufficiently low that the selective advantage obtained through recombination leads to the dominance of the strategy .
in fact , at a given replication rate and for a fixed environment volume , sexual replication is selected for in high populations because of the reduced time spent finding a reproductive partner . |
distributions of transverse momenta of charged leptons ( @xmath5 ) produced in drell yan ( dy ) processes are important observables in hadron collider experiments .
their sensitivity to the values of the standard model ( sm ) parameters and to the polarisation of @xmath0 and @xmath6 bosons produced in the dy process can be used in precision tests of the standard model ( sm ) and in searches for new physics .
the measurement of the sm parameters : @xmath0-boson mass ( @xmath7 ) , its charge asymmetry ( @xmath8 ) and width ( @xmath9 ) at the lhc was investigated in refs .
it was shown that in order to reach a precision at the level of @xmath10mev , the @xmath5 distributions must be controlled experimentally to a comparable accuracy .
since the monte carlo ( mc ) event generators are indispensable tools to derive the values of the sm parameters from the measured distributions , they must predict these distributions to even higher precision .
mc event generators for drell yan processes developed so far can be divided into two categories .
the first one includes generators which are based on the precise calculations of matrix elements ( me ) for hard processes : those including qcd effects to the nlo or even nnlo level , and those including qed and electroweak ( ew ) corrections . in these generators ,
called hereafter the me generators , differential cross sections for the hard process are convoluted with the universal parton distribution functions ( pdfs ) of the hadron beams .
these pdfs depend , apart from the factorisation scale , only on the longitudinal momenta of partons , @xmath11 . in this type of the mc generators partons entering the hard process
are assumed to be parallel to hadron beam direction .
the second category of mc generators includes the so - called parton - shower ( ps ) generators , such as @xcite , @xcite , etc .
they generate initial - state multi - parton radiation in form of the lo - type qcd / qed parton cascade and then perform hadronisation as well as some particle decays . in the ps generators , partons entering the hard process
are no longer collinear with the hadron - beam directions , but acquire non - zero transverse momenta .
the hard process itself is described by these generators usually at the lo level .
thus , as long as the me precision is the key factor determining the overall accuracy of the measurement of a selected observable , they are inferior with respect to the me generators . however , in the remaining cases they often provide better description of the hadronic energy flow associated with the dy process , in particular for not to high transverse momenta of the @xmath0 and @xmath6 bosons .
a very important , and at the same time difficult issue is how to combine these two types of generators , avoiding , on the one hand , a double counting of qcd corrections and , on the other , possible gaps in phase space present in some ps algorithms . as the state - of - the - art practical solutions to this problem for the qcd effects are regarded the @nlo @xcite and @xcite generators .
they match the nlo qcd me calculations with the parton shower generators , albeit in different ways .
@nlouses or @xcite for parton shower generation , while is more universal , in principle it can use an arbitrary parton - shower generator . in the case of me generators that include qed / ew radiative corrections
the situation is simpler as combining them with ps generators usually does not lead to double counting .
one of such generators is the mc program @xcite .
it includes higher - order qed effects for the final - state radiation ( fsr ) and initial - final state interferences in the yennie
suura ( yfs ) exclusive exponentiation scheme @xcite together with the @xmath12 ew corrections for the full charged - current dy process @xcite . for the qcd ( and qed isr )
effects it is interfaced with 6.4 @xcite .
however , we do not use the les houches accord ( lha ) scheme @xcite but our own interface , in which the routines for parton - shower generation and hadronisation are called directly from the program .
the principal reason is that the lha is not general enough to transmit the full information contained in the spin density matrix of the @xmath0 and @xmath6 bosons between these two generators .
moreover , from the purely technical perspective , we avoid writing / reading events into / from disk files , which makes event generation easier and much more efficient .
for example , in our studies presented in refs.@xcite and requiring generation of @xmath13 events , its efficiency was one of the principal optimisation targets . in any interface which extends the lha scheme to processes involving the spin-1 ew bosons as intermediate particles a particular care
must be taken of the proper matching of the me - type kinematics with the ps - type kinematics . in the following
we show that this is particularly important to describe the charged lepton transverse momenta distributions in the drell yan processes at the lhc . in particular , we find that using the original effective momenta of incoming quarks in the above matching results in strongly biased @xmath5 distributions , which is particularly visible in their charge asymmetries .
we propose solutions to this problem that seem to reproduce well the nlo qcd predictions for these asymmetries , as obtained e.g. from @nlo .
the goal of this exercise is to try to achieve the nlo qcd precision for the description of the leptonic observables in the dy process with the suitably matched lo qcd generator which incorporates the state - of - the - art ew corrections .
the paper is organised as follows . in the next section
we describe in detail the matching of the me kinematics with ps kinematics as it is realised in the mc event generator . in section 3
we present numerical results illustrating the above issues , discuss their meaning for the lhc physics and propose our solutions that match the nlo qcd predictions .
finally , in section 4 we summarise the paper .
the @xcite mc event generator is dedicated to precision modelling of single @xmath0-boson production with leptonic decays , i.e. the charged - current drell yan processes , in proton proton , proton antiproton and ion ion collisions , with the main emphasis on the qed effects and electroweak corrections .
it uses fully massive spin amplitudes to evaluate the hard process matrix elements .
they can be computed in an arbitrary reference frame , in particular they can be used to calculate polarised @xmath0-boson cross sections - boson processes in several reference frames . ] . in terms of the perturbative qcd
the current version includes the lo hard - process matrix element .
the qcd effects enter only through scaling - violating pdfs , taken from the lhapdf library @xcite .
therefore , in incoming quarks producing a @xmath0-boson are collinear with hadron ( ion ) beams ; their longitudinal momenta are given by the two @xmath11-variables which are generated according two pdfs and subsequently convoluted with the hard - process differential cross section . at this stage
is a me - type mc generator .
its full event kinematics ( i.e. all four - momenta of initial , intermediate and final state particles ) is constructed for incoming partons collinear with the colliding beams .
let us call it the _ me kinematics_. at this level has been cross - checked numerically to a high precision with independent calculations @xcite .
events of the _ me kinematics _ do not look very realistic from the experimental point of view for the following two reasons .
firstly , qcd radiation affect not only longitudinal momenta of partons but also their transverse momenta .
therefore , using purely collinear pdfs for the description of the qcd effects is not sufficient .
secondly , partons are not observed experimentally . what can be observed
are the products of their hadronisation and decays .
therefore , in a realistic mc generator , to be used in an experimental data analysis , the above effects must be taken into account . in this
is done through the interface to the 6.4 generator which performs the initial - state lo - type qcd ( and qed ) parton shower , appropriate proton - remnant treatment , and necessary hadronisation / decays . in , partons entering the hard process are not collinear with the hadron beams . in the case of the charged - current dy processes ,
provides , in its event record , the momenta of the two effective on - mass - shell quarks producing @xmath0-bosons .
a vectorial sum of these momenta gives a momentum of an appropriate @xmath0-boson .
such a @xmath0-boson , in contrast to the case of a me - type mc generator , carries the transverse momentum , being a vectorial sum of the quarks transverse momenta .
this has to be taken into account in constructing the hard - process event kinematics .
let us call the kinematics in which the incoming partons are not parallel to the beams ( as a result of the aforementioned effects ) the _ ps kinematics _ , and a corresponding mc generator the ps generator .
while interfacing the me - type generator with the ps generator one has to take care of the appropriate matching of the _ me kinematics _ with the _
ps kinematics_. this is particularly important for processes in which particles with non - zero spin , e.g. @xmath14-bosons , are produced as intermediate states . in the following , we describe in detail how such a matching is performed in the interface to 6.4 .
then we discuss possible pitfalls of the kinematical matching of these two types of mc generators . in the interface to
the final hard - process event kinematics is constructed through the following steps : 1 .
the _ me kinematics _ of a given mc event is generated in
the four - momenta of : the incoming quarks , the intermediate @xmath0-boson , the final - state leptons , and the radiative photons are constructed in the frame in which quarks are collinear with the hadron beams , where the @xmath15 axis is the direction of one of the beams at the collision point .
all the above four - momenta are then lorentz - boosted along the @xmath0-boson direction to the @xmath0-boson rest frame , which is also the centre - of - mass frame of the incoming quarks . in this frame quarks
are still aligned along the @xmath16 axis .
the ps kinematics is generated by , in which the effective on - mass - shell quarks producing the @xmath0-boson are non - collinear with the beams as a results of the initial - state qcd / qed parton shower .
their four - momenta are given in the lab frame with the @xmath15 axis along one of the hadron beams .
4 . the above -quarks four - momenta
are lorentz - boosted along the sum of their momenta to their centre - of - mass frame , which is also the @xmath0-boson rest frame and along @xmath17 .
we have checked numerically that both methods are fully equivalent . ] .
contrary to the quarks in point 2 , their momenta , although back - to - back , are not aligned with the @xmath16 axis in this frame . instead , their direction is rotated with respect to the @xmath16 axis by the polar angle @xmath18 , and with respect to the @xmath11 axis by the azimuthal angle @xmath19 .
5 . from the -quarks momenta , specified in the above frame
, we calculate the angles @xmath20 and then we perform rotations of all the momenta specified in the @xmath0-rest frame ( point 2 ) using these rotation angles . after such rotations
the quarks are aligned with the quarks .
6 . finally , the whole event is lorentz - boosted from the above frame to the lab frame along the sum of the original -quarks momenta ( the boost is opposite to the one in point 4 ) . in our opinion
all the above steps are needed for a proper matching of the _ me kinematics _ with the _ ps kinematics _ in any interface between the me - type mc event generator and the ps generator .
this is particularly important for the production of the @xmath0-bosons which are spin-@xmath21 particles with @xmath22 couplings to fermions .
in such a case respecting all the spin correlations in the above matching is obligatory .
the above kinematical matching relies on the correct generation of the incoming `` effective '' quarks momenta .
their spacial orientation is crucial for the spin correlations which , in turn , influence the angular distributions of the @xmath0-decay leptons and , as a consequence , their @xmath5 distribution . at this point
one may ask if such effective on - shell quarks make sense at all .
it has been known for some time that a cross section corresponding to the real nlo qcd emission can be expressed as a linear combination of the lo matrix elements @xcite .
the latter can be calculated using some effective incoming on - shell partons four - momenta , e.g. with the help the spin amplitudes .
however , care must be taken while constructing these effective four - momenta for each lo matrix element individually in order not to spoil spin correlations .
coefficients of these matrix elements can be expressed as functions of variables related to the radiated partons ( e.g. their momenta fractions and polar angles in an appropriate frame ) and they are generally different for each lo matrix element . in a monte carlo approach , computations of the nlo cross section can be done with the use of the so - called branching algorithm , where in each branch a single lo matrix element is evaluated and a particular branch is picked up with a probability proportional to the coefficient of this matrix element .
what is also interesting , flavours of these effective partons depend only on the respective lo process and can be different than flavours of partons initiating the nlo process . recently , such a method has been adapted to implement the drell yan processes in the generator @xcite . for the dy processes
the differential cross section corresponding to the nlo real - parton radiation can be expressed as the following combination of the lo matrix elements : @xmath23 where @xmath24 are the coefficients depending on the radiated parton variables , e.g. @xmath25 , where @xmath11 and @xmath26 are the momentum fraction and the polar angle of this parton .
the effective four - momenta @xmath27 and @xmath28 of the incoming on - shell quark @xmath29 and antiquark @xmath30 entering the lo matrix element are constructed in the nlo - process cm frame in such a way that the momentum of the `` spectator '' of the radiation is only rescaled without changing its direction , while the four - momentum of the `` emitter '' is calculated as a difference of the electroweak boson four - momentum and the `` spectator '' effective four - momentum , so its direction is different from that of the original `` emitter '' parton .
the `` emitter spectator '' assignment to the incoming partons is done on the feynman - diagramatic basis , details can be found e.g. in @xcite . in order to compute the appropriate lo matrix elements and to match the lo kinematics with the nlo one ,
appropriate lorentz transformations should be performed for the effective on - shell quarks four - momenta .
actually , they are analogous to the ones described above for the matching , see @xcite . based on this analogy we believe that our procedure for the kinematical matching between and should be correct , at least up to the nlo .
of course , the parton shower provides only a leading - log ( ll ) type approximation of the nlo qcd corrections , but in the exact nlo matrix elements for real - parton emission can be taken into account through appropriate correcting weights .
if this is done , the predictions of for the leptonic distributions in the dy processes should be exact at the nlo for the hard process , except for the normalisation .
the latter will not be correct because does not include the nlo virtual corrections
this could be easily fixed by applying the nlo @xmath31-factor .
for the lepton charge asymmetries this @xmath31-factor is even not needed because it cancels out between numerators and denominators . in the discussed to interface these nlo matrix element corrections
are included by default .
however , the predictions will not be exactly the same as from the fixed - order nlo calculations , because generates also the higher - order ll - type qcd corrections through parton showers .
they will lead to additional distortions of the leptonic distributions , in particular that of @xmath5 , but should not change them drastically .
an observable which is most sensitive to details of the kinematical matching between the me - type mc generator and the ps generator in the charged - current dy process is the final - state charged lepton transverse momentum @xmath5 .
this is because its distribution is a strongly varying function with a sharp jacobian peak .
its shape is considerably affected by the non - zero @xmath0-boson transverse momentum @xmath17 , see e.g. @xcite . moreover , since @xmath0 is a vector boson and its coupling to fermions are of the @xmath22 type , the angular distributions of its decay leptons in the @xmath0-rest frame are highly asymmetric : @xmath32 where @xmath33 is , in the limit of massless quarks , the angle between the outgoing charged lepton and incoming quark directions , and @xmath34 is the @xmath0-boson electric charge in the units of the positron charge . because of that , @xmath5 depends strongly not only on @xmath17 but also on the individual momenta of the quark and antiquark . in our numerical comparisons
we have used the following mc programs : 1.35 @xcite , 6.401 @xcite , @nlo 4.03 @xcite and mcfm 5.8 @xcite .
the lepton charge asymmetry observables are used to scrutinize the differences between the @xmath35 and @xmath36 mediated processes in these generators . for a given kinematical variable @xmath37 , the charge asymmetry @xmath38 is defined as : @xmath39 where @xmath40 and @xmath41 refer to the electric charge of the @xmath0 boson ( or the final - state charged lepton ) and @xmath42 is the differential cross section of an observable @xmath37 .
the asymmetry distributions have been obtained for the proton proton collisions at @xmath43tev using the cteq 6.1 pdf set @xcite and the particles properties from the pdg 2011 publication @xcite , for the following two cases : ( 1 ) without any kinematical restrictions for the outgoing lepton and ( 2 ) with the kinematical cuts @xmath44 in fig .
[ ptlas : wpm ] we show the charge asymmetry distributions as a function of @xmath5 for electrons obtained from interfaced with .
the kinematical matching described the previous section was applied with the effective on - shell quarks four - momenta as provided by .
these distributions are compared with the ones coming directly from 6.4 ( left plots ) and the ones obtained from @nlo ( right plots ) .
lower plots show the differences of the distributions presented in the upper plots .
a very good agreement between and shows that all the technical aspects of the kinematical matching in their interface were done correctly .
however , we see that the and results differ considerably from the @nlo results , in particular we observe the opposite behaviour of the @xmath5 asymmetry above the jacobian peak ( @xmath45gev ) .
the region around the jacobian peak is crucial for the @xmath0 mass measurements at the lhc , see e.g. refs .
the source of this discrepancy must be understood to hope for any improvement of the present precision of the the @xmath0-boson mass , width and their charge asymmetries at the lhc . in fig .
[ ptlas : wm ] we present the comparisons of the @xmath5 charge asymmetry distributions for muons without kinematical cuts and with the cuts of eq .
( [ eq : cuts ] ) between ( with the same kinematical matching as in fig .
[ ptlas : wpm ] ) and @nlo .
similar discrepancies as for the electrons are observed for the fully inclusive distributions . in the presence of cuts
the asymmetry distribution changes its shape and the differences between the two programs are smaller but still unacceptable .
we have also compared the charge asymmetry distributions as a function of : @xmath46 and @xmath47 , where : @xmath48 and found a good agreement between and @nlo . in fig .
[ etalas : wm ] we present charge asymmetries for @xmath47 . except for the large values of @xmath47 ,
i.e. except for the region which is beyond the measurement domain of the atlas and cms experiments , the agreement between the two programs is good .
the large discrepancies between the two programs in the restricted phase - space , as specified by eq .
( [ eq : cuts ] ) , are thus important only for the charge asymmetries of the transverse lepton momentum distributions .
one may argue that these discrepancies result from differences in the shape of the @xmath17 distributions between and @nlo .
indeed , we have found that the @xmath17 distributions differ for and @nlo , but mainly at low @xmath17 ( @xmath49gev ) , while for higher values their ratio is flat . in the ps algorithm used by the @xmath17 distribution at low values
is affected mainly but the so - called intrinsic partonic @xmath50 which is generated from a gaussian distribution with an adjustable width .
we have used this dependence and generated the samples of events with amplified differences of the @xmath17 distributions between the above two generators in the range which is well beyond the present measurement uncertainties .
we have observed that the corresponding charge asymmetries of the @xmath5 distributions remained hardly changed .
we have also compared these asymmetries for @xmath51gev , where the ratio of the @xmath17 distributions from and @nlo is flat , and found similar results . finally ,
have checked that , in spite of differences in the absolute @xmath17 distributions , their charge asymmetries agree very well between the two programs . this proofs that the differences in @xmath17 do not explain the large discrepancies in the @xmath5 asymmetries between and @nlo .
thus , the latter must be attributed to the differences in the effective polarisation of the and @nlo @xmath0-bosons
. can we find simple physical arguments to explain the shape of the @xmath5 charge asymmetry distribution ?
can we say which program is right and which is wrong ? in order to try to answer the above two questions , we first produced the distributions for the case of @xmath52 , i.e. without the parton shower in ( using purely beam - collinear quarks from the standard pdfs ) .
these distributions are shown in fig .
[ ptlas - ptw0:wm ] and compared with the previous @nlo result . as one can expect , for the @xmath52 case the charge asymmetry distribution is flat below the jacobian peak position ( @xmath53gev ) , and then rises very slowly with increasing @xmath5 .
its average value below the peak position reflects the difference between the total cross sections for positively and negatively charged dy processes which is driven by the effective excess of the @xmath54-quarks with respect to the @xmath55-quarks producing the @xmath0-bosons at the lhc . for the @xmath5 values above the jacobian peak @xmath0-bosons must be off - shell if @xmath52 .
since higher invariant mass prefers harder quarks and since @xmath54 is on average harder than @xmath55 , the relative number of produced @xmath56 rises with respect to @xmath57 . in the following , this effect will be called the isospin effect . in the presence of the kinematical cuts we see a good agreement between the two programs below the jacobian peak , and
the discrepancy begins above this peak but is smaller than without cuts and also than when is used .
the sharp cut at @xmath58gev for the results comes from the cut on @xmath59 ( in the case of @xmath60 they are equivalent ) .
it is rather striking that below the jacobian peak the charge asymmetry of the @xmath5 distribution is generated mainly by the cut on the lepton pseudorapidity , @xmath47 , and is hardly sensitive to the @xmath0-bosons transverse momentum spectrum .
what is the reason for the observed shape of the distribution for @xmath60 ? in the @xmath0-rest frame the events with high @xmath5 correspond to @xmath61 , while the ones with low @xmath5 to large positive and large negative @xmath47 .
if we take the @xmath15 axis along the quark momentum , then the @xmath0-bosons will have preferably positive rapidity in such a frame , since the quarks are on average harder than the antiquarks .
thus , when we perform a boost to the lab frame in the presence of symmetric cuts on @xmath47 , the events with negative @xmath47 will migrate in while the ones with positive @xmath47 will migrate out of the selected kinematical region ( the @xmath0-boson rapidity just adds to the lepton pseudorapdity ) .
since for @xmath57 charged leptons are emitted preferably along its direction , then more events with low @xmath5 will move out than move in , while for @xmath56 it will be opposite .
this is why we observe the decrease of the asymmetry with increasing @xmath5 up to the value close to the jacobian peak position .
close to the peak position majority of leptons must have @xmath61 in the @xmath0-rest frame and the migration mechanism discussed above can be neglected .
for such @xmath5 values and above the discussed earlier mechanism related to the relative hardness of the distributions of the @xmath54 and @xmath55 quarks takes over and the asymmetry rises .
having understood the influence of the migration and the quark - isospin effects on the charge asymmetry distribution , let us try to answer our main question : do we understand the @xmath5 charge asymmetry when @xmath62 ? as discussed before , the shape of the distribution below the jacobian peak position is determined by the migration mechanism and is hardy dependent on the underlying distribution of @xmath17 . therefore
, in this region , our previous analysis holds .
we thus concentrate on the region of the large @xmath5 ( above the position of the jacobian peak ) .
the main difference here with respect to the @xmath60 case is that in addition to the isospin effect another effect come into play and become dominant : the effect of hard qcd radiation which influences the effective polarisation of the @xmath0-bosons . in the discussion presented below the @xmath0-polarisation
is specified in the reference frame in which the spin quantisation axis is parallel to the direction of the @xmath0-boson .
it has been shown recently that for the processes of @xmath63 production at the lhc left - handedly polarized @xmath0s dominate over the right - handedly polarized @xmath0s @xcite .
for the left - handed @xmath0s the charged leptons are emitted preferably in the @xmath57 direction and opposite to the @xmath56 direction ( and _ vice versa _ for the right - handed @xmath0s ) .
therefore , the non - zero @xmath17 increases , on the average , the transverse momentum of the negatively charged lepton and decreases it for the positively charged one .
this is what we observe in the left plot of fig .
[ ptlas - ptw0:wm ] where for @nlo the asymmetry decreases for high @xmath5 .
there is , of course , some contribution from longitudinally polarized @xmath0s , but it never dominates @xcite and , what is more important , charged lepton angular distributions are in this case identical for @xmath56 and @xmath57 .
moreover , the isospin effect , which could potentially counterbalance such a decrease , is sizeably smaller in magnitude due to a steeply falling breit wigner distribution .
therefore , the @nlo results do have a rather convincing explanation of the @xmath5 charge asymmetry behaviour while the results do not . in order to check the validity of the above reasoning
we have generated events using the monte carlo program mcfm @xcite which calculates the fixed - order qcd corrections to the hard process convoluted with the collinear pdfs .
the charge asymmetry distribution for @xmath64 and @xmath65 are presented in fig .
[ ptlas : mcfm ] and compared with the ones from with the standard parton - shower matching .
the @xmath5 asymmetries predicted by mcfm are close to those from @nlo , which supports the conclusion that the @nlo predictions on the @xmath5 asymmetries are more likely to be correct than those of .
but can we find the simple reason why the predictions are so grossly wrong ?
from our numerical tests and discussion presented above it becomes obvious that the problem must be related to the modelling of the effective polarisation of @xmath0-bosons . in the lo approximation and for on - shell
partons the @xmath0-polarisation is uniquely driven by the asymmetry in the distributions of the momenta of the effective quark and antiquark entering the dy processes ( see eq .
( [ eq : dsig - nlo ] ) ) , rather than by their sum which determines @xmath66 and @xmath17 . inspecting the 6.4 manual @xcite we have found that the construction of these effective on - shell partons momenta should agree at nlo with that of ref . @xcite .
, of course , generates through the parton - shower more than a single nlo emission , however they should not change considerably ( or even revert ) the nlo effective partons momenta as such additional emissions are mainly soft and collinear .
therefore , the predictions for the @xmath5 asymmetries should not differ much from the nlo ones . at this point
we started searching not only for possible conceptual but also for the technical errors affecting the spatial orientation of the quark and antiquark momenta .
we have made several technical checks of the generator along this line .
one of the checks done was to swap the transverse momenta of the effective on - shell quark and antiquark .
to our great surprise , once this was done on the event - by - event basis , we have obtained a very good agreement with the @nlo charge asymmetry distribution , both in the full phase - space and in the restricted kinematical region .
the comparisons are shown in fig .
[ ptlas : wm - ktswap_ptl ] for the @xmath5 and in fig .
[ ptlas : wm - ktswap_etal ] for the @xmath47 dependence of the lepton charge asymmetry .
this agreement may be accidental but it may also suggest that the transverse momenta are , perhaps , not correctly assigned to the effective quark and antiquark in . whether or not such a hypothesis is true
can however be verified only by the authors of the generator . on the conceptual side
we have investigated the mechanism which drives the effective lo polarisation of the @xmath0 bosons in the dy process involving on - shell quarks .
we have found that skipping the rotations of leptons momenta in the @xmath0-boson rest frame , described in point 5 of the previous section , gives better agreement of the @xmath5 charge asymmetry with the @nlo one .
the results are shown in fig .
[ ptlas : wm-2boosts ] .
the agreement with @nlo is slightly worse than in fig .
[ ptlas : wm - ktswap_ptl ] , however much better than in fig .
[ ptlas : wm ] .
note that skipping these rotations is equivalent to retaining the ps - initial ( parton - shower unaffected ) effective on - shell quark helicities rather than those corresponding the ps - final ones ( following the parton - shower ) .
we have implemented the above two options in the new version of @xcite .
these versions can not replace the future state - of - the - art nlo programs with the nlo ps and the full set of ew radiative corrections .
however , as long as such programs are not available , they may be of use in the initial phase of the measurement of the lepton charge asymmetries at the lhc .
first of all , they can be of help in the unfolding of the measured charge lepton asymmetries in the experimental procedures where the precision of the ew and the real photon radiative correction matters .
more importantly , the above versions , providing the simplified lo picture of the effective polarisation of @xmath0-bosons at the lhc , may help in designing the new polarisation - dependent observables for the studies of the electroweak symmetry breaking ( ewsb ) mechanism .
in this paper we have discussed the generic problem of kinematical matching of a parton shower generator with a matrix element generator for the drell - yan processes involving spin 1 intermediate particles . we have argued that the les houches accord must be extended to take into account the spin correlations at all the stages of the event generation .
we have described in detail our kinematical matching procedure which is used in the interface of our generator with the 6.4 generator .
we have demonstrated that the momentum vectors of the on - shell quark and antiquark , carrying in the lo approximation the full information on the @xmath0-boson polarisation , must be well defined .
any error in directions of these vectors has dramatic consequences for the @xmath5 dependence of the charge asymmetries at the lhc .
in particular , using the transverse momenta of the effective quarks provided in 6.4 leads to completely different behaviour of the above asymmetry than predicted by the nlo ( and beyond ) calculations , e.g. @nlo and mcfm .
we have found that simple swapping of the effective quark and antiquark transverse momenta in , or skipping the rotation of the outgoing lepton momenta , results in the @xmath5 charged asymmetries that match the nlo predictions of @nlo .
we have implemented the corresponding matching schemes in the new version of the generator .
the issue of the proper matching between the matrix element calculations and the parton - shower generators respecting the spin correlations is important not only for the charged - current drell
yan processes but also for any process of production and decay of non - zero spin particles at the lhc .
it needs to be readdressed in the more general context of matching the nlo matrix elements with the nlo parton shower such that a handle is given to the experimentalists to control the relative contributions of all the spin density matrix elements of the decaying particles , thus allowing for an experimental verification of the implemented monte carlo mechanism which drives the polarisation of non - zero spin particles at the lhc .
t. sjostrand , s. mrenna , and p. skands , jhep * 05 * , 026 ( 2006 ) , hep - ph/0603175 .
g. corcella _ et al .
_ , jhep * 01 * , 010 ( 2001 ) , hep - ph/0011363 . s. frixione and b. r. webber , jhep * 06 * , 029 ( 2002 ) , hep - ph/0204244 .
p. nason , jhep * 11 * , 040 ( 2004 ) , hep - ph/0409146 .
m. bahr _ et al .
j. * c58 * , 639 ( 2008 ) , 0803.0883 .
w. paczek and s. jadach , : the monte carlo event generator for single @xmath0-boson production with leptonic decays in hadron collisions , available from http://cern.ch/placzek/winhac .
d. bardin , s. bondarenko , s. jadach , l. kalinovskaya , and w. placzek , acta phys . polon . *
b40 * , 75 ( 2009 ) , 0806.3822 .
j. alwall _ et al .
_ , comput .
. commun . * 176 * , 300 ( 2007 ) , hep - ph/0609017 .
m. r. whalley , d. bourilkov , and r. c. group , ( 2005 ) , hep - ph/0508110 . c. m. carloni calame , s. jadach , g. montagna , o. nicrosini , and w. placzek , acta phys .
* b35 * , 1643 ( 2004 ) , hep - ph/0402235 .
r. kleiss , phys.lett .
b180 * , 400 ( 1986 ) . m. h. seymour , nucl.phys .
* b436 * , 443 ( 1995 ) , hep - ph/9410244 .
k. hamilton , p. richardson , and j. tully , jhep * 0810 * , 015 ( 2008 ) , 0806.0290 .
t. sjostrand , s. mrenna , and p. skands , , . | charged lepton transverse momenta in the drell yan processes play an important role at the lhc in precision measurements of the standard model parameters , such as the @xmath0-boson mass and width , their charge asymmetries and @xmath1 .
therefore , their distributions should be described as accurate as possible by the monte carlo event generators . in this paper
we discuss the problem of matching the hard - process kinematics of the monte carlo generator with the parton - shower kinematics of the 6.4 generator while interfacing these two programs .
we show that improper assignment of the quark and antiquark effective momenta in the lo matrix element computations may affect considerably the predicted lepton transverse momenta and even completely reverse their charge asymmetries at the lhc .
we propose two matching schemes in which the nlo qcd distributions of the leptonic kinematical variables can be well reproduced by the lo generator . * charge asymmetries of lepton transverse momenta in drell
yan processes at the lhc * * m. w. krasny@xmath2 and w. paczek@xmath3 * + _ to be submitted to acta physica polonica b _
@xmath4the work is partly supported by the program of the french polish co - operation between in2p3 and copin no . 05 - 116 , and by the polish national centre of science grant no .
dec-2011/03/b / st2/00220 . |
nuclei under extreme conditions ( temperature , spin , neutron to proton ratio n / z , density ) have been the focus of experimental and theoretical investigations .
the fusion process between heavy ions colliding at incident energies around the coulomb barrier is well suited to study hot and rotating compound nuclei .
these excited nuclei decay through a variety of channels such as neutrons , light charged particles , fission and production of intermediate mass fragments ( imf s ) with charge z@xmath33 .
the neutron enrichment of the compound nuclei is expected to influence the competition between these decay modes . in this contribution
we present new data on the production of fragments with 6@xmath0z@xmath028 formed in @xmath1kr+@xmath2ca fusion reactions at 5.5 mev / a incident energy .
the experiments were performed at the ganil facility and @xmath1kr beams were used to bombard a self - supported @xmath2ca target of 1mg / cm2 in thickness .
fragments were measured with the 4@xmath4-indra array @xcite . for 3@xmath5 45
, each module comprises an ensemble made of ionization chamber , silicon detector and csi . for fragments with 6@xmath0z@xmath028
, this angular range covers the forward hemisphere in the centre of mass frame .
the energy calibration was deduced from the kr+ca elastic scattering and energy loss tables . for both systems ,
the centre of mass average velocity of fragments is roughly constant as a function of the emission angle and agrees rather well with the viola systematic@xcite .
the angular distribution d@xmath6/d@xmath7 of fragments is almost isotropic in the measured angular range and the sum of the charges of the two biggest fragments is close to the charge of the compound nucleus .
all these features are compatible with binary emission from a compound nucleus .
the absolute cross - sections were obtained from the normalization to the elastic scattering measured at a laboratory angle for which the scattering is given by the rutherford formula .
the measured charge distributions are shown in fig.1a for @xmath8kr+@xmath2ca .
the cross - section distributions present a shape with a maximum around half of the charge of the compound nuclei .
this indicates that elements with z@xmath314 come mainly from a symmetric fission process induced by high transferred spins .
besides this feature , for light fragment ( z@xmath010 ) a strong even - odd staggering is visible , and this effect is still present for higher z with a smaller amplitude .
similar features are observed for the @xmath9kr+@xmath2ca reaction .
however the yields for symmetric splitting are about 30% higher than for the @xmath8kr+@xmath2ca system .
cross section for odd - z fragments is higher for the neutron rich nuclei system and for even - z fragments it is higher for the neutron poor nuclei system .
1b shows the charge dependence of the ratio r= @xmath10 .
r decreases roughly from 1.25 for z=6 down to 0.75 for z=7 . for odd z ,
the ratio increases up to z=21 and reaches a kind of plateau .
for even z , r decreases from z=6 to z=10 and then increases to reach the same kind of plateau as for odd - z .
the excitation energy and the maximum angular momentum stored in the compound nucleus are expected to be very similar in both reactions .
thus , the observed effects are probably linked to the difference in the neutron enrichment of the compound nucleus .
it is worth noticing that the z=6 cross section is comparable to the one infered in @xcite
statistical decay calculations were performed using the monte carlo code gemini @xcite .
all channels are considered within the hauser - feshbach ( for z@xmath02 ) and transition state ( for z@xmath33 ) approaches .
the key ingredient is the conditional saddle configuration ( or conditional barrier ) which is constrained by the mass asymmetry .
the saddle conditional energy for different mass ( or charge ) asymmetry was deduced from sierk s model @xcite . in the present work , the fusion evaporation cross section is not yet available but , for this kind of studied reaction , we know that the fission cross section is sensitive to the highest angular momentum leading to fusion mechanism . thus , the maximum angular momentum @xmath11 was considered as a free parameter to reproduce the yields around the symmetric splitting .
all the calculations presented here have been performed using the standard parameters of the gemini code and a level density parameter a = a/8 .
results are reported in fig.1a for the @xmath8kr+@xmath2ca and fig .
1b for the ratio @xmath10 . in both figures
, the line represents the best fit obtained with the indicated @xmath11 .
the shape of the charge distribution around symmetric fission is relatively well reproduced , while the model predicts a smooth behaviour for z@xmath014 at variance with the data . for fragments with 6@xmath0z@xmath012 , the calculation overestimates the measured cross - sections .
various attempts at changing the @xmath11 value lead to the conclusion that the present version of the code fails to explain the relative population of the imf s and symmetric fission .
the key ingredient governing the competition between channels is the barrier . besides the even - odd effect
, the observed disagreement would indicate a failure of the sierk barriers and/or the transition state picture for large asymmetry and large angular momentum .
indeed , in the transition state picture the asymmetry at saddle configuration is frozen all along the path to scission . for medium - mass compound nuclei at high spin as those studied in this work , data presented in fig.1a suggest that saddle and scission configurations are no longer degenerate .
then , the surface of the potential energy for high angular momentum would be narrower , and the charge distribution too , leading to an increase of the relative yields between the imf s and symmetric fission . coming back to the even - odd effects in cross - sections ,
the gemini code allows to include the washing out of the shell and pairing effects with increasing temperature .
standard calculations , as such those shown in fig.1 , correspond to exclude any structure effect for z@xmath33 .
we made an attempt assuming a phenomenological description of pairing . in that case , the even - odd effect predicted by the model has a smaller amplitude with respect to the data .
moreover , the ratio @xmath10 reproduces experimental data for light fragments with odd z and strongly underestimates the yields of even z. the agreement becomes satisfactory for z@xmath314 .
it is worth noticing that for z@xmath314 the shape of the z distribution is well reproduced . from these comparisons
, we can conclude that the way the pairing is included requires more investigations .
as noticed above , the calculated yields of the imf s are overestimated , possibly due to an underestimation of the barrier .
thus , in the model , the available excitation energy available after the splitting would be higher , leading to higher excitation energy in the imf s .
the main consequence would be a decreasing of the even - odd effect due to an overestimation of the calculated temperature of the imf s . to explore this scenario we performed some calculations using of a different approach of the statistical model .
the measured even - odd effect in imf yields suggests that they are emitted relatively cold or close to the threshold of particle emission otherwise the picture would have been blurred by secondary emission . in the busco model
@xcite the decay of compound nuclei includes all open channels from z=0 to z=20 .
partial widths are calculated within the hauser - feshbach framework .
ground and excited states of each emitted fragment were included .
an attempt was made assuming @xmath11=45@xmath12 . from the comparison to the data the model fails to reproduce the z distribution , the calculated cross section decreasing strongly as z increases at variance with data . fixing a value for @xmath11 in order to reproduce cross section for carbon leads to underestimate the yields of z=10 by about three orders of magnitude .
the calculated ratio @xmath10 is completely wrong and is not shown in fig.1b .
further experimental investigations will be performed to analyse the light charged particles emitted in coincidence with imf s .
this will inform on the excitation energy stored in the imf s and on the level density parameter .
measurement of the evaporation residue cross section will allows to deduce the maximum angular momentum for fusion process . to improve the modelization , refinements are needed to describe the disintegration process in an excitation range where structure effects influence strongly the phase space available for a statistical description .
0 j. pouthas et al . , _ nucl .
methods phys .
_ @xmath13 , 418 ( 1995 ) . c. beck and a.
szanto de toledo _ phys .
rev . _ c @xmath14 , 1989 ( 1996 ) : see also c. beck and a. szanto de toledo , arxiv : * nucl - th/9512025 * ( 1995 ) and reference therein .
m. la commara et al . , _ nucl .
_ @xmath15 , 43 ( 2000 ) .
r.j . charity et al , _ nucl .
_ @xmath16 , 43 ( 1988 ) .
sierk , _ phys .
_ c @xmath17 2039 ( 1986 ) .
j. gomez del campo et al , _ phys .
_ @xmath18 , 290 ( 1988 ) . | cross sections , kinetic energy and angular distributions of fragments with charge 6@xmath0z@xmath028 emitted in @xmath1kr+@xmath2c at 5.5 mev / a reactions were measured at the ganil facility using the indra apparatus .
this experiment aims to investigate the influence of the neutron enrichment on the decay mechanism of excited nuclei .
data are discussed in comparison with predictions of transition state and hauser - feshbach models . |
this is the fifth in a series of seven papers describing the pan - starrs1 surveys , the data reduction techiques and the resulting data products .
this paper ( paper v ) describes the final calibration process , and the resulting photometric and astrometric quality .
* paper i ) provides an overview of the pan - starrs system , the design and execution of the surveys , the resulting image and catalog data products , a discussion of the overall data quality and basic characteristics , and a brief summary of important results .
* paper ii ) describes how the various data processing stages are organised and implemented in the imaging processing pipeline ( ipp ) , including details of the the processing database which is a critical element in the ipp infrastructure .
* paper iii ) describes the details of the pixel processing algorithms , including detrending , warping , and adding ( to create stacked images ) and subtracting ( to create difference images ) and resulting image products and their properties .
* paper iv ) describes the details of the source detection and photometry , including point - spread - function and extended source fitting models , and the techniques for forced " photometry measurements .
* paper vi ) describes the details of the resulting catalog data and its organization in the pan - starrs database .
* paper vii ) describes the medium deep survey in detail , including the unique issues and data products specific to that survey .
the medium deep survey is not part of data release 1 .
( dr1 ) the pan - starrs1 filters and photometric system have already been described in detail in @xcite .
from may 2010 through march 2014 , the pan - starrs science consortium used the 1.8 m telescope to perform a set of wide - field science surveys .
these surveys are designed to address a range of science goals included the search for hazardous asteroids , the study of the formation and architecture of the milky way galaxy , and the search for type ia supernovae to measure the history of the expansion of the universe .
the wide - field telescope consists of a 1.8 meter diameter @xmath2/4.4 primary mirror with an 0.9 m secondary , producing a 3.3 degree field of view @xcite .
the optical design yields low distortion and minimal vignetting even at the edges of the illuminated region .
the optics , in combination with the natural seeing , result in generally good image quality : the median image quality for the 3@xmath3 survey is fwhm = ( 1.31 , 1.19 , 1.11 , 1.07 , 1.02 ) arcseconds for ( ) , with a floor of @xmath4 arcseconds .
the camera @xcite is a mosaic of 60 edge - abutted @xmath5 pixel back - illuminated ccid58 orthogonal transfer arrays manufactured by lincoln laboratory @xcite .
the ccds have 10 @xmath6 m pixels subtending 0.258 arcsec and are 70@xmath6 m thick .
the detectors are read out using a stargrasp ccd controller , with a readout time of 7 seconds for a full unbinned image @xcite .
the active , usable pixels cover @xmath7% of the fov .
nightly observations are conducted remotely from the advanced technology research center in kula , the main facility of the university of hawaii s institute for astronomy operations on maui . during the science survey , images obtained by the system were stored first on computers at the summit , then copied with low latency via internet to the dedicated data analysis cluster located at the maui high performance computer center in kihei , maui .
images obtained by are automatically processed in real time by the image processing pipeline ( ipp , * ? ? ?
real - time analysis goals are aimed at feeding the discovery pipelines of the asteroid search and supernova search teams .
the data obtained for the science survey has also been used in three additional complete re - processing of the data : processing versions 1 , 2 , and 3 ( pv1 , pv2 , and pv3 ) .
the real - time processing of the data is considered `` pv0 '' . except
as otherwise noted , the pv3 analysis of the data is used for the purpose of this article .
the data processing steps are described in detail by @xcite and @xcite . in summary ,
individual images are detrended : non - linearity and bias corrections are applied , a dark current model is subtracted and flat - field corrections are applied .
the -band images are also corrected for fringing : a master fringe pattern is scaled to match the observed fringing and subtracted .
mask and variance image arrays are generated with the detrend analysis and carried forward at each stage of the ipp processing .
source detection and photometry are performed for each chip independently . as discussed below , preliminary astrometric and photometric calibrations
are performed for all chips in a single exposure in a single analysis .
chip images are geometrically transformed based on the astrometric solution into a set of pre - defined pixel grids covering the sky , called skycells .
these transformed images are called the warp images .
sets of warps for a given part of the sky and in the same filter may be added together to generate deeper ` stack ' images .
psf - matched difference images are generated from combinations of warps and stacks ; the details of the difference images and their calibration are outside of the scope of this article .
astronomical objects are detected and characterized in the stacks images .
the details of the analysis of the sources in the stack images are discussed in @xcite , but in brief these include psf photometry , along with a range of measurements driven by the goals of understanding the galaxies in the images . because of the significant mask fraction of the gpc1 focal plane , and the varying image quality both within and between exposures , the effective psf of the ps1 stack images is highly variable .
the psf varies significantly on scales as small as a few to tens of pixels , making accurate psf modelling essentially infeasible .
the psf photometry of sources in the stack images is thus degraded significantly compared to the quality of the photometry measured for the individual chip images . to recover most of the photometric quality of the individual chip images ,
while also exploiting the depth afforded by the stacks , the pv3 analysis make use of forced photometry on the individual warp images .
psf photometry is measured on the warp images for all sources which are detected in the stack images images .
the positions determined in the stack images are used in the warp images , but the psf model is determined for each warp independently based on brighter stars in the warp image .
the only free parameter for each object is the flux , which may be insignificant or even negative for sources which are near the faint limit of the stack detections .
when the fluxes from the individual warp images are averaged , a reliable measurement of the faint source flux is determined .
the details of this analysis are described in detail in magnier et al @xcite . in this article , we discuss the photometric calibration of the individual exposures , the stacks , and the warp imags .
we also discuss the astrometric calibration of the individual exposures and the stack images .
three somewhat distinct astrometric models are employed within the ipp at different stages .
the simplest model is defined independently for each chip : a simple tan projection as described by @xcite is used to relate sky coordinates to a cartesian tangent - plane coordinate system .
a pair of low - order polynomials are used to relate the chip pixel coordinates to this tangent - plane coordinate system .
the transforming polynomials are of the form : @xmath8 where @xmath9 are the tangent plane coordinates , @xmath10 are the coordinates on the 60 gpc1 chips , and @xmath11 are the polynomial coefficients for each order . in the analysis , @xmath12 where the order of the fit , @xmath13 , may be 1 to 3 , under the restriction that sufficient stars are needed to constrain the order . a second form of astrometry model which yields somewhat higher accuracy consists of a set of connected solutions for all chips in a single exposure .
this model also uses a tan projection to relate the sky coordinates to a locally cartesian tangent plane coordinate system .
a set of polynomials is then used to relate the tangent plane coordinates to a focal plane coordinate system , @xmath14 : @xmath15 this set of polynomial accounts for effects such as optical distortion in the camera and distortions due to changing atmospheric refraction across the field of the camera .
since these effects are smooth across the field of the camera , a single pair of polynomials can be used for each exposure . like in the chip analysis about , the code restricts the exponents with the rule @xmath12 where the order of the fit , @xmath16 , may be 1 to 3 , under the restriction that sufficient stars are needed to constrain the order for each chip , a second set of polynomials describes the transformation from the chip coordinate systems to the focal coordinate system : @xmath17 a third form of the astrometry model is used in the context of the calibration determined within the dvo database system .
we retain the two levels of transformations ( chip @xmath18 focal plane @xmath18 tangent plane ) , but the relationship between the chip and focal plane is represented with only the linear terms in the polynomial , supplemented by a course grid of displacements , @xmath19 sampled across the coordinate range of the chip .
this displacement grid may have a resolution of up to @xmath20 samples across the chip .
the displacement for a specific chip coordinate value is determined via bilinear interpolation between the nearest sample points .
thus , the chip to focal - plane transformation may be written as : @xmath21 * wcs keywords * when this polynomial representation is written to the output files , a set of wcs keywords are used to define the astrometric transformation elements .
it is necessary to transform the simply polynomials above into an alternate form : @xmath22
as images are processed by the data analysis system , every exposure is calibrated individually with respect to a photometric and astrometric database .
the goal of this calibration step is to generate a preliminary astrometric calibration , to be used by the warping analysis to determine the geometric transformation of the pixels , and preliminary photometric transformation , to be used by the stacking analysis to ensure the warps are combined using consistent flux units .
the program used for the real - time calibration , , loads the measurements of the chip detections from their individual -format files .
it uses the header information populated at the telescope to determine an initial astrometric calibration guess based on the position of the telescope boresite right ascension , declination and position angle as reported by the telescope & camera subsystems . using the initial guess , loads astrometric and photometric data from the reference database . during the course of the ps1sc survey ,
several reference databases have been used . for the first 20 months of the survey , used a reference catalog with synthetic ps1 photometry generated by the pan - starrs ipp team based on based combined photometry from tycho ( b , v ) , usno ( red , blue , ir ) , and 2mass @xmath23 .
the astrometry in the database was from 2mass .
after 2012 may , a reference catalog generated from internal re - calibration of the pv0 analysis of ps1 photometry and astrometry was used for the reference catalog .
coordinates and calibrated magnitudes of stars from the reference database are loaded by . a model for the positions of the 60 chips in the focal plane is used to determine the expected astrometry for each chip based on the boresite coordinates and position angle reported by the header .
reference stars are selected from the full field of view of the gpc1 camera , padded by an additional 25% to ensure a match can be determined even in the presence of substantial errors in the boresite coordinates .
it is important to choose an appropriate set of reference stars : if too few are selected , the chance of finding a match between the reference and observed stars is diminished .
in addition , since stars are loaded in brightness order , a selection which is too small is likely to contain only stars which are saturated in the gpc1 images . on the other hand ,
if too many reference stars are chosen , there is a higher chance of a false - positive match , especially as many of the reference stars may not be detected in the gpc1 image .
the seletion of the reference stars includes a limit on the brightest and fainted magnitude of the stars selected .
the astrometric analysis is necessarily performed first ; after the astrometry is determined , an automatic byproduct is a reliable match between reference and observed stars , allowing a comparison of the magnitudes to determine the photometric calibration .
the astrometric calibration is performed in two major stages : first , the chips are fitted independently with independent models for each chip .
this fit is sufficient to ensure a reliable match between reference stars and observed sources in the image .
next , the set of chip calibrations are used to define the transformation between the focal plane coordinate system and the tangent plane coordinate system .
the chip - to - focal plane transformations are then determined under the single common focal plane to tangent plane transformation .
the first step of the analysis is to attempt to find the match between the reference stars and the detected objects .
uses 2d cross correlation to search for the match .
the guess astrometry calibration is used to define a predicted set of @xmath24 values for the reference catalog stars . for all possible pairs between the two lists , the values of @xmath25 are generated .
the collection of @xmath26 values are collected in a 2d histogram with sampling of 50 pixels and the peak pixel is identified . if the astrometry guess were perfect , this peak pixel would be expected to lie at ( 0,0 ) and contain all of the matched stars
. however , the astrometric guess may be wrong in several ways .
an error in the constant term above , @xmath27 shifts the peak to another pixel , from which @xmath27 can easily be determined .
an error in the plate scale or a rotation will smear out the peak pixel potentially across many pixels in the 2d histogram . to find a good match in the face of plate scale and rotation errors ,
the cross correlation analysis above is performed for a series of trials in which the scale and rotation are perturbed from the nominal value by a small amount . for each trial ,
the peak pixel is found and a figure of merit is measured .
the figure of merit is defined as @xmath28 where @xmath29 are the second moment of @xmath30 for the star pairs associated with the peak pixel , and @xmath31 is the number of star pairs in the peak .
this figure of merit is thus most sensitive to a narrow distribution with many matched pairs . for the ps1 exposures , rotation offsets of ( -1.0 , -0.5 , 0.0 , 0.5 , 1.0 ) degrees , and plate scales of
( + 1% , 0 , -1% ) of the nominal plate scale are tested .
the best match among these 15 cross - correlation tests is selected and used to generate a better astrometry guess for the chip . the astrometry solution from the cross correlation step above
is again used to selected matches between the reference stars and observed stars in the image .
the matching radius starts off quite large , and a series of fits is performed to generate the transformation between chip and tangent plane coordinates .
three clipping iterations are performed , with outliers @xmath32 rejected on each pass , where here @xmath33 is determined from the distribution of the residuals in each dimension ( x , y ) independently .
after each fit cycle , the matches are redetermined using a smaller radius and the fit re - tried .
the astrometry solutions from the independent chip fits are used to generate a single model for the camera - wide distortion terms .
the goal is to determine the two stage fit ( chip @xmath18 focal plane @xmath18 tangent plane ) .
there are a number of degenerate terms between these two levels of transformation , most obviously between the parameters which define the constant offset from chip to focal plane ( @xmath34 ) and those which define the offset from focal plane to tangent plane ( @xmath35 ) .
we limit ( @xmath35 ) to be 0,0 to remove this degeneracy .
the initial fit of the astrometry for each chip follows the distortion introduced by the camera : the apparent plate scale for each chip is the combination of the plate scale at the optical axis of the camera , modified by the local average distortion . to isolate the effect of distortion
, we choose a single common plate scale for the set of chips and re - define the chip @xmath18 sky calibrations as a set of chip @xmath18 focal plane transformation using that common pixel scale .
we can now compare the observed focal plane coordinates , derived from the chip coordinates , and the tangent plane coordiantes , derived from the projection of the reference coordinates .
one caveat is that the chip reference coordinates are also degenerate with the fitted distortion . in order to avoid being sensitive to the exact positions of the chips at this stage , we measure the local gradient between the focal plane and tangent plane coordinate systems .
we then fit the gradient with a polynomial of order 1 less than the polynomial desired for the distortion fit .
the coefficients of the gradient fit are then used to determine the coefficients for the polynomials representing the distortion .
once the common distortion coming from the optics and atmosphere have been modeled , determines polynomial transformations from the 60 chips to the focal plane coordinate system . in this stage ,
5 iterations of the chip fits are performed . before each iteration ,
the reference stars and detected objects are matched using the current best set of transformations .
these fits start with low order ( 1 ) and large matching radius .
as the iterations proceed , the radius is reduced and the order is allowed to increaes , up to 3rd order for the final iterations .
after the astrometric calibration has finished , the photometric calibration is performed by . when the reference stars are loaded , the apparent magnitude in the filter of interest is also loaded .
stars for which the reference magnitude is brighter than ( ) = ( 19 , 19 , 18.5 , 18.5 , 17.5 ) are used to determine the zero points by comparison with the instrumental magnitudes . for the pv3 analysis ,
an outlier - rejecting median is used to measure the zero point .
for early versions of the analysis , when the reference catalog used synthetic magnitudes , it was necessary to search for the blue edge of the distribution : the synthetic magnitude poorly predicted the magnitudes of stars in the presence of significant extinction or for the very red stars , making the blue edge somewhat more reliable . note that we do not include an airmass correction in this zero point analysis : the airmass correction is folded into the observed zero point .
the zero point may be measured separately for each chip or as a single value for the entire exposure ; the latter option was used for the pv3 analysis .
the calibrations determined by as saved as part of the header information in the output fits tables .
a single multi - extension fits table is written using the format . in these files , the measurements from each chip are written as a separate fits table . a second fits extension for each chip is used to store the header information from the original chip image .
the original chip header is modified so that the extension corresponds to an image with no pixels data : is set to 0 , even though and are retained with the original dimensions of the chip . a pixel - less primary header unit ( phu ) is generated with a summary of some of the important and common chip - level keywords ( e.g. , ) .
the astrometric transformation information for each chip is saved in the corresponding header using standard ( and some non - standard ) wcs keywords .
for the two - level astrometric model , the phu header carries the astrometric transformation related to the projection and the camera - wide distortions .
photometric calibrations are written as a set of keywords to individual chip headers , and if the calibration is performed at the exposure - level , to the phu .
the photometry calibration keywords are : * : the nominal zero point for this filter * : the measured zero point for this chip / exposure * : the measured error on * : the number of stars used to measure * : minimum reference magnitude included in analysis * : maximum reference magnitude included in analysis the keyword is used to set the initial zero point when the data from the exposure are loaded into the dvo database .
data from the gpc1 chip images , the stack images , and the warp images are loaded into dvo using the real - time analysis astrometric calibration to guide the association of detections into objects . after the full pv3 dvo database was constructed , including all of the chip , stack , and warp detections ,
several external catalogs were merged into the database .
first , the complete 2mass psc was loaded into a stand - alone dvo database , which was then merged into the pv3 master database .
next the dvo database of synthetic photometry in the ps1 bands ( see section [ sec : synthdb ] ) was merged in .
next , the full tycho database was added , followed by the allwise database .
after the gaia release in august 2016 @xcite , we generated a dvo database of the gaia positional and photometric information and merged that into the master dvo database .
the photometric calibration of the dvo database starts with the `` ubercal '' analysis technique as described by @xcite .
this analysis is performed by the group at harvard , loading data from the files into their instance of the large scale database ( lsd , * ? ? ?
* ) , a system similar to dvo used to manage the detections and determine the calibrations .
photometric nights are selected and all other exposures are ignored . each night
is allowed to have a single fitted zero point and a single fitted value for the airmass extinction coefficient per filter .
the zero points and extinction terms are determined as a least squares minimization process using the repeated measurements of the same stars from different nights to tie nights together .
flat - field corrections are also determined as part of the minimization process . in the original ( pv1 ) ubercal analysis , @xcite determined flat - field corrections for @xmath36 sub - regions of each chip in the camera and four distinct time periods ( `` seasons '' ) .
later analysis ( pv2 ) used an @xmath37 grid of flat - field corrections to good effect .
the ubercal analysis was re - run for pv3 by the harvard group .
for the pv3 analysis , under the pressure of time to complete the analysis , we chose to use only a @xmath36 grid per chip as part of the ubercal fit and to leave higher frequency structures to the later analysis . a 5th flat - field season
consisting of nearly the last 2 years of data was also included for pv3 .
in retrospect , as we show below , the data from the latter part of the survey would probably benefit from additional flat - field seasons . by excluding non - photometric data and only fitting 2 parameters for each night ,
the ubercal solution is robust and rigid .
it is not subject to unexpected drift or sensitivity of the solution to the vagaries of the data set .
the ubercal analysis is also especially aided by the inclusion of multiple medium deep field observations every night , helping to tie down overall variations of the system throughput and acting as internal standard star fields .
the resulting photometric system is shown by @xcite to have reliability across the survey region at the level of ( 8.0 , 7.0 , 9.0 , 10.7 , 12.4 ) millimags in ( ) . as we discuss below , this conclusion is reinforced by our external comparison .
the overall zero point for each filter is not naturally determined by the ubercal analysis ; an external constraint on the overall photometric system is required for each filter .
@xcite used photometry of the md09 medium deep field to match the photometry measured by @xcite on the reference photometric night of mjd 55744 ( ut 02 july 2011 ) . @xcite and @xcite have re - examined the photometry of calspec standards observed by ps1 .
@xcite reject 2 of the 7 stars used by @xcite and add photometry of 5 additional stars . @xcite further reject measurements of calspec standards obtained close to the center of the camera field of view where the psf size and shape changes very rapidly .
the result of this analysis modifies the over system zero points by 20 - 35 millimags compared with the system determined by @xcite .
the ubercal analysis above results in a table of zero points for all exposures considered to be photometric , along with a set of low - resolution flat - field corrections .
it is now necessary to use this information to determine zero points for the remaining exposures and to improve the resolution of the flat - field correction .
this analysis is done within the ipp dvo database system .
the ubercal zero points and the flat - field correction data are loaded into the pv3 dvo database using the program .
this program converts the reported zero point and flat field values to the dvo internal representation in which the zero point of each image is split into three main components : @xmath38 where @xmath39 and @xmath40 are static values for each filter representing respectively the nominal zero point and the slope of the trend with respect to the airmass ( @xmath41 ) for each filter .
these static values are listed in table [ tab : zpts ] .
when was run , these static zero points have been adjusted by the calspec offsets listed in table [ tab : zpts ] based on the analysis of calspec standards by @xcite .
these offsets bring the photometric system defined by the ubercal analysis into alignment with @xcite .
the value @xmath42 is the offset needed by each exposure to match the ubercal value , or to bring the non - ubercal exposures into agreement with the rest of the exposures , as discussed below .
the flat - field information is encoded in a table of flat - field offsets as a function of time , filter , and camera position .
each image which is part of the ubercal subset is marked with a bit in the field : when applies the ubercal information to the image tables , it also updates the individual measurements associated with those images . in the dvo database schema , the normalized instrumental magnitude , @xmath43
are stored for each measurement .
the value of 25.0 is an arbitrary ( but fixed ) constant offset to place the instrumental magnitudes into approximately the correct range .
associated with each measurement are two correction magnitudes : @xmath44 and @xmath45 , along with the airmass for the measurement , calculated using the altitude of the individual detection as determined from the right ascension , declination , the observatory latitude , and the sidereal time . for a camera with the field of view of the ps1 gpc1 ,
the airmass may vary significantly within the field of view , especially at low elevations . in the worst cases , at the celestial pole , the airmass range within a single exposure is xxx - xxx .
the complete calibrated ( ` relative ' ) magnitude is determined from the stored database values as : @xmath46 the calibration offsets , @xmath44 and @xmath45 , represent the per - exposure zero point correction and the slowly - changing flat - field correction respectively .
these two values are split so the flat - field corrections may be determined and applied independently from the time - resolved zero point variations . note that the above corrections are applied to each of the types of measurements stored in the database , psf , aperture , kron .
the calibration math remains the same regardless of the kind of magnitude being measured .
also note that for the moment , this discussion should only be considered as relevant to the chip measurements .
below we discuss the implications for the stack and warp measurements .
when the ubercal zero points and flat - field data are loaded , updates the @xmath44 values for all measurements which have been derived from the ubercal images .
these measurements are also marked in the field with the bit . at this stage , also updates the values of @xmath45 for all gpc1 measurements in the appropriate filters .
relative photometry is used to determine the zero points of the exposures which were not included in the ubercal analysis .
the relative photometry analysis has been described in the past by @xcite .
we review that analysis here , along with specific updates for pv3 .
as described above , the instrumental magnitude and the calibrated magnitude are related by arithmetic magnitude offsets which account for effects such as the instrumental variations and atmospheric attenuation .
@xmath47 from the collection of measurements , we can generate an average magnitude for a single star ( or other object ) : @xmath48 we find that the color difference of the different chips can be ignored , and set the value of @xmath49 to 0.0 .
note that we only use a single mean airmass extinction term for all exposures the difference between the mean and the specific value for a given night is taken up as an additional element of the atmospheric attenuation .
we write a global @xmath50 equation which we attempt to minimize by finding the best mean magnitudes for all objects and the best @xmath44 offset for each exposure : @xmath51 + zp + k \zeta + m_{clouds}[i ] - m_{ave}[j ] ) w_{i , j } / \sum_{i , j } w_{i , j}\ ] ] if everything were fitted at once and allowed to float , this system of equations would have @xmath52 unknowns .
we solve the system of equations by iteration , solving first for the best set of mean magnitudes in the assumption of zero clouds , then solving for the clouds implied by the differences from these mean magnitudes .
even with 1 - 2 magnitudes of extinction , the offsets converge to the milli - magnitude level within 8 iterations . only brighter
, high quality measurements are used in the relative photometry analysis of the exposure zero points .
we use only the brighter objects , limiting the density to a maximum of 4000 objects per square degree ( lower in areas where we have more observations ) . when limiting the density , we prefer objects which are brighter ( but not saturated ) , and those with the most measurements ( to ensure better coverage over the available images ) .
there are a few classes of outliers which we need to be careful to detect and avoid .
first , any single measurement may be deviant for a number of reasons ( e.g. , it lands in a bad region of the detector , contamination by a diffraction spike or other optical artifact , etc ) .
we attempt to exclude these poor measurements in advance by rejecting measurements which the photometric analysis has flagged the result as suspcious .
we reject detections which are excessively masked ; these include detections which are too close to other bright objects , diffraction spikes , ghost images , or the detector edges .
however , these rejections do not catch all cases of bad measurements .
after the initial iterations , we also perform outlier rejections based on the consistency of the measurements . for each star ,
we use a two pass outlier clipping process .
we first define a robust median and sigma from the inner 50% of the measurements .
measurements which are more than 5@xmath33 from this median value are rejected , and the mean & standard deviation ( weighted by the inverse error ) are recalculated .
we then reject detections which are more than 3@xmath33 from the recalculated mean .
suspicious stars are also exclude from the analsis .
we exclude stars with reduced @xmath50 values more than 20.0 , or more than 2@xmath53 the median , whichever is larger .
we also exclude stars with standard deviation ( of the measurements used for the mean ) greater than 0.005 mags or 2@xmath53 the median standard deviation , whichever is greater . similarly for images , we exclude those with more than 2 magnitudes of extinction or for which the deviation greater of the zero points per star are than 0.075 mags or 2@xmath53 the median value , whichever is greater .
these cuts are somewhat conservative to limit us to only good measurements .
the images and stars rejected above are not used to calculate the system of zero points and mean magnitudes .
these cuts are updated several times as the iterations proceed .
after the iterations have completed , the images which have been reject are calibrated based on their overlaps with other images .
we overweight the ubercal measurements in order to tie the relative photometry system to the ubercal zero points .
ubercal images and measurements from those images are not allowed to float in the relative photometry analysis .
detections from the ubercal images are assigned weights of 10x their default ( inverse - variance ) weight .
the calculation of the formal error on the mean magnitudes propagates this additional weight , so that the errors on the ubercal observations dominates where they are present .
@xmath54 @xmath55 the calculation of the relative photometry zero points is performed for the entire @xmath0 data set in a single , highly parallelized analysis . as discussed above ,
the measurement and object data in the dvo database are distributed across a large number of computers in the ipp cluster : for pv3 , 100 parallel hosts are used .
these machines by design control data from a large number of unconnected small patches on the sky , with the goal of speeding queries for arbitrary chunks of the sky . as a result , this parallelization is entirely inappropriate as the basis of the relative photometry analysis . for the relative photometry calculation ( and later for relative astrometry calculation ) , the sky is divided into a number of large , contiguous regions each bounded by lines of constant ra & dec , 73 regions in the case of the pv3 analysis .
a separate computer , called a `` region host '' is responsible for each of these regions : that computer is responsible for calculating the mean magnitudes of the objects which land within its region and for determining the exposure zero points for exposures for which the center of the exposure lands in the region of responsibility .
the iterations described above ( calculate mean magnitudes , calculate zero points , calculate new measurements ) are peformed on each of the 73 region hosts in parallel .
however , between certain iteration steps , the region hosts must share some information .
after mean object magnitudes are calculated , the region hosts must share the object magnitudes for the objects which are observed by exposures controlled by neighboring region hosts .
after image calibrations have been determined by each region host , the image calibrations must be shared with the neighboring region hosts so measurement values associated with objects owned by a neighboring region host may be updated .
the completely work flow of the all - sky relative photometry analysis starts with an instance of the program running on a master computer .
this machine loads the image database table and assigns the images to the 73 region hosts .
a process is then launched on each of the region hosts which is responsible for managing the image calibration analysis on that host .
these processes in turn make an intial request of the photometry information ( object and measurement ) from the 100 parallel dvo partition machines . in practice ,
the processes on the the region hosts are launched in series by the master process to avoid overloading the dvo partition machines with requests for photometry data from all region hosts at once .
once all of the photometry has been loaded , the region hosts perform their iterations , sharing the data which they need to share with their neighbors and blocking while they wait for the data they need to receive from their neighbors .
the management of this stage is performed by communication between the region host . at the end of the iterations ,
the regions hosts write out their final image calibrations .
the master machine then loads the full set of image calibrations and then applies these calibrations back to all measurements in the database , updating the mean photometry as part of this process .
the calculations for this last step are performed in parallel on the dvo partition machines . with the above software , we are able to perform the entire relphot analysis for the full 3@xmath3 region at once , avoiding any possible edge effects .
the region host machines have internal memory ranging from 96 gb to 192 gb .
regions are drawn , and the maximum allowed density was chosen , to match the memory usage to the memory available on each machine .
a total of 9.8 tb of ram was available for the analysis , allowing for up to 6000 objects per square degree in the analysis .
consistency of photometry measurements across the sky .
each panel shows a map of the standard deviation of photometry residuals for stars in each pixel .
the median value of the measure standard deviations across the sky is @xmath56 millimags .
these values reflect the typical single - measurement errors for bright stars . ] for pv3 , the relphot analysis was performed two times .
the first analysis used only the flat - field corrections determined by the ubercal analysis , with a resolution of 2x2 flat - field values for each gpc1 chip ( corresponding to 2400 pixels ) , and 5 separate flat - field seasons. however , we knew from prior studies that there were significant flat - field structures on smaller scales .
we used the data in dvo after the initial relphot calibration to measure the flat - field residual with much finer resolution : 124 x 124 flat - field values for each gpc1 chip ( 40x40 pixels per point ) .
we then used to apply this new flat - field correction , as well as the ubercal flat - field corrections , to the data in the database . at this point , we re - ran the entire relphot analysis to determine zero points and to set the average magnitudes .
figure [ fig : photflat ] shows the high - resolution photometric flat - field corrections applied to the measurements in the dvo database .
these flat - fields make low - level corrections of up to 0.03 magnitudes .
several features of interest are apparent in these images .
first , at the center of the camera is an important structure caused by the telescope optics which we call the `` tent '' . in this portion of the focal plane ,
the image quality degrades very quickly .
the photometry is systematically biased because the point spread function model can not follow the real changes in the psf shape on these small scales .
as is evident in the image , the effect is such that the flux measured using a psf model is systematically low , as expected if the psf model is too small .
the square outline surrounding the `` tent '' is due to the 2@xmath532 sampling per chip used for the ubercal flat - field corrections .
the imprint of the ubercal flat - field is visible throughout this high - resolution flat - field : in regions where the underlying flat - field structure follows a smooth gradient across a chip , the ubercal flat - field partly corrects the structure , leaving behind a saw - tooth residual .
the high - resolution flat - field corrects the residual structures well .
especially notable in the bluer filters is a pattern of quarter circles centered on the corners of the chips .
these patterns are similar to the `` tree rings '' reported by the des team and others ( g. berstein ref & refs ) .
the details of these tree rings are beyond the scope of this article , and will be explored in future work . unlike the tree ring features discussed by these other authors ,
the features observed in the gpc1 photometry are not caused by an interaction of the flat - field with the effective pixel geometry .
instead , these photometric features are due to low - level changes in the psf size which we attribute to variable charge diffusion ( magnier in prep ) .
other features include some poorly responding cells ( e.g. , in xy14 ) and effects at the edges of chips , possibly where the psf model fails to follow the changes in the psf . for stacks and warps ,
the image calibrations were determined after the relative photometry was performed on the individual chips . each stack and each warp
was tied via relative photometry to the average magnitudes from the chip photometry . in this case , no flat - field corrections were applied . for the stacks , such a correction would not be possible after the stack has been generated because multiple chip coordinates contribute to each stack pixel coordinate . for the warps ,
it is in principle possible to map back to the corresponding chip , but the information was not available in the dvo database , and thus it was not possible at this time to determine the flat - field correction appropriate for a given warp .
this latter effect is one of several which degrade the warp photometry compared to the chip photometry at the bright end .
figure [ fig : allsky.photom.sigma ] shows the standard devitions of the mean residual photometry for bright stars as a function of position across the sky . for each pixel in these images
, we selected all objects with ( 14.5 , 14.5 , 14.5 , 14.0 , 13.0 ) @xmath57 ( @xmath58 ) @xmath57 ( 17 , 17 , 17 , 16.5 , 15.5 ) , with at least 3 measurements in @xmath59-band ( to reject artifacts detected in a pair of exposures from the same night ) , with @xmath60 ( to reject excessively - masked objects ) , and with @xmath61 ( to reject galaxies ) .
we then generated histograms of the difference between the average magnitude and the apparent magnitude in an individual image for each filter for all stars in a given pixel in the images . from these residual histograms
, we can then determine the median and the 68%-ile range to calculate a robust standard deviation .
this represents the bright - end systematic error floor for a measurement from a single exposure .
the standard deviations are then plotted in figure [ fig : allsky.photom.sigma ] .
the 5 panels in figure [ fig : allsky.photom.sigma ] show several features .
the galactic bulge is clearly seen in all five filters , with the impact strongest in the reddest bands .
we attribute this to the effects of crowding and contamination of the photometry by neighbors .
large - scale , roughly square features 10 degrees on a side in these images can be attributed to the vagaries of weather : these patches correspond to the observing chunks .
these images include both photometric and non - photometric exposures .
it seems plausible that the non - photometric images from relatively poor quality nights elevate the typical errors . on small scales , there are circular patterns 3 degrees in diameter corresponding to individual exposures ; these represent residual flat - fields structures not corrected by our stellar flat - fielding .
the median of the standard deviations in the five filters are @xmath62 millimagnitudes .
map of the amplitude of the koppenhfer effect on chips across the focal plane . in the affected chips ,
bright stars are up to 0.2 deviant from their expected positions . *
bottom left * x - direction before correction . * bottom right * y - direction before correction . *
top left * x - direction after correction .
* top right * y - direction after correction . ]
once the full pv3 dataset loaded into the master pv3 dvo database , along with supporting databases , and the photometric calibrations were performed , relative astrometry could be performed on the database to improve the overall astrometric calibration . in many respects
the relative astrometric analysis is similar to the relative photometric analysis : the repeated measurements of the same object in different images are used to determine a high quality average position for the object .
the new average positions are then used to determine improved astrometric calibrations for each of the images .
these improved calibrations are used to set the observed coordinates of the measurements from those images , which are in turn used to improve the average positions of the objects .
the whole process is repeated for several iterations . like the photometric analysis
, the astrometric analysis is performed in a parallel fashion with the same concept that specific machines are responsible for exposures and objects which land within their regions of responsibility , defined on the basis of lines of constant ra and dec . between iteration steps
, the astrometric calibrations are shared between the parallel machines as are the improved positions for objects controlled by one machine but detect in images controlled by another machine . like the photometric analysis ,
the entire sky is processed in one pass . however , there are some important differences in the details .
first , the astrometric calibration has a larger number of systematic effects which must be performed .
these consist of : 1 ) the koppenhffer effect , 2 ) differential chromatic refraction , 3 ) static deviations in the camera .
we discuss each of these in turn below .
the koppenhffer effect was first identified in february 2011 by johannes koppenhffer ( mpe ) as part of the effort to search for planet transists in the stellar transit survey data .
he noticed that the astromety of bright stars and faint stars disagreed on overlapping chips at the boundary between the sts fields . after some exploration
, it was determined that the x coordinate of the brightest stars was offset from the expected location based on the faint stars for a subset of the gpc1 chips .
the essence of the effect was that a large charge packet could be drawn prematurely over an intervening negative serial phase into the summing well , and this leakage was proportionately worse for brighter stars .
the brighter the star , the more the charge packet was pushed ahead on the serial register .
the amplitude of the effect was at most @xmath63 , corresponding to a shift of about one pixel .
this effect was only observed in 2-phase ota devices , with 22 / 30 of these suffering from this effect . by adjusting the summing well high voltage down from a default + 7 v to + 5.5v on the 2-phase devices ,
the effect was prevented in exposures after 2011 - 05 - 03 .
however , this left 101,550 exposures ( 27% ) already contaminated by the effect .
we measured the koppenhffer effect by accumulating the residual astrometry statistics for stars in the database . for each chip
, we measured the mean x and y displacements of the astrometric residuals as function of the instrumental magnitude of the star divided by the fwhm@xmath64 .
we measured the trend for all chips in a number of different time ranges and found the effect to be quite stable , in the period where it was present .
the effect only appeared in the serial direction .
figure [ fig : koppenhoefer ] shows the ke trend for a typical affected chip both before and after the correction .
for the pv3 dataset , we re - measured the ke trends using stars in the galactic pole regions after an initial relative astrometry calibration pass : the galactic pole is necessary because the real - time astrometric calibration relies largely on the fainter stars which are not affected by the ke .
the trend is then stored in a form which can be applied to the database measurements .
differential chromatic refraction ( dcr ) affects astrometry because the reference stars used the calibrate the images are not the same color ( sed ) as the rest of the stars in the image .
for a given star of a color different from the reference stars , as exposures are taken at higher airmass , the apparent position of the star will be biased along the parallactic angle .
while it is possible to build a model for the dcr impact based on the filter response functions and atmospheric refraction , we have instead elected to use an empirical correction for the dcr present in the pv3 database .
we have measured the dcr trend using the astrometric residuals of millions of stars after performing an initial relative astrometry calibration .
we define a blue dcr color ( @xmath65 ) to be used when correcting the filters , , , and a red dcr color ( @xmath66 ) to be used when correcting the filters @xmath67 . in the process of performing the relative astrometry calibration ,
we record the median red and blue colors of the reference stars used to measure the astrometry calibration for each image . as we determine the astrometry parameters for each object in the database , we record the median red and blue reference star colors for all images used to determine the astrometry for a given object . for each star in the database , we know both the color of the star and the typical color of the reference stars used to calibrate the astrometry for that star .
we measure the mean deviation of the residuals in the parallactic angle direction and the direction perpendicular to the parallactic angle . for each filter , we determine the dcr trend as a function of the difference between the star color and the reference star color , using the red or blue color approriate to the particular filter , times the tangent of the zenith distance .
figure [ fig : dcr ] shows the dcr trend for the 5 filters , as well as the measured displacement in the direction perpendicular to the parallactic angle .
we represent the trend with a spline fitted to this dataset .
example of the dcr trend in the g - band . *
top : * dcr trend in the parallactic direction * bottom : * dcr trend perpendicular to the parallactic angle . ] the amplitude of the dcr trend in the five filters is @xmath68 arcsec airmass@xmath69 magntiude@xmath69 .
we saturate the dcr correction if the term @xmath70 for a given measurement is outside a range where the dcr correction is well measured .
the maximum dcr correction applied to the five filters is @xmath71 arcseconds .
after correction for both ke and dcr , we observe persistent residual astrometric deviations which depend on the position in the camera .
we construct an astrometric `` flat - field '' response by determining the mean residual displacement in the x and y ( chip ) directions as a function of position in the focal plane .
we have measured the astrometric flat using a sampling resolution of 40x40 pixels , matching the photometric flat - field correction images .
figures [ fig : astroflat.gri ] and [ fig : astroflat.zy ] show the astrometric flat - field images for the five filters in each of the two coordinate directions .
these plots show several types of features . the dominant pattern in the astrometric residual is roughly a series of concentric rings .
the pattern is similar to the pattern of the focal surface residuals measured by ( ref ) , which also has a concentric series of rings with similar spacing .
the `` tent '' in the center of the focal surface reflected in these astrometry residual plots .
our interpretation of the structure is that the deviations of the focal plane from the ideal focal surface introduces small - scale psf changes , presumably coupled to the optical aberrations , which result in small changes in the centroid of the object relative to the psf model at that location . since the psf model shape parameters are only able to vary at the level of a 6x6 grid per chips , the finer structures are not included in the psf model .
the pv2 analysis shows the ring structure more clearly , with a pattern much more closely following the focal surface deviations . in the pv2 analysis , the psf model used at most a 3x3 grid per chip to follow the shape variations ,
so any changes caused by the optical aberrations would be less well modeled in the pv2 analysis , as we observe .
a second pattern which is weakly seen in several chips consists of consistent displacements in the x ( serial ) direction for certain cells .
this effect can be seen most clearly in chips xy45 and xy46 . in the pv2 analysis ,
this pattern is also more clearly seen . in this case
, the fact that the astrometric model used polynomials with a maximum of 3rd order per chip means the deviation of individual cells can not be followed by the astrometric model .
a third effect is seen at the edge of the chips , where there appears to be a tendency for the residual to follow the chip edge .
the origin of this is unclear , but likely caused by the astrometry model failing to follow the underlying variations because of the need to extrapolate to the edge pixels .
finally , we also mention an interesting effect _ not _ visible at the resolution of these astrometric flat - field images .
fine structures are observed at the 10 pixel scale similar to the `` tree rings '' reported by the des team and others ( g. berstein ref & refs ) .
the details of these tree rings are beyond the scope of this article , and will be explored in future work .
unfortunately , we discovered a problem with the astrometric flat - field correction too late to be repaired for dr1 .
as can be seen by inspection of figures [ fig : astroflat.gri ] and [ fig : astroflat.zy ] , there is significant pixel - to - pixel noise in the the astrometric flat - field images .
this pixel - to - pixel noise is caused by too few stars used in the measuremnt of the flat - field structure for the high - resolution sampling . as a result ,
the astrometric flat - field correction reduces systematic structures on large spatial scales , but at the expense of degrading the quality of an individual measurement .
only @xmath59-band has sufficient signal - to - noise per pixel to avoid significantly increasing the per - measurement position errors .
figure [ fig : allsky.astrom.sigma ] shows the standard devitions of the mean residual astrometry in @xmath72 for bright stars as a function of position across the sky . for each pixel in these images , we selected all objects with @xmath73 , with at least 3 measurements in @xmath59-band ( to reject artifacts detected in a pair of exposures from the same night ) , with @xmath60 ( to reject excessively - masked objects ) , and with @xmath74 ( to reject galaxies ) .
we then generated histograms of the difference between the object position predicted for the epoch of each measurement ( based on the proper motion and parallax fit ) and the observed position of that measurement , in both the right ascension and declination directions ( in linear arcseconds ) , for all stars in a given pixel in the images . from these residual histograms
, we can then determine the median and the 68%-ile range to calculate a robust standard deviation .
this represents the bright - end systematic error floor for a measurement from a single exposure .
the standard deviations are then plotted in figure [ fig : allsky.photom.sigma ] .
the median value of the standard deviations across the sky is @xmath75 milliarcseconds .
the galactic plane is clearly apparently in these images . like photometry
, we attribute this to failure of the psf fitting due to crowding .
the celestial north pole regions have somewhat elevated errors in both r.a . and dec .
this may be due to the larger typical seeing at these high airmass regions , but without further exploration this is interpretation uncertain .
several features can be seen which appear to be an effect of the tie to the gaia astrometry : the stripes near the center of the dec image and the right side of the r.a .
the mesh of circular outlines is due to the outer edge of the focal plane where the astrometric calibration is poorly determined . as discussed above ,
the median values in the images are higher than expected based on our pv2 analysis of the astrometry : the median per - measurement error floor of 22 mas is significantly worse than the 17 mas value in that earlier analysis .
we attribute this degradation to the noise introduced by the astrometric flat - field .
this noise can likely be addressed before the dr2 release of the individual measurement data .
consistency of photometry measurements across the sky .
each panel shows a map of the standard deviation of astrometry residuals for stars in each pixel .
the median value of the standard deviations across the sky is @xmath75 milliarcseconds .
these values reflect the typical single - measurement errors for bright stars .
see discussion regarding the astrometric flat which is likely responsible for these elevated value . ] after the initial analysis to measure the ke corrections , dcr corrections , and astrometric flat - field corrections , we applied these corrections to the entire database . within the schema of the database , each measurement has the raw chip coordinates ( ) as well as the offset for that object based on each of these three corrections : .
the offsets are calculated for each measurement based on the observed instrumental chip magnitudes and fwhm for the koppenhoffer effect , on the average chip colors and the altitude & azimuth of each measurement for the dcr correction , and on the chip coordinates for the astrometric flat - field corrections .
the corrections are combined and applied to the raw chip coordinates and saved back in the database in the fields . at this point , we are ready to run the full astrometric calibration .
the initial analysis of the pv2 astrometry used the 2mass positions as an inertial constraint : the 2mass coordiates were included in the calculation of the mean positions for the objects in the database , with weight corresponding to the reported astrometric errors . in this analysis , the object positions used to determine the calibrations of the image parameters ignored proper motion and parallax . after the image calibrations were determined , then individual objects were fitted for proper motion and possibly parallax , as discussed in detail below . using the pv2 analysis of the astrometry calibration
, we discovered large - scale systematic trends in the reported proper motions of background quasars .
this motion had an amplitude of 10 - 15 milliarcseconds per year and clear trends with galactic longitude .
we also observed systematic errors of the mean positions with respect to the icrf milliarcsecond radio quasar positions , with an amplitude of 60 milliarcseconds , again with trends associated with galactic longitude .
since the 2mass data were believed to have minimal average deviations relative to the icrf quasars , this latter seemed to be a real effect .
we realized that both the proper motion and the mean position biases could be caused by a single common effect : the proper motion of the stars used as reference stars between the 2mass epoch ( 2000 ) and ps1 epoch ( 2012 ) .
since we are fitting the image calibrations without fitting for the proper motions of the stars , we are in essencence forcing those stars to have proper motions of 0.0 .
the background quasars would then be observed to have proper motions corresponding to the proper motions of the reference stars , but in the opposite direction .
we demonstrated that the observed quasar proper motions agreed well with the distribution expected if the median distance to our reference stars was 500 pc . for pv3
, we desired to address this bias by including our knowledge about the distances to the reference stars and the expected typical proper motions for stars at those distances . with some constraint on the distance to each star , we can determine the expected proper motion based on a model of the galactic rotation and solar motions .
we can then calculate the mean positions for the objects keeping the assumed proper motion fixed .
when calibrating a specific image , the reference star mean position is then translated to the expected position at the epoch of that image .
the image calibration is then performed relative to these predicted postions .
this process naturally accounts for the proper motion of the reference stars . in order to make the calibrations consistent with the observed coordinates of an external inertial reference
, we perform the iterative fits using the technique as described , but assign very high weights in the initial iterations to the inertial reference , and reduce the weights as the astrometric calibration iterations proceed . in order to perform this analysis , we need estimated distances for every reference star used in the analysis .
green et al ( ref ) performed sed fitting for 800 m stars in the 3@xmath3 region using pv2 data .
the goal of this work was to determine the 3d structure of the dust in the galaxy . by fitting model seds to stars meeting a basic data quality cut , they determined the best spectral type , and thus @xmath76 , absolute @xmath77-band magnitude , distance modulus , and extinction @xmath78 ( the desired output and used to determine the dust extinction as a function of distance throughout the galaxy ) .
we use the distance modulus determined in this analysis to predict the proper motions . to convert the distances to proper motions
, we use the galactic rotation parameters ( @xmath79 ) = ( 14.82,-12.37 ) km sec@xmath69 pc@xmath69 and solar motion parameters ( @xmath80 ) = ( 9.32 , 11.18 , 7.61 ) km sec@xmath69 as determined by feast & whitelock ( ref ) using hipparchos data .
proper motions are determined from the following : @xmath81 where @xmath82 is the distance and @xmath83 are the galactic coordintes of the star .
note that the proper motion induced by the galactic rotation is independent of distance while the reflex motion induced by the solar motion decreases with increasing distance .
also note that this model assumes a flat rotation curve for objects in the thin disk ; any reference stars which are part of the halo population will have proper motions which are not described by this model ; the mostly random nature of the halo motions should act to increase the noise in the measurement , but should not introduce detectable motion biases .
also , if the distance modulus is not well determined , we can assume the object is simply following the galactic rotation curve and set a fixed proper motion .
if we do not have a distance modulus from the green et al analysis , we assume a value of 500pc .
after the full relative astrometry analysis was performed for the pv3 database , the gaia data release 1 became available .
this afforded us the opportunity to constrain the astrometry on the basis of the gaia observations .
gaia dr1 objects which are bright enough to have proper motion and parallax solutions are in general saturated in the ps1 observations .
thus , we are limited to using the gaia mean positions reported for the fainter stars .
we extracted all gaia sources not marked as a duplicate from the gaia archive and generated a dvo database from this dataset .
we then merged the gaia dvo into the pv3 master dvo database .
we re - ran the complete relative astrometry analysis using gaia as an additional measurement .
we applied the analysis described above , applying the estimated distances to determine preliminary proper motions .
the gaia mean epoch is reported as 2015.0 , so all gaia measurements were assigned this epoch .
we wanted to ensure the gaia measurements dominated the astrometric solutions , so we made the weight very high for the gaia points : 1000@xmath53 the nominal weight in the initial fits ( to lock down the reference frame ) , decreasing to 100@xmath53 the nominal weight for the last fits .
we also retained the 2mass measurements in the analysis , but gave them somewhat lower weights than gaia : while the 2mass data does not have the accuracy of gaia , the coverage is known to be quite complete , while the gaia dr1 has clear gaps and holes . having 2mass , even at a lower weight ,
helps to tile over those gaps .
figure [ fig : gaia.photom ] shows a comparison between the pan - starrs photometry in @xmath84 and the gaia photometry in the @xmath85-band . to compare the ps1 photometry to the very broadband gaia g filter ,
we have determined a transformation based on a 3rd order polynomial fit to @xmath86 and @xmath65 colors .
this transformation reproduces gaia photometry reasonably well for stars which are not too red . for a comparison
, we have seleted all ps1 stars with gaia measurements meeting the following criteria : @xmath87 , with at least 10 total measurements , within a modest color range @xmath88 .
we also restricted to objects with @xmath89 , using the average @xmath59 magnitudes determined from the individual exposures . for figure [ fig : gaia.photom ]
, we calculate the difference between the estimated @xmath85-band magnitude based on ps1 @xmath84 photometry and the @xmath85-band photometry reported by gaia . for each pixel
, we determine the histogram of these differences and calculate the median and the 68%-ile range . in figure
[ fig : gaia.photom ] , these values are plotted as a color scale .
the galactic plane is clearly poorly matched between the two photometry systems .
this may in part be due to the difficulty of predicting @xmath85-band magnitudes for stars which are significantly extincted : the @xmath85-band includes significant flux from the ps1 @xmath90-band which was not used in our transformation .
many other large scale feature in the median differences have structures similar to the gaia scanning pattern ( large arcs and long parallel lines .
there are also structures related to the ps1 exposure footprint .
these show up as a mottling on the 3 degree scale ( e.g. , lower right below the galactic plane ) .
the amplitude of the residual structures is fairly modest .
the standard devition of the median difference values is 7 millimagnitudes .
this number gives an indication of the overall photometric consistency of both gaia and ps1 and implies that the systematic error floor for each survey is less than 7 millimags .
figure [ fig : gaia.astrom ] shows a comparison between the pan - starrs mean astrometry positions in @xmath91 and the gaia astrometry . for this comparison
, we have seleted all ps1 stars with gaia measurements with @xmath87 and with at least 10 total measurements .
for figure [ fig : gaia.astrom ] , we calculate the difference between the position predicted by ps1 at the gaia epoch ( using the proper motion and parallax fit ) and the position reported by gaia . for each pixel
, we determine the histogram of these differences in the r.a . and dec directions , and calculate the median and the 68%-ile range . in figure
[ fig : gaia.astrom ] , these values are plotted as a color scale .
there is good consistency between the ps1 and gaia astrometry .
there are patterns from the galactic plane ( though not very strongly at the bulge ) .
there are also clear features due to the ps1 exposure footprint ( ring structure on 3 degree scales ) . in the plots of the scatter , there are patterns which are related to the gaia scanning rule .
these are presumably regions with relatively low signal to noise in gaia ; they were also apparent in the plots of the statisics of the per - exposure measurement residuals ( figure [ fig : allsky.astrom.sigma ] .
the standard deviations of the median differences are ( @xmath92 milliarcseconds .
the pan - starrs1 surveys ( ps1 ) have been made possible through contributions of the institute for astronomy , the university of hawaii , the pan - starrs project office , the max - planck society and its participating institutes , the max planck institute for astronomy , heidelberg and the max planck institute for extraterrestrial physics , garching , the johns hopkins university , durham university , the university of edinburgh , queen s university belfast , the harvard - smithsonian center for astrophysics , the las cumbres observatory global telescope network incorporated , the national central university of taiwan , the space telescope science institute , the national aeronautics and space administration under grant no .
nnx08ar22 g issued through the planetary science division of the nasa science mission directorate , the national science foundation under grant no .
ast-1238877 , the university of maryland , and eotvos lorand university ( elte ) and the los alamos national laboratory . | the pan - starrs1 @xmath0 survey has produced photometry and astrometry covering the 30,000 square degrees @xmath1 .
this article describes the photometric and astrometric calibration of this survey . |
the study of high - frequency collective dynamics in liquids and glasses at microscopic length scales and terahertz frequency region has been a subject of intense investigations in the past decade . in particular , recent development of the inelastic x - ray scattering ( ixs )
technique has renewed the interest in a long - standing issue of sound propagation @xcite .
it is now generally accepted that a well - defined sound - like oscillatory mode called the high - frequency or fast sound is supported in liquids outside the strict hydrodynamic region , down to wavelengths of a few interparticle distances , but with a sound velocity larger than the hydrodynamic value .
traditionally , such increase of the sound velocity upon increase of the wave number called the anomalous or positive dispersion has been interpreted within the so - called viscoelastic model @xcite . in this model ,
relevant memory kernel in the zwanzig - mori or memory - function equation for the dynamics structure factor @xmath0 is modeled by an exponential function with a single time constant @xmath1 reflecting structural relaxation in liquids .
an increase of the sound velocity is predicted at the wave number @xmath2 where the condition @xmath3 is fulfilled with the resonance frequency @xmath4 of the dynamic structure factor .
this condition marks the transition from the low - frequency viscous behavior to the high - frequency elastic behavior of the liquid , and thus , the positive dispersion effect is ascribed to the transition between a liquid - like to a solid - like response of the system . according to the viscoelastic model , the absence of the positive dispersion
is naturally expected in glasses where the structural relaxation is basically frozen .
however , molecular dynamics ( md ) simulations have demonstrated the presence of the positive dispersion in model glasses @xcite .
a first experimental confirmation has been reported recently from an ixs study on vitreous silica @xcite .
the observed anomalous dispersion implies the presence of some additional `` relaxation '' process which is active even in glasses , and stimulates an extension of the viscoelastic model to fully account for the anomalies present both in liquids and glasses .
such extension was also motivated by recent ixs studies on simple liquid metals near the melting temperature @xcite . from a detailed line - shape analysis of @xmath0 ,
the viscoelastic model was found to be unable to account for inelastic peaks in these systems .
it was found , instead , that a two - time - scale model for the memory kernel successfully describes the observed spectral shapes .
thus , a convincing experimental demonstration is provided for the simultaneous presence two relaxation processes with slow and fast characteristic time scales , termed structural ( labeled @xmath5 ) and microscopic ( labeled @xmath6 ) processes .
furthermore , it was pointed out that the faster @xmath6 process is the major controller of the positive dispersion observed in these systems since it was always found that @xmath7 for the slower time scale @xmath8 , whereas @xmath9 becomes close to unity at some wave number for the faster time scale @xmath10 . according to this experimental observation
, the positive dispersion can not fully be ascribed to the structural relaxation .
the nature of these two processes have been studied in more detail in ref .
@xcite based on the md simulation for a model of lithium . from the analysis of the simulated @xmath0 with the two - time - scale model , a strong temperature dependence was observed for the time scale @xmath8 , corroborating that the slower process is associated with the structural relaxation . on the other hand ,
the faster @xmath6 process was found to persist both in liquid and glass phases with almost no temperature dependence of its relaxation time @xmath10 .
the persistence of the @xmath6 relaxation accounts for the presence of the positive dispersion in glasses .
of course , a natural question would be what is relaxing at that low @xmath11 , i.e. , what is the microscopic nature of the @xmath6 dynamics .
this was studied in ref .
@xcite based on the computer simulation for a harmonic glass .
there , it was suggested that the origin of this process can be ascribed to the topological disorder , i.e. , the decay of the memory kernel is due to the dephasing of different oscillatory components in the force fluctuations .
besides the peak associated with the high - frequency sound , the md @xcite and ixs @xcite works showed the presence of additional low - frequency excitation in the dynamic structure factors @xmath0 or in the related longitudinal current spectra @xmath12 .
this second low - frequency excitation also called the boson peak exhibits some general characteristics : it appears in the spectra at @xmath2 values larger than some fraction of the structure - factor - peak position , and its resonance position is weakly @xmath2-dependent .
it was conjectured that the low - frequency excitation reflects the transversal dynamics , and the appearance of the transversal mode in the longitudinal current spectra was interpreted as being due to `` mixing '' phenomena caused by microscopic disorder @xcite .
the main evidence supporting this conjecture lies in the simulation results that the low - frequency excitation appears at the same resonance frequency in both of the transversal and longitudinal current spectra , but its intensity is more enhanced in the former .
another support was inferred from studies on the density dependence . the computational work in ref .
@xcite demonstrated that the resonance frequency of the low - frequency excitation in the transversal current spectra increases with increasing density .
experimentally , it is observed that upon densification the boson - peak energy shifts to higher energy and its intensity strongly decreases @xcite .
this parallel of the density dependence was claimed to also suggest that the boson - peak - like low - frequency excitation arises from the transversal dynamics . in this paper , we present a microscopic and unified explanation of the mentioned features of the high - frequency dynamics in liquids and glasses the anomalous dispersion of the high - frequency sound velocity and the nature of the low - frequency excitation which have so far been investigated somewhat independently .
this will be done based on the mode - coupling theory ( mct ) for ideal liquid - glass transition @xcite .
originally , the mct was developed to deal with structural - relaxation processes which evolve as precursor of the glass transition .
on the other hand , the glass transition also modifies the short - time or the high - frequency dynamics .
the study of these modifications was the main subject in ref .
it was shown there that the strong interaction between density fluctuations at microscopic length scales and the arrested glass structure causes an anomalous - oscillation peak ( aop ) , which exhibits the properties of the boson peak . as will be demonstrated in the present study , the aop persists in liquid states as well as in glass states .
it will be discussed how the sound mode interferes with other low - lying modes present in the system .
thereby , we provide a systematic explanation of the anomalous sound - velocity dispersion in systems ranging from high temperature liquid down to deep inside the glass state in terms of the contributions from the structural - relaxation processes and from the vibrational modes building the aop . a possibility of observing negative dispersion
the _ decrease _ of the sound velocity upon increase of the wave number is argued when the sound - velocity dispersion is dominated by the contribution from the vibrational dynamics .
we also show that the low - frequency excitation , observable in both of the glass - state longitudinal and transversal current spectra at the same resonance frequency , is the manifestation of the aop .
this seems to be an alternative interpretation to the mixing phenomena mentioned above . as a natural consequence of the presence of the aop in the transversal current spectra
, it is predicted that the transversal sound velocity also exhibits the anomalous dispersion .
the paper is organized as follows . in sec .
[ sec : theory ] , the basic mct equations of motion are formulated , and the model for the demonstration of our theoretical results is specified .
section [ sec : sound ] deals with the anomalous sound - velocity dispersion of the high - frequency sound , while in sec .
[ sec : transversal ] we discuss features of glass - state longitudinal and transversal current spectra at low frequencies . section [ sec : conclusions ] concludes the paper .
the basic quantity describing the equilibrium structure of a simple system is the static structure factor @xmath13 defined in terms of the density fluctuations for wave vector @xmath14 , @xmath15 . here
, @xmath16 is the total number of particles in the system distributed with the average density @xmath17 , and @xmath18 denotes the position of the @xmath19th particle . for homogeneous and isotropic system ,
which we assume throughout the paper , the static structure factor depends only on the modulus @xmath20 .
the most relevant variables characterizing the structural changes as a function of time @xmath21 are the density correlators @xmath22 which are normalized to unity at @xmath23 .
the short - time asymptote of these functions is given by @xmath24 , where @xmath25 denotes the square of the characteristic frequency @xmath26 @xcite . here
@xmath27 with boltzmann s constant @xmath28 denotes the thermal velocity of particle of mass @xmath29 at temperature @xmath11 . in the small wave - vector limit ,
one obtains @xmath30 with the isothermal sound velocity @xmath31 . within the zwanzig - mori formalism @xcite
one can derive the following exact equation of motion [ eq : gle - phi ] @xmath32 in which the relaxation kernel @xmath33 is a correlation function of fluctuating forces .
let us introduce fourier - laplace transformations to map @xmath34 , @xmath35 , and similar functions from the time domain onto the frequency domain according to the convention @xmath36 .
equation ( [ eq : gle - phi - a ] ) is equivalent to the representation @xmath37 \ , \}. \label{eq : gle - phi - b}\ ] ] within the mct , the kernel @xmath33 is given by a mode - coupling functional of density correlators describing the cage effect of dense systems , [ eq : mct - phi ] @xmath38 .
\label{eq : mct - phi - a}\ ] ] here the mode - coupling functional reads @xmath39 = 1/(2\pi)^{3 } \int d{\vec k } \ , v({\vec q } \ , ; { \vec k } , { \vec p } \ , ) \ , \tilde{f}_{k } \tilde{f}_{p}$ ] with @xmath40 .
the coupling coefficients @xmath41 are determined by the equilibrium structure . within the convolution approximation for triple correlations ,
they are given by @xmath42^{2 } \ , / \ , ( 2 q^{4})$ ] , where @xmath43 denotes the direct correlation function related to @xmath44 via the ornstein - zernike equation , @xmath45 @xcite . for details ,
the reader is referred to ref .
@xcite . in the following
, wave - vector integrals will be approximated by riemann sums . the mode - coupling functional
can then be written as a quadratic polynomial @xmath46 = \sum_{k , p } v_{q , kp } \ , \tilde{f}_{k }
\tilde{f}_{p } , \label{eq : mct - phi - b}\ ] ] where the indices run over the discretized wave - number grids .
the coefficients @xmath47 are trivially related to @xmath48 .
the @xmath49 limit of the functional reads @xmath50 = \sum_{k } v_{k } \tilde{f}_{k}^{2 } , \label{eq : mct - phi - c}\ ] ] with @xmath51 $ ] .
equations ( [ eq : gle - phi ] ) and ( [ eq : mct - phi ] ) are closed , provided equilibrium static quantities are known as input . in the following , the results will be demonstrated for the lennard - jones ( lj ) system : the interaction between two particles is given by the lj interaction @xmath52 .
the first mct work for the lj system has been done in refs .
@xcite with the optimized random - phase approximation for @xmath44 @xcite . in the present study
, @xmath44 is evaluated within the percus - yevick approximation @xcite .
wave numbers will be considered up to a cutoff value @xmath53 , and they are discretized to @xmath54 grid points . from here on , all quantities are expressed in reduced units with the unit of length @xmath55 , the unit of energy @xmath56 ( setting @xmath57 ) , and the unit of time @xmath58 . the dynamics by varying @xmath11 shall be considered for a fixed density @xmath59 , except for sec .
[ sec : transversal-2 ] where the dynamics at a lower density @xmath60 is discussed .
( we notice that the triple point of the lj system is at @xmath61 and @xmath62 @xcite . )
the specified model exhibits a fold bifurcation @xcite . for small coupling constants @xmath47 , the correlators @xmath63 and the kernels @xmath33 decay to zero for long times , @xmath64 and @xmath65 . in this case , the spectra @xmath66 and @xmath67 are continuous in @xmath68 . in particular , there hold @xmath69 and @xmath70 .
density fluctuations created at time @xmath71 disappear for long times and the same holds for the force fluctuations as expected for an ergodic liquid . for large coupling constants , on the other hand , there is arrest of density fluctuations for long times : @xmath72 , @xmath73 .
thus , nonergodic dynamics is obtained in which the perturbed system does not return to the equilibrium state .
similarly , there is arrest of the force fluctuations , @xmath74 .
the nonergodicity parameters @xmath75 and @xmath76 are connected via @xcite @xmath77 .
\label{eq : dw}\ ] ] for this strong - coupling solution , the kernel exhibits a zero - frequency pole , @xmath78 . the fluctuation spectrum @xmath79 exhibits a strictly elastic peak : @xmath80 .
this is the signature for a solid with @xmath75 denoting its debye - waller factor .
hence , the strong coupling solution deals with a disordered solid ; it is a model for an ideal glass state .
if one increases the coupling constants smoothly from small to large values , one finds a singular change of the solution from the ergodic liquid to the nonergodic glass state , i.e. , an idealized liquid - glass transition . for simple - liquid models ,
the transition occurs upon cooling at some critical temperature @xmath81 or upon compression at some critical density @xmath82 @xcite . for the lj model under study where the temperature is considered as a control parameter
, one finds @xmath83 for @xmath59 within the percus - yevick approximation for @xmath44 . if @xmath11 decreases below @xmath81 , @xmath75 increases above its value at the critical point , called the critical nonergodicity parameter or the plateau @xmath84 , as demonstrated in fig .
[ fig : fq - sq ] .
and one - tenth of the static structure factor @xmath44 at @xmath59 for @xmath85 ( solid lines ) , @xmath86 ( dashed lines ) , and @xmath87 ( dotted lines ) . ]
as precursor of the ideal liquid - glass transition , mct predicts the evolution of the glassy dynamics which are stretched over many decades .
such mct scenario has been comprehensively discussed for the hard - sphere system in refs .
this subsection compiles the mct universal predictions for the dynamics near the liquid - glass transition to an extent which is necessary for understanding the present article .
we start from introducing some concepts to describe the mct - liquid - glass - transition dynamics @xcite . in the space of control parameters , a smooth function @xmath88 is defined near the transition point , called the separation parameter .
glass states are characterized by @xmath89 , liquid states by @xmath90 , and @xmath91 defines the transition point . when only the temperature @xmath11 is considered as a control parameter near the transition point
, one can write for small distance parameters @xmath92 : @xmath93 , @xmath94 .
in addition , the transition point is characterized by a time scale @xmath95 and by a number @xmath96 , @xmath97 .
the scale @xmath95 specifies properties of the short - time transient dynamics , and @xmath96 is called the exponent parameter .
the latter determines a certain number @xmath98 , the critical exponent @xmath99 , and the von - schweidler exponent @xmath100 . for the specified lj model at @xmath59 , we have @xmath101 , @xmath102 , @xmath103 , @xmath104 , @xmath105 , and @xmath106 .
let us consider the correlator @xmath107 of some variable @xmath108 coupling to the density fluctuations .
its nonergodicity parameter @xmath109 exhibits a square - root singularity near the transition , @xmath110 the critical nonergodicity parameter @xmath111 and the critical amplitude @xmath112 are equilibrium quantities to be calculated from the relevant mode - coupling functionals at the critical point @xmath81 . at the transition ,
the correlator exhibits a power - law decay that is specified by the critical exponent @xmath113 . in a leading - order expansion in @xmath114 ,
one gets @xmath115 for small values of @xmath116 , there is a large time interval , where @xmath117 is close to its critical nonergodicity parameter @xmath111 .
solving the equations of motion asymptotically for this plateau regime , one gets in leading order in the small quantities @xmath118 the factorization theorem : [ eq : beta - correlator ] @xmath119 the function @xmath120 , called the @xmath121 correlator , is the same for all variables @xmath108 , and describes the complete dependence on time and control parameter via the first scaling law @xmath122 here , @xmath123 is the first critical time scale .
the master functions @xmath124 are determined by @xmath96 @xcite . for @xmath125 ,
the master functions are the same on both sides of the transition point , and one gets the power law @xmath126 so that eq .
( [ eq : critical - x ] ) is reproduced for fixed large @xmath21 and @xmath88 tending to zero . for @xmath127 ,
one gets for glasses @xmath128 , so that eq . (
[ eq : plateau - x ] ) is reproduced . on the other hand , for liquids
, one finds for the large rescaled time @xmath129 : @xmath130 . substituting this result into eqs .
( [ eq : beta - correlator ] ) , one obtains von schweidler s law for the decay of the liquid correlator below the plateau @xmath111 : @xmath131 the control - parameter dependence is given via the second critical time scale @xmath132 with @xmath133 . the dynamical process described by the cited results
is called the @xmath121-process .
the @xmath121 correlator @xmath120 describes in leading order the decay of the correlator towards the plateau @xmath111 within the interval @xmath134 . for @xmath135 , the glass correlator arrests at @xmath136 , while the liquid correlator exhibits von schweidler s law . the decay of the liquid correlator below the plateau @xmath111 is called the @xmath5-process . for this process
, there holds in leading order for @xmath137 the second scaling law , @xmath138 with @xmath139 , which is also referred to as the superposition principle .
the control - parameter - independent shape function @xmath140 is to be evaluated from the mode - coupling functionals at the critical point . for short rescaled times @xmath141 ,
one gets @xmath142 , so that eq .
( [ eq : phi - x - von ] ) is reproduced .
the ranges of applicability of the first and the second scaling laws overlap ; both scaling laws yield von schweidler s law for @xmath143 . for later convenience
, we introduce a reformulation of the mct equations of motion which is more appropriate in handling the dynamics in glass states @xcite .
we map the density correlators @xmath34 to new ones @xmath144 by @xmath145 this amounts to dealing only with the decay relative to the frozen amorphous structure described by @xmath75 . introducing new characteristic frequencies @xmath146 by @xmath147 one obtains for short times @xmath148 in analogy to the one for @xmath63 . in the small wave - vector limit , there holds @xmath149 with @xmath150 .
the modification of the sound velocity reflects the reduced compressibility in nonergodic glass states .
substitution of these results into eq .
( [ eq : gle - phi - a ] ) reproduces the mct equations of motion with @xmath34 , @xmath151 , and @xmath35 replaced by @xmath144 , @xmath146 , and @xmath152 , respectively . here
, the new relaxation kernel is related to the original one by @xmath153 the new correlator has a vanishing long - time limit , @xmath154 , and correspondingly the fourier - laplace transform exhibits a regular zero - frequency behavior @xmath155 . combining eqs .
( [ eq : dw ] ) and ( [ eq : gle - hat - phi - c ] ) , one also concludes that @xmath156 and @xmath157 . since the equation of motion does not change its form , one gets as the analog of eq .
( [ eq : gle - phi - b ] ) @xmath158 \ , \}. \label{eq : gle - hat - phi - d}\ ] ] the reformulated equation of motion for @xmath159 can be closed by finding an expression for the new kernel @xmath160 as a new mode - coupling functional @xmath161 .
one finds combining eqs .
( [ eq : gle - hat - phi - a ] ) and ( [ eq : gle - hat - phi - c ] ) with eqs .
( [ eq : mct - phi - a ] ) and ( [ eq : mct - phi - b ] ) that @xmath161 is given by a sum of linear and quadratic terms [ eq : mct - hat - phi ] @xmath162 = \hat{\cal f}^{(1)}[\hat{\phi}(t ) ] + \hat{\cal f}^{(2)}[\hat{\phi}(t ) ] , \label{eq : mct - hat - phi - a}\ ] ] where @xmath163 = \sum_{k } \hat{v}_{q , k } \tilde{f}_{k } , \quad \hat{\cal f}_{q}^{(2)}[\tilde{f } ] = \sum_{k , p } \hat{v}_{q , kp } \tilde{f}_{k } \tilde{f}_{p},\ ] ] with renormalized coefficients @xmath164 correspondingly , one finds from eq . (
[ eq : mct - phi - c ] ) for the zero wave - number limit of the kernel @xmath165 [ eq : mct - hat - phi - qnull ] @xmath166 in which linear and quadratic terms are given by @xmath167 with new coefficients @xmath168 as a result , equations of motion are produced , which are of the same form as the original mct equations ( [ eq : gle - phi ] ) and ( [ eq : mct - phi ] ) .
but in addition to the quadratic mode - coupling term , there appears a linear term , and the original mode - coupling coefficients are renormalized to hatted ones .
the linear term describes interactions of density fluctuations with the arrested amorphous structure , while the quadratic one deals with two - mode decay processes .
current correlators are also relevant variables describing the dynamics of dense systems , in particular , when one is interested in vibrational properties .
these are defined in terms of the current fluctuations , @xmath169 , where @xmath5 ( @xmath170 , @xmath171 , or @xmath172 ) refers to the spatial component and @xmath173 denotes the velocity of the @xmath19th particle .
since @xmath174 , one gets the static correlation @xmath175 . the current correlators for isotropic system
can be represented by two independent functions which depend only on the modulus @xmath2 : @xmath176 \ , \phi_{q}^{\rm t}(t ) .
\label{eq : l - t - def}\ ] ] here @xmath177 denotes the @xmath5 component of the unit vector along @xmath14 . the functions @xmath178 and @xmath179 are called the longitudinal and transversal current correlators , respectively , and they are normalized to unity at @xmath71 . the density fluctuations and the longitudinal current fluctuations are related via the continuity equation , @xmath180 , and there holds @xmath181 .
their fourier - laplace transforms therefore satisfy @xmath182 = \omega_{q}^{2 } \phi_{q}^{\rm l}(\omega)$ ] , and one obtains from eq .
( [ eq : gle - phi - b ] ) for the longitudinal current spectrum @xmath183 @xmath184^{2 } + [ \omega \ , \omega_{q}^{2 } \ , m_{q}''(\omega)]^{2 } } , \label{eq : gle - l - b}\ ] ] with @xmath185 .
the continuity equation holds also in glass states , @xmath186 in terms hatted variables defined in eqs .
( [ eq : gle - hat - phi - a ] ) and ( [ eq : gle - hat - phi - b ] ) , and their spectra are related via @xmath187 we therefore have the same expression for @xmath183 for glass states as in eq .
( [ eq : gle - l - b ] ) but with all quantities in the right - hand side replaced by hatted ones : @xmath188^{2 } + [ \omega \ , \hat{\omega}_{q}^{2 } \ , \hat{m}_{q}''(\omega)]^{2 } } \quad \mbox{(glass ) } , \label{eq : gle - l - glass}\ ] ] with @xmath189
. thus , the same relaxation kernel @xmath35 or @xmath152 as the one for the density correlator describes the dynamics of the longitudinal current .
the mct equations of motion for the transversal current correlator @xmath179 have been derived in ref .
the zwanzig - mori equation is given by [ eq : gle - t ] @xmath190 and its fourier - laplace transform reads @xmath191 with @xmath192 .
the relaxation kernel @xmath193 is a correlation function of transversal fluctuating forces . within the mct ,
the transversal fluctuating forces are approximated by their projection onto the subspace spanned by the density products , and with the factorization approximation the kernel is given by a mode - coupling functional @xmath194 of the density correlators [ eq : mct - t ] @xmath195 , \quad { \cal f}_{q}^{\rm t}[\tilde{f } ] = \sum_{k , p } v_{q , kp}^{\rm t } \
, \tilde{f}_{k } \tilde{f}_{p}. \label{eq : mct - t - a}\ ] ] here @xmath196^{2 } \ , / \ , ( 2 q^{2})$ ] with @xmath197 denoting a unit vector orthogonal to @xmath14 . the @xmath49 limit of the functional is given by @xmath198 = \sum_{k } v_{k}^{\rm t } \tilde{f}_{k}^{2 } , \label{eq : mct - t - b}\ ] ] with @xmath199^{2}$ ] .
these mct equations of motion for @xmath179 can be solved with the knowledge of the static quantities and of the density correlators . in glass states
, @xmath200 acquires a zero - frequency pole since the transversal force fluctuations do not decay to zero for long times , @xmath201 .
we therefore write in analogy to eq .
( [ eq : gle - hat - phi - c ] ) @xmath202 , \label{eq : mct - t - glass}\ ] ] and introduce @xmath203 for the description of the glass - state transversal current dynamics .
the new kernel satisfies @xmath204 and @xmath205 . substituting the fourier - laplace transform of eq .
( [ eq : mct - t - glass ] ) into eq .
( [ eq : gle - t - b ] ) yields for the transversal current spectrum @xmath206^{2 } + [ \omega ( \hat{\omega}_{q}^{\rm t})^{2 } \hat{m}_{q}^{{\rm t}''}(\omega)]^{2 } } , \label{eq : gle - t - glass - b}\ ] ] where @xmath207 and @xmath208 .
the reformulated equation of motion for the glass - state transversal current dynamics can be closed by finding an expression for @xmath203 as a new mode - coupling functional @xmath209 .
one finds from eqs .
( [ eq : gle - hat - phi - a ] ) , ( [ eq : mct - t - a ] ) , and ( [ eq : mct - t - glass ] ) that @xmath210 in analogy to eq .
( [ eq : mct - hat - phi - a ] ) . correspondingly , there holds @xmath211 for @xmath212 .
explicit expressions for the new mode - coupling functionals shall be omitted here for brevity : they take the same form as in eqs .
( [ eq : mct - hat - phi ] ) and ( [ eq : mct - hat - phi - qnull ] ) , but with coupling coefficients replaced by the ones for the transversal current correlators . the mathematical form of eq .
( [ eq : gle - t - glass - b ] ) for glass states is identical to that for the longitudinal current given in eq .
( [ eq : gle - l - glass ] ) .
in particular , @xmath213 for glass states vanishes proportional to @xmath214 for small frequencies in contrast to the one in liquid states .
thus , the transversal current correlators exhibit a drastic difference between liquid and glass states @xcite : the long wave length correlators in liquids describe diffusive processes , while the ones in the glass states describe the propagation of transversal sound waves as expected for an isotropic elastic continuum .
finally , we briefly summarize here for later discussion the mct equations of motion for the normalized velocity correlator @xmath215 defined with the velocity @xmath216 of a tagged particle ( labeled @xmath217 ) . an exact equation for this quantity
is given by @xmath218 and the mct expression for the kernel @xmath219 reads @xmath220 with @xmath221 @xcite . here
@xmath222 with @xmath223 denotes tagged - particle density correlator . the zwanzig - mori equation for @xmath224 has the same form as eq .
( [ eq : gle - phi - a ] ) with @xmath34 , @xmath35 , and @xmath25 replaced by @xmath225 , @xmath226 , and @xmath227 , respectively , and the mct kernel is given by the functional @xmath228 with @xmath229 determined by @xmath44 @xcite . for glass states , the mct equations for @xmath230 can be reformulated in terms of the hatted kernel @xmath231 defined in analogy to eq .
( [ eq : mct - t - glass ] ) via @xmath232 ( circles ) of the longitudinal current spectra @xmath233 as a function of @xmath2 at @xmath59 for temperatures indicated in each panel referring to liquid states .
dashed lines denote the hydrodynamic dispersion law @xmath234 with the isothermal sound velocity @xmath235 .
dash - dotted lines show the linear dispersion law @xmath236 with the infinite - frequency sound velocity @xmath237 .
the insets exhibit apparent sound velocities @xmath238 ( circles ) as a function of @xmath2 along with horizontal lines denoting @xmath235 ( dashed lines ) and @xmath237 ( dash - dotted lines ) .
dotted lines in ( c ) refer to the linear dispersion law @xmath239 and the velocity @xmath240 discussed in the text ( _ cf .
[ sec : sound-4 ] ) .
long - dashed lines through circles in ( c ) are from the solution to eq .
( [ eq : lowest - order - ghd ] ) , while solid lines are based on eqs .
( [ eq : positive - dispersion ] ) . ]
circles in fig . [ fig : dispersion - liquids ] show the dispersion relation the peak positions @xmath241 of the longitudinal current spectra @xmath233 as a function of wave number @xmath2 for three representative temperatures referring to liquid states .
( we notice that , in experimental and computer - simulation studies for the dispersion relation cited in sec .
[ sec : introduction ] , it is the peak positions of the longitudinal current spectra that are usually reported .
the peak positions of the dynamic structure factors @xmath242 and of the longitudinal current spectra @xmath183 nearly coincide as can be inferred , e.g. , from fig .
[ fig : sqw - cl - ct ] . in the present article , we are primarily interested in the wave - number regime @xmath243 where @xmath241 increases with @xmath2 . in this pseudo first brillouin zone , the peaks @xmath241
are associated with the collective mode . )
the dashed line in each panel denotes the hydrodynamic dispersion law @xmath234 with the isothermal sound velocity @xmath235 @xcite , whereas the dash - dotted line exhibits the linear dispersion law @xmath236 defined in terms of the infinite - frequency sound velocity @xmath237 ( see below ) .
the inset in each panel shows an apparent sound velocity @xmath244 ( circles ) as a function of @xmath2 , along with horizontal lines denoting @xmath235 ( dashed line ) and @xmath237 ( dash - dotted line ) .
the increase of the sound velocity @xmath245 from @xmath235 towards @xmath237 upon increase of the wave number @xmath2 , which is observable for @xmath246 in all the insets of fig .
[ fig : dispersion - liquids ] , is called the anomalous or positive dispersion
. there are subtle variations in the anomalies depending on the temperature . at high temperature @xmath247 , the whole increase of
the sound velocity is observable in fig .
[ fig : dispersion - liquids](a ) starting from @xmath235 at small @xmath2 up to close to @xmath237 at @xmath248 .
this feature is altered if the temperature is decreased . for @xmath249 ( fig .
[ fig : dispersion - liquids](b ) ) , the sound velocity @xmath245 for small wave numbers is still located above the hydrodynamic value @xmath235 . for @xmath250 ( fig .
[ fig : dispersion - liquids](c ) ) , this `` gap '' becomes larger , and the sound velocity @xmath245 in the small-@xmath2 limit seems to approach another limit denoted by the dotted line which will be defined below . , but here temperatures refer to glass states , and the hydrodynamic dispersion law @xmath251 is determined by @xmath252 defined in the text .
long - dashed lines through circles in ( a ) are from the solution to eq .
( [ eq : lowest - order - ghd ] ) , while solid lines are based on eqs .
( [ eq : positive - dispersion ] ) . ] corresponding results for representative glass states are shown in fig .
[ fig : dispersion - glasses ] . but here , the hydrodynamic dispersion law @xmath251 is determined by @xmath252 introduced in connection with eq .
( [ eq : gle - hat - phi - b ] ) ( see also below ) .
it is clear from fig .
[ fig : dispersion - glasses ] that the anomalous sound - velocity dispersion is present also in glass states .
again , there are subtle variations in the appearance of the anomalies . for @xmath253 ( fig .
[ fig : dispersion - glasses](b ) ) , the sound velocity @xmath245 approaches the hydrodynamic value @xmath252 rather rapidly for @xmath49 .
on the other hand , for @xmath254 ( fig .
[ fig : dispersion - glasses](a ) ) , the approach towards @xmath252 is retarded . for the deep - in - glass state @xmath85 ( fig .
[ fig : dispersion - glasses](c ) ) , a new feature shows up , and `` negative '' dispersion is observable for @xmath255 , where the sound velocity @xmath245 is smaller than the hydrodynamic value @xmath252 .
the results shown in figs . [ fig : dispersion - liquids](b ) , [ fig : dispersion - liquids](c ) , and [ fig : dispersion - glasses](a ) indicate that the increase of the sound velocity from the hydrodynamic value has already started at smaller wave - number regime which can not be adequately resolved with the linear @xmath2 axis .
in addition , the presence of the negative dispersion shown in fig .
[ fig : dispersion - glasses](c ) implies that the anomalous sound - velocity dispersion can not be accounted for solely in terms of relaxation processes since those processes would lead only to the positive dispersion as known , e.g. , from the viscoelastic model . in the following , a microscopic and unified understanding of the mentioned anomalies shall be attempted .
this will be done based on the generalized hydrodynamic description @xcite . in this description ,
the memory kernel @xmath256 is approximated by its @xmath49 limit , while the full @xmath68 dependence of @xmath257 is retained . in this manner
, one gets from eq .
( [ eq : gle - l - b ] ) @xcite @xmath258^{2 } + [ \omega \ , \omega_{q}^{2 } \ , m''(\omega)]^{2 } } , \label{eq : gle - l - ghd}\ ] ] with @xmath259 . in lowest order , the resonance frequency @xmath241 can be obtained as a solution to @xmath260 the long - dashed lines through circles in figs . [
fig : dispersion - liquids](c ) and [ fig : dispersion - glasses](a ) show the dispersion law and the sound velocity based on the solution to this equation .
one understands that the lowest - order solution describes the results from the full mct solutions ( circles ) fairly well for the whole @xmath2-range shown in the figure .
since we are interested in the small wave - number region @xmath246 where the anomalous sound - velocity dispersion is present , a further simplification is possible : one can replace @xmath151 by its leading - order contribution , @xmath261 .
this yields [ eq : positive - dispersion ] @xmath262 which can be rewritten for the the sound velocity @xmath245 as @xmath263 the solid lines in figs .
[ fig : dispersion - liquids](c ) and [ fig : dispersion - glasses](a ) exhibit the dispersion law and the sound velocity based on the solutions to these equations .
one understands that no serious error is introduced with the use of eqs .
( [ eq : positive - dispersion ] ) as far as the regime @xmath246 of interest is concerned . before proceeding ,
let us show that eqs .
( [ eq : positive - dispersion ] ) reproduce known results @xcite in the limit of small and large resonance frequencies .
the hydrodynamic sound is obtained by taking the @xmath264 limit of @xmath265 in eqs .
( [ eq : positive - dispersion ] ) .
one obtains the dispersion law @xmath266 since @xmath267 . for liquids where the memory kernel relaxes to zero for long times ,
@xmath268 , the dispersion law @xmath269 is obtained with the isothermal sound velocity @xmath235 @xcite . for glasses , on the other hand , the kernel does not decay to zero for long times , @xmath270 , leading to the dispersion law @xmath271 with @xmath272 .
( this agrees with @xmath273 given below eq .
( [ eq : gle - hat - phi - b ] ) because of eq .
( [ eq : dw ] ) . )
these results have been included with dashed lines in figs .
[ fig : dispersion - liquids ] and [ fig : dispersion - glasses ] . in the high - frequency limit , one obtains , using @xmath274 , the dispersion law @xmath275 with the infinite - frequency sound velocity @xmath276 .
this holds for both liquid and glass states , and has been included with dash - dotted lines in figs .
[ fig : dispersion - liquids ] and [ fig : dispersion - glasses ] . thus , on the basis of eq .
( [ eq : positive - dispersion - b ] ) , it is clear that the anomalous sound - velocity dispersion from the low - frequency hydrodynamic value ( @xmath235 for liquids and @xmath252 for glasses ) towards the infinite - frequency one @xmath237 upon increase of the wave number @xmath2 is controlled by the detailed frequency dependence of @xmath265 .
this in turn is determined by the relaxation of the memory kernel @xmath277 .
within the mct , @xmath277 is given in terms of the mode - coupling functional as @xmath278 $ ] [ _ cf . _ eq .
( [ eq : mct - phi - c ] ) ] . in the following two subsections , the details of the relaxation of @xmath277 and of the frequency dependence of @xmath265 will be investigated .
these results will be employed in sec .
[ sec : sound-4 ] for a systematic study of the anomalies . , @xmath279 , as a function of @xmath280 at @xmath59 for temperatures indicated in the figure .
curves with labels @xmath281 and 3 are calculated for distance parameters @xmath282 for liquids ( @xmath283 ) and for glasses ( @xmath284 ) .
dotted line with label @xmath285 denotes the kernel at @xmath81 .
horizontal line marks the plateau height @xmath286 .
dashed lines show the mct asymptotic - law results based on eq .
( [ eq : first - m ] ) for @xmath250 and @xmath287 ( @xmath283 ) .
the kernels for @xmath85 , 0.6 , and 0.8 refer to the right vertical axis . ]
figure [ fig : zm - qnull ] exhibits the evolution of the dynamics of @xmath277 upon decrease of the temperature @xmath11 . the curve for @xmath247 , which is quite higher than @xmath83 , decays rapidly and nearly exponentially to zero . by lowering the temperature ,
the dynamics becomes slower and stretched due to the development of the cage effect . for temperatures close to but above @xmath81
, the kernel @xmath277 exhibits a two - step relaxation : the dynamics towards the plateau @xmath286 , followed by the final relaxation from the plateau to zero .
the dynamics that occurs near the plateau is referred to as the @xmath121 process , whereas the final relaxation is called the @xmath5 process ( _ cf .
[ sec : theory-3 ] ) .
the @xmath121 process for small @xmath288 can be well described by the mct asymptotic formula ( [ eq : beta - correlator - a ] ) specialized to @xmath277 @xmath289 as exemplified with the dashed lines in fig .
[ fig : zm - qnull ] .
( @xmath290 and @xmath291 for the model under study . )
at the critical point @xmath81 , the kernel approaches the plateau in a stretched manner , as shown by the dotted curve with label c in fig .
[ fig : zm - qnull ] .
this process , called the critical decay , is described by a power law [ _ cf . _ eq .
( [ eq : critical - x ] ) ] @xmath292 decreasing @xmath11 below the critical temperature @xmath81 , the values for the long - time limits @xmath293 increase .
these limits are approached exponentially fast for @xmath294 @xcite .
if @xmath11 is far below @xmath81 , one observes enhanced oscillatory features for times near @xmath295 as can be seen from the curves for @xmath86 , 0.6 , and 0.5 in fig .
[ fig : zm - qnull ] .
such vibrational dynamics reflects the anomalous - oscillation peak ( aop ) discussed in ref .
it is caused by strong interaction between density fluctuations at microscopic length scales and the arrested glass structure .
features of the aop will be discussed in the next subsection . for the memory kernels
@xmath277 shown in fig .
[ fig : zm - qnull ] for liquid states ( a ) and for glass states ( b ) .
the insets exhibit the double logarithmic representation of the spectra for @xmath11 close to @xmath81 .
the meaning of the labels @xmath281 and @xmath287 is the same as in fig .
[ fig : zm - qnull ] .
dotted line with label @xmath285 denotes the susceptibility spectrum at @xmath81 .
dashed line in the upper inset shows the mct asymptotic - law result based on eq .
( [ eq : first - susceptibility ] ) for @xmath287 ( @xmath283 ) .
dash - dotted lines in the insets exhibit the hydrodynamic linear law @xmath296 . ]
figure [ fig : susceptibility - zm - qnull ] shows the susceptibility spectra @xmath297 of the kernels shown in fig . [ fig : zm - qnull ] .
the spectrum at @xmath247 exhibits only one peak located at high frequencies , to be called a microscopic peak . by lowering @xmath11 ,
an additional peak emerges at low frequencies reflecting the evolution of the structural - relaxation processes .
the lower frequency peak , whose position decreases strongly by lowering @xmath11 , is due to the @xmath5 process , and is called the @xmath5 peak .
it is separated from the microscopic peak by a minimum , called the @xmath121 minimum .
the appearance of the minimum is caused by the crossover from a power law @xmath298 to another one @xmath299 which occurs in the @xmath121 regime ( _ cf .
[ sec : theory-3 ] ) . for small @xmath288 ,
the @xmath121 minimum is well described by the formula which follows from eq .
( [ eq : first - m ] ) , @xmath300 as exhibited by the dashed line in the inset of fig .
[ fig : susceptibility - zm - qnull](a ) . at @xmath254 ,
the critical decay ( [ eq : critical - m ] ) leads to a sublinear susceptibility variation @xmath301 as shown by the dotted line with label c. here , @xmath302 denotes the gamma function .
below @xmath81 , the final exponential decay towards @xmath293 @xcite implies the hydrodynamic law in the small-@xmath68 limit , @xmath303 .
but , near @xmath81 , there is a frequency interval where the spectrum is described by the sublinear critical spectrum ( [ eq : critical - susceptibility ] ) .
this produces a `` knee '' at some frequency where the crossover from the linear low - frequency spectrum @xmath304 to the sublinear critical one @xmath305 occurs , as can be inferred from the inset of fig .
[ fig : susceptibility - zm - qnull](b ) .
if @xmath11 is far below @xmath81 , the knee disappears , and the spectrum is dominated by a single peak reflecting the aop ( see the next subsection ) .
this holds for @xmath86 , 0.6 , and 0.5 .
it is interesting to notice that there is only a weak temperature dependence in the position @xmath306 of the high - frequency peak from high-@xmath11 liquids down to deep - in - glass states .
this feature will also be studied in the next subsection . for the memory kernels
@xmath277 shown in fig .
[ fig : zm - qnull ] for liquid states ( a ) and for glass states ( b ) .
the insets exhibit the double logarithmic representation of @xmath265 for @xmath11 close to @xmath81 .
the meaning of the labels @xmath281 and @xmath287 is the same as in fig .
[ fig : zm - qnull ] .
dotted line with label @xmath285 denotes @xmath265 at @xmath81 .
horizontal lines mark the plateau height @xmath307 ( see text ) . dashed lines in the upper inset show the mct asymptotic - law results based on eq .
( [ eq : delta - beta ] ) for @xmath250 and @xmath287 ( @xmath283 ) , while dashed line in the lower inset exhibits the curve based on eq .
( [ eq : critical - delta ] ) .
the open ( @xmath308 ) and filled ( @xmath309 ) diamonds refer to the frequencies @xmath310 defined in the text . ]
figure [ fig : delta - zm - qnull ] shows the reactive part of the normalized longitudinal moduli @xmath259 .
since @xmath311 and @xmath312 are related via the kramers - kronig relation @xcite , features for @xmath265 can be understood from those for the susceptibility spectra just mentioned .
in particular , a peak in @xmath297 implies a strong increase of @xmath265 near that peak frequency .
this way , one understands an almost single - step increase of @xmath265 at @xmath306 for @xmath313 and @xmath314 , reflecting the presence of only the high - frequency microscopic peak in @xmath297 for those temperatures ( _ cf_. fig .
[ fig : susceptibility - zm - qnull ] ) . for @xmath315 ,
on the other hand , the presence of two peaks in @xmath297 explains the two - step increase of @xmath265 shown in fig .
[ fig : delta - zm - qnull](a ) . because of the strong @xmath11 dependence of the @xmath5-peak frequency , the position where the low-@xmath68 variation of @xmath265 occurs depends strongly on the temperature .
in addition , the sublinear susceptibility variation @xmath305 , present both in liquid and glass states near @xmath81 , also leads to the enhancement of @xmath265 .
indeed , it follows from eq .
( [ eq : critical - m ] ) that @xmath316 holds near the @xmath121 minimum for liquids or near the knee position for glasses . the curve based on this formula ( dashed line )
is compared with the memory kernel at the critical point ( dotted line ) in the inset of fig .
[ fig : delta - zm - qnull](b ) . for later convenience ,
let us define the @xmath5 and @xmath121 regimes for @xmath265 .
we start from liquid states .
since the memory kernel @xmath277 decays from the plateau @xmath286 to zero in the @xmath5 regime , corresponding @xmath5 regime for @xmath265 shall be defined as the frequency interval @xmath317 in which @xmath318 satisfies @xmath319 .
thus , the @xmath5 regime for @xmath265 is below the horizontal bar marking @xmath307 in fig .
[ fig : delta - zm - qnull](a ) . in the @xmath121 regime ,
the kernel @xmath277 is well described by the mct asymptotic formula ( [ eq : first - m ] ) ( _ cf . _
[ fig : zm - qnull ] ) . the corresponding @xmath121 regime shall be defined as the one where @xmath265 can be well described by the asymptotic formula @xmath320 which follows from eq .
( [ eq : first - m ] ) .
the dashed lines in fig .
[ fig : delta - zm - qnull](a ) show curves based on this formula .
let us introduce @xmath310 as the frequencies where @xmath265 ( solid lines ) differ from the mct asymptotic formula ( [ eq : delta - beta ] ) ( dashed lines ) by 2% .
these frequencies are marked by open ( @xmath308 ) and filled ( @xmath309 ) diamonds in fig .
[ fig : delta - zm - qnull](a ) , and the frequency interval @xmath321 shall be considered as the @xmath121 regime for @xmath265 . as mentioned in sec .
[ sec : theory-3 ] , there is an overlap between the @xmath5 and @xmath121 regimes . for glass states , the @xmath121 regime for @xmath265
shall be defined as the frequency interval @xmath322 where @xmath309 marks the point at which @xmath265 differs from the mct asymptotic formula ( [ eq : critical - delta ] ) by 2% .
the frequency @xmath309 at the critical point is marked by the filled diamond in the inset of fig .
[ fig : delta - zm - qnull](b ) .
in contrast to the structural - relaxation processes , there is only a weak temperature dependence in the time scale for the short - time dynamics in @xmath277 occurring at @xmath323 ( _ cf_. fig .
[ fig : zm - qnull ] ) .
correspondingly , the high - frequency peaks in the susceptibility spectra @xmath297 are located nearly at the same frequencies @xmath306 from high-@xmath11 liquid down to deep - in - glass state ( _ cf_. fig .
[ fig : susceptibility - zm - qnull ] ) , leading to a common increase of @xmath265 around this frequency regime ( _ cf_. fig .
[ fig : delta - zm - qnull ] ) . in this section ,
nature of such high - frequency microscopic process shall be discussed
. for @xmath59 and @xmath85 , and dashed and dotted lines respectively refer to its one - mode and two - mode contributions , @xmath324 and @xmath325 .
dash - dotted line shows the sga spectrum @xmath326 based on eqs .
( [ eq : harmonic - app ] ) .
( b ) the same as in ( a ) , but for @xmath253 .
( c ) comparison of the memory - kernel spectrum @xmath327 ( solid line ) with @xmath328 ( dashed line ) for the transversal current correlator and @xmath329 ( dotted line ) for the velocity correlator for @xmath59 and @xmath85 . ]
we start from temperatures far below @xmath81 , @xmath314 , called stiff - glass states , which are characterized by the debye - waller factors close to unity as exemplified by the curve for @xmath85 shown in fig .
[ fig : fq - sq ] .
one understands from eqs .
( [ eq : mct - hat - phi - b ] ) ( [ eq : mct - hat - phi - c ] ) , ( [ eq : mct - hat - phi - e ] ) , and ( [ eq : mct - hat - phi - f ] ) that the renormalized coupling coefficients become very small for the stiff - glass states , and they decrease towards zero in the limit @xmath330 . furthermore , the two - mode contributions to the kernels get suppressed relative to the one - mode contributions , in particular @xmath331 .
figure [ fig : hat - memories - t050](a ) shows the hatted spectrum @xmath332 for @xmath85 ( solid line ) along with its one - mode @xmath324 ( dashed line ) and two - mode @xmath325 ( dotted line ) contributions ( _ cf . _
( [ eq : mct - hat - phi - d ] ) ) .
it is seen that the one - mode contribution @xmath324 indeed provides the dominant contribution to the total memory - kernel spectrum @xmath332 .
the role played by the two contributions @xmath333 and @xmath334 is utterly different .
the former describes the aop , while the latter provides a background spectrum .
further properties of the aop have been discussed in ref .
@xcite based on a stiff - glass approximation ( sga ) , which is obtained by dropping the two - mode contributions to the memory kernels .
( see also ref .
@xcite on the sga . ) in particular , it has been found that harmonic oscillations of the particles within their cages are a good description of the relevant modes , and the density correlator as well as the memory kernel in the stiff - glass states can be described as a superposition of independent harmonic oscillator spectra where the distribution of oscillator frequencies is caused by the distribution of sizes and shapes of the cages . according to the cited mct result for the aop , one obtains @xmath335 with [ eq : harmonic - app ] @xmath336 .
\label{eq : harmonic - app - a}\ ] ] here @xmath337 is given by a superposition of undamped harmonic - oscillator spectra @xmath338 .
\label{eq : harmonic - app - d}\end{aligned}\ ] ] the weight distribution @xmath339 for the oscillators with frequency @xmath340 extends from the low - frequency threshold @xmath341 to the high - frequency one @xmath342 , where @xmath343 with an integrated coupling coefficient @xmath344 ( _ cf . _ eq .
( [ eq : mct - hat - phi - e ] ) ) .
@xmath345 denotes some average value of @xmath146 defined in eq .
( [ eq : gle - hat - phi - b ] ) over the @xmath2 range @xmath346 @xcite . here , @xmath347 denotes the debye wave number , and @xmath348 for the density @xmath59 . since the @xmath2 range around the first peak position @xmath349 of the static structure factor @xmath44 is known to constitute the dominant contribution to the kernel @xcite , @xmath345 in the present study is taken as an average of the first maximum and first minimum of @xmath146 , which occurs at @xmath350 and @xmath349 , respectively ( _ cf . _
[ fig : hat - omega ] ) .
a friction term @xmath351 is necessary when one wants to incorporate the neglected two - mode contributions in a perturbative manner @xcite , but one may read the formulas with @xmath352 unless stated otherwise .
figure [ fig : hat - memories - t050](a ) demonstrates that eqs .
( [ eq : harmonic - app ] ) with @xmath353 ( dash - dotted line ) describe qualitative features of the aop of @xmath332 for @xmath354 . by increasing @xmath11 , one observes enhanced intensity in @xmath332 at low frequencies as shown for @xmath253 in fig .
[ fig : hat - memories - t050](b ) .
this is a precursor phenomenon of the glass melting at the transition point . near @xmath81 ,
such enhancement for decreasing frequencies can be described by a critical power law @xmath355 ( _ cf . _
( [ eq : critical - susceptibility ] ) ) . by substituting the critical decay of the density correlators @xmath356 into eq .
( [ eq : mct - hat - phi - qnull - b ] ) , one understands that it is the linear term @xmath357 that provides the critical decay of @xmath358 , whereas the quadratic term @xmath359 serves only as a correction @xmath360 .
this explains why in fig .
[ fig : hat - memories - t050](b ) the strong increase of the intensity is dominated by @xmath324 ( dashed line ) .
thus , the modes building the aop are buried under the tail of the central peak , and they show up only as a shoulder .
nevertheless , it is remarkable that the sga spectrum @xmath326 ( dash - dotted line ) succeeds to extract the buried aop portion as demonstrated in fig .
[ fig : hat - memories - t050](b ) .
the shift of the aop to lower frequencies upon increase of the temperature reflects the widened size of the cages so that particles `` bounce '' in their cages with smaller average frequencies .
indeed , the particles are localized in such tight cages at @xmath85 that the square root @xmath361 of the long - time limit of the mean - squared displacement @xcite is only 5.0% of the particle s lj diameter , whereas it is increased to @xmath362 at @xmath253 .
the shift is accompanied by an increase of the inelastic spectrum including the aop , reflecting the decrease of the elastic contribution @xmath363 .
( solid lines ) at indicated temperatures taken from fig . [
fig : susceptibility - zm - qnull ] , but plotted here on the linear-@xmath68 axis .
dashed lines denote the sga susceptibility spectra @xmath364 based on eq .
( [ eq : susceptibility - harmonic - app ] ) . ]
the analysis presented in fig .
[ fig : hat - memories - t050](b ) also implies a possibility to extract aop portion in the spectrum on the basis of eqs .
( [ eq : harmonic - app ] ) .
so , let us examine to what extent the high - frequency microscopic peaks in the susceptibility spectra @xmath297 at higher temperatures have such `` harmonic '' character .
this is done in fig .
[ fig : susceptibility - memory - harmonic - app ] where representative susceptibility spectra taken from fig .
[ fig : susceptibility - zm - qnull ] are compared with the sga susceptibility spectra @xmath365 which follows from eqs .
( [ eq : gle - hat - phi - c ] ) and ( [ eq : harmonic - app - a ] ) . to focus on the high - frequency regime , the comparison is done in fig .
[ fig : susceptibility - memory - harmonic - app ] with the linear-@xmath68 axis . in evaluating the sga susceptibility spectra for liquids , nonergodicity parameters @xmath75 and @xmath293 in the equations and quantities
involved are replaced by the ones at the critical point . as discussed in connection with figs .
[ fig : hat - memories - t050](a ) and [ fig : hat - memories - t050](b ) , the sga spectra describe the aop portion of the memory - kernel spectra for @xmath85 and @xmath253 fairly well , and this is reflected in the susceptibility spectra shown in the two bottom panels of fig .
[ fig : susceptibility - memory - harmonic - app ] . somewhat large deviation seen at @xmath366 for @xmath253 is due to the neglected two - mode contribution , which is more enhanced than the deviation in fig .
[ fig : hat - memories - t050](b ) due to the presence of the factor @xmath68 in @xmath297 . on the other hand , because of this factor , the deviation at small frequencies seen in fig .
[ fig : hat - memories - t050](b ) is suppressed for @xmath297 .
one understands from the other panels of fig .
[ fig : susceptibility - memory - harmonic - app ] that the harmonic approximation based on eq .
( [ eq : susceptibility - harmonic - app ] ) reasonably accounts for overall features
peak position and strength of the microscopic peaks in @xmath297 not only near @xmath81 but even at high temperatures such as @xmath247 .
the large deviations for low frequencies at @xmath249 and 10 simply reflect the fact that at such high @xmath11 the structural - relaxation contributions also enter into the high - frequency regime ( _ cf . _ fig .
[ fig : susceptibility - zm - qnull ] ) .
thus , considerable portion of the short - time or high - frequency microscopic process , from high-@xmath11 liquids down to deep - in - glass states , can be described as a superposition of harmonic vibrational dynamics .
this explains the weak @xmath11 dependence of the microscopic process shown in figs .
[ fig : zm - qnull]-[fig : delta - zm - qnull ] . at @xmath59 for temperatures referring to liquid states ( a ) and glass states ( b ) . ]
the persistence of the vibrational dynamics inside the cage in the liquid state is not surprising . to see this ,
we show in fig .
[ fig : vacf ] the velocity correlator @xmath230 for representative liquid and glass states .
it is well known that the cage effect in dense liquids manifests itself by oscillatory variations with a decay of @xmath230 to negative values @xcite , and fig .
[ fig : vacf](a ) demonstrates this phenomenon . that such oscillatory dynamics exhibited by @xmath230 reflects the aop can be understood from the observations ( i ) in the stiff - glass state @xmath85 the spectrum @xmath329 of the relaxation kernel for @xmath230 exhibits an aop whose
spectral features threshold frequencies and maximum position and height nearly coincide with those for the aop of @xmath332 as demonstrated in fig .
[ fig : hat - memories - t050](c ) , and ( ii ) at higher @xmath11 including liquid states there is a correlation between the peak position in the sga susceptibility spectra @xmath364 shown in fig .
[ fig : susceptibility - memory - harmonic - app ] and the time at which the velocity correlator @xmath230 reaches the first minimum .
( solid line ) for @xmath59 and @xmath85 taken from fig .
[ fig : delta - zm - qnull ] , but plotted here on the linear-@xmath68 axis . dashed
line exhibits the sga modulus @xmath367 based on eq .
( [ eq : delta - sga ] ) .
horizontal dotted line denotes @xmath368 .
vertical arrows mark the threshold frequencies @xmath369 and @xmath370 defined in the text . ]
there is another interesting feature caused by the dominant harmonic nature of the dynamics in stiff - glass states . from @xmath265 for @xmath85 shown in fig .
[ fig : delta - zm - qnull](b ) , one observes the decrease of @xmath265 from @xmath368 at @xmath371 . one can show that such behavior is also well described by the harmonic approximation : it follows from eqs .
( [ eq : gle - hat - phi - c ] ) and ( [ eq : harmonic - app - a ] ) that the reactive part of the modulus within the sga is given by @xmath372 \ , \}.
\label{eq : delta - sga}\ ] ] figure [ fig : delta - harmonic - app ] compares @xmath265 for @xmath85 with the curve based on eq .
( [ eq : delta - sga ] ) on the linear-@xmath68 axis .
one understands that decrease of @xmath265 from @xmath368 observable at @xmath373 is due to the harmonic modes near the threshold frequency @xmath369 .
similarly , the bump at @xmath374 reflects the harmonic modes near @xmath370 . here
we provide an explanation of the anomalous sound - velocity dispersion shown in figs .
[ fig : dispersion - liquids ] and [ fig : dispersion - glasses ] based on eq .
( [ eq : positive - dispersion - b ] ) and on the features of @xmath265 explored in the previous two subsections . at @xmath247 ,
the whole relaxation of the memory kernel @xmath277 occurs on the short time scale @xmath323 ( fig .
[ fig : zm - qnull ] ) .
this yields a susceptibility spectrum @xmath297 consisting only of the microscopic peak located at @xmath306 ( fig .
[ fig : susceptibility - zm - qnull ] ) , and the entire increase of @xmath265 occurs in the frequency regime @xmath375 ( fig . [ fig : delta - zm - qnull ] ) .
it is therefore sufficient to use the linear-@xmath68 axis to describe the whole @xmath68 variation in @xmath265 , and this explains via eq .
( [ eq : positive - dispersion - b ] ) the observation in fig .
[ fig : dispersion - liquids](a ) that the entire positive dispersion of the sound velocity @xmath245 starting from @xmath235 up to close to @xmath237 occurs within the linear-@xmath2 axis . by decreasing the temperature to @xmath249 ,
the relaxation of @xmath277 becomes slower and stretched due to the development of the cage effect ( fig .
[ fig : zm - qnull ] ) , leading to the evolution of low - frequency contributions in @xmath297 and @xmath265 ( figs .
[ fig : susceptibility - zm - qnull ] and [ fig : delta - zm - qnull ] ) . translated to @xmath245 via eq .
( [ eq : positive - dispersion - b ] ) , this means that the approach of @xmath245 towards the hydrodynamic value @xmath235 becomes delayed compared to the one at @xmath247 .
this explains why @xmath245 for small wave numbers shown in fig .
[ fig : dispersion - liquids](b ) is still located above @xmath235 .
( circles ) for @xmath59 and @xmath250 , taken from the inset of fig .
[ fig : dispersion - liquids](c ) , but plotted here with the @xmath376 axis .
long - dashed line is from the solution to eq .
( [ eq : lowest - order - ghd ] ) , while solid line is based on eq .
( [ eq : positive - dispersion - b ] ) .
horizontal lines denote @xmath235 ( dashed line ) , @xmath237 ( dash - dotted line ) , and @xmath240 ( dotted line ) .
the contributions @xmath377 , @xmath378 , and @xmath379 to the sound - velocity increase defined in the text are indicated by vertical arrows .
the open and filled diamonds respectively mark @xmath380 and @xmath381 defined in the text .
( b ) the same as in ( a ) , but here the sound velocity refers to @xmath254 , taken from the inset of fig .
[ fig : dispersion - glasses](a ) , and horizontal dashed line denotes @xmath272 . ] for temperatures close to but above @xmath81 , the memory kernel @xmath277 exhibits the glassy structural relaxations ( fig .
[ fig : zm - qnull ] ) , leading to low-@xmath68 variations in @xmath297 and @xmath265 which can be adequately described only with the @xmath382 axis ( figs .
[ fig : susceptibility - zm - qnull ] and [ fig : delta - zm - qnull ] ) .
let us analyze the sound - velocity increase for @xmath250 in detail .
for this purpose , we replot the inset of fig . [
fig : dispersion - liquids](c ) in fig .
[ fig : velocity - dispersion - t180](a ) , but with the @xmath376 axis .
this is necessary to fully appreciate the small-@xmath2 variation of the sound velocity @xmath245 , which on the basis of eq .
( [ eq : positive - dispersion - b ] ) originates from the glassy low-@xmath68 variations in @xmath265 .
notice that the circles from the full mct solutions for @xmath183 can be obtained only up to the minimum value of the discretized wave - number grids . on the other hand
, there is no such restriction in the solution ( solid line ) to the lowest - order equation ( [ eq : positive - dispersion - b ] ) of the generalized hydrodynamic description .
the difference between the circles and the solid line in the @xmath2 range where the circles are available is due to the approximations involved , but there should be no problem to use the approximate solution for understanding the essence of the anomalies .
let us quantify the contributions to the sound velocity increase due to the @xmath5 process ( to be denoted as @xmath377 ) , due to the @xmath121 process ( @xmath378 ) , and due to the microscopic process ( @xmath379 ) in the following way based on the @xmath5 and @xmath121 regimes for @xmath265 defined at the end of sec .
[ sec : sound-2 ] . as mentioned there
, @xmath265 varies from @xmath383 to @xmath319 in the @xmath5 regime .
one therefore obtains from eq .
( [ eq : positive - dispersion - b ] ) : @xmath384 with @xmath385 .
the notation comes from the fact that @xmath240 is the high - frequency limit of the sound velocity in the @xmath5 regime . the horizontal dotted line in fig .
[ fig : velocity - dispersion - t180](a ) marks @xmath240 , and corresponding linear dispersion law @xmath239 and the velocity @xmath240 have been included with dotted lines in fig .
[ fig : dispersion - liquids](c ) .
the @xmath121-relaxation contribution @xmath378 can be obtained in terms of the frequencies @xmath310 also introduced at the end of sec . [ sec : sound-2 ] : one obtains from eq .
( [ eq : positive - dispersion - b ] ) @xmath386 with @xmath387 .
these velocities are marked by open ( @xmath380 ) and filled ( @xmath381 ) diamonds in fig .
[ fig : velocity - dispersion - t180](a ) .
the rest of the sound - velocity increase towards @xmath237 is due to the microscopic process . as mentioned in the previous subsection , considerable portion of the microscopic process
is due to the harmonic vibrational dynamics .
the contributions @xmath377 , @xmath378 , and @xmath379 are indicated by vertical arrows in fig .
[ fig : velocity - dispersion - t180](a ) . since there is an overlap between the @xmath5 and @xmath121 regimes ( _ cf .
[ sec : theory-3 ] ) , there is a corresponding overlap between @xmath377 and @xmath378 .
the full sound - velocity variation for @xmath250 can be understood in this way in terms of the contributions from @xmath5- and @xmath121-relaxation processes and from the microscopic vibrational dynamics .
notice that the structural - relaxation contributions can be adequately described only with the logarithmic @xmath2 axis .
this explains why in fig .
[ fig : dispersion - liquids](c ) the approach of @xmath245 towards the hydrodynamic value @xmath235 could not be resolved , but only the approach towards @xmath240 can be observed .
thus , the sound - velocity variation shown in fig .
[ fig : dispersion - liquids](c ) mostly reflects the microscopic contribution @xmath379
. in glass states , the @xmath5 process is arrested , and the hydrodynamic sound velocity is altered to @xmath272 .
there are at most @xmath121 and @xmath6 contributions to the sound - velocity increase from this hydrodynamic value . at and near @xmath81 , the @xmath121 contribution , @xmath388 ,
can not be neglected , where @xmath389 with @xmath309 for glasses defined at the end of sec .
[ sec : sound-2 ] .
this feature can be understood from fig .
[ fig : velocity - dispersion - t180](b ) , a redrawing of the inset of fig .
[ fig : dispersion - glasses](a ) with the logarithmic @xmath2 axis , where the @xmath121 contribution @xmath378 and the microscopic contribution @xmath379 ( defined as the rest of the sound - velocity increase towards @xmath237 ) are marked by vertical arrows .
the @xmath121-relaxation contribution @xmath378 can not be resolved with the linear-@xmath2 axis , and this explains why in fig .
[ fig : dispersion - glasses](a ) the approach of @xmath245 towards @xmath252 with decreasing @xmath2 is retarded . for temperatures further away from @xmath81 ,
the @xmath121-relaxation contribution becomes smaller , and the sound - velocity increase from @xmath252 is entirely due to the microscopic process .
this explains why in fig .
[ fig : dispersion - glasses](b ) the approach of @xmath245 towards @xmath252 can be observed for @xmath253 even with the linear-@xmath2 axis . by further decreasing the temperature to @xmath85
, a new feature shows up reflecting the dominant harmonic nature in the stiff - glass - state dynamics . as discussed in connection with fig .
[ fig : delta - harmonic - app ] , there appears a frequency interval where @xmath265 becomes smaller than @xmath390 . according to eq .
( [ eq : positive - dispersion - b ] ) , this gives rise to a `` negative '' dispersion , and explains the result for @xmath85 shown in fig .
[ fig : dispersion - glasses](c ) .
the evolution of the aop is related to the arrest of density fluctuations caused by the cage effect , and the aop is the result of a mapping of the cage distribution on the frequency axis ( _ cf .
[ sec : sound-3 ] ) .
since the arrest of density fluctuations is driven by the ones with a wave number @xmath2 near the structure - factor - peak position @xcite , this interpretation of the aop suggests that it appears in spectra of all probing variables that couple to density fluctuations of short wavelengths . but different probing variables will weight the oscillating complexes differently , and therefore the shape and the peak position of the aop will depend somewhat on the probe .
two such examples have been analyzed in ref .
@xcite : the tagged - particle density correlators and the velocity correlator @xmath230 .
in particular , it has been shown there how the aop in the memory - kernel spectrum @xmath329 for @xmath230 ( _ cf . _
[ fig : hat - memories - t050](c ) ) accounts for the excess intensity in the density of states @xmath391 with respect to the debye model . in this section ,
we investigate how the aop manifests itself in spectral features of the longitudinal and transversal current spectra .
this issue is relevant since in recent computational studies @xcite two excitations have been observed in the longitudinal current spectra , but the low - frequency boson - peak - like excitations have been assigned to the transversal mode . since there is no direct cross correlation between the longitudinal and transversal current dynamics ,
the appearance of the transversal mode in the longitudinal spectra is interpreted as being due to `` mixing '' phenomena @xcite .
this is in contrast to the above interpretation of the aop , according to which possible low - frequency excitations in both of the longitudinal and transversal current spectra should be the manifestation of the aop .
indeed , the aop is present also in the memory kernel for the transversal current correlators , and its spectral features are quite similar to the one for the longitudinal current correlators as can be inferred from fig .
[ fig : hat - memories - t050](c ) . to corroborate our statement
, we will also study the density dependence of spectral features , because such study was claimed to support the interpretation in terms of the mixing of the longitudinal and transversal modes ( _ cf .
[ sec : introduction ] ) . , ( b ) longitudinal current spectrum @xmath183 , and ( c ) transversal current spectrum @xmath213 at wave number @xmath392 for @xmath59 and @xmath85 .
solid lines are based on the full mct equations of motion described in sec .
[ sec : theory ] .
dash - dotted lines show curves within the generalized hydrodynamic description , eqs .
( [ eq : decomposition - basic ] ) .
dashed lines denote the hfs part based on eqs .
( [ eq : decomposition - hfs ] ) , whereas dotted lines refer to the aop portion from eqs .
( [ eq : decomposition - aop-1 ] ) .
curves for @xmath393 are obtained from those for @xmath183 via eq .
( [ eq : continuity ] ) . ]
we present with solid lines in fig .
[ fig : sqw - cl - ct ] typical density fluctuation spectrum @xmath393 , longitudinal current spectrum @xmath183 , and transversal current spectrum @xmath213 at deep in the glass state , @xmath394 .
these results are based on the full mct equations of motion described in sec .
[ sec : theory ] . the wave number @xmath395 chosen for the demonstration is about half of the structure - factor - peak position ( _ cf . _ fig .
[ fig : fq - sq ] ) .
the fluctuation spectrum @xmath393 shown in fig .
[ fig : sqw - cl - ct](a ) exhibits two peaks for @xmath396 , and illustrates a hybridization of the high - frequency sound with the modes building the aop @xcite : the narrow peak located at higher frequency is due to the high - frequency sound propagation , and a broad lower - frequency excitation reflects the aop studied in sec .
[ sec : sound-3 ] . on the other hand , there is apparently only one peak in the longitudinal current spectrum @xmath183 shown in fig .
[ fig : sqw - cl - ct](b ) .
it is due to the high - frequency sound , and the peak position @xmath241 is nearly the same as the one for the fluctuation spectrum .
it is not clear how the aop manifests itself in the spectral shape of @xmath183 since , compared to @xmath393 , the low frequency part is suppressed due to the presence of the factor @xmath214 ( _ cf . _ eq .
( [ eq : continuity ] ) ) . the transversal current spectrum @xmath213 shown in fig .
[ fig : sqw - cl - ct](c ) also exhibits only one peak .
its peak frequency @xmath397 is located at a lower frequency than @xmath241 reflecting a smaller transversal sound velocity @xmath398 compared to the longitudinal one .
again , it is not clear whether there is a contribution to the spectral shape of @xmath213 from the aop .
it is the purpose of this subsection to clarify this point . for the longitudinal sound ( solid line ) and @xmath399 for the transversal sound ( dashed line ) at @xmath85 for @xmath59 ( a ) and @xmath60 ( b ) .
horizontal dotted lines denote the threshold frequencies @xmath400 ( _ cf . _ fig .
[ fig : hat - memories - t050 - 2 ] ) , and arrows mark the debye wave number @xmath401 and the structure - factor - peak position @xmath349 . ] to this end , we will derive approximate formulas specialized to describe each of the high - frequency - sound ( hfs ) and low - frequency - aop portions of the spectra .
the starting point is the following expression under the generalized hydrodynamic description ( _ cf . _
( [ eq : gle - l - ghd ] ) and citation @xcite ) , obtained from eq .
( [ eq : gle - l - glass ] ) with the hatted memory kernel @xmath402 approximated by its @xmath49 limit , @xmath403 : [ eq : decomposition - basic ] @xmath404^{2 } + [ \omega \hat{\omega}_{q}^{2 } \hat{m}''(\omega)]^{2}}. \label{eq : decomposition - basic-1}\ ] ] here @xmath405 . corresponding description for the fluctuation spectrum @xmath393 is derived via eq .
( [ eq : continuity ] ) .
the formula ( [ eq : decomposition - basic-1 ] ) describes hybridization of two oscillations : one is a bare sound with dispersion @xmath146 , whose wave - number dependence is shown in fig .
[ fig : hat - omega](a ) , and the other is represented by the aop spectrum in @xmath332 .
the aop region in the spectrum @xmath332 shall be defined as @xmath406 in terms of the threshold frequencies @xmath400 introduced via the sga spectrum in sec .
[ sec : sound-3 ] .
these frequencies are marked in fig .
[ fig : hat - memories - t050 - 2](a ) .
the hybridization occurs in the range @xmath407 indicated in fig .
[ fig : hat - omega](a ) where the two oscillations overlap .
the generalized hydrodynamic description for the transversal current can be introduced similarly on the basis of eq .
( [ eq : gle - t - glass - b ] ) : @xmath408^{2 } + [ \omega ( \hat{\omega}_{q}^{\rm t})^{2 } \hat{m}_{\rm t}''(\omega)]^{2}}. \label{eq : decomposition - basic-2}\ ] ] here @xmath409 and @xmath410 .
since @xmath411 as shown in fig .
[ fig : hat - memories - t050](c ) , one understands that the hybridization of the bare transversal sound of dispersion @xmath399 with the aop occurs also in the range @xmath412 indicated in fig .
[ fig : hat - omega](a ) . figure [ fig : sqw - cl - ct ] demonstrates that the generalized hydrodynamic description ( dash - dotted lines ) reproduce main features of all spectra fairly well . in particular , the subtle hybridization of the high - frequency sound and the aop for @xmath393 is treated semiquantitatively correctly , justifying the use of eqs .
( [ eq : decomposition - basic ] ) as starting equations . for @xmath59 and @xmath85 , while dash - dotted line refers to the sga spectrum based on eqs .
( [ eq : harmonic - app ] ) .
arrows mark the threshold frequencies @xmath400 defined below eqs .
( [ eq : harmonic - app ] ) .
dotted line denotes the frequency dependence of the function @xmath413 given in eq .
( [ eq : decomposition - aop - l-1 ] ) in arbitrary units , and its peak frequency @xmath414 is marked by an arrow .
long dashed line denotes the corresponding result based on eq .
( [ eq : decomposition - aop - l-2 ] ) .
( b ) the same as in ( a ) , but for @xmath60 @xcite . ] the hfs portion of @xmath183 shall be approximated by a lorenzian ( multiplied by @xmath214 ) with the width determined at the peak frequency @xmath241 : [ eq : decomposition - hfs ] @xmath415^{2 } + [ \omega \hat{\omega}_{q}^{2 } \hat{m}''(\omega_{q}^{\rm l \ , max } ) ] ^{2}}. \label{eq : decomposition - hfs-1}\ ] ] the dashed curve in fig .
[ fig : sqw - cl - ct](b ) shows the result based on this formula , and the corresponding hfs curve for the fluctuation spectrum @xmath416 is included in fig .
[ fig : sqw - cl - ct](a ) .
one understands that the formula ( [ eq : decomposition - hfs-1 ] ) well describes the hfs portion of these spectra .
thus , the rest of the spectral intensity in the low - frequency regime should be due to the aop .
corresponding expression for the hfs part of the transversal current spectrum reads @xmath417^{2 } + [ \omega ( \hat{\omega}_{q}^{\rm t})^{2 } \hat{m}_{\rm t}''(\omega_{q}^{\rm t max } ) ] ^{2}}. \label{eq : decomposition - hfs-2}\ ] ] figure [ fig : sqw - cl - ct](c ) shows that this formula ( dashed line ) fairly catches the main peak of the spectrum , but the predicted width is considerably narrower than the one of the full mct solution ( solid line ) .
this feature can be understood in terms of the aop contribution as follows .
we first notice that @xmath418 for @xmath392 is located in the central region of the aop ( _ cf . _
[ fig : hat - memories - t050 - 2](a ) ) .
let us introduces frequencies @xmath419 in this central regime so that @xmath420 . since there holds @xmath421 as shown in fig .
[ fig : hat - memories - t050](c ) , the kramers - kronig relation implies @xmath422 .
one therefore understands from fig .
[ fig : delta - harmonic - app ] that @xmath423 in the central regime of the aop increases with @xmath68 , i.e. , @xmath424 . since in lowest order @xmath397
is the solution to the equation @xmath425 ( _ cf . _ eq .
( [ eq : lowest - order - ghd ] ) ) , one has @xmath426 .
combined with @xmath427 , this leads to an inequality @xmath428^{2 } > [ \tilde{\omega}_{\pm}^{2 } - ( \hat{\omega}_{q}^{\rm t})^{2 } \delta_{\rm t}(\tilde{\omega}_{\pm})]^{2}$ ] .
we also notice from fig .
[ fig : hat - memories - t050](c ) that there is only a weak @xmath68 dependence in @xmath328 in the central regime of the aop .
all this together , one finds on the basis of eqs .
( [ eq : decomposition - basic-2 ] ) and ( [ eq : decomposition - hfs-2 ] ) : @xmath429 . this way
, one understands that the residual intensity in fig .
[ fig : sqw - cl - ct](c ) which can not be described as the hfs portion ( dashed line ) is caused by the frequency dependence of @xmath423 near @xmath397 , which in turn is due to the aop . from the mentioned reasoning ,
it is clear that such deviation from the lorenzian shape of the spectrum occurs whenever the sound resonance frequency is located in the central regime of the aop .
this explains why such deviation is small for the longitudinal current spectrum shown in fig .
[ fig : sqw - cl - ct](b ) , whose resonance frequency @xmath430 exceeds @xmath370 and is not located in the central regime of the aop .
such larger resonance frequency @xmath241 than expected from the bare dispersion @xmath146 for @xmath392 shown in fig .
[ fig : hat - omega](a ) is due to the positive dispersion discussed in sec .
[ sec : sound ] . in view of the results presented so far
, one understands that there are residual spectral intensities due to the aop in the frequency regime @xmath431 . in the following
, we will derive approximate formulas extracting such aop portion based on eqs .
( [ eq : decomposition - basic ] ) . for this purpose ,
let us consider the denominator on the right - hand side of eq .
( [ eq : decomposition - basic-1 ] ) , @xmath432^{2 } + [ \omega \ , \hat{\omega}_{q}^{2 } \ , \hat{m}''(\omega)]^{2}$ ] .
the appearance of the hfs peak is mainly due to the decrease of the first term towards zero approaching the resonance position @xmath241 .
since we are interested in the off - resonant regime @xmath433 , this term will be approximated by its the small-@xmath68 limit , @xmath432^{2 } \approx \hat{\omega}_{q}^{4}$ ] .
correspondingly , only the leading - order contribution for small @xmath68 shall be retained in the second term , @xmath434^{2 } \approx [ \ , \omega \ , \hat{\omega}_{q}^{2 } \ , \hat{m}''(\omega=0 ) \ , ] ^{2}$ ] .
this results in the following expression for possible aop portion in @xmath183 : [ eq : decomposition - aop-1 ] @xmath435^{2}}. \label{eq : decomposition - aop - l-1}\ ] ] according to the derived formula , the @xmath2 and @xmath68 dependences are factorized .
this implies that the peak position of the aop portion when it is visible is @xmath2 independent .
the frequency dependence of the formula ( [ eq : decomposition - aop - l-1 ] ) is shown as dotted line in fig .
[ fig : hat - memories - t050 - 2](a ) , and its peak frequency , to be denoted as @xmath414 , is marked by an arrow .
furthermore , one infers from the @xmath2 dependence of @xmath146 shown in fig .
[ fig : hat - omega ] that the intensity of the aop portion is predicted to increase upon increase of the wave number for @xmath436 , whereas this trend becomes reversed for @xmath346 .
corresponding formula for the transversal current spectrum can be derived starting from eq .
( [ eq : decomposition - basic-2 ] ) : @xmath437^{2}}. \label{eq : decomposition - aop - t-1}\ ] ] since @xmath421 ( _ cf . _ fig .
[ fig : hat - memories - t050](c ) ) , eqs .
( [ eq : decomposition - aop-1 ] ) imply that the peak positions of the aop portions of both longitudinal and transversal current spectra when they are visible are located at the same frequency @xmath414 .
furthermore , a stronger intensity of the aop portion is predicted for the transversal current spectra in the pseudo first brillouin zone since there holds @xmath438 as can be seen from fig .
[ fig : hat - omega](a ) .
finally , we notice that it is reasonable to further approximate the above formulas as [ eq : decomposition - aop-2 ] @xmath439^{2 } } , \label{eq : decomposition - aop - l-2 } \\
\phi_{q}^{{\rm t}\mbox{-}{\rm aop}''}(\omega ) = \frac{1}{(\hat{\omega}_{q}^{\rm t})^{2 } } \frac{\omega^{2 } \hat{m}_{\rm t}^{(1)''}(\omega ) } { 1 + [ \omega \hat{m}_{\rm t}^{(2)''}(\omega = 0)]^{2 } } , \label{eq : decomposition - aop - t-2}\end{aligned}\ ] ] since it is the one - mode contribution which provides the dominant contribution to the memory kernel and which describes the aop , whereas the two - mode contribution provides a background spectrum as discussed in sec .
[ sec : sound-3 ] . figure [ fig : hat - memories - t050 - 2](a ) demonstrates that indeed the difference between the curves based on eqs .
( [ eq : decomposition - aop-1 ] ) and ( [ eq : decomposition - aop-2 ] ) is small . our motivation for introducing eqs .
( [ eq : decomposition - aop-2 ] ) will become clear in the next subsection .
( solid lines ) at @xmath59 and @xmath85 for indicated wave numbers .
dashed lines denote the hfs part based on eq .
( [ eq : decomposition - hfs-1 ] ) , whereas dotted lines refer to the aop portion obtained from eq .
( [ eq : decomposition - aop - l-1 ] ) . ]
( solid lines ) at @xmath59 and @xmath85 for indicated wave numbers .
dashed lines denote the hfs part based on eq .
( [ eq : decomposition - hfs-2 ] ) , whereas dotted lines refer to the aop portion obtained from eq .
( [ eq : decomposition - aop - t-1 ] ) . ] of @xmath183 ( open circles ) and @xmath397 of @xmath213 ( filled circles ) as a function of @xmath2 for @xmath59 and @xmath85 .
solid and dashed lines denote the bare dispersions @xmath146 and @xmath399 , respectively .
dotted lines exhibit the hydrodynamic dispersion laws @xmath251 and @xmath440 .
peak positions of the aop portion in @xmath183 are denoted by open diamonds .
horizontal arrows mark @xmath400 and @xmath414 taken from fig .
[ fig : hat - memories - t050 - 2](a ) .
( b ) full width at half maximum ( fwhm ) for @xmath183 ( circles ) as a function of @xmath2 .
dashed line denotes the lorenzian width @xmath441 predicted by eq .
( [ eq : decomposition - hfs-1 ] ) .
dotted line refers to the hydrodynamic width , @xmath442 .
vertical arrow marks the wave number where @xmath443 holds .
( c ) the same as in ( b ) , but for @xmath213 for which the lorenzian width ( dashed line ) is given by @xmath444 and the hydrodynamic width ( dotted line ) by @xmath445 .
vertical arrow marks the wave number where @xmath446 holds . ]
the dotted lines in fig .
[ fig : sqw - cl - ct ] are based on eqs .
( [ eq : decomposition - aop-1 ] ) .
for the density fluctuation spectrum ( fig .
[ fig : sqw - cl - ct](a ) ) , it is seen that the dotted line , obtained from eq .
( [ eq : decomposition - aop - l-1 ] ) via the relation ( [ eq : continuity ] ) , extracts the aop portion quite reasonably .
it is also seen from fig .
[ fig : sqw - cl - ct](b ) that eq .
( [ eq : decomposition - aop - l-1 ] ) provides a reasonable account of the residual low - frequency intensity in the longitudinal current spectrum located at @xmath447 .
notice that the aop appears differently in @xmath393 and @xmath183 reflecting the difference of the factor @xmath214 . to understand the result shown in fig .
[ fig : sqw - cl - ct](c ) for the transversal current spectrum , we notice that , by construction , the applicability of the formula ( [ eq : decomposition - aop - t-1 ] ) is limited to the small-@xmath68 regime considerably lower than the sound resonance position @xmath397 . therefore , the dotted line in fig . [ fig : sqw - cl - ct](c ) in the region @xmath448 has no physical meaning .
nevertheless , it is seen that the dotted line accounts for the residual intensity in the low - frequency regime which can not be described as the hfs portion ( dashed line ) .
thus , eq . ( [ eq : decomposition - aop - t-1 ] ) describes the low - frequency part of the deviation from the lorenzian shape explained above . the variation of the longitudinal and transversal current spectra ( solid lines ) with changes of wave number @xmath2 is presented in figs .
[ fig : cl-1093-t050 ] and [ fig : ct-1093-t050 ] , along with their decomposition into the hfs part ( dashed lines ) and possible aop portion ( dotted lines ) . the dispersion relation and the full width at half maximum ( fwhm ) as a function of @xmath2
are summarized in fig .
[ fig : dispersion - t050-rho-1093 ] .
we start from the discussion on the longitudinal current spectra .
when @xmath2 is small so that @xmath449 , there is practically no effect from the aop on the spectral shape .
( however , the aop affects the resonance frequency @xmath241 near @xmath369 as discussed in sec .
[ sec : sound-4 ] . )
therefore , the spectrum for @xmath450 shown in fig .
[ fig : cl-1093-t050 ] is dominated by the hfs portion ( dashed line ) , and there is no physical meaning in the dotted line for this @xmath2 value .
thus , the spectrum for @xmath450 is very close to lorenzian ( multiplied by @xmath214 ) , and its width agrees with the one from the hydrodynamic prediction @xmath442 as demonstrated in fig . [ fig : dispersion - t050-rho-1093](b ) . as @xmath2 increases so that @xmath241 becomes larger than @xmath369 , effects from the aop set in .
these effects increase the spectral width , as exemplified for @xmath451 and 2.1 in fig .
[ fig : cl-1093-t050 ] , in two ways .
first , the width becomes larger compared to the hydrodynamic width because @xmath452 when the resonance frequency @xmath241 is located in the aop regime , @xmath453 ( _ cf .
[ fig : hat - memories - t050 - 2](a ) ) .
this shows up as the deviation of the dashed line from the dotted line in fig .
[ fig : dispersion - t050-rho-1093](b ) .
in addition , the frequency dependence of @xmath454 , which again is caused by the aop , also enlarges the width as explained above in connection with fig .
[ fig : sqw - cl - ct](c ) .
this latter effect leads to a the non - lorenzian shape of the spectrum , i.e. , the deviation of the solid line from the dashed line for @xmath451 and 2.1 shown in fig .
[ fig : cl-1093-t050 ] , and also explains the difference between corresponding circles and dashed line in fig .
[ fig : dispersion - t050-rho-1093](b ) . the dotted lines in fig .
[ fig : cl-1093-t050 ] account for the low - frequency part of the non - lorenzian spectra as discussed above .
as @xmath2 is increased further , the positive dispersion effect becomes important , and it pushes @xmath241 considerably above frequencies expected from the bare dispersion @xmath146 as shown in fig . [
fig : dispersion - t050-rho-1093](a ) . thereby , there opens a frequency window , where @xmath455 holds and the appearance of the aop itself can be observed in the spectra .
this feature holds @xmath456 as shown in fig .
[ fig : cl-1093-t050 ] .
as mentioned above , there is no @xmath2 dependence in @xmath414 , which results in flat dispersion curve for the low - frequency excitations as demonstrated with open diamonds in fig .
[ fig : dispersion - t050-rho-1093](a ) .
corresponding results for the transversal current spectra shown in figs .
[ fig : ct-1093-t050 ] and [ fig : dispersion - t050-rho-1093 ] can be explained similarly . but
here , the positive dispersion of @xmath397 from the hydrodynamic law @xmath440 with @xmath457 , obtained from the leading - order expansion of @xmath399 , is not so strong as in the longitudinal case .
this is because the bare transversal sound dispersion @xmath399 for @xmath436 is located below the maximum frequency @xmath458 of the the susceptibility spectrum @xmath459 ( _ cf . _ the bottom panel for @xmath85 in fig .
[ fig : susceptibility - memory - harmonic - app ] ) , where the positive dispersion effect is most significant as explained in sec .
[ sec : sound ] .
thus , there is no frequency regime where @xmath460 holds for the transversal current spectra , which can be inferred from fig .
[ fig : dispersion - t050-rho-1093](a ) .
therefore , only the width of @xmath213 is affected by the aop , which is summarized in fig .
[ fig : dispersion - t050-rho-1093](c ) .
however , this feature is altered when the density is decreased , as we will see in the next subsection . in this subsection ,
effects of varying density on spectral features of the longitudinal and transversal current spectra shall be investigated . specifically , a density @xmath60 lower than 1.093 studied so far will be considered , but with temperature @xmath85 fixed . , but for a lower density @xmath60 , and dotted lines referring to the aop portion are based on eq .
( [ eq : decomposition - aop - l-2 ] ) as discussed in text . ] , but for a lower density @xmath60 , and dotted lines referring to the aop portion are based on eq .
( [ eq : decomposition - aop - t-2 ] ) as discussed in text . ] , but for a lower density @xmath60 . in ( a )
, filled diamonds are added denoting the peak positions of the aop portion in @xmath213 .
dash - dotted lines are added in ( b ) and ( c ) denoting @xmath461 and @xmath462 , respectively ( see text for details ) . ] upon decrease of the density , the bare sound dispersions @xmath146 and @xmath399 and the memory - kernel spectrum @xmath332 are shifted to lower frequencies as demonstrated in figs .
[ fig : hat - omega](b ) and [ fig : hat - memories - t050 - 2](b ) , respectively .
notice the striking similarity between the change in @xmath332 upon decrease of the density at fixed @xmath11 and the change upon increase of the temperature at fixed @xmath17 found in fig .
[ fig : hat - memories - t050](b ) .
indeed , the critical temperature is decreased to @xmath463 for @xmath60 , and lowering the density at fixed @xmath85 drives the system closer to the critical point . as discussed with fig .
[ fig : hat - memories - t050](b ) , a strong central peak is formed when the system approaches the critical point , and the modes building the aop get buried under its tail . but again , fig .
[ fig : hat - memories - t050 - 2](b ) demonstrates that the sga kernel @xmath326 extracts the buried aop portion quite reasonably @xcite .
thus , in analogy to what is found in fig .
[ fig : hat - memories - t050](b ) , the aop shifts towards lower frequencies upon decrease of the density , accompanied by an increase of its spectral intensity . as a related problem
, we found that the use of eqs .
( [ eq : decomposition - aop-1 ] ) to extract the aop portion from current spectra leads to unplausible results .
this is because of the mentioned development of the central peak , leading to quite large value of @xmath464 . as will be demonstrated in the following ( _ cf .
[ fig : dispersion - t050-rho-950](b ) and [ fig : dispersion - t050-rho-950](c ) ) , such large value of @xmath464 caused by the critical decay near @xmath81 is not appropriate for handling the high - frequency dynamics .
we will therefore use eqs .
( [ eq : decomposition - aop-2 ] ) instead of eqs .
( [ eq : decomposition - aop-1 ] ) , which essentially amounts to replacing @xmath464 with @xmath465 . the curves based on eqs .
( [ eq : decomposition - aop-2 ] ) are not much affected by the critical dynamics because ( i ) only corrections to the critical decay enter into @xmath325 as explained in connection with fig .
[ fig : hat - memories - t050](b ) , and ( ii ) the critical - decay contribution in @xmath324 is suppressed in eq .
( [ eq : decomposition - aop - l-2 ] ) due to the presence of the factor @xmath214 in its denominator . as demonstrated in fig .
[ fig : hat - memories - t050 - 2](a ) , there is practically no difference between the use of eqs .
( [ eq : decomposition - aop-1 ] ) and ( [ eq : decomposition - aop-2 ] ) for @xmath59 , and we expect that physics is not altered with the use of the latter for @xmath60 . longitudinal and transversal current spectra for @xmath60 ( solid lines ) with changes of wave number @xmath2 are presented in figs .
[ fig : cl-950-t050 ] and [ fig : ct-950-t050 ] , along with their decomposition into the hfs part ( dashed lines ) and possible aop portion ( dotted lines ) . as mentioned above ,
the dotted lines in these figures are based on eqs .
( [ eq : decomposition - aop-2 ] ) . the dispersion relation and the fwhm as a function of @xmath2
are summarized in fig .
[ fig : dispersion - t050-rho-950 ] .
it is seen from figs .
[ fig : dispersion - t050-rho-950](b ) and [ fig : dispersion - t050-rho-950](c ) that indeed the hydrodynamic prediction for the width with @xmath466 does not work for @xmath60 ; instead , the corresponding prediction with @xmath467 reasonably accounts for the width in the small-@xmath2 regime .
it is seen by comparing figs .
[ fig : cl-1093-t050 ] and [ fig : cl-950-t050 ] that the aop portion of the longitudinal current spectra for the smaller density @xmath60 is located at lower frequencies and its intensity is more enhanced .
these differences simply reflect the shift of the aop to lower frequencies accompanied by an increase of its spectral intensity for @xmath60 .
in addition , the positive dispersion of the sound velocity and the deviation from the lorenzian shape of the spectra start at smaller @xmath2 for @xmath60 , as can be inferred by comparing figs .
[ fig : dispersion - t050-rho-1093 ] and [ fig : dispersion - t050-rho-950 ] .
these features are also due to the shift of the aop to lower frequencies upon decrease of the density .
furthermore , there is no `` negative '' dispersion ( _ cf .
[ sec : sound-4 ] ) for @xmath60 since the memory kernel for @xmath85 at this density is not dominated by the harmonic contributions ( _ cf . _
[ fig : hat - memories - t050 - 2](b ) ) . otherwise , the variation of the longitudinal current spectra with changes of @xmath2 for @xmath60 is quite similar to the one found for @xmath59 , and the discussion shall not be repeated .
the variation of the transversal current spectra with changes of @xmath2 for @xmath60 is also quite similar to the one found for @xmath59 , as can be understood by comparing figs .
[ fig : ct-1093-t050 ] and [ fig : ct-950-t050 ] . but
here , the resonance frequency @xmath397 of the transversal sound exhibits a strong positive dispersion as shown in fig .
[ fig : dispersion - t050-rho-950](a ) .
again , this is caused by the shift of the aop to lower frequencies .
the positive dispersion of the transversal sound velocity due to the contribution from the aop can be discussed quite similarly to what is presented in sec .
[ sec : sound ] for the longitudinal sound .
thereby , there opens a frequency window also for the transversal current spectra , where @xmath460 holds and the appearance of the aop itself can be observed in the spectra .
such aop portion can be observed for @xmath468 in fig .
[ fig : ct-950-t050 ] , yielding a flat dispersion curve for the low - frequency excitations of the transversal current spectra as demonstrated with filled diamonds in fig .
[ fig : dispersion - t050-rho-950](a ) .
this is in contrast to the result found for @xmath59 , where @xmath469 holds for the whole @xmath2 range and the effects from the aop showed up only in the width of the transversal current spectra .
in this paper , we presented a theoretical investigation on the high - frequency collective dynamics in liquids and glasses at microscopic length and time scales based on the mct for ideal liquid - glass transition .
we focused on recently investigated issues from ixs and computer - simulation studies for dynamic structure factors and longitudinal and transversal current spectra : the anomalous dispersion of the high - frequency sound velocity and the nature of the low - frequency excitation called the boson peak .
the results of the theory are demonstrated for a model of the lj system .
the mct explains the evolution of the structural - relaxation processes as precursor of the glass transition at a critical temperature @xmath81 , which is driven by the mutual blocking of a particle and its neighbors ( cage effect ) @xcite . on the other hand
, the theory also predicts the development of the harmonic vibrational excitations in stiff - glass states @xmath314 , called the aop .
the aop is caused by the strong interaction between density fluctuations at microscopic length scales and the arrested glass structure , and exhibits the properties of the boson peak @xcite . as demonstrated in sec .
[ sec : sound-3 ] , the aop persists also near and above @xmath81 , and considerable portion of the microscopic process in systems ranging from high-@xmath11 liquids down to deep - in - glass states can be described as a superposition of harmonic vibrational dynamics .
we investigated in sec .
[ sec : sound ] how the interference of the sound mode with these structural - relaxation processes and vibrational excitations shows up as anomalies in the sound - velocity dispersion . in the vicinity of @xmath81 ,
the structural - relaxation contributions to the sound - velocity dispersion show up at very small wave numbers , which can fully be appreciated only with the logarithmic @xmath2 axis ( _ cf . _ fig .
[ fig : velocity - dispersion - t180 ] ) .
therefore , with the linear-@xmath2 axis adopted in conventional studies on the dispersion relation , the sound velocity @xmath245 in the small-@xmath2 limit is located above the hydrodynamic value @xmath235 ( _ cf . _
[ fig : dispersion - liquids](c ) ) , and this even at @xmath249 ( _ cf . _ fig .
[ fig : dispersion - liquids](b ) ) which is more than twice of the critical temperature @xmath81 .
we notice that this temperature is also higher than the freezing temperature @xmath470 which is estimated based on the so - called hansen - verlet criterion @xcite .
thus , except for very high temperature like @xmath247 ( _ cf . _
[ fig : dispersion - liquids](a ) ) , the variation of the sound velocity @xmath245 which is visible in the linear @xmath2 axis basically reflects the contribution from the microscopic process .
this theoretical result is in contrast to the traditional picture based on the viscoelastic model according to which the anomalous dispersion is fully ascribed to the structural relaxation , but is in agreement with finding from recent ixs studies on simple liquid metals near the melting temperature @xcite .
the description of the microscopic process in terms of the aop a superposition of harmonic oscillations of particles inside their cages implies that the decay of the memory kernel at microscopic times @xmath323 ( _ cf .
[ fig : zm - qnull ] ) reflects the dephasing of different oscillatory components in the force fluctuations .
this is in consistent with the implication from the computer - simulation study on harmonic glass @xcite . such microscopic process , despite arrested structural relaxations , accounts for the positive dispersion in glass states ( _ cf .
[ fig : dispersion - glasses ] ) , whose presence has been reported from recent md and ixs studies @xcite . that the microscopic process due to the aop persists both in liquid and glass states with only weak temperature dependence of its characteristic time or frequency ( _ cf . _
[ fig : susceptibility - memory - harmonic - app ] ) is in accord with the computer - simulation result presented in ref .
@xcite . as a function of @xmath11 at @xmath59 for @xmath471 ( solid line ) and @xmath450 ( dash - dotted line ) , which are based on eq .
( [ eq : positive - dispersion - b ] ) with @xmath265 determined for each temperature ( _ cf . _ fig .
[ fig : delta - zm - qnull ] ) .
hydrodynamic sound velocities @xmath472 for @xmath473 and @xmath474 ( see text ) for @xmath475 are denoted as open circles and filled squares , respectively .
@xmath476 for @xmath473 , given by @xmath477 , are also denoted as filled squares .
dotted line shows the square - root singularity translated to @xmath476 , i.e. , @xmath478 with @xmath479 for @xmath480 and @xmath481 for @xmath473 .
dashed line exhibits the @xmath121-relaxation contribution to @xmath482 which is based on eq .
( [ eq : positive - dispersion - b ] ) with @xmath265 determined from the mct asymptotic formula ( [ eq : delta - beta ] ) . ] to see further implications of the theory , we show in fig .
[ fig : vq - vs - t ] the sound velocities @xmath482 as a function of @xmath11 for fixed @xmath471 ( solid line ) and @xmath450 ( dash - dotted line ) .
( the argument @xmath11 shall be added here to emphasize the @xmath11 dependence . )
the density is fixed to @xmath59 , and @xmath482 are obtained from eq .
( [ eq : positive - dispersion - b ] ) with @xmath265 determined for each temperature ( _ cf . _ fig .
[ fig : delta - zm - qnull ] ) . assuming that lj particles under study are of argon size ( @xmath483 nm ) , the wave number @xmath450 ( 1.5 nm@xmath484 ) is in the lowest momentum - transfer range in ixs measurements , and @xmath471 ( 0.03 nm@xmath484 ) is a typical momentum transfer in brillouin light scattering ( bls ) experiments . as shown in fig .
[ fig : dispersion - glasses](c ) for a stiff - glass state @xmath85 , a negative dispersion is observable for @xmath255 where the sound velocity @xmath482 is smaller than the hydrodynamic ( @xmath49 ) value .
this is due to the dominant harmonic nature of the dynamics at @xmath85 , because of which there appears a frequency interval where @xmath265 is smaller than its @xmath264 limit ( _ cf .
[ fig : delta - harmonic - app ] ) .
the presence of the negative dispersion is reflected in fig .
[ fig : vq - vs - t ] as the emergence of the low-@xmath11 region where the ixs sound velocity ( @xmath482 for @xmath450 ) is smaller than the bls sound velocity ( @xmath471 )
. such @xmath11 dependence of the ixs and bls sound - velocity data in the low-@xmath11 region has been reported for glycerol @xcite ( see fig .
5(b ) of ref .
thus , there is an experimental result which is consistent with the presence of the negative dispersion predicted by the theory . that the difference between the ixs and bls sound - velocity data in the low-@xmath11 region is smaller in fig .
[ fig : vq - vs - t ] than the one found in fig .
5(b ) of ref .
@xcite for glycerol and that such difference is not clearly observable in orthoterphenyl @xcite might be rationalized by the stronger influence of vibrational modes ( boson peaks ) in the dynamics of network glass former glycerol than in fragile glass formers like the lj system and orthoterphenyl @xcite . according to this reasoning ,
the presence of the negative dispersion is also expected in strong glass formers like silica where the influence of vibrational dynamics is even stronger @xcite . the sound velocity data for vitreous silica reported in fig
. 3 of ref . @xcite seems to indicate this possibility , but large error bars do not allow ones to draw a decisive conclusion .
it would therefore be of value to re - analyze the data for silica with improved resolution .
hydrodynamic sound velocities @xmath472 for @xmath485 and @xmath486 for @xmath475 are included in fig .
[ fig : vq - vs - t ] as open circles and filled squares , respectively .
@xmath472 increases with increasing @xmath11 because in our study @xmath11 is varied with density fixed , leading to increased pressures at elevated temperatures . on the other hand ,
the @xmath11 dependence of @xmath487 is dominated by that of @xmath488 ( _ cf . _
[ fig : fq - sq ] ) , and this explains the increase of @xmath487 with decreasing @xmath11 . here
a side remark shall be added concerning the notion of the `` hydrodynamic '' sound velocity @xmath487 for glass states referring to @xmath475 .
the mct in its idealized form , adopted in the present study , predicts the structural arrest at the critical temperature @xmath81 which is located above the calorimetric glass - transition temperature @xmath489 . in reality
there are slow dynamical processes referred to as hopping processes which restore ergodicity for @xmath475 . these processes , which are studied within the extended version of the mct @xcite , change the ideal elastic peaks @xmath490 of density fluctuation spectra @xmath66 into quasielastic @xmath5 peaks of nonzero width also for @xmath475 .
@xmath491 in real systems therefore has to be interpreted as an effective debye - waller factor @xcite , which can be measured as area under the quasielastic @xmath5 peak in @xmath66 or by determining the plateau of the @xmath63-versus-@xmath492 curve .
correspondingly , @xmath487 should be interpreted as the extension of @xmath476 the sound velocity in the high - frequency limit of the @xmath5 regime introduced in sec .
[ sec : sound-4 ] to @xmath475 .
@xmath476 for @xmath473 are given by @xmath477 ( _ cf .
[ sec : sound-4 ] ) , and are also included with filled squares in fig . [
fig : vq - vs - t ] : @xmath476 defined below and above @xmath81 merge at @xmath81 .
the mentioned dependence of @xmath476 on @xmath488 has been utilized to extract experimentally the zero wave - number limit of the debye - waller factor via the relation @xmath493^{2}$ ] @xcite and to test the square - root singularity as predicted by the mct , @xmath479 for @xmath480 and @xmath481 for @xmath473 ( _ cf . _ eq .
( [ eq : plateau - x ] ) ) . the dotted line in fig . [ fig : vq - vs - t ] shows the square - root singularity translated to @xmath476 . in our system , there is no possibility to detect the singularity based on the ixs sound velocity ( @xmath482 for @xmath450 ) , which basically reflects the contribution from the microscopic process ( _ cf . _
[ fig : velocity - dispersion - t180 ] ) .
on the other hand , the bls sound velocity ( @xmath471 ) probes much lower momentum - transfer range , and is thus affected also by the structural - relaxation processes .
in particular , the @xmath121-relaxation contribution to @xmath482 plays an important role in the vicinity of @xmath81 ( _ cf . _
[ fig : velocity - dispersion - t180 ] ) , as can also be inferred by comparing the solid and dashed lines in fig .
[ fig : vq - vs - t ] , the latter being obtained based on eq .
( [ eq : positive - dispersion - b ] ) with @xmath265 determined from the mct asymptotic formula ( [ eq : delta - beta ] ) for the @xmath121 relaxation .
thus , the @xmath121 relaxation must be included for reliable determination of @xmath488 via the sound velocity data , which is in full agreement with the conclusion drawn in ref .
( _ cf . _ fig .
13 of ref .
@xcite where bls data for cakno@xmath494 corresponding to solid line and filled squares in fig .
[ fig : vq - vs - t ] are shown . )
we studied in sec .
[ sec : transversal ] how the aop manifests itself in the spectral shape of the glass - state longitudinal and transversal current spectra at low frequencies . for small wave numbers where the sound resonance frequencies are located below the low - frequency threshold of the aop , @xmath495
, there is practically no effect from the aop , and the longitudinal and transversal current spectra are dominated by the sound excitations of the lorenzian spectral shape . by increasing the wave numbers
so that the sound resonance frequencies enter the aop region , @xmath496 , the hybridization of the sound mode with the aop becomes important , and the shape of the current spectra deviates from lorenzian . as the wave number is increased further , the interference of the sound mode with the aop leads to the positive dispersion of the sound velocities , pushing @xmath497 considerably above frequencies expected from the bare dispersion .
thereby , there opens a frequency window , where the appearance of the aop itself can be observed in the spectra at low frequencies .
furthermore , our approximate formulas ( [ eq : decomposition - aop-1 ] ) predict that ( i ) there is no @xmath2 dependence in the peak frequency @xmath414 of the aop portions in the current spectra , ( ii ) @xmath414 is nearly the same for both of the longitudinal and transversal current spectra since @xmath411 as shown in fig .
[ fig : hat - memories - t050](c ) , and ( iii ) the intensity of the aop portion is stronger for the transversal current spectra than for the longitudinal ones in the pseudo first brillouin zone .
these theoretical predictions , demonstrated in figs .
[ fig : cl-1093-t050 ] to [ fig : dispersion - t050-rho-950 ] for the lj system , are in agreement with previous findings from md and ixs studies @xcite . thus , the low - frequency excitations observable at the same resonance frequency @xmath414 in both of the glass - state longitudinal and transversal current spectra are the manifestation of the aop . usually , the boson peaks are discussed below @xmath489 . on the other hand , the glass state in the present paper refers to @xmath475 .
since @xmath81 is located above @xmath489 , our theoretical results for glasses apply also for some ranges of the liquid state in the conventional terminology . according to our theoretical results ,
it is essential to have the positive dispersion of the transversal sound velocity for the appearance of the low - frequency excitations in the transversal current spectra , whose presence has never been discussed so far .
this comes out from our study on the density dependence of the dispersion relation , shown in figs .
[ fig : dispersion - t050-rho-1093](a ) and [ fig : dispersion - t050-rho-950](a ) for the higher and the lower densities , respectively . such density dependence of the dispersion relation is in accord with the one reported from the computer simulation for vitreous silica ( _ cf . _
fig . 4 of ref .
it would therefore be of value if simulation studies could test whether or not our theoretical prediction agrees with their data .
we notice that the intensity of the low - frequency excitations for the lj system studied in this paper is weaker than the one , e.g. , known from computer - simulation results on silica @xcite : the low - frequency excitations in the present study are shoulders rather than peaks . again
, this feature may be ascribed to the fact that the lj system can be classified as a fragile glass former , in which the boson peak is less pronounced than in strong glass formers like silica @xcite . and ( b ) the sga spectrum @xmath326 obtained with the cutoff @xmath498 ( solid lines ) , 80 ( dashed lines ) , and 120 ( dotted lines ) for the hard - sphere system at packing fraction @xmath499 .
as in ref .
@xcite , the static structure factor @xmath44 has been evaluated within the percus - yevick approximation , and the units of length and time have been chosen so that the hard - sphere diameter @xmath500 and the thermal velocity @xmath501 . ] finally , let us make a comment on the cutoff problem mentioned in ref . @xcite .
as already noticed there , the mct results for the hard - sphere system in stiff - glass states change if the cutoff @xmath502 of the wave - number grids used in the calculations is varied .
this is demonstrated in fig .
[ fig : hss - cutoff](a ) showing the @xmath502 dependence of the memory - kernel spectrum @xmath332 for the hard - sphere system at packing fraction @xmath499 studied in ref .
it is seen that the intensity of @xmath332 increases with increasing @xmath502 . furthermore , the broad harmonic - oscillation `` peak '' reflecting the aop , observable in @xmath332 for the cutoff @xmath498 adopted in ref .
@xcite , disappears for @xmath503 ; it shows up only as a shoulder .
such @xmath502 dependence reflects the slow decrease towards zero of the direct correlation function @xmath504 of the hard - sphere system for @xmath505 tending to infinity , because of which the relevant coupling coefficient @xmath506 ( _ cf . _
( [ eq : mct - phi - c ] ) ) does not approach zero even in the @xmath507 limit .
however , this does not mean that there is no aop in the hard - sphere system .
as discussed in sec .
[ sec : sound-3 ] , the aop portion of @xmath332 , even when it is buried under the tail of the quasielastic peak , is well extracted by the sga spectrum @xmath326 . figure [ fig : hss - cutoff](b ) demonstrates that there is no @xmath502 dependence in @xmath326 provided a sufficiently large @xmath502 is chosen . in this sense
, the aop is well defined in the hard - sphere system .
but , it does not show up as a peak in @xmath332 due to the two - mode contributions , which are neglected within the sga .
thus , the increase of the intensity of @xmath332 , in particular , the smearing out of the `` gap '' for @xmath508 , with increasing @xmath502 is due to the development of the two - mode contributions . because of
the strong anharmonic effects , no harmonic - oscillation peak in the spectrum @xmath332 is expected for the hard - sphere system . on the other hand ,
introducing the cutoff is equivalent to softening the hard - sphere potential .
indeed , we have confirmed that there is no cutoff problem in the lj system considered in this paper , and that all the essential results presented in ref .
@xcite can be reproduced for this system .
in particular , one observes harmonic - oscillation peak for the lj system in stiff - glass states as shown in fig .
[ fig : hat - memories - t050](a ) , which is free from the mentioned cutoff problem . thus , the whole physics discussed in ref .
@xcite for stiff - glass states remain valid for systems where particles interact with regular interaction potentials .
the author thanks w. gtze for valuable comments .
this work was supported by grant - in - aids for scientific research from the ministry of education , culture , sports , science and technology of japan ( no .
17740282 ) . | we present theoretical investigation on the high - frequency collective dynamics in liquids and glasses at microscopic length scales and terahertz frequency region based on the mode - coupling theory for ideal liquid - glass transition .
we focus on recently investigated issues from inelastic - x - ray - scattering and computer - simulation studies for dynamic structure factors and longitudinal and transversal current spectra : the anomalous dispersion of the high - frequency sound velocity and the nature of the low - frequency excitation called the boson peak .
it will be discussed how the sound mode interferes with other low - lying modes present in the system .
thereby , we provide a systematic explanation of the anomalous sound - velocity dispersion in systems ranging from high temperature liquid down to deep inside the glass state in terms of the contributions from the structural - relaxation processes and from vibrational excitations called the anomalous - oscillation peak ( aop ) . a possibility of observing negative dispersion the _ decrease _ of the sound velocity upon increase of the wave number
is argued when the sound - velocity dispersion is dominated by the contribution from the vibrational dynamics .
we also show that the low - frequency excitation , observable in both of the glass - state longitudinal and transversal current spectra at the same resonance frequency , is the manifestation of the aop . as a consequence of the presence of the aop in the transversal current spectra
, it is predicted that the transversal sound velocity also exhibits the anomalous dispersion .
these results of the theory are demonstrated for a model of the lennard - jones system . |
it is currently believed that pulsars are among the few objects in our galaxy that are candidate sources of ultrarelativistic charged cosmic rays .
relativistic particles within the magnetosphere emit @xmath0-rays at energies up to several gev in various processes such as curvature radiation , synchrotron radiation and inverse compton ( ic ) scattering .
thus , observations in the multi - gev @xmath0-ray domain allow one to study the acceleration sites in the magnetosphere of a pulsar . predicted sites where particle acceleration can take place are , for example , above the polar cap of the neutron star ( e.g. *
* ; * ? ? ?
* ) and in the so - called outer gap of the magnetosphere ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
furthermore , particle acceleration can take place outside the magnetosphere in the region where the pulsar wind interacts with the interstellar medium . if electrons are accelerated in these shocks , they could give rise to ic - scattered photons from , for example , the cosmic microwave background , synchrotron radiation , or thermal origin .
psr b1951 + 32 was detected first at radio frequencies by @xcite , and is one of the six rotation - powered high energy pulsars whose gev emission was detected by egret @xcite . among @xmath0-ray pulsars , psr b1951 + 32 is the only source observed to emit up to @xmath8gev with no cutoff being evident in the differential energy spectrum .
the spectrum shows a hard spectral index of 1.8 between @xmath9mev and @xmath8gev .
the pulsar has an apparent high efficiency ( @xmath10 ) of converting its rate of rotational energy loss , @xmath11ergs@xmath12s@xmath5 , into @xmath0-rays above @xmath9mev ( assuming a distance of @xmath13kpc to the pulsar ) .
moreover , the @xmath0-ray luminosity at @xmath14gev is comparable to that of the crab pulsar @xcite .
as inferred from its rotational parameters , the spin - down age of psr b1951 + 32 is @xmath15yr @xcite , that is , about 100 times older than the crab pulsar .
the magnetic field strength of @xmath16 g @xcite is lower than that in most rotation - powered pulsars . because of the lower magnetic field
, curvature @xmath0-rays emitted near the stellar surface , as predicted in polar - cap models , are less affected by magnetic pair production . compared with younger , more strongly magnetized pulsars ,
the spectral cutoff energy is thereby shifted to higher energies , up to a few tens of gev ( @xcite ; see also @xcite for a discussion of low - field millisecond pulsars ) . on the contrary ,
if the @xmath0-rays are emitted in the outer magnetosphere , as predicted in outer gap models , the potential drop in the outer gap of psr b1951 + 32 is expected to be comparable to that of young pulsars ( see eq . [ 12 ] of @xcite and eq . [ 2.1 ] of @xcite ) .
therefore , the cutoff energy , which reflects the maximum lorentz factor of the electrons or positrons accelerated in the outer gap , is expected to be around 10gev @xcite .
thus , features in the predicted spectral shape of weakly magnetized pulsars at energies above 10gev are strongly dependent on the emission altitudes . in order to discriminate between emission models , psr b1951 + 32
is a prime candidate for observation by ground - based @xmath0-ray detectors with low energy thresholds such as the imaging air cherenkov telescope magic .
this pulsar is located in the core of the radio nebula ctb 80 , which is thought to be physically associated with the pulsar . in x
- rays the nebula shows a cometary shape @xcite , being confined by a bow shock that is produced by the pulsar s high proper motion ( @xmath17 km s@xmath5 ) @xcite . @xcite
predict an over-@xmath18gev flux from the nebula at a level of @xmath19 of the crab s flux , by assuming that high - energy leptons can accumulate inside the well - localized nebula for long periods of time , as observed in the case of the crab nebula .
the current tightest constraint on the emission above @xmath9gev from the pulsar and its nebula , obtained by the whipple collaboration @xcite , puts an upper limit of @xmath2gev on the cutoff energy of the pulsed emission and an upper limit of @xmath20@xmath4s@xmath5 , on the steady emission above @xmath21gev .
the latter is within a factor @xmath22 of the prediction of @xcite . in polar - cap models ,
the cutoff energy is determined by the attenuation of @xmath0-rays due to magnetic pair production and hence by the emission altitude of @xmath0-rays . as a consequence ,
the energy spectrum above the cutoff energy is superexponentially attenuated .
if the emission altitude in the polar cap model shown in figure [ 1951_pulsar ] changes from 1 to 2 stellar radii , the cutoff energy will increase from 20gev to 60gev , which is , according to our observations , the maximum allowed cutoff energy for a superexponentially shaped cutoff . on the contrary , in outer - gap models ,
the cutoff is determined by the maximum lorentz factor of the accelerated positrons and electrons . as a consequence ,
the cutoff of the @xmath0-ray spectrum is smoother , resulting in an exponential cutoff .
if the magnetic field lines near the light cylinder are straighter than assumed for the outer - gap spectrum in figure [ 1951_pulsar ] , the predicted flux below 60gev will increase . for more precise predictions of the cutoff energy in polar - cap models ,
multidimensional and self - consistent electrodynamics have to be examined from first principles , whereas a three - dimensional magnetic field configuration has to be investigated in the outer - gap model .
assuming that these improvements in theory will be achieved in the near future , measurements with higher statistics around @xmath23gev , for example , by glast , or measurements by future ground - based experiments with lower thresholds than magic , for example , magic ii or the cherenkov telescope array , will be needed in order to distinguish between models .
the predicted ic flux at tev energies in the outer - gap model ( figure [ 1951_pulsar ] _ solid black line _ ) appears to be inconsistent with our upper limits .
nevertheless , it must be noted that the ic flux is obtained by assuming that all the magnetospheric soft photons illuminate the equatorial region of the magnetosphere in which the gap - accelerated positrons are migrating outwards .
therefore , the predicted ic flux as a function of energy specifies an upper boundary to the possible pulsed tev emission .
the open poloidal magnetic field lines could have a single - signed curvature within 1.8 light - cylinder radii , as the solution of the time - dependent force - free electrodynamics of an oblique rotator indicates @xcite .
if this is the case , soft photons emitted inside the light cylinder along the convex magnetic field lines will not efficiently illuminate the magnetic field lines , which are slightly convex even outside the light cylinder . as a result
, the predicted ic flux at tev energies will be significantly reduced .
this problem will be solved in future when the self - consistent gap electrodynamics @xcite and the three - dimensional force - free electrodynamics are combined .
we are grateful for the preparation of the ephemeris of psr b1951 + 32 by andrew lyne , thus enabling us to perform the pulsed analysis .
alice harding was so kind as to provide us with her polar - cap predictions of psr b1951 + 32 .
we also would like to thank the instituto de astrofisica de canarias for the excellent working conditions at the observatorio del roque de los muchachos , in la palma . the support of the german bundesministerium fr bildung und forschung and the max - planck - gesellschaft , the italian istituto nazionale de fisica nucleare , the spanish comisin interministerial de ciencias y tecnologa , eth research grant th 34/04 3 , and grant 1p03d01028 from the polish ministerstwo nauki
i informatyzacji are gratefully acknowledged . | we report on very high energy @xmath0-observations with the magic telescope of the pulsar psr b1951 + 32 and its associated nebula , ctb 80 .
our data constrain the cutoff energy of the pulsar to be less than @xmath1gev , assuming the pulsed @xmath0-ray emission to be exponentially cut off .
the upper limit on the flux of pulsed @xmath0-ray emission above @xmath2gev is @xmath3photons @xmath4 sec@xmath5 , and the upper limit on the flux of steady emission above @xmath6gev is @xmath7photons @xmath4 sec@xmath5 .
we discuss our results in the framework of recent model predictions and other studies . |
there is a rapidly expanding use of low temperature platforms ( @xmath1 k ) , for fundamental science , sensitive instrumentation , and new technologies of potentially significant commercial impact .
the availability of commercial `` push button '' cryogen - free dilution refrigerators is a key game - changing recent development , in which european companies have played the leading role .
the availability of cryogen - free commercial platforms to 1 mk is a future prospect @xcite .
precise measurement of the thermodynamic temperature of these low temperature platforms is crucial for their operation .
however , despite the sophisticated nature of the cooling technology , reliable and convenient thermometers are unavailable . in the temperature range below 1
k the dissemination of the kelvin is usually realised using secondary thermometers and superconducting reference points @xcite .
examples include : resistance thermometers ; paramagnetic susceptibility ; vibrating wire / tuning fork viscometry .
these thermometers do not reliably cover the temperature range 1 k to 10 mk ( a fortiori to 1 mk ) .
the calibration of these thermometers against the plts-2000 @xcite is impractically time consuming and expensive .
resistance thermometers are intrinsically subject to shifts in calibration and prone to self heating and pick - up of electromagnetic interference .
paramagnetic - salt susceptibility thermometers are subject to degradation with time , and viscometric methods measure temperature inside the mixing chamber and not on the typical sample mounting platform .
the dependence of @xmath2he melting pressure on temperature has been used as the laboratory standard in specialist laboratories for many years .
the plts-2000 scale is based on the melting pressure of @xmath2he but this is viewed as difficult to implement for the rapidly expanding community of users of ultra - low temperature platforms . @xmath3co nuclear orientation thermometry is a primary technique that relies on the known anisotropy of the decay of radioactive nuclei as a function of temperature but is only effective in a narrow temperature range , and is slow and user unfriendly .
nuclear magnetic susceptibility thermometry is likewise insensitive above 20 mk , and user unfriendly .
reliable and traceable thermometry below 1 k thus offers demanding challenges even to the specialist , and , despite the sophisticated low temperature platforms now readily available with `` push - button '' operation , a solution which is reliable , traceable and user - friendly has yet to be developed .
a quote attributed to prof o v lounasmaa , founder of the low temperature laboratory at helsinki university of technology ( now aalto university ) in the context of the t < 1 k regime reads `` a person with one thermometer knows what the temperature is ; a person with two will forever have doubt '' .
the challenge is to rectify this state of affairs .
the need to have a thermometer that measures thermodynamic temperature outside of international standards laboratories has led to the development of thermometers with a simple relationship between the measured parameter and the boltzmann constant , @xmath4 , for example thermometers that couple noise to a squid either directly @xcite or indirectly via the magnetic johnson noise @xcite and coulomb blockade thermometers @xcite based on electric conductance characteristics of tunnel junction arrays . in this paper
we describe recent advances in current sensing noise thermometer , exploiting low noise and high bandwidth dc squid amplifiers , for the high precision and rapid measurement of absolute temperature , covering the entire temperature range of interest from 4k to below 100 microk .
the current sensing noise thermometer was pioneered by giffard et al.@xcite and developed by webb and washburn @xcite .
the principle is to measure the thermal johnson noise @xcite in a resistor to determine its temperature .
the performance of the thermometers developed in @xcite was limited by the noise of the rf squid amplifiers ; in @xcite the typical device noise temperature was a few mk . using a low transition temperature dc squid as the front end amplifier
the current sensing technique becomes a practical approach for measuring absolute temperature from 4.2 k down to sub mk temperatures , as shown in @xcite ; this work used commercial dc squids .
the improved sensitivity of dc squids fabricated at ptb @xcite allows the use of sensor resistors with much larger resistance , thereby increasing the speed of the thermometer . in this paper
we demonstrate experimentally an improvement in speed by orders of magnitude and we demonstrate that the percentage precision is independent of temperature . the principle of operation of a current sensing noise thermometer is discussed in detail in our earlier paper @xcite but is summarised below .
the mean square open circuit noise voltage per unit bandwidth across the terminals of a resistor , @xmath5 , at temperature @xmath6 is given by the nyquist expression : @xmath7 this is a white noise spectrum over the frequency range of interest . in our current sensing noise thermometer , the resistor whose absolute temperature is to be measured is connected in series with the superconducting input coil of the readout squid through a superconducting twisted pair .
the mean square noise current flowing in the squid input coil per unit bandwidth arising from the thermal noise in the resistor is then given by : @xmath8 where @xmath9 .
the time constant @xmath10 , where the total inductance in the input circuit @xmath11 includes the squid input coil inductance @xmath12 and any additional stray inductance including that of the twisted pairs . by increasing the resistance , it is possible to increase the bandwidth of the noise spectrum .
however , this results in a reduced noise power from the resistor at @xmath13 , which must be kept much greater than the squid noise in order to ensure a reasonable device noise temperature , @xmath14 , much less than the minimum temperature we wish to measure .
the squid is operated in flux - locked loop ( fll ) mode where the gain depends only on fixed , known parameters .
thus , if the value of the resistive sensor and the squid gain can be measured , a current noise spectrum can be used to determine absolute temperature using equation [ eq:2 ] .
can be at a different temperature to @xmath15 , the temperature of the squid . ]
figure [ fig:1 ] shows a schematic of the measurement set - up . the squid sensor used in this work is a ptb two - stage squid @xcite with an input coil inductance of @xmath16 @xmath17h and a mutual inductance @xmath18 between the squid and the input coil equivalent to @xmath19 @xmath20a/@xmath21 , where @xmath21 is the flux quantum . when operated with a 10 k@xmath0 feedback resistor , the fll gain is 0.42 v/@xmath21 .
the squid is located remotely from the sensor resistor , allowing the resistor to be placed in environments hostile to squid operation . in this case
the squid is mounted on the continuous heat - exchanger plate of the dilution fridge at approximately 100 mk . at this temperature ,
the white flux noise floor of the squid is @xmath22hz@xmath23 , equivalent to a coupled energy sensitivity , @xmath24 in units of planck s constant , where : @xmath25 and @xmath26 is the flux noise power .
this can be compared to @xmath27 for the commercial quantum design @xcite squid operated at 4.2 k used in our previous work @xcite .
this lower noise floor corresponds to a lower amplifier noise temperature @xmath14 .
thus through choice of resistor the user can minimise the noise temperature for absolute precision or increase the bandwidth of the measurement for increased speed .
the primary focus of this work was to establish the optimum choice of resistance value for operation of the thermometers on a dilution refrigerator ( with a typical base temperature of 5 mk ) .
in addition in order to investigate the @xmath2he melting curve fixed points on the provisional low temperature scale plts-2000 we need a thermometer that can be reliably operated below 1 mk .
for these reason we chose sensor resistance values of 0.2 m@xmath0 , 0.14 @xmath0 and 1.290 @xmath0 .
the 0.2 m@xmath0 resistor was fabricated with a copper foil as described in @xcite . using this material it is not practical to make a sensor with a resistance greater than @xmath28 10 m@xmath0 .
a higher resistivity material was needed for the 0.14 @xmath0 and 1.290 @xmath0 sensors . for this work 50
@xmath17 m diameter platinum tungsten wire ( pt92%/w8% ) @xcite was used .
ptw was chosen as it has a weakly temperature dependant resistivity as well as having a very small magneto - resistance @xcite .
figure [ fig:2 ] is a schematic diagram of the sensor resistor and the sample plate .
the required length of ptw wire is achieved by hard soldering it to two pieces of 100 @xmath17 m thick copper foil separated by a well - defined gap , using johnson matthey easyflo2 solder .
one foil is electrically grounded and consequently heat sunk to a large annealed copper holder with a copper screw contact . the other foil is electrically isolated from the holder .
the sensor resistor is then connected to a pair of niobium screw terminals via short sections of cuni clad nbti wire .
the screw terminals are connected to the remote squid input coil through a shielded nbti superconducting twisted pair .
the resistor and copper former are enclosed in a niobium can for shielding and heat sunk to a copper sample plate with an m4 screw and cone joint as depicted in figure [ fig:2 ] .
for the tests of the 1.290 @xmath0 sensor the sample plate was mounted at the mixing chamber plate of the dilution fridge , the 0.2 m@xmath0 and 0.14 @xmath0 sensors were mounted directly on the top plate of a copper nuclear demagnetisation stage .
.,scaledwidth=75.0% ] in principle , the current sensing noise thermometer is a primary thermometer ; if the resistance of the sensor and the transfer function of the squid are known accurately . in practice however , it is more convenient to stabilise at a known reference temperature and determine the parameters @xmath5 and @xmath11 from a fit of the output noise power . for this single point calibration of our thermometers ,
the sample plate was held at 500 mk , measured by the melting curve thermometer .
the temperature stabilisation was realised with a digital pid temperature controller to within 5 @xmath17k .
the squid output voltage is read out using room temperature magnicon xxf electronics @xcite and is captured with a pxi digitiser card @xcite .
time traces were captured with sampling frequencies of 50 ks / s , 500 ks / s , 2.5 ms / s for 0.2 m@xmath0 , 0.14 @xmath0 and 1.290 @xmath0 sensors respectively , with a record length of 2@xmath29 points .
1000 time domain traces were captured for each sensor .
each captured time trace is individually fourier transformed and the power spectral density is averaged in the frequency domain for improved precision . before fitting , it was necessary to remove the noise power contribution of the squid , which was measured at high frequencies and assumed to be frequency independent over the range used for the fits .
the current noise spectra obtained in this way are shown in figure [ fig:3 ] .
fits were performed on the resulting current noise spectrum using equation [ eq:2 ] , with @xmath30 mk .
the resulting values for the resistance of the noise sensor @xmath5 and total inductance of the circuit @xmath11 are summarised in table [ tab:1 ] .
these parameters were constrained in the fitting routine to determine the temperature of subsequent noise traces .
an alternative method to obtain the single fixed point calibration is to integrate a superconducting fixed point directly into the input circuit of the thermometer .
an additional coil can be added to the superconducting flux transformer , wound around a sample with a known superconducting transition temperature ; thermalized at the same point as the resistive sensor .
the total inductance , @xmath11 , of the input circuit is then sensitive to the state of the superconducting fixed point .
figure [ fig:4 ] shows the temperature dependence of the inductance measured in a current sensing noise thermometer in which we have included an in - situ tantalum @xcite fixed point device ( @xmath31 = 4.45 k ) .
here 12 turns of nbti wire are wound around a tantalum bobbin forming part of the superconducting flux transformer .
the sharp transition of the tantalum is clearly visible in the noise thermometer measurement . , of the noise thermometer input circuit as a function of temperature measured on the calibrated germanium resistance thermometer .
a sharp decrease in inductance is observed when the noise thermometer cools through the superconducting transition of the integrated tantalum fixed point device.,scaledwidth=75.0% ] llll resistance & total inductance & @xmath14 & bandwidth ( 3db ) + 0.20436 m@xmath0 & 1.5016 @xmath17h & @xmath321 @xmath17k & 22 hz + 0.13781 @xmath0 & 1.5483 @xmath17h & 40 @xmath17k & 14.5 khz + 1.290 @xmath0 & 1.315 @xmath17h & 400 @xmath17k & 156 khz +
he melting curve thermometer , a calibrated germanium resistance thermometer and a pulsed platinum nmr thermometer over a wide temperature range . the linear fit ( solid line ) to high temperature data
is also shown .
the decoupling of the thermometer at the lowest temperatures is modelled by fits that assume a metallic thermal link to the nuclear stage with a constant heat leak into the noise thermometer ( dotted lines).,scaledwidth=75.0% ] the current sensing noise thermometers were operated over a wide temperature range from 4.2 k down to 200 @xmath20k . over this range there is not a single thermometer against which a comparison can be made . in this work we used a @xmath33he melting curve thermometer to determine the provisional low temperature scale , plts-2000 ( between 1 and 700 mk ) ; a calibrated germanium thermometer from 50 mk to 4.2 k and a pulsed nmr platinum thermometer , calibrated by the @xmath33he melting curve thermometer , was used below 1 mk .
figure [ fig:5 ] shows this comparison for all three thermometers . at each temperature above 6 mk the temperature was stabilised for 1 hour .
excellent agreement between the thermometers was observed over a wide temperature range .
deviations from the reference thermometer are observable up to progressively higher temperature for higher noise sensor resistance .
this can be accounted for by a heat leak to the resistor resulting in a temperature gradient across the resistor .
reducing the size of the heat leak through improving the heat sinking and filtering of the wires to the ` floating ' side of the sensor resistor will be the subject of future development .
figure [ fig:6 ] shows the percentage deviation of the comparison between the 0.14 @xmath0 sensor and the melting curve thermometer . here
an agreement to better than 1 % is achieved down to the lowest temperatures attainable on a commercial dilution refrigerator .
it is clear that below 1 mk a relatively low sensor resistance is desirable to ensure good thermalization .
the non - contact measurements , which form the basis of the magnetic field fluctuation thermometer @xcite are an alternative method for @xmath34k . at the lowest temperature
obtained , 165 @xmath17k on the platinum thermometer , the 0.2 m@xmath0 sensor cooled to 180 @xmath17k .
current sensing noise thermometers with the @xmath2he melting curve thermometer , down to the lowest temperatures typically reached using a dilution refrigerator .
plotted as a percentage of @xmath35.,scaledwidth=75.0% ] another figure of merit for a thermometer is the measurement time required to obtain a given precision of temperature measurement .
if we consider capturing a time trace of @xmath36 points with a time @xmath37 between each point , for each point to be an independent measure of the square of the noise current , @xmath37 must be of order the correlation time of the noise or greater ( which for this system is the time constant @xmath38 ) .
therefore in a measuring time @xmath39 one can make approximately @xmath40 independent measurements . a rough estimate for the precision in a given measuring time
is thus given by : @xmath41 the percentage precision in a given measuring time is thus independent of temperature and determined by the noise bandwidth .
the noise bandwidth for the sensors used is given in table [ tab:1 ] , and can be seen in figure [ fig:3 ] . increasing the resistance of the sensor results in an improved precision in a given time , which was the motive behind using ptw resistors with a larger resistance value . to experimentally determine the precision of the 1.29 @xmath0 ptw thermometer ,
the sample plate was stabilised at a known fixed temperature .
single shot noise traces were taken consecutively .
approximately 1000 individual spectra were captured for accurate statistics .
a sampling frequency of 2.5 ms / s with 2@xmath29 points was used , giving a measurement time of 419 ms per trace .
figure [ fig:7]a shows data taken with the ptw thermometer stabilised at 493 mk using the germanium thermometer .
this follows a normal distribution as shown by fig.[fig:7]b , where a gaussian fit to a histogram of temperatures is shown .
the precision is calculated using @xmath42 , where @xmath43 is the standard deviation of the data and @xmath6 is the mean temperature . in a measurement time of 419 ms , a precision of 0.50 % was obtained .
this precision measurement was repeated at 1.5 k , 100 mk and at 13 mk ; the percentage precision was confirmed to be independent of temperature . reducing the measurement time to 105 ms , resulted in a temperature independent precision of 1.15 % .
we contrast this with the pioneering work on current sensing noise thermometers @xcite , which required a measurement time of 2500 s to achieve a similar performance .
the 1.29 @xmath0 ptw thermometer is about two orders of magnitude faster than our previous noise thermometers @xcite , magnetic field fluctuation thermometers @xcite and on - chip current sensing noise thermometers @xcite , which all have achieved 1 % precision in times of the order of 10 s. ultra - low temperature measurements using a contact free cross - correlation noise thermometer @xcite have been reported down to 45 @xmath17k , here using two squids the cross - correlation technique allows them to reduce the noise contribution from squid amplifiers , and the reported measurement time was 400 seconds .
ptw current sensing noise thermometer , stabilised with a germanium thermometer at 493 mk .
the x axis shows the real time taken to capture the traces .
each temperature reading was acquired in a measurement time of 419 ms .
the horizontal line of best fit is the mean temperature .
b : a histogram of temperatures showing the scatter follows a normal distribution.,scaledwidth=75.0% ] figure [ fig:8 ] illustrates the dependence of the precision obtained from all three of the current sensing noise thermometers as a function of the measurement time at a temperature of 100 mk . to vary the measurement time ,
the sampling frequency was kept constant at 50 ks / s , 500 ks / s , 2.5 ms / s for 0.2 m@xmath0 , 0.14 @xmath0 and 1.290 @xmath0 respectively ; 2@xmath29 points were captured for each trace and the number of averages was changed to give total measurement times of between 0.1 and 250 seconds .
the best fit lines shown are proportional to the inverse of the square root of the measurement time , as in equation [ eq:4 ] .
the measured precision of 0.03 % obtained with the 1.29@xmath0 , sensor in a measurement time of 100 s , is approaching the limit of our ability to stabilise the temperature at 100 mk . sensors .
, scaledwidth=75.0% ]
in conclusion , a series of fast , precise current sensing noise thermometers based on a range of resistive sensors has been tested . using a single temperature calibration ,
the noise thermometers can be used to accurately measure a wide range of temperatures .
we have presented data on a current sensing noise thermometer that has been optimised for speed , taking advantage of the improvements in squid noise and bandwidth .
this is two orders of magnitude faster than previous noise thermometers , achieving 1.15 % precision in just 0.1 seconds .
the 0.14 @xmath0 sensor tested is in agreement with the @xmath2he melting curve thermometer to within better than 1 % over the operating temperature range of a dilution refrigerator , and in this case less than 2 seconds of measurement time is required for 1 % precision .
finally the lowest value of resistance chosen , 0.2 m@xmath0 resulted in a thermometer with a noise temperature of 1 @xmath17k that showed reasonable agreement with a pulsed platinum nmr thermometer to below 200 @xmath17k . the current sensing noise thermometers described here constitute practical devices for disseminating thermodynamic temperature .
the system is user - friendly and could be installed on any commercial ultra - low temperature platform , to provide a reliable and traceable measure of thermodynamic temperature , with high precision ( 0.1 % ) and high accuracy ( better than 1% ) in a reasonable measurement time ( @xmath32 10 s).the potential provision of absolute temperature to all users promises to be a revolutionary technical development in a temperature range expected to be of growing technological and commercial significance .
this work was supported by the european metrology research program ( emrp ) .
the emrp is jointly funded by the emrp participating countries within euramet and the european union .
additional support was through the microkelvin project .
we acknowledge the support of the european community - research infrastructures under the fp7 capacities specific programme , microkelvin project number 228464 .
g. batey , a. casey , m. n. cuthbert , a. j. matthews , j. saunders , a. shibahara , accepted for new j. phys .
o. v. lounasmaa , _ experimental principles and methods below 1 k _ , academic press ( 1974 ) . c. enss , s. hunklinger , _ low temperature physics _ , springer ( 2005 ) .
f. pobell , _ matter an methods at low temperatures _ , springer ( 2007 ) .
bipm , procs - verbaux des sances du comit international des poids et mesures * 68 * , 128 ( 2001 ) .
r. l. rusby , m. durieux , a. l. reesink , r. p. hudson , g. schuster , m. khne , w. e. fogle , r. j. soulen , d. e. adams , j. low temp . phys . * 126 * , 633 ( 2002 ) .
c. p. lusher , junyun li , v. a. maidanov , m. e. digby , h. dyball , a. casey , j. nyki , v. v. dmitriev , b. p. cowan and j. saunders , meas .
* 12 * , 1 ( 2001 ) . t. varpula and h. sepp , rev .
instrum . *
64*(6 ) , 1593 ( 1993 ) .
j. beyer , d. drung , a. kirste , j. engert , a. netsch , a. fleischmann , and c. enss , ieee trans .
17 , 2 ( 2007 ) .
j. engert , j. beyer , d. drung , a. kirste , d. heyer , a. fleischmann , c. enss and h .- j .
barthelmess , j. phys .
c 150 , 12012(2009 ) .
d. rothfu , a. reiser , a. fleischmann , c. enss , appl .
* 103 * , 052605 ( 2013 ) .
j. p. pekola , k. p. hirvi , j. p. kauppinen , and m. a. paalanen , phys . rev
. lett . * 73 * , 2903 ( 1994 ) .
r. p. giffard , r. a. webb , j. c wheatley , j. low temp . phys .
* 6 * , 533 - 610 ( 1972 ) .
r. a. webb , r. p. giffard , j. c. wheatley , j. low temp . phys . * 13 * , 383 - 429 ( 1973 ) .
r. a. webb , s. washburn , aip conference proceedings , * 103 * , 453 - 466 ( 1983 ) j. b. johnson , phys . rev . * 32 * , 97 ( 1928 ) .
d. drung , c. amann , j. beyer , a. kirste , m. peters , f. ruede , th .
schurig , ieee trans .
* 17 * , 699 ( 2007 ) .
quantum design , 11578 sorrento valley road , san diego , ca 92121 - 1311 , usa .
a. casey , b. p. cowan , h. dyball , junyun li , c. p. lusher , v. maidanov , j. nyki , j. saunders , dm .
shvarts , physica b * 329 - 333 * , 1556 ( 2003 ) .
goodfellow cambridge ltd , ermine business park , huntingdon , england pe29 6wr .
g. s. cieloszyk , p. j. cote , g. l. salinger , and j. c. williams , rev .
. instrum .
* 46 * , 1182 ( 1975 ) .
magnicon gmbh , lemsahler landstr .
17 , hamburg , germany .
national instruments corporation , 1105 n mopac expressway , austin , tx 78759 - 3504 , usa . j. g. c. milne , phys . rev . *
122 * , 387388 ( 1961 ) . | we describe the design and performance of a series of fast , precise current sensing noise thermometers .
the thermometers have been fabricated with a range of resistances from 1.290 @xmath0 down to 0.2 m@xmath0 .
this results in either a thermometer that has been optimised for speed , taking advantage of the improvements in superconducting quantum interference device ( squid ) noise and bandwidth , or a thermometer optimised for ultra - low temperature measurement , minimising the system noise temperature . with a single temperature calibration point
, we show that noise thermometers can be used for accurate measurements over a wide range of temperatures below 4 k. comparisons with a melting curve thermometer , a calibrated germanium thermometer and a pulsed platinum nuclear magnetic resonance thermometer are presented . for the 1.290 @xmath0 resistance we measure a 1 % precision in just 100 ms , and
have shown this to be independent of temperature .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
if the galactic dark halo were composed entirely of massive compact halo objects ( machos ) , the instantaneous probablity of microlensing ( the `` optical depth '' ) towards the large magellanic cloud ( lmc ) would have been @xmath15 ( pacznski 1986 ) .
the most recent observational result for machos with masses up to about 1 @xmath5 is @xmath16 , which comes from an analysis of 5.7 years of lmc microlensing survey data ( alcock et al .
thus an all - macho galactic dark halo is ruled out . however , the observed optical depth is significantly higher than the estimate for known galactic and lmc stellar populations , @xmath17 ( alcock et al .
the excess lmc microlensing signal may be telling us the macho fraction of the galactic dark halo or revealing gaps in our understanding of the essential structure of the lmc and galaxy .
critical discussions of the microlensing optical depth from known stellar populations have developed along two lines .
one debate has been about the possible existence of a `` new '' stellar population which would account for the lmc microlensing result .
some example suggestions include an intervening dwarf galaxy or tidal stream ( zhao 1998 ; zaritsky & lin 1997 , zaritsky et al .
1999 ) , and a very warped galactic disk ( evans et al . 1998 ) .
these hypotheses have been tested , and in some cases ruled out ( alcock et al .
1997 ; beaulieu & sackett 1998 ; bennett 1998 ; gould 1998 , 1999 ; gyuk , flynn , & evans 1999 ; ibata , germaint , & beaulieu 1999 ; johnston 1998 ) .
the second debate has been about the importance of lmc star - star `` self - lensing '' ( sahu 1994 ; wu 1994 ) . in particular , although the lmc is well represented by an exponential disk model which often serves as the basis for the self - lensing optical depth calculation , it may exhibit some important detailed structures .
our paper is motivated by this lmc self - lensing problem .
the line - of - sight velocity dispersion of lmc stars yields a strong constraint on star - star self - lensing from the virialized lmc disk : @xmath18 ( gould 1995 ) .
gould s elegant limit is probably uncertain by no more than a factor of @xmath72 .
however , several lmc models have been devised which would increase the self - lensing optical depth over the gould limit .
one example of a detailed structure that might increase the self - lensing optical depth is a highly - inclined and flared lmc disk ( zhao 1999 ) .
it has also been suggested that the oldest lmc disk stars have a very large characteristic scale height , or that the lmc harbors an as yet unseen but massive stellar spheroid 2 - 4 for the former , and a very large
m / l for the latter ) , and the density profiles . ]
( e.g. aubourg et al
. 1999 ; salati et al . 1999 ; evans & kerins 2000 ) . finally , it is possible that a nonvirialized stellar component ( i.e. , a shroud of tidal debris ; weinberg 2000 ) acts to increase the self - lensing optical depth ( zhao 1998 ) .
kinematic studies play a critical role in testing these various proposed lmc models . for example
, the flare of the lmc disk may be inferred from its velocity dispersion at different radii , and the existence of a non - equilibrium stellar component lying near the lmc might be proven with a large kinematic survey ( zhao 1999 ; graff et al . 1999 ; see also zaritsky et al .
1999 ; ibata et al .
in addition , one could search directly for an lmc stellar population with a velocity dispersion of @xmath750 km s@xmath1 , the prediction for a spheroid or very thick disk .
these latter components must be identified before their importance to the self - lensing optical depth can be assessed .
the lmc rotation curve provides a framework for detailed kinematic studies such as these .
the decomposition of the lmc rotation curve into disk and dark halo components has several implications for microlensing .
first , as emphasized by gyuk et al .
( 1999 ) , the total mass of the lmc is an important constraint on self - lensing optical depth calculations .
a high - mass model will typically predict a high self - lensing optical depth , with an additional dependence on whether the mass lies mostly in the disk or in a halo .
unfortunately , recent analyses of the lmc rotation curve have lead to mass estimates that range over a factor of @xmath74 , which can be attributed primarily to different assumptions about a dark halo ( kim et al .
1998 , kunkel et al .
1997 , schommer et al .
second , a massive lmc dark halo might significantly affect the lmc disk kinematics ( e.g. bahcall 1984 ) , with possible consequence for constraints on self - lensing or determining the flare of the disk
. the influence of an lmc dark halo on the lmc disk kinematics has not been investigated in detail .
we note that the interplay between the disk and dark halo is also of general interest for studies of galaxy formation and evolution .
finally , if the lmc has a dark halo with a physical makeup similar to that of the galactic dark halo ( i.e. composed partly of machos ) , the macho fraction of the galactic dark halo implied by the observed optical depth would be lowered ( alcock et al .
2000 ) . motivated by the above considerations , we present a new analysis of the lmc rotation curve . in this work
we are particularly concerned with the flare of the lmc disk . however , other issues pertinent to microlensing are discussed as our calculations and analyses permit .
we refer extensively to the rotation curve recently presented by kim et al .
( 1998 ) and the impressive radial velocity dataset for lmc carbon stars summarized by kunkel , demers , irwin & albert ( 1997 ; hereafter kdia ) .
a significant subset of these latter data are public ( kunkel , irwin & demers 1997 ) , and archived electronically at the cds .
we also refer extensively to the analysis of the lmc space motion presented by kroupa & bastian ( 1997 ) .
our paper is organized as follows . in 2 , we present the impetus for this work , a comparison of constant - thickness exponential disk models ( with no dark halos ) to the lmc disk velocity dispersions reported by kdia . in 3 , we reanalyze the lmc rotation curve and velocity dispersion curve using archived carbon - star radial - velocity data . in 4 , we present a multi - mass component kinematic model for the lmc . in 5 , we compare our carbon - star rotating disk solution to our model .
we discuss the microlensing implications of our analyses in 6 and conclude in 7 .
disks in galaxies often have exponential surface brightness profiles from which we infer exponential mass surface density profiles of the form @xmath19 by assuming constant mass - to - light ratios ( @xmath20 ) . in equation
( 1 ) , @xmath21 is the central mass surface density , @xmath22 is the radial scale length , and @xmath9 is a cylindrical radial coordinate . for an infinitely thin disk described by equation ( 1 ) , freeman ( 1970 )
calculated that the rotation curve reaches a maximum circular velocity at @xmath23 @xmath24 where @xmath25 is the total mass of the disk , which is related to the central surface density by @xmath26 for the case of a disk with a modest finite thickness , @xmath27 will be decreased by @xmath28 ( van der kruit & searle 1982 ) .
a useful model with a finite thickness is the exponential disk ( also known as a double - exponential disk ) , which has a density distribution @xmath29 in equation ( 4 ) , the variable @xmath8 is the vertical , or what we will sometimes refer to as the `` z '' scale height . if we choose the spatial density normalization @xmath30 , the exponential disk projects to the surface density of the infinitely thin freeman disk , i.e. equation ( 1 ) .
the z - velocity dispersion , @xmath31 , for the exponential disk is given by wainscoat , freeman and hyland ( 1989 ) , @xmath32^{1/2}\ ] ] which assumes that the disk is virialized .
equation ( 5 ) is strictly valid for a radially infinite disk .
the radial dependence ( i.e. , the @xmath33 term ) in equation ( 5 ) scales the local density normalization , as prescribed by others ( van der kruit & searle 1982 ; van der kruit & freeman 1984 ) .
a model extensively discussed by van der kruit & searle ( 1981 , 1981b , 1982 ) is the isothermal disk .
the vertical distribution of stars in an isothermal disk is described by the sech@xmath34 function , as originally derived by spitzer ( 1942 ; see also 4.2 of this paper ) .
the spatial density for this disk model is , @xmath35 which projects to the surface density of the infinitely thin freeman disk if @xmath36 .
assuming virialization , the z - velocity dispersion of the isothermal disk is proportional to the z - scale height and independent of z : @xmath37^{1/2}\ ] ] where the radial dependence scales the local density normalization .
equation ( 7 ) is strictly valid for a radially infinite disk .
we note that the `` z - scale height '' in the isothermal disk is not strictly a scale height ( and not the same as @xmath8 in the double - exponential model ) , but it is a similar parameter .
further comparisons of the exponential and isothermal disk models are found in wainscoat , freeman & hyland ( 1989 ) , and van der kruit ( 1988 ) .
the systematic error in the velocity dispersion predicted by the radially infinite isothermal disk model has been shown to be only @xmath710% in the range @xmath38 using numerical studies of truncated isothermal disk models ( van der kruit & searle 1982 ) .
we will use the isothermal disk model in our analyses , because these numerical calculations provide us with an estimate of the model accuracy , and will later allow us to approximately account for the finite radial extent of the lmc disk ( see 4.3 of this paper ) .
observations of edge - on spirals indicate that both the exponential and isothermal models are viable representations of real disks ( de grijs , peletier , & van der kruit 1997 )
. the specific choice of model does not affect the main results of this work .
it is reasonable to compare the model z - velocity dispersion to the observed velocity dispersion at different projected radii in the lmc for the case of uniform anisotropy ( @xmath39 constant at different radii ; van der kruit & freeman 1984 ) .
although we do not know if this is true for the lmc , the assumption of uniform anisotropy is supported by kinematic studies of the old galactic disk ( lewis & freeman 1989 ) .
gould ( 1995 ) notes that if @xmath40 1 as found elsewhere , the line of sight velocity dispersion measured for an inclined disk overestimates @xmath41 .
therefore , our model z - velocity dispersion is likely an upper - limit on the true z - velocity dispersion . combining equations ( 2 ) , ( 3 ) and ( 7 ) yields @xmath42 which is most accurate for disks in the range @xmath38 .
the z - scale height has been observed to be constant over a large range of @xmath9 in numerous disk galaxies ( van der kruit & freeman 1984 ; de grijs , peletier , & van der kruit 1997 ) , implying that @xmath43 decreases with increasing @xmath9 in these galaxies .
indeed , this trend of decreasing velocity dispersion has been observed in over a dozen spiral / disk galaxies ( bottema 1993 ) , including our own . in summary , if the disk velocity dispersions are the same at different radii , the disk is likely flared ( e.g. zhao 1999 )
however , this inference assumes that ( 1 ) the disk is virialized , ( 2 ) the radial scale length is constant , ( 3 ) the disk velocity dispersions exhibit uniform anisotropy , and ( 4 ) the galaxy has no dark halo ( or that the dark halo has a negligible affect on the disk velocity dispersions ) .
finally , one must be careful to compare the observed velocity dispersions to the simple model described above over a restricted range of radii , because real galactic disks are finite in radial extent .
we compare to the lmc in the next section . in this section , we will adopt a plausible value for the product @xmath44 in equation ( 8) , and compare that model to lmc disk velocity dispersion data .
following kdia and many others , we adopt a distance to the lmc of 50.1 kpc , and an inclination of @xmath14 = 33 degrees . for comparison ,
wide - field uv polarimetric image data have yielded a precise but model - dependent estimate of the inclination of the lmc disk : @xmath45 degrees ( cole , wood , & nordsieck 1999 ) .
westerlund ( 1997 ) summarizes a number of other inclination measurements .
we adopt @xmath46 kpc for the radial scale length of the lmc disk .
the surface brightness profile of the lmc disk is well - fit by an exponential with a scale length @xmath47 kpc ( de vaucouleurs 1957 , bothun & thompson 1988 ) .
moreover , the surface density profile of intermediate - age long period variables ( lpvs ) in the lmc is also fit by an exponential with a scale length of @xmath48 kpc ( hughes , wood & reid 1991 ) .
this result directly associates our kinematic dataset with an lmc population that follows the @xmath22 = 1.6 kpc exponential profile , because some of the intermediate - age lpvs are also carbon stars .
finally , the rr lyrae stars in the lmc also appear to lie in a @xmath47 kpc exponential disk ( alcock et al .
since the rr lyraes are presumably much older than the carbon stars , the carbon stars were likely born into the same disk .
we note that no stellar population in the lmc has yet been shown to be inconsistent with a @xmath46 kpc exponential profile .
we assume that the carbon stars observed by kdia properly represent this lmc disk .
the kdia dataset consists of radial velocity measurements for 759 carbon stars spanning a true lmc radius of 2 to 10 kpc .
kdia summarize the results of 11 different zonal solutions for a rotating disk . in figure 1
, we plot the kdia velocity dispersions as a function of radial distance , which we take as the central value for each zonal solution given in kdia s table 1 .
we show data only for @xmath9 = 3 to 6 kpc .
this is the range of radii where equation ( 8) will most accurately predict the velocity dispersions . the dashed line in figure 1
is the mean of the plotted , observed velocity dispersions , @xmath49 13.7 km s@xmath1 . assuming a constant value of @xmath43 equal to the mean , we find @xmath50 = 0.4 , consistent with the data .
the solid line plotted in figure 1 shows the prediction of equation ( 8) assuming values of @xmath51 km s@xmath1 and @xmath8 = 0.35 kpc at @xmath52 .
this model predicts @xmath53 km s@xmath1 at @xmath54 kpc .
the fit for this model distribution is @xmath50 = 8.7 , which can be rejected with high significance .
models with larger scale heights intersecting the data at larger @xmath9 give similarly poor fits are not strictly apppropriate because the zonal solutions given by kdia are not all independent .
however , subsets of the @xmath55 measurements which are independent also give poor fits to the distribution given by equation ( 7 ) . ] .
one possible interpretation of the observed constancy of @xmath43 with @xmath9 is that the z - scale height increases as @xmath56 , i.e. _ the lmc disk is flared . _
an alternate interpretation of figure 1 might be a constant - thickness disk under the dynamical influence of a dark halo ( bahcall 1984 ) . indeed , kim et al .
( 1998 ) compared their rotation curve with the kdia carbon star rotation curve and , on this basis , argued that an lmc dark halo is dynamically significant at a radius of @xmath74 kpc .
however , kdia favored a different interpretation ; they attributed the rising outer portion of their carbon star rotation curve to tidal effects , and not a dark halo curve in figure 1 , supporting this aspect of our analysis in 2 . ] . before concluding that the lmc disk is flared , it is worth testing disk plus dark halo kinematic models in a self - consistent manner . with this model comparison in mind ,
let us consider the kdia analysis in greater detail .
we have several specific concerns with the kdia analysis .
first , and perhaps most importantly , kdia excluded some carbon stars from their solutions for their association with a polar ring .
this procedure might have artificially lowered the disk velocity dispersions . by retaining these stars in our solutions , we guarantee that our results will be properly comparable to our model ( which will not include a polar ring ) .
kdia also corrected for the lmc transverse motion by forcing the position angle of the kinematic line of nodes to match the photometric line of nodes in each zonal solution .
this is a questionable procedure .
moreover , it is inconsistent with the analysis by kim et al . (
we will correct the carbon - star velocities for the transverse motion of the lmc in a manner consistent with the rotation curve analysis by kim et al .
in fact , we will make our carbon star rotation curve analysis similar in other ways to the rotation curve analysis ( i.e. , by adopting the same kinematic center ) , in order to lend maximum credibility to detailed comparisons of the two rotation curves .
we assemble positions ( right ascension and declination ; equinox 1950 ) and galactocentric radial velocities for magellanic cloud carbon stars from kunkel , irwin & demers ( 1997 ) .
we discard all small magellanic cloud ( smc ) carbon stars , `` inter - cloud '' carbon stars ( their table 15 ) , and carbon stars located near the center of the lmc ( their table 17 ; see also 3.3 ) .
we also discard one rogue carbon star ( c0433@xmath576607 ) for its highly discrepant velocity .
the resulting homogeneous dataset contains 422 carbon stars , representing a significant subset of the data analyzed in kdia .
ongoing observational campaigns to obtain carbon star radial velocities will likely increase the sample size in the coming years ( suntzeff 1998 ) .
the archival dataset we have assembled is sufficient for the purposes of this study .
feast , thackeray & wesselink ( 1961 ) have discussed the apparent rotation induced by the transverse motion of the lmc .
each of our 422 carbon star velocities are corrected for this projected radial velocity gradient as follows .
we adopt the lmc space motion calculated by kroupa & bastian ( 1997 ) , which is derived from an average of the lmc proper motion they measure with _ hipparcos _ data and the lmc proper motion measured by jones , klemola , & lin ( 1994 ) .
the _ hipparcos _ and jones et al .
( 1994 ) measurements of the lmc proper motion agree within their respective error bars .
we note that kim et al .
( 1998 ) make their transverse motion correction using the proper motion measurement by jones et al .
( 1994 ) .
it is useful to establish a galactic coordinate system for the vector analysis that follows .
following kroupa & bastian ( 1997 ) , the galactocentric coordinates used here are such that the galactic center and the sun are located at 0 and @xmath58 kpc along the @xmath59-axis , respectively .
the positive @xmath60-axis points toward the north galactic pole , and the positive @xmath61-axis points in the direction of galactic rotation at the position of the sun . in these coordinates ,
the position vectors of the sun and lmc are @xmath62 @xmath63 in units of kpc .
the velocity vector of the lmc is @xmath64 in units of km s@xmath1 ( kroupa & bastian 1997 ) . the galactocentric radial velocity of the lmc is the vector dot product @xmath65 km s@xmath1 , where @xmath66 is a normalized unit vector .
the vector between the sun and lmc is @xmath67 , and the radial velocity component of the lmc seen from the local standard of rest ( accounting for the sun s peculiar velocity and galactic rotation ) at the position of the sun is @xmath68 km s@xmath1 .
the vector @xmath69 points toward the `` center '' of the lmc , which is assumed to be at @xmath70 , @xmath71 ( kroupa & bastian 1997 ) .
the unit vector connecting the sun to any position in the sky in galactocentric coordinates is @xmath72 the transverse motion velocity correction that we seek is the difference between the projected radial velocity toward the lmc center and the projected radial velocity toward each carbon star .
if we define the vector @xmath73 , the apparent rotation induced by the lmc space motion is @xmath74 .
we make our correction by subtracting @xmath75 from the observed carbon star velocities .
small differences between the kroupa & bastian ( 1997 ) model used here for the transverse motion radial velocity correction and assumptions ( i.e. , for the lmc center and distance ) in subsequent sections have a negligible effect on our results .
first , we convert each carbon star s right ascension and declination into spherical coordinates in units of radians .
we designate these @xmath76 and @xmath77 , respectively .
we adopt the center of the lmc at right ascension @xmath78 and declination @xmath79 ( equinox 2000 ) following kim et al .
we designate the lmc center with @xmath80 and @xmath81 ( also in units of radians and precessed to equinox 1950 ) .
the distribution of carbon stars is shown in figure 2 ( where north is up and east is to the right ) .
the smc is located well outside of the boundaries of the figure in a south - westerly direction ( toward the lower left ) .
we remind that @xmath76 and @xmath77 are spherical coordinates plotted on a rectilinear scale .
therefore , the distribution of carbon stars shows a @xmath7cos(@xmath77 ) distortion .
our solution proceeds as follows .
we calculate the position angle ( @xmath82 ) and radial distance from the center of the lmc ( @xmath83 in units of radians ) for each carbon star using the formulae @xmath84\ ] ] @xmath85\ ] ] @xmath86 equation ( 15 ) assigns the correct sign and phase to the position angle depending on which quadrant each star resides in figure 2 .
we have defined @xmath82 to increase west of north .
we deproject each radial distance on the sky to true lmc radius ( @xmath9 in units of kpc ) using , @xmath87^{1/2 } \
\right ) \right|\ ] ] we take the absolute value of the inverse tangent in equation ( 16 ) to guarantee a positive value for @xmath9 .
we accommodate a twisting disk model by introducing @xmath88 .
the parameter @xmath88 allows for a changing kinematic line of nodes in different radial zone solutions .
we note that @xmath88 is sensitive to the adopted lmc transverse velocity ( meatheringham et al .
we define our different zones by making cuts in true radius .
each zonal solution is derived by fitting the function @xmath89 to the distribution of galactocentric radial velocities , @xmath90^{1/2 } } \ + \ v_{sys } \right)\ ] ] we must choose the correct sign for @xmath89 at the appropriate position angle , which is the meaning of the ` @xmath0 ' in equation ( 17 ) .
we choose ` @xmath91 ' for @xmath92 = @xmath93 to @xmath94 , and ` @xmath57 ' for @xmath92 = @xmath94 to @xmath95
. the fitted model could be generalized to account for a warped and decentered disk by letting @xmath14 and ( @xmath80 , @xmath81 ) vary in each zonal solution .
however , we fix @xmath14 , @xmath80 , and @xmath81 to the values given above in all solutions .
we derive solutions in four radial zones which span @xmath9 = 2.5 - 5 , 5 - 6 , 6 - 7 , and 7 - 13 kpc .
these were chosen to provide an adequate number of stars and phase coverage in each zone .
the inner - most and outer - most boundaries were chosen to be inclusive of the entire dataset when @xmath88 = 0 .
we begin by solving for @xmath96 , @xmath88 , and @xmath97 that yield a minimum dispersion about @xmath89 in each of our zones .
we average the values of @xmath96 , excluding the outermost zone ( which was sometimes poorly constrained ) . in a second round of fitting ,
we fix @xmath96 to the three - zone average value , and search only for values of @xmath88 and @xmath97 that minimize the velocity dispersion .
this procedure was easily reproduced , and facilitated our bootstrap and monte carlo error analyses ( described below ) .
the mean radii in each final zonal solution are @xmath98 = 4.0 , 5.6 , 6.5 , and 7.8 kpc .
the minimum velocity dispersions occurred at @xmath88 = @xmath5737 , @xmath5722 , @xmath578 , and 3 degrees .
we find circular velocities @xmath97 = 72 , 68 , 65 , and 65 km s@xmath1 and velocity dispersions @xmath43 = 17.7 , 14.9 , 19.3 and 20.6 km s@xmath1 , respectively ( in order of increasing radius ) .
the four zonal solutions are shown in figure 3 .
we indicate the carbon stars with small open circles .
each carbon star is plotted twice at @xmath82 and ( @xmath82 + @xmath99 ) .
the best fit curve is plotted as a solid line .
the mean radius , circular velocity , and velocity dispersion are also labeled in each panel .
our first estimate of the errors is made with a bootstrap resampling analysis ( barrow , sonada , & bhavsar 1984 ) .
we create 250 artificial datasets by randomly removing 1 - 10% of the carbon stars from our full dataset , and then find new rotating disk solutions for each of these . in a bootstrap reasampling analysis
, the variances in the mean values of interest averaged over all of the artificial datasets are taken to represent the internal variances of the full dataset .
we find errors in @xmath97 of 1.3 , 1.7 , 2.3 , and 2.4 km s@xmath1 for the four radial zones ( in order of increasing radii ) .
we find an error in @xmath43 of 0.3 , 0.2 , 0.4 , and 0.5 km s@xmath1 , and errors in @xmath88 of 1.1 , 1.0 , 2.4 , and 1.5 degrees , respectively . our second calculation is designed to estimate the errors associated with our choice of zonal boundaries and the movement of stars between zones due to the changing @xmath88 parameters . for each of the four zones
, we find new solutions for 64 different variations in the zonal boundaries .
we varied each boundary by no more than @xmath0 0.25 kpc , which yielded roughly comparable numbers of stars and mean radii in each zonal solution ( except for the outer boundary of the outer - most zone which was varied by 1.5 kpc in steps of 0.25 kpc ) .
the variances of @xmath97 , @xmath43 , and @xmath88 in the zonal boundary solutions are similar to those found with the bootstrap analysis . moreover
, we find no systematic trends with boundary choice , except for the outermost zone . in the outermost zone
, @xmath97 decreases systematically ( from @xmath760 to 30 km s@xmath1 ) as the inner boundary is varied from @xmath76.7 to 7.2 kpc .
finally , we perform a monte carlo calculation to propagate the error on the lmc space motion to the rotation curve .
we create 500 artificial datasets , where the lmc space motions are drawn randomly from a sample that reproduces the best fit values and errors in equation ( 11 ) .
we assume gaussian distributions for the space motion errors .
this calculation yields errors in @xmath97 of 7.1 , 8.7 , 10.4 and 15.8 km s@xmath1 , errors in @xmath43 of 0.1 , 0.2 , 0.3 and 0.4 km s@xmath1 , and errors in @xmath88 of 6.6 , 4.6 , 7.6 and 8.4 degrees , in the four zones ( in order of increasing radius ) .
similar values were found with as few as @xmath7250 artificial datasets , indicating that our calculation is robust . for most of the parameters in our rotating disk solution ,
the error from the lmc space motion dominates .
our final solutions and adopted errors are summarized in table 1 .
we note that there would be no stars in common between the different solutions if all zones had the same @xmath88 , but in our final solutions , @xmath75% of the stars are located in two zones ( see table 1 ) .
however , since we have accounted for the movement of small numbers of stars across the zonal boundaries in our error analysis , we may consider each zonal solution as independent in subsequent analyses .
we previously discarded 22 radial velocity measurements for carbon stars very near the center of the lmc ( kunkel , irwin & demers 1997 ; their table 17 ) because this sample of stars did not span an adequate range of @xmath82 to yield an interesting zonal rotation curve solution .
however , the velocity dispersion of these carbon stars is of considerable interest .
this sample of stars at @xmath100 kpc has a velocity dispersion of @xmath101 km s@xmath1 .
guided by our error analyses above , we adopt an error of 1 km s@xmath1 .
the average velocity dispersion of carbon stars in our five zones is @xmath43 = 18.9 km s@xmath1 .
we defer further discussion and interpretation of the lmc rotation curve and disk velocity dispersions until 5 , after we present our theoretical kinematic model .
we will consider two representations of the contribution of the luminous mass in the lmc disk to the rotation curve .
the first is derived empirically by kim et al .
( 1998 ) from the @xmath9-band surface brightness data of de vaucouleurs ( 1958 ) . for two assumptions of the disk mass - to - light ratio
, we present this stellar rotation curve in table 2 . the second representation of the disk rotation curve we will consider is the finite - thickness , truncated disk model of van der kruit and searle ( 1982 ) .
we adopt their model no . 2 ( truncated at @xmath9 = 4@xmath22 ) from their appendix a.1 .
table 3 summarizes this disk model rotation curve , which is given in units of the radial scale length ( @xmath22 ) and the equivalent maximum circular velocity for an infinitely thin exponential disk ( @xmath27 ; see 2.1 of this paper ) .
this model will also be used to calculate a correction to the scale heights inferred from the disk velocity dispersions at small radii in the lmc ( see 4.3 of this paper ) . when comparing either of these two disk rotation curves to our carbon star data points , we will add in quadrature the small additional contribution of gas in the lmc disk to the rotational velocity .
kim et al . ( 1998 ) calculated the gas rotation curve using the single - dish data of luks & rolfs ( 1992 ) , and assuming a 30% contribution from .
the gas rotation curve is summarized in table 4 .
finally , we will account for a pseudo - isothermal dark halo in the traditional manner , assuming a density profile , @xmath102 where the circular velocity is a function of the integrated total mass @xmath103\ ] ] the psuedo - isothermal dark halo is parameterized by a central density , @xmath104 , and core radius , @xmath105 .
we add the circular velocities of the disk ( @xmath106 ) and psuedo - isothermal halo ( @xmath107 ) in quadrature in order to calculate the rotation curve of our model : @xmath108 .
the effect of a spherical dark halo on the velocity dispersion of an embedded disk has been extensively discussed by bahcall ( 1984 ) , bahcall & casertano ( 1984 ) and bottema ( 1993 ) .
the vertical motions of stars in a disk and halo system are governed by poisson s equation @xmath109 where @xmath110 and the first moment of the boltzmann equation @xmath111 the effective halo density , @xmath112 , includes the contribution from the radial deriviative of the total circular velocity , @xmath97 . if the rotation curve is fairly flat , we may discard the radial deriviative term in equation ( 22 ) and equate @xmath113 .
this formulation of the problem ( bahcall 1984 ) neglects the @xmath114 cross terms in poisson s equation . with the following redefinitions ( bahcall 1984 ) @xmath115 and @xmath116 we rewrite the equations governing the vertical motions of the disk stars as @xmath117 with the boundary conditions @xmath118 the solution for @xmath119 = 0 was first published by spitzer ( 1942 ) : @xmath120 the solution to equation ( 25 ) for @xmath121 is @xmath122 the differential equation ( 25 ) must be solved numerically for intermediate values of @xmath119 .
the @xmath119 parameter define in equation ( 24 ) is the ratio of effective halo density to the disk density in the plane of the disk .
it is particularly useful parameter because it relates ratio of the velocity dispersion in a disk plus halo system , @xmath123 , to the velocity dispersion in a disk with no halo , @xmath124 .
following bottema ( 1993 ) , and suppressing the subscript @xmath60 for clarity , @xmath125 where @xmath76 is the surface density @xmath126 in the limit of @xmath121 , @xmath127 we have solved for @xmath128 as a function of @xmath119 numerically ; these values are listed in table 5 . in the limit of large @xmath119
, these values confirm equation ( 31 ) .
the error in @xmath76 is @xmath730% at @xmath129 and less than 20% at @xmath130 , in the sense that equation ( 31 ) overestimates the numerically calculated value of @xmath76 .
finally , we note that the density profile of the dark halo may be derived from the observed velocity dispersion in the disk at large @xmath9 , if the disk scale height is constant .
equating @xmath113 and @xmath131 , and combining equations ( 6 ) , ( 7 ) , ( 24 ) , ( 29 ) , and ( 31 ) gives @xmath132 where the approximate density profile of the dark halo ( at large @xmath9 ) follows from equation ( 18 ) .
thus , if a galactic disk is embedded in a dark halo , one expects @xmath133 at large @xmath9 with the slope proportional to @xmath134 .
the velocity dispersion in the lmc bar will yield an accurate estimate of the scale height only if we make a numerical correction to account for the finite extent of the lmc .
we note that applying a pure disk model at small radii in the lmc is reasonable because the lmc has no bulge , and the surface brightness profile ( even in the bar ) is purely exponential ( bothun & thompson 1988 ) .
we appeal to the numerical integrations by van der kruit & searle ( 1982 ) to estimate a `` correction factor '' for the disk velocity dispersions given by equation ( 7 ) . in table
a1 of van der kruit & searle ( 1982 ) , model no . 2 , the ratio of the actual z - force ( @xmath135 ) to the model z - force is 0.55 , 0.82 , 0.95 , 1.07 , 1.06 , 0.64 , and 0.77 at @xmath136 = 0 , 1 , 2 , 3 , 4 , 5 and 6 , respectively .
we have simply averaged over the tabulated values of @xmath137 ( various other weighted - average schemes give similar values ) .
the correction to the model velocity dispersons is the square root of the z - force ratios ( van der kruit & searle 1982 ) , and is @xmath725% near the lmc center .
the scale height inferred from the velocity dispersion in the lmc bar would be underestimated by a factor of @xmath72 without the correction . here
we summarize the theoretical calculations presented thus far . in
2.1 , we introduced the basic formulae that describe an exponential disk with no dark halo , and a prediction for the velocity dispersion perpendicular to the plane of the disk .
the latter calculation was made by assuming a radially infinite disk . for a disk truncated at @xmath136
= 4 or 5 , the error in this predicted velocity dispersion is less than 10% at radii of a few scale lengths ( van der kruit & searle 1982 ) . in 4.1 , we described the rotation curve of an exponential disk embedded in a psuedo - isothermal dark halo .
we presented a stellar rotation curve derived from lmc surface brightness data , and a gas rotation curve , both from kim et al .
( 1998 ) . we presented a model rotation curve for a finite - thickness disk truncated at @xmath136 = 4 , from van der kruit & searle ( 1982 ) . in 4.2 , we calculated the velocity dispersion perpendicular to the plane of a disk that is embedded in a psuedo - isothermal dark halo . at large radii , the dynamical influence of the dark halo can dominate . in this limit
, we showed that the disk velocity dispersion can be predicted by an analytic formula .
we estimated the error of this approximation using numerical integrations .
in 4.3 , we give a numerical correction to the velocity dispersion perpendicular to the plane of the disk ( the prediction from 2.1 ) that accounts for truncation of the disk at @xmath136 = 4 , which is adopted from van der kruit and searle ( 1982 ) .
although the lmc disk may not be truncated at precisely four radial scale lengths , by using the same model throughout this work , our analysis is self - consistent .
in figure 4 , we present the lmc rotation curve and a `` maximal '' disk decomposition .
our four zonal solutions are indicated with bold dots and error bars .
data points on the rotation curve from kim et al .
( 1998 ) are shown with bold cross symbols .
for comparison , we plot the carbon star zonal solutions of kdia with open circles .
also shown in figure 4 are the gas and stellar disk rotation curves .
we assume m / l = 2.2 for the stellar rotation curve , which is similar to that found for the galactic disk ( bahcall , flynn & gould 1992 ) .
disk mass - to - light ratios of @xmath72 are favored by dynamical stability arguments ( bottema 1993 ) .
finally , we plot the sum of the gas and stellar disk rotation curves as a bold solid line .
our four zonal solutions and the solutions of kdia agree within our respective errors .
the model rotation curve and our data points show quite good agreement .
in fact , although the overall scaling of the model is a free parameter ( the m / l ratio of the stellar disk ) , the variation with radius is reproduced remarkably well .
the disagreement between the model and data points would appear less severe if error bars similar to those on the carbon star data points were appropriate .
if the lower envelope of the carbon star error bars were more representative of the true lmc rotation curve , then one could adopt a slightly smaller m / l ratio and the agreement with the data points would improve .
however , since the data points appear to deviate from a smooth curve , most notably the dip at @xmath138 3 kpc , it is possible that all of the data points lying below the model curve represent a significant failure of the model .
we speculate that a bar in the lmc may be responsible , although we have not attempted any modeling of this effect . in figure 5
, we show the same carbon star and data points as in figure 4 .
we now replace the stellar rotation curve derived from surface brightness data with the finite - thickness , truncated model disk rotation curve .
the gas rotation curve , and the sum of the gas and model disk rotation curves are also plotted .
the model curve assumes a scale length of 1.6 kpc and an infinitely thin disk equivalent maximum circular velocity .
a finite - thickness disk has a maximum circular velocity approximately 5% smaller than an equivalent infinitely thin disk ( see table 3 ) . in this decomposition
, we adopt the finite - thickness disk model , but account for the contribution from gas ( approximately 5% near maximum ) , and quote here the maximum circular velocity for an equivalent infinitely thin disk . ]
@xmath27 = 71 km s@xmath1 .
the model disk rotation curve is clearly consistent with the data , and fits the run of and carbon star data points even better than the ( semi - empirical ) stellar rotation curve shown in figure 4 .
the small differences between the model disk rotation curve , the stellar disk rotation curve , and the data points as illustrated in figures 4 and 5 are not critical to our conclusions .
we estimate the mass of the lmc disk using @xmath27 = 71 km s@xmath1 and the formulae in 2.1 , which yields @xmath139 @xmath3 @xmath140 .
the error is estimated by adopting the uncertainty of the maximum observed circular velocity in our carbon star solutions ( @xmath07 km sec@xmath1 ; see table 1 ) for the uncertainty of @xmath27 .
this maximal disk model has a surface density normalization of @xmath141 @xmath5 pc@xmath142 .
kim et al . ( 1998 ) estimate the total mass of gas in the lmc to be 0.5 @xmath3 @xmath140 .
thus , the total mass of the lmc is @xmath143 @xmath3 @xmath140 . in figure 6 , we present a `` minimal '' disk decomposition of the lmc rotation curve .
we plot the same carbon star and data points as in figures 4 and 5 .
we plot the stellar rotation curve assuming m / l = 1 , the gas rotation curve , and the sum of the stellar and gas rotation curves .
we also plot the contribution of a pseudo - isothermal dark halo , and the sum of the disk and halo rotation curves . in this decomposition , we calculate @xmath144 @xmath3 @xmath140 and @xmath145 @xmath5 pc@xmath142 . the shallow and declining run of carbon star data points favors a small core radius for the lmc dark halo . the dark halo shown has @xmath105 = 1 kpc and @xmath104 = 0.10 @xmath5 pc@xmath12 . by assuming m
l = 1 in the disk , this decomposition has a maximal halo , and yields a total lmc mass of @xmath75 @xmath3 @xmath140 ( corresponding to an upper limit on the lmc global mass - to - light ratio of @xmath74 ) . in figure 7
, we plot our lmc disk velocity dispersions as a function of true radius with bold dots and error bars .
we plot the results of kdia with open circles ( see also fig . 1 ) .
the prediction of equation ( 8) for our maximal disk model is shown as a dotted line and labeled .
we plot this model prediction again , but now corrected for the truncation of the lmc disk ( the `` @xmath146 force correction '' ) as a bold solid line .
we have assumed a constant scale height of @xmath147 kpc , which normalizes the latter curve to intersect the carbon star data point in the lmc bar .
a constant - thickness disk model can not be reconciled with the data by any choice of @xmath8 . for our minimal disk decomposition ,
we must account for the effect of the lmc dark halo on the disk velocity dispersions . by combining equations ( 6 ) , ( 18 ) , and ( 24 ) , and equating @xmath148
, we may estimate the @xmath119 parameter , @xmath149 assuming a constant scale height of @xmath150 kpc , the spatial density normalization of the minimal disk is @xmath151 = 0.068 @xmath5 pc@xmath12 . adopting the parameters of the maximal lmc dark halo from above , we find @xmath152 1.1 , 1.0 , 1.2 , 1.9 and 3.4 at integer steps of @xmath136 = 1 to 5 . for these values of @xmath119 ,
the error associated with equation ( 31 ) is too large , and thus we appeal directly to our numerical integrations in table 5 . for this decomposition , we begin with the velocity dispersion according to equation ( 8) , and then make the @xmath153 force correction .
we calculate @xmath154 using equation ( 33 ) , and use spline interpolations of the numerical data in table 5 to calculate @xmath76 .
the model disk velocity dispersions are corrected according to equation ( 29 ) by a factor of @xmath128 ( equating @xmath155 ) .
these various curves are plotted in figure 8 .
our choice of @xmath8 = 0.4 kpc normalizes the dark halo - corrected curve to intersect the data point in the lmc bar .
a constant - thickness disk in the presence of a maximal dark halo can not be reconciled with the observed velocity dispersions by any choice of @xmath8 . for comparison ,
a fit to the velocity dispersion data with the minimal disk and an arbitrary pseudo - isothermal dark halo implies a maximum circular velocity in the rotation curve of @xmath7500 km s@xmath1 , which is about 7 times higher than observed .
our maximal disk model yields @xmath8 = 0.25 , 0.93 , and 1.62 kpc at @xmath9 = 0.5 , 4.0 , and 5.6 kpc , respectively . for comparison , the minimal disk yields @xmath8 = 0.55 , 1.62 , and 2.61 kpc at the same radii . at larger radii ( @xmath10 6 kpc ) , where the disk velocity dispersions begin to rise , the implied scale heights for the maximal disk model are @xmath8 = 5.1 and 28.4 kpc ( at @xmath9 = 6.5 and 8.2 kpc , respectively ) .
the scale height inferred from the last data point is much larger than the tidal radius of the lmc ( weinberg 2000 ) , and obviously wrong .
we will return to the interpretation of these high velocity dispersions at large radii in a moment .
a linear regression on the three data points with @xmath156 kpc yields : @xmath157 where @xmath9 is in units of kpc , and @xmath55 is in units of km s@xmath1 .
we adopt this regression to represent the flare of the lmc disk .
the change of scale height is represented by a function of the form @xmath158 where @xmath8 and @xmath9 are in units of kpc .
we find @xmath159 , and @xmath160 kpc .
equation ( 35 ) predicts the scale height to within 0.1 kpc of the values estimated above . assuming a constant mass - to - light ratio
, our flared disk model must project to a surface density that varies as @xmath33 in order to reproduce the observed surface brightness profile of the lmc ( de vaucouleurs 1957 , bothun & thompson 1988 ) .
therefore , the spatial density of the flared disk , @xmath161 , must change with an effective radial scale length that compensates for the variation of scale height with @xmath9 .
this effective radial scale length is easily calculated by considering @xmath162 where we require that @xmath163 .
substituting @xmath164 = 1.4 yields @xmath165 = 0.583 ( note that @xmath165 and @xmath164 are dimensionless scale factors ) .
it also follows that @xmath166 = @xmath167 = 1.064 @xmath5 pc@xmath12 . in summary ,
our flared disk model has the following spatial density , @xmath168 where @xmath169 is given in equation ( 35 ) .
the dynamical influence of the galactic dark halo is calculated following the discussion in 4.2 of this paper ( see also aubourg et al .
we designate the mean density of the galactic dark halo at the distance of the lmc as @xmath170 .
if we substitute @xmath170 for @xmath171 in equation ( 20 ) , then @xmath119 from equation ( 24 ) yields @xmath76 from table 5 , and thus the correction to the disk velocity dispersions through equation ( 29 ) .
we estimate @xmath170 following griest ( 1991 ; see his eqn . [ 3 ] ) .
we assume that the galactic dark halo has a core radius of 3 kpc , and a solar neighborhood spatial density normalization of 0.0079 @xmath5 pc@xmath12 .
we adopt the distance from the sun to the galactic center of 8.5 kpc , and the distance from the sun to the lmc of 50.1 kpc , which yields @xmath170 = 0.00025 @xmath5 pc@xmath12 .
we estimate @xmath119 = 0.001 , 0.007 , 0.040 , 0.224 and 1.246 at integer steps of @xmath136 = 1 to 5 . in figure 9
, we plot the same data points as in figures 7 and 8 .
we indicate the regression of equation ( 34 ) as a solid line , and this regression corrected for the effect of the galactic dark halo as a bold solid line .
although the fit is not perfect , the agreement with the last data point is quite good .
thus the influence of the galactic dark halo on our flared disk model is of the correct magnitude to account for the outer - disk velocity dispersions .
in addition , the radius where the galactic dark halo begins to have a significant dynamical effect ( i.e. , where the velocity dispersions begin to rise in the disk ) is also reasonably reproduced .
a more detailed modeling of the outer - disk velocity dispersions is beyond the scope of this paper .
we find that @xmath119 is quite small at all radii of interest for the constant - thickness maximal disk model ( accounting for the galactic dark halo but no lmc dark halo ) , and for the constant - thickness minimal disk model ( accounting for both the galactic and lmc dark halos ) . in summary
, constant - thickness exponential disk models can not be reconciled with the observed run of lmc disk velocity dispersions . as discussed above
, we adopted the mean density of the galactic dark halo at a distance of 50 kpc of @xmath170 = 0.00025 @xmath5 pc@xmath12 , or @xmath172 ( griest 1991 ) .
it is worth considering the range of allowed galactic dark halo profiles . assuming flat rotation curves at large radii , many authors have made galaxy models with pseudo - isothermal density profiles ( e.g. caldwell & ostriker 1981 ; bahcall , schmidt , & soneira 1982 ) .
typical core radii range from 2 to 8 kpc .
the average value of the density at 50 kpc predicted by these models is @xmath173 , with a standard deviation of 0.2 dex . as a final check
, we compare the n - body cold dark matter simulation of the galactic dark halo presented by dubinski ( 1994 ) .
the initial density profile for this dark halo model was similar to an ellipsoidal hernquist potential , which is like the universal dark halo profile found in dark matter - dominated galaxies ( kravstov et al .
dubinski s ( 1994 ) initial density profile was then compressed by the galactic potential ( e.g. blumenthal et al .
1986 ) and in its final form predicts @xmath172 .
the density profile of the galactic dark halo assumed for machos ( alcock et al .
2000 ) is also pseudo - isothermal ; thus our analysis of the lmc disk kinematics supports this dark halo profile over the range of interest ( @xmath174 to 50 kpc ) .
it has been suggested that the lmc is embedded in a `` shroud '' of tidal debris ( weinberg 2000 ) .
such debris is unlikely to contribute significantly to the lmc self - lensing optical depth because too little mass is involved , and it is likely to be located close to the disk ( weinberg 2000 ; see 6.2 ) .
nonetheless , it is worth considering the possibility that our sample of true disk carbon stars is contaminated by tidal debris , which might reconcile a constant - thickness disk model with the disk kinematic data . in this case
, tidal debris would also affect our interpretation of the outer - disk velocity dispersions .
the notion of tidal debris surrounding the lmc is somewhat similar to the suggestion by kdia that the lmc harbors a polar ring , also of tidal origin .
( we note that kdia analysed a larger sample of carbon stars , in which they claim the kinematic signature of a polar ring is evident . )
we have thus far presented an interpretation of the radial velocity data for 422 carbon stars in the lmc that does not require a polar ring , or tidal debris .
the nature of tidal debris ( in a polar ring or otherwise ) is investigated as follows .
if we assume that a constant - thickness maximal disk model represents the true lmc disk , the `` excess '' velocity dispersion in each zonal solution may be calculated by subtracting the model contribution from the observed dispersions in quadrature , which yields @xmath175 = 15.0 , 13.7 , 18.8 and 20.5 km s@xmath1 at @xmath9 = 4.0 , 5.6 , 6.4 and 8.2 kpc , respectively .
next , if we assume that the contaminating debris has a uniform velocity dispersion of @xmath750 km s@xmath1 , the ratio of contaminating stars to total stars in each zonal solution would be of order @xmath710 - 20% .
( this ratio is inversely proportional to the velocity dispersion assumed for the tidal debris . ) in each of our zones , we estimate the number of @xmath750 km s@xmath1 tidal debris stars would be 12 , 17 , 10 , and 8 ( in order of increasing radius ) , if the lmc disk were of constant thickness .
we calculate that our four disk zones subtend relative sky areas of 0.40/0.16/0.16/1.00 , in order of increasing zone radius .
a regression of the relative numbers of tidal debris stars with the relative zone areas shows an anti - correlation , significant at the @xmath71@xmath43 level . assuming that the true disk and tidal debris carbon stars are similarly affected by incompleteness in the data , this is not what we naively expect from the tidal debris scenario described by weinberg ( 2000 ) .
the two outermost zones are illustrative .
they have area ratios that differ by a factor of @xmath76 , yet the estimated numbers of contaminating tidal debris stars are 10 and 8 .
therefore , the contaminating debris must have a non - uniform radial distribution of surface density , velocity dispersion , or both .
it is also possible , but even more contrived , that a spatially - dependent incompleteness in the carbon star dataset has conspired with a non - uniform spatial distribution of tidal debris to yield the observed velocity dispersions . in light of this analysis ,
we prefer an interpretation of the lmc rotation curve and disk velocity dispersions that invokes a negligible contamination from tidal debris ( i.e. , our maximal flared disk model ) . the unambiguous detection of nonvirialized lmc stars and the accurate characterization of their spatial distribution are clearly desirable .
observations of such a population / structure would be relevant to our interpretation of the lmc rotation curve and disk kinematics , with possible additional implications for lmc microlensing ( e.g. zhao 1999 ; graff et al . 1999 ; weinberg 2000 ) .
here we calculate the lmc self - lensing optical depth of our maximal flared disk model by directly integrating the spatial density of stars given in equation ( 37 ) .
we begin with an integral of the form @xmath176 which yields the optical depth to any source star in the lmc at position @xmath177 , a cylindrical coordinate system in the plane of the disk . the position angle @xmath178 is defined to be zero at the near side if the inclined disk .
the integration variable is the line of sight distance between the source and lens .
we use the subscripts `` @xmath179 '' and `` @xmath180 '' to denote the coordinates of the sources and lenses , respectively . the approximation in equation ( 38 )
is that @xmath61 is always much smaller than the distance to the source ( @xmath181 50.1 kpc ) , which is reasonable for lmc self - lensing .
note that @xmath61 is simply related to @xmath182 , the distance from the observer to the lens , and @xmath183 , the distance from the observer to the source @xmath184 where @xmath14 is the inclination angle of the disk ( measured from the plane of the sky ) . for completeness
, we provide the following geometric identity relating the source and lens positions to @xmath61 : @xmath185 it is necessary to integrate over the line of sight a second time , in order to calculate a density - weighted average of @xmath186 over the distribution of source stars @xmath187 which may be properly compared to the observed optical depth .
the integration over @xmath183 is made in practice by transforming to a variable in our lmc cylindrical coordinate system ( i.e. , @xmath188 ) .
we truncate the lmc disk at a conservative radius of 10 kpc .
our results are insensitive to the choice of truncation radius at the few percent level .
we calculate @xmath189 for multiple lines of sight , designated by ( @xmath190 ) in the plane of the disk .
our calculation is the case of `` pure '' self - lensing ( gould 1995 ) , which yields an upper limit . the effects of dust obscuration and nonlensing mass in the plane of the disk will lower the real value of the optical depth for a fixed total mass of the lmc . we have calculated the self - lensing optical depths in each of the 30 survey fields in the 5.7-year lmc microlensing analysis by alcock et al .
( 2000 ; see also gyuk et al .
1999 ) . in the central - most fields ,
we find @xmath191 , while in the fields lying at the largest true radii , we find @xmath192 .
the field - averaged optical depth for lmc disk self - lensing is : @xmath193 .
the lmc inclination dominates the uncertainty in @xmath194 .
first , we note that the inclination is probably not a serious concern for the rotation curve and disk kinematics because the shape of the @xmath89 function ( eqn .
[ 15 ] ) is fairly insensitive to @xmath14 over the range of plausible values .
however , gould ( 1995 ) and gyuk et al .
( 1999 ) show that , to first order , @xmath195 is proportional to the square of the mass - weighted velocity dispersion and an inclination factor of sec@xmath196 . therefore , we simply rescale our calculation of @xmath194 for different inclinations , @xmath197 which should be quite accurate for the lmc inclinations typically found . as noted in 2.2 , cole et al .
( 1999 ) find @xmath45 .
thus their @xmath72@xmath43 upper limit , @xmath198 , yields @xmath199 . for comparison ,
bothun & thompson ( 1988 ) find @xmath200 , which corresponds to @xmath201 .
the contribution to the uncertainty in @xmath194 from the velocity dispersion follows from gould s ( 1995 ) formula and our equation ( 34 ) ; it is of order @xmath202 , or a negligible 0.03% .
finally , the @xmath710% uncertainty in @xmath203 yields a @xmath720% uncertainty on the total lmc mass , but this corresponds to a small uncertainty in the factor of @xmath204 in equation ( 42 ) .
the self - lensing optical depth on the near and far sides of the minor axis at the distances spanned by the 30 fields in alcock et al . ( 2000 ) varies by @xmath75% due to the inclination of the disk ( e.g. gould 1994 ) .
thus the flaring of the lmc disk has only a minor effect on the spatial distribution of the self - lensing optical depth , which is a potential diagnostic of the lens population ( e.g. alcock et al .
the estimate of a @xmath75% minor axis asymmetry in the optical depth would increase for larger adopted values of the inclination .
gyuk et al .
( 1999 ) recently reviewed different self - lensing calculations in the literature . to first order ,
self - lensing from the virialized lmc disk is proportional to the following combinations of derivable quantities ( gyuk et al .
1999 ) , @xmath205 our model accounts for `` second order '' effects , such as the finite radial extent and flare of the lmc disk .
we have shown that uncertainties in the lmc proper motion contribute a @xmath71% uncertainty to @xmath43 and a @xmath720% uncertainty to @xmath25 ( via @xmath203 ) , assuming a truncated and flared maximal disk model . assuming this model , our numerical calculation of the self - lensing optical depth ( eqn . [ 42 ] ) is appropriate .
the maximal disk model is consistent with lmc dynamics .
however , if a model of the lmc with a dark halo is preferred , the halo will probably contribute to lmc self - lensing in addition to the disk . in this case , our equation ( 42 ) is not appropriate ( see gyuk et al .
we have adopted a minimal lmc disk , and provided an example maximal dark halo which is consistent with the rotation curve .
measurements of the lmc disk mass to light ratio would better constrain the minimal lmc disk model ( our model is extreme ) . for the more complicated case of an lmc with a dark halo , constraints on
self - lensing are weaker .
this is due partly to the uncertainty that the lmc proper motion contributes to the rotation curve , which constrains the total lmc mass .
the lmc self - lensing optical depth may depend systematically on the age of the stars whose kinematics are studied .
historically , it has been difficult to prove that old lmc stellar populations have different velocity dispersions .
for example , the ch stars in the lmc were originally thought to trace the elusive lmc halo population , by analogy with ch stars in the galaxy ( hartwick & cowley 1988 , cowley & hartwick 1991 ) .
however , the ch stars in the lmc are now believed to represent a much younger population ( suntzeff et al .
we note that our average carbon star velocity dispersion ( 18.9 km s@xmath1 ) is similar to that found for the ch stars ( 20 km s@xmath1 ) .
according to aubourg et al .
( 1999 ) and salati et al .
( 1999 ) , the lmc could harbor an old population with a large velocity dispersion .
these authors invoke wielen s ( 1977 ) age - velocity dispersion calibration from the solar neighborhood .
we note that weilen s calibration was derived for stars younger than @xmath73 gyr , and extrapolation beyond this age may not be appropriate .
moreover , weinberg ( 2000 ) predicts that the velocity dispersions in the lmc disk , under the influence of the galactic tidal field , will remain constant ( or decrease ) over time .
extant observational studies have not yielded a clear picture of the variation of velocity dispersion with age in the lmc .
for example , the old globular clusters in the lmc , with ages of @xmath712 gyrs ( olsen et al .
1998 ) , exhibit a velocity dispersion of @xmath721 - 24 km s@xmath1 ( schommer et al .
1992 ; see also freeman , illingworth , & oemler 1982 ) .
thus , the lmc carbon stars and ancient clusters support the hypothesis of a constant disk velocity dispersion ( for stellar populations older than @xmath73 gyr ) , in agreement with weinberg s ( 2000 ) prediction .
however , the velocity dispersion of old lpvs in the lmc , with ages of @xmath78 gyr ( hughes , wood , & reid 1991 ; olszweski , sunzteff , & mateo 1996 ) , is @xmath735 km s@xmath1 ( hughes et al .
further observational studies are needed , particularly given the provocative old lpv result .
last , we suggest that a shroud of tidal debris or a polar ring , if they exist , would probably make a negligible contribution to the lmc self - lensing optical depth because the total amount of mass involved is small .
weinberg ( 2000 ) reached a similar conclusion .
the rotation of the disk of the lmc has been derived from the radial velocities of 422 carbon stars ( kunkel , irwin , & demers 1997 ) .
we have propogated the uncertainty in the lmc space motion to the lmc rotation curve with a monte carlo calculation .
the associated disk velocity dispersions found in each rotating disk zonal solution are less sensitive to systematic uncertainties from the lmc space motion than the circular velocities obtained .
we note that our carbon star rotation curve is derived in a manner consistent with the rotation curve analysis by kim et al .
( 1998 ) ; thus our respective results are properly comparable .
we have fit our carbon star rotation curve and the rotation curve with a truncated , maximal lmc disk model , yielding a total lmc mass of @xmath143 @xmath3 @xmath140 .
we also conclude that the disk of the lmc is flared . extrapolating the flare inferred at small radii ( where tidal perturbations are small ) to the outer - disk , and accounting for the influence of the galactic dark halo , we are able to approximately reproduce the observed disk kinematics .
this model favors an isothermal density profile for the galactic dark halo out to a distance of 50 kpc .
our truncated and flared maximal disk model yields a limit on the spatially - averaged lmc self - lensing optical depth of @xmath206 .
for plausible values of the lmc inclination , this low self - lensing rate compared to the measured microlensing rate allows for the existence of a dark lens population in the galactic halo ( alcock et al .
2000 ) .
finally , we caution that we have not included in our analysis ( 1 ) the dynamical effects of an lmc bar , ( 2 ) large - scale non - circular motions , ( 3 ) non - uniform anisotropy of the disk velocity dispersions , or ( 4 ) arbitrary spatial distributions of tidal debris ( i.e. , a polar ring or stars out of virial equilibrium ) . despite these shortcomings , our truncated and flared maximal disk model successfully accounts for the general dynamical characteristics of the lmc , lending inferences from the model high weight .
d.r.a . acknowledges howard e. bond for support of work performed at the space telescope science institute , and kem h. cook for support of work performed at the lawrence livermore national laboratory .
nasa research grant nag5 - 6821 `` uv , visible , and gravitational astrophysics research and analysis '' and doe contract w7405-eng-48 are recognized .
c.a.n . acknowledges support from national physical science consortium graduate student fellowship .
we thank ken freeman , david bennett , tim axelrod , howard bond , kailash sahu , geza gyuk , greg bothun , and the anonymous referee for their helpful discussions and comments .
lccccclccccc 2.5 - 5.0 & 91 & 4.0 & @xmath207 & 72 @xmath0 7.1 & 17.6 @xmath0 0.3 5.0 - 6.0 & 139 & 5.6 & @xmath208 & 68 @xmath0 8.7 & 14.9 @xmath0 0.2 6.0 - 7.0 & 122 & 6.4 & @xmath209 & 65 @xmath0 10.4 & 19.3 @xmath0 0.4 7.0 - 13.0 & 73 & 8.2 & 3 & 65 @xmath0 15.8 & 20.6 @xmath0 0.5 ccc|ccc 0.5 & 14.0 & 30.8 & 5.5 & 28.6 & 62.7 1.0 & 23.5 & 51.7 & 6.0 & 28.3 & 62.1 1.5 & 28.6 & 62.7 & 6.5 & 28.0 & 61.5 2.0 & 30.8 & 67.7 & 7.0 & 28.0 & 61.5 2.5 & 31.4 & 68.9 & 7.5 & 28.0 & 61.5 3.0 & 31.4 & 68.9 & 8.0 & 27.7 & 60.9 3.5 & 31.4 & 68.9 & 8.5 & 27.4 & 60.3 4.0 & 30.8 & 67.7 & 9.0 & 27.4 & 60.3 4.5 & 30.2 & 66.4 & 9.5 & 27.4 & 60.3 5.0 & 29.1 & 64.0 & 10.0 & 27.2 & 59.7 cc|cc 0.5 & 5.0 & 5.5 & 22.8 1.0 & 9.0 & 6.0 & 22.5 1.5 & 10.0 & 6.5 & 21.0 2.0 & 12.0 & 7.0 & 20.0 2.5 & 16.0 & 7.5 & 19.0 3.0 & 22.0 & 8.0 & 18.5 3.5 & 24.2 & 8.5 & 17.0 4.0 & 23.8 & 9.0 & 16.0 4.5 & 23.0 & 9.5 & 15.0 5.0 & 23.0 & 10.0 & 14.5 ll|ll 0.000 & 1.000 & 6.000 & 0.339 0.250 & 0.865 & 7.000 & 0.317 0.500 & 0.776 & 8.000 & 0.298 1.000 & 0.658 & 9.000 & 0.283 1.500 & 0.582 & 10.000 & 0.269 2.000 & 0.528 & 12.000 & 0.248 2.500 & 0.486 & 15.000 & 0.223 3.000 & 0.453 & 20.000 & 0.194 4.000 & 0.404 & 30.000 & 0.160 5.000 & 0.367 & 50.000 & 0.124 | the rotation of the disk of the large magellanic cloud ( lmc ) is derived from the radial velocities of 422 carbon stars ( kunkel , irwin , & demers 1997 , a&as , 122 , 463 ) .
new aspects of this analysis include the propagation of uncertainties in the lmc proper motion with a monte carlo , and a self - consistent modeling of the rotation curve and disk kinematics .
the lmc rotation curve reaches a maximum circular velocity of 72 @xmath0 7 km s@xmath1 at @xmath2 kpc and then declines .
the rotation curve is well fit by a truncated , finite - thickness exponential disk model with no dark halo , implying a total mass of 5.3 @xmath0 1.0 @xmath3 @xmath4 @xmath5 .
the velocity dispersion in concentric radial bins from @xmath6 to 5.6 kpc decreases from 22 to 15 km s@xmath1 , then increases to @xmath720 km s@xmath1 at larger radii .
constant - thickness disk models in virial equilibrium can not be reconciled with the data even if the effects of lmc or galactic dark halos are included .
if the disk is virialized , the scale height rises from @xmath8 = 0.3 to 1.6 kpc over the range of @xmath9 = 0.5 to 5.6 kpc . thus the lmc disk is flared .
we model the velocity dispersion at large radii ( @xmath10 6 kpc ) as a maximal flared disk under the influence of the galactic dark halo , which favors a mean density for the latter of @xmath11 @xmath5 pc@xmath12 at the lmc distance . lmc stellar kinematics play an important role in elucidating the nature of machos , a dark population inferred from lmc microlensing .
we have constructed a truncated and flared maximal disk model for the lmc which is kinematically based .
our model does not include a nonvirialized component such as tidal debris .
the instantaneous probability of microlensing from lmc stars in our model is @xmath13 , where @xmath14 is the disk inclination .
our upper limit on the self - lensing optical depth is in good agreement with that obtained from less sophisticated models , and is an order of magnitude too small to account for the macho microlensing signal . |
during the last two decades models of pedestrian behavior have found notable interest for several reasons . first , there are some striking analogies with gases @xcite and fluids @xcite .
second , all model quantities like places @xmath0 and velocities @xmath1 of pedestrians @xmath2 are measurable and , therefore , comparable with empirical data .
third , there already exists a considerable amount of data material like flow measurements or ( video ) films @xcite .
fourth , pedestrian models can provide valuable tools for designing and planning pedestrian areas , subway or railroad stations , big buildings , shopping malls , etc . @xcite . in the 1970s
henderson @xcite has compared measurements of pedestrian flows with navier - stokes equations with considerable success .
his phenomenological fluid - dynamic approach has been improved and mathematically founded by helbing in the 1990s on the basis of a pedestrian specific gaskinetic ( boltzmann - like ) model @xcite .
this model is closely related to some gaskinetic @xcite and fluid - dynamic @xcite _ traffic _ models .
recently , much attention has been attracted by _ ` microscopic ' _ approaches for vehicular traffic @xcite . in the following ,
we will introduce a _ social force model _ for pedestrian motion .
first ideas of this ` microscopic ' modeling concept were formulated in @xcite . a short overview of ` microscopic ' , ` mesoscopic ' and ` macroscopic ' descriptions of pedestrian behavior and their interrelations is given in ref .
many people have the feeling that human behavior is ` chaotic ' or at least very irregular and not predictable .
this is probably true for behaviors that are found in complex situations .
however , at least for relatively simple situations _ stochastic _ behavioral models may be developed if one restricts to the description of behavioral _ probabilities _ that can be found in a huge population ( resp .
group ) of individuals ( see @xcite ) .
this idea has been followed by the _
gaskinetic _ pedestrian model @xcite .
another approach for modeling behavioral changes has been suggested by lewin @xcite . according to his idea
behavioral changes are guided by socalled _ social fields _ resp .
_ social forces_. in the following we will examine if this idea could be applied to the description of pedestrian behavior .
figure 1 illustrates a scheme of the processes that lead to behavioral changes . according to this ,
a sensory _ stimulus _ causes a behavioral _ reaction _ that depends on the personal aims and is chosen from a set of behavioral alternatives with the objective of utility maximization .
table 1 suggests a classification of stimuli into simple or standard situations that are well predictable , and complex or new situations that may be modelled with probabilistic models .
however , since a pedestrian is used to the situations he / she is normally confronted with , his / her reaction is usually rather automatic , and determined by his / her experience of which reaction will be the best .
it is therefore possible to put the rules of pedestrian behavior into an equation of motion . according to this equation the systematic temporal changes @xmath3 of the _ prefered velocity _
@xmath4 of a pedestrian @xmath2 are described by a vectorial quantity @xmath5 that can be interpreted as _
social force_. clearly , this force must represent the effect of the environment ( e.g. other pedestrians or borders ) on the behavior of the described pedestrian .
however , the social force is not excerted by the environment on a pedestrian s body .
it is rather a quantity that describes the concrete _ motivation to act_. in the case of pedestrian behavior this motivation evokes the physical production of an acceleration or deceleration force as reaction to the perceived information that he / she obtains about his / her environment ( see figure 1 ) . in summary , one can say that a pedestrian acts _ as if _ he / she would be subject to external forces . this idea has been mathematically founded in ref .
in the following the main effects that determine the motion of a pedestrian @xmath2 will be introduced : * he / she wants to reach a certain destination @xmath6 as comfortable as possible .
therefore , he / she normally takes a way without detours , i.e. , the shortest possible way .
this way will usually have the shape of a polygon with edges @xmath7 . if @xmath8 is the next edge of this polygon to reach , his / her _ desired direction _
@xmath9 of motion will be @xmath10 where @xmath11 denotes the _ actual position _ of pedestrian @xmath2 at time @xmath12 .
exactly speaking , the goals of a pedestrian are usually rather _ gates _ or _ areas _ than points @xmath8 . in this case , he / she will at every time @xmath12 steer for the _ nearest _ point @xmath13 of the corresponding gate / area .
+ if a pedestrian s motion is not disturbed , he / she will walk into the desired direction @xmath9 with a certain _ desired speed _
@xmath14 . a deviation of the _ actual velocity _
@xmath15 from the _ desired velocity _
@xmath16 due to necessary deceleration processes or avoidance processes leads to a tendency to approach @xmath17 again within a certain _ relaxation time _ @xmath18 .
this can be described by an _ acceleration term _ of the form @xmath19 * the motion of a pedestrian @xmath2 is influenced by other pedestrians .
especially , he / she keeps a certain distance from other pedestrians that depends on the pedestrian density and the desired speed @xmath14 .
here , the _ privat sphere _ of each pedestrian , which can be interpreted as _ territorial effect _ @xcite , plays an essential role .
a pedestrian normally feels increasingly incomfortable the closer he / she gets to a strange person , who may react in an aggressive way .
this results in _ repulsive effects _ of other pedestrians @xmath20 that can be represented by vectorial quantities @xmath21 \ , .\ ] ] we will assume that the repulsive potential @xmath22 is a monotonic decreasing function of @xmath23 with equipotential lines having the form of an ellipse that is directed into the direction of motion .
the reason for this is that a pedestrian requires space for the next step which is taken into account by other pedestrians .
@xmath23 denotes the semi - minor axis of the ellipse and is given by @xmath24 where @xmath25 .
@xmath26 is of the order of the step width of pedestrian @xmath20 . despite the simplicity of this approach
, it describes avoidance maneuvers of pedestrians quite realistically .
+ a pedestrian also keeps a certain distance from _
borders _ of buildings , walls , streets , obstacles , etc . @xcite .
he / she feels the more incomfortable the closer to a border he / she walks since he / she has to pay more attention to avoid the danger of getting hurt , e.g. by accidentally touching a wall .
therefore , a border @xmath27 evokes a _ repulsive effect _ that can be described by @xmath28 with a repulsive and monotonic decreasing potential @xmath29 . here
, the vector @xmath30 has been introduced , where @xmath31 denotes the location of that piece of border @xmath27 that is nearest to pedestrian @xmath2 . *
pedestrians are sometimes attracted by other persons ( friends , street artists , etc . ) or objects ( e.g. window displays ) .
these _ attractive effects _
@xmath32 at places @xmath33 can be modelled by attractive , monotonic increasing potentials @xmath34 in a similar way like the repulsive effects : @xmath35 ( @xmath36 ) .
the main difference is that the _ attractiveness _
@xmath37 is normally decreasing with time @xmath12 since the interest is declining .
the attractive effects are , e.g. , responsible for the formation of pedestrian groups ( that are comparable to molecules ) .
however , the formulas above for attractive and repulsive effects only hold for situations that are perceived in the desired direction @xmath9 of motion .
situations located behind a pedestrian will have a weaker influence @xmath38 with @xmath39 . in order to take this effect of perception ( i.e. of the _ effective angle @xmath40 of sight _ ) into account
we have to introduce the direction dependent weights @xmath41 in summary , the repulsive and attractive effects on a pedestrian s behavior are given by @xmath42 we can now set up the equation for a pedestrian s total motivation @xmath5 .
since all the previously mentioned effects influence a pedestrian s decision at the same moment , we will assume that their total effect is given by the sum of all effects , like this is the case for forces .
this results in @xmath43 the _ social force model _ is now defined by @xmath44 here , we have added a _ fluctuation term _ that takes into account random variations of the behavior .
these fluctuations stem , on the one hand , from ambiguous situations in which two or more behavioral alternatives are equivalent ( e.g. if the utility of passing an obstacle on the right or left hand side is the same ) . on the other hand , fluctuations arise from accidental or deliberate deviations from the usual rules of motion . in order to complete the model of pedestrian dynamics a relation between the actual velocity @xmath15 and the prefered velocity @xmath4 must be introduced . since the actual speed is limited by a pedestrian s _ maximal acceptable speed _
@xmath45 , we will assume that the _ realized _
motion is given by @xmath46 with @xmath47 note that the pedestrian model ( [ sozforce ] ) , ( [ sozdyn ] ) has the form of nonlinearly coupled langevin equations .
a simplified version of it can be extended to an _ active walker model _ @xcite for the self - organization of systems of ways @xcite .
this is related to the mathematical description of trunk trail formation by ants @xcite .
the model of pedestrian dynamics developed in section iii has been simulated on a computer for a large number of interacting pedestrians confronted with different situations . despite the fact that the the proposed model is very simple
it describes a lot of observed phenomena very realistically . in the following
, two examples will be presented showing the _ self - organization _ of collective phenomena of pedestrian behavior .
the simulations assumed that the desired speeds @xmath48 are gaussian distributed @xcite with mean @xmath49 and standard deviation @xmath50 @xcite .
speeds were limited to @xmath51 .
pedestrians enter the walkway at the ends at random positions .
those intending to walk from the left to the right hand side are represented by full circles , whereas pedestrians intending to move into the opposite direction are represented by empty circles .
the diameter of a circle is a measure for the actual speed @xmath52 of a pedestrian @xmath2 . for simplicity , no attractive effects @xmath32 or fluctuations were taken into account .
the repulsive potentials were assumed to decrease exponentially , i.e. @xmath53 with @xmath54 , @xmath55 and @xmath56 , @xmath57 .
potentials with a hard core would be more realistic , of course , but they will not yield other results since the simulated pedestrians always keep enough distance .
walkways were chosen 10 m wide . for @xmath58 in formula ( [ for4 ] ) we took @xmath59 , and for the relaxation times we used @xmath60 .
smaller values of @xmath18 let the pedestrians walk more aggressive . finally , the effective angle of sight ( which also takes into account head movements ) was set to @xmath61 .
situations outside the angle of sight were assumed to have an influence of @xmath62 .
the model parameters introduced above were chosen in a way that is compatible with empirical data .
figure 2 shows the empirically confirmed @xcite development of dynamically varying _ lanes _ consisting of pedestrians who intend to walk into the same direction .
periodic boundary conditions in transversal direction would stabilize these lanes since they were not any more destroyed at the ends of the walkway by randomly entering pedestrians @xcite .
figure 3 shows the number of forming lanes in dependence of the width of the walkway for a pedestrian density of @xmath63 .
the segregation effect of lane formation is _ not _ a result of the initial pedestrian configuration but a consequence of the pedestrians interactions .
nevertheless , it normally leads to a more effective pedestrian flow since time - consuming avoidance maneuvers occur less frequently .
figures 4 and 5 depict different moments of two pedestrian groups that try to pass a narrow door into opposite directions .
the corresponding simulation shows the following : once a pedestrian has passed the door , other pedestrians intending to move into the same direction are able to follow him / her easily ( see figure 4 ) .
however , the stream of passing pedestrians is stopped by the pressure of the opposing group after some time .
subsequently , the door is captured by pedestrians who pass the door into the opposite direction ( see figure 5 ) .
this change of the passing direction may occur several times and is well - known from observations .
it has been shown that pedestrian motion can be described by a simple social force model for individual pedestrian behavior .
computer simulations of pedestrian groups demonstrated 1 .
the development of lanes consisting of pedestrians who walk into the same direction , 2 .
oscillatory changes of the walking direction at narrow passages .
these spatio - temporal patterns arise due to the _ nonlinear interactions _ of pedestrians .
they are _ not _ the effect of strategical considerations of the individual pedestrians since they were assumed to behave in a rather ` automatical ' way .
presently , the social force model is extended by a model for the route choice behavior of pedestrians .
as soon as the computer program is completed it will provide a comfortable tool for town- and traffic - planning .
the investigation of pedestrian behavior is an ideal starting point for the development of other or more general quantitative behavioral models , since the variables of pedestrian motion are easily measurable so that corresponding models are comparable with empirical data
. a further step could be the application of the social force concept to the description of opinion formation , group dynamics , or other social phenomena @xcite .
for this purpose , an abstract behavioral space has to be introduced .
the authors want to thank w. weidlich , m. r. schroeder , and w. scholl for valuable and inspiring discussions .
d. helbing , complex systems * 6 * , 391 ( 1992 ) .
d. helbing , _ physikalische modellierung des dynamischen verhaltens von fugngern _
( diplom thesis , georg - august university , gttingen , germany , 1990 ) .
d. helbing , _ stochastische methoden , nichtlineare dynamik und quantitative modelle sozialer prozesse _
( shaker , aachen , germany , 1993 ) .
l. f. henderson , nature * 229 * , 381 ( 1971 ) .
l. f. henderson , transp .
res . * 8 * , 509 ( 1974 ) .
d. oeding , _ verkehrsbelastung und dimensionierung von gehwegen und anderen anlagen des fugngerverkehrs _
( straenbau und straenverkehrstechnik , heft 22 , bonn , germany , 1963 ) . f. p. d. navin and r. j. wheeler , traffic engineering * 39 * , 30 ( 1969 ) .
13 ( transportation research board , special report 209 , washington , d.c . , 1985 ) .
g. p. gipps and b. marksj , math .
* 27 * , 95 ( 1985 ) .
a. borgers and h. timmermans , geographical analysis * 18 * , 115 ( 1986 ) .
g. g. lvs , in proceedings of the 1993 european simulation multiconference , lyon , france , june 79 , 1993 .
i. prigogine and r. herman , _ kinetic theory of vehicular traffic _
( american elsevier , new york , 1971 ) .
s. l. paveri - fontana , transp
. res . * 9 * , 225 ( 1975 ) .
b. s. kerner and p. konhuser , phys .
e * 48 * , 2335 ( 1993 ) .
b. s. kerner and p. konhuser , phys .
e * 50 * , 54 ( 1994 ) .
d. helbing , an improved fluid - dynamic model for vehicular traffic , phys .
e. ( submitted ) .
o. biham , a. a. middleton , and d. levine , phys .
a * 46 * , 6124 ( 1992 ) .
k. nagel and m. schreckenberg , j. phys .
i france * 2 * , 2221 ( 1992 ) .
t. nagatani , phys .
e * 48 * , 3290 ( 1993 ) .
j. a. cuesta , f. c. martnez , j. m. molera , and a. snchez , phys .
e. * 48 * , 4175 ( 1993 ) .
d. helbing , behavioral science * 36 * , 298 ( 1991 ) .
d. helbing , in _ natural structures .
principles , strategies , and models in architecture and nature _ , part ii ( sonderforschungsbereich 230 , stuttgart , germany , 1992 ) .
w. weidlich and g. haag , _ concepts and models of a quantitative sociology _
( springer , berlin , 1983 ) .
w. weidlich , physics reports * 204 * , 1 ( 1991 ) .
k. lewin , _ field theory in social science _
( harper & brothers , new york , 1951 ) .
d. helbing , physica a * 196 * , 546 ( 1993 ) .
a. e. scheflen and n. ashcraft , _ human territories : how we behave in space - time _
( prentice - hall , englewood cliffs , 1976 ) .
r. d. freimuth and l. lam , in _ modeling complex phenomena _ , edited by l. lam and v. naroditsky ( springer , new york , 1992 ) .
d. r. kayser , l. k. aberle , r. d. pochy , and l. lam , physica a * 191 * , 17 ( 1992 ) .
f. schweitzer and l. schimansky - geier , physica a * 206 * , 359 ( 1994 ) .
d. helbing , p. molnr , and f. schweitzer , in _ evolution of natural structures _
( sonderforschungsbereich 230 , stuttgart , germany , 1994 ) .
f. schweitzer , k. lao , and f. family , active walker simulate trunk trail formation by ants , adaptive behavior ( submitted ) .
u. weidmann _
transporttechnik der fugnger , _ pp .
8788 ( schriftenreihe des instituts
fr verkehrsplanung , transporttechnik , straen- und eisenbahnbau nr .
90 , eth zrich , 1993 ) . | it is suggested that the motion of pedestrians can be described as if they would be subject to ` social forces ' .
these ` forces ' are not directly exerted by the pedestrians personal environment , but they are a measure for the internal motivations of the individuals to perform certain actions ( movements ) .
the corresponding force concept is discussed in more detail and can be also applied to the description of other behaviors . in the presented model of pedestrian behavior
several force terms are essential : first , a term describing the acceleration towards the desired velocity of motion .
second , terms reflecting that a pedestrian keeps a certain distance to other pedestrians and borders .
third , a term modeling attractive effects .
the resulting equations of motion are nonlinearly coupled langevin equations .
computer simulations of crowds of interacting pedestrians show that the social force model is capable of describing the self - organization of several observed collective effects of pedestrian behavior very realistically . |
quantum key distribution ( qkd ) is one of the most interesting topics in quantum information processing , which allows the sender ( alice ) and the receiver ( bob ) to share a secret key with negligibly leakage of information to an eavesdropper ( eve ) .
this task is accomplished by making use of the properties of quantum mechanics , such as no - cloning theorem @xcite , indistinguishability of nonorthogonal quantum states , entanglement , and so on @xcite . the b92 protocol @xcite , where only two nonorthogonal states are used , is thought to be secure due to the fact that eve can not clone and distinguish the states deterministically .
although the qualitative reason of the security is simple , it is quite hard to show the quantitative security against so - called coherent attack , which is the most general attack allowed by quantum mechanics . in @xcite ,
the security of the bb84 protocol @xcite is proven by showing the relationship between qkd and other important protocols in quantum information , such as the entanglement distillation protocol ( edp ) @xcite and the calderbank - shor - steane ( css ) quantum error correcting codes @xcite . recently in @xcite ,
the unconditional security of the b92 protocol has been proved by relating the b92 protocol to an entanglement distillation protocol ( edp ) @xcite initiated by local filtering @xcite .
the assumptions of the proof are that alice and bob s devises are perfect and the quantum channel is loss - free .
to study the unconditional security of b92 protocol over a lossy and noisy channel is interesting from the point of view of the quantum information theory , since the channel loss directly affects the security of the b92 protocol .
the proof based on the equivalence between the security and edps gives us an idea how to distill the epr state from states that arise from the lossy and noisy channel . in this paper
, we study the unconditional security of the b92 protocol .
we assume that alice has a perfect single photon source , bob s polarization measurement is perfect and he has a perfect single photon counter that discriminates between single photon states on one hand and vacuum state or multi - photon states on the other hand . under these assumptions ,
we show the unconditional security of the b92 protocol .
the proof follows the idea in @xcite .
this paper is organized as follows . in sec .
[ review ] , we propose the protocol based on edp with local filtering and css codes @xcite , and we reduce this to the b92 protocol . in sec .
[ estimation ] we give the formula for the error estimations for edp based on css code . finally , in sec . [ conclusion ] we give the examples of the resulting performance , and the summary and discussion follow .
in this section , we introduce the protocol that will turn out to be secure and will reduce to the b92 protocol . in this protocol , alice first prepares a two - qubit nonmaximally entangled states that can be written as @xmath0 and sends system b to bob .
here @xmath1 , ( @xmath2 ) , where @xmath3 , @xmath4 , and @xmath5 is a basis ( @xmath6-basis ) of the qubit .
@xmath6-basis and @xmath7-basis is related through the relationship as @xmath8/\sqrt{2}$ ] . after receiving the state
, bob first performs the qnd measurements that can be described by the povm ( positive operator valued measure ) @xcite @xmath9 where @xmath10 is the projector onto the single photon space and `` v '' means the detection of vacuum state or multi - photon states .
thanks to this qnd measurements , even if eve sends multi - photon states or vacuum state to bob , alice and bob can work only on the qubit pairs , provided that the outcome is `` s '' .
we also assume that , after this measurement , if the outcome is `` s '' , bob performs the `` local filtering operation '' @xcite on qubit b , described by the kraus operator @xmath11 where @xmath12 $ ] .
note that if the channel is loss - free and noiseless or eve does nothing , this operation yields the maximally entangled state ( epr state ) @xmath13 with probability @xmath14 due to the fact that the initial state is also written as @xmath15 .
thus , the filtering operation with @xmath16 is optimum one in this case .
when channel is lossy and noisy , or eve is present , the filtered states may include bit errors , represented by the subspace spanned by @xmath17 , and a phase error , represented by the subspace spanned by @xmath18 . in order to distill the epr pairs ,
alice and bob run the edp based on css code @xcite and after this edp alice and bob measure each in the @xmath7-basis of the epr pairs so that they obtain a secure key .
note that the qnd measurement followed by the filtering @xmath19 and @xmath7-basis measurement is equivalent to a measurement @xmath20 described by @xmath21 , @xmath22 , @xmath23 , and @xmath24 , where @xmath25 .
thus @xmath26 is the same measurement as in the b92 with perfect photon detector , and @xmath20 corresponds to the generalized b92 protocol . for convenience ,
let @xmath27 be the measurements described by the set of the povm elements @xmath28 .
the following is the protocol that we want to be shown secure and reduce to the generalized b92 protocol .
_ protocol : _ ( 1 ) alice prepares the state @xmath29 , and she sends the system b to bob over a quantum channel .
( 2 ) bob performs the qnd measurement described by @xmath30 , and he publicly announces the outcome .
they take notes of the outcome .
( 3 ) alice and bob repeats the step ( 1 ) and ( 2 ) @xmath31 times .
let @xmath32 be the number of outcomes that resulted in `` v '' .
( 4 ) alice and bob discard all the pairs whose outcomes of the qnd measurement are `` v '' . by public discussion , alice and
bob randomly permute the positions of the remaining @xmath33 pairs .
( 5 ) for the first @xmath34 pairs ( check pairs ) , alice measures system a in the @xmath7-basis , and bob performs measurement @xmath27 on his system . by public discussion
, they determine the number @xmath35 of errors in which alice found @xmath36 and bob s outcome was @xmath37 , or alice found @xmath37 with bob s outcome @xmath36 .
( 6 ) for the second @xmath34 pairs ( data pairs ) , bob performs the filtering @xmath19 on each of his qubits , and announces the total number @xmath38 and the positions of the qubits that have passed the filtering .
( 7 ) from @xmath39 , @xmath35 and @xmath38 , they estimate an upper bound for the number of bit errors @xmath40 , and an upper bound for the number of phase errors @xmath41 , in the @xmath38 pairs .
if these bounds are too large , they abort the protocol .
( 8) they run an edp that can produce @xmath42 nearly perfect epr pairs if the estimation is correct .
( 9 ) alice and bob each measures the epr pairs in @xmath7-basis to obtain a shared secret key . in the following
, we will reduce our protocol to the b92 protocol . before this reduction
, we should briefly mention what alice and bob do experimentally in the b92 protocol . in the b92 protocol , alice randomly prepares the single photon polarization in the state @xmath43 or @xmath44 and sends the state to bob .
after receiving the state , bob performs measurement described by @xmath20 . when @xmath45 and bob detects a single photon state , this measurement effectively performs a polarization measurement on the single photon state in a basis randomly chosen from @xmath46 and @xmath47 . in the case where @xmath16 and bob detects a single photon state , this measurement effectively performs a general measurement which describes the optimum unambiguous state discrimination between @xmath43 and @xmath44 @xcite .
thanks to bob s perfect detector , if the outcome of the qnd measurement is `` s '' , then bob s states are projected onto the single photon space .
this projection requires two more steps in our protocol compared to the _ protocol 1 _ in @xcite .
the benefit of this projection is that for the states projected onto qubit - pairs alice and bob can use the one - way edp based on css code @xcite . in the context of qkd
, this edp allows alice to perform @xmath7-basis measurements immediately after she has prepared the state @xmath29 and bob to perform @xmath7-basis measurements immediately after he has performed the filtering .
now bob s measurement is described by @xmath27 .
since bob performs the qnd measurement before the measurement @xmath27 , his measurement is equivalent to the measurement @xmath48 . as a whole
, the protocol has been reduced into the protocol where alice prepares and sends @xmath43 or @xmath44 randomly and bob performs the @xmath20 measurement .
this completes the reduction of our protocol into the generalized b92 protocol over lossy and noisy quantum channel . due to this equivalence , it is sufficient to prove the security of the protocol for the security analysis of the b92 .
the most important part in the quantitative security analysis of our protocol is to estimate phase errors and bit errors in the step ( 7 ) .
if this estimation is not exponentially reliable , then in the protocol alice and bob can not distill the state sufficiently close to epr state so that the protocol is not secure @xcite . in the next section , we will give the formula to estimate the bit errors and phase errors .
in this section , we consider the error estimations in the step ( 7 ) of the protocol .
thanks to bob s perfect detector , the filtered states are qubit states , which allows us to directly apply some inequalities from @xcite .
first , note that bit errors @xmath40 in the data pairs could be determined if alice measures those pairs in @xmath7-basis , bob performs measurement @xmath27 onto those pairs , and they compare their results .
thus , the bit error rate @xmath35 in the check pairs and bit errors @xmath40 in the data pairs are determined by the same measurements .
furthermore thanks to the random permutation in step ( 4 ) , we can apply the classical probability estimation so that we obtain @xmath49 for any strategy by eve , the probability of violating this inequality is asymptotically less than @xmath50 $ ] .
since the povm elements that give the bit error rate on the filtered states @xmath51 and phase error rate on the filtered states @xmath52 do not commute @xcite , the classical probability argument can not be applied to the estimation of the phase errors from the bit errors . for
the estimation , let us perform the gedanken measurements in the @xmath53 basis @xmath54 , and let the numbers @xmath55 of the pairs found in the state @xmath53 .
here we define @xmath56 as the outcome where alice obtains the bit value @xmath57 and bob obtains vacuum or multi - photon states . as far as we consider the measurement outcomes of the povm elements that are diagonal with respect to @xmath53 basis , we are allowed to apply the classical argument for the estimation .
thus , we have the following inequalities , @xmath58 which are violated with probability asymptotically less than @xmath59 $ ] and @xmath60 $ ] , respectively . similarly , since neither the noisy channel nor eve can touch the qubits held by alice , we have @xmath61 with probability of violation asymptotically less than @xmath62 . in order to represent the bit error in the check pairs in terms of gedanken measurements , we introduce the global gedanken measurements in the basis @xmath63 @xmath64 and the number of the corresponding measurement outcomes @xmath65 , where @xmath66 . here
the summation is taken modulo 2 .
since the povm element for the bit error is diagonal with respect to the @xmath63 basis @xcite , we have @xmath67 which is violated with probability asymptotically less than @xmath68
. since @xmath69 and @xmath70 span the same subspace , we can relate @xmath71 and @xmath72 by the classical probability estimate as @xmath73 which is violated with probability asymptotically less than @xmath74 $ ] . for the estimate of the phase error rate in the data qubit from the bit error rate in the check qubit
, we need to relate @xmath75 to @xmath76 .
this relation can not be obtained by classical arguments , since @xmath77 and @xmath78 are nonorthogonal . in @xcite ,
the authors consider the qubits that are invariant under the random permutation in the limit of large number .
then they showed that when we consider the measurement outcomes onto such qubits , we are allowed to regard the qubits as if they arose from the independently and identically distributed quantum source . their argument can be directly applied to our case , since in step ( 4 ) of the protocol we have also applied the random permutation .
thus , we can make use of the exponentially reliable inequality derived in @xcite as @xmath79 for @xmath80 , where all the angles are defined in @xmath81 $ ] by the relations @xmath82 , @xmath83 , @xmath84 , @xmath85 , and @xmath86 . the eqs .
( [ ineq1])([lastineq ] ) , together with the relation @xmath87 and @xmath88 can be used to derive the formula for the estimation of the phase error from bit error . in the following we set @xmath89 , @xmath90 by taking the limit @xmath91 so that eqs .
( [ ineq1])([ineq6 ] ) are now linear equations .
we first consider the special case for the b92 where bit error rate in the filtered qubits is zero while the phase error rate in the qubits is not zero .
eve can accomplish this by using channel loss . from @xmath92 ,
i.e. , @xmath93 , one can easily find that for @xmath94 @xmath95 the limiting loss rate @xmath96 coincides with the case where eve employs optimum unambiguous state discrimination measurement @xcite . in this attack ,
if she obtains the conclusive outcome , then she sends the corresponding state to bob , otherwise she does not .
next we consider the general case .
the relationship can be obtained by solving the equations that we have mentioned just above .
after straight forward calculations , we find the implicit constraint @xmath97 where @xmath98/(\beta^2-\alpha^2 ) \nonumber\\ x&\equiv&\left(\beta^2/\gamma^2\right ) 2 n_{\rm ph}/n-(\beta^2-\alpha^2)\delta \nonumber\\ l_0&\equiv&\alpha^2 n_{1{\rm v}}/n+\beta^2 n_{0{\rm v}}/n \nonumber\\ l_1&\equiv&\alpha^2 n_{0{\rm v}}/n+\beta^2 n_{1{\rm v}}/n\end{aligned}\ ] ] here , @xmath99 and @xmath100 are equal to @xmath101 and @xmath102 respectively so that they must be positive .
combining these positivity constraints with the positivities inside the square roots , we have @xmath103 .
we can easily confirm that these equations are the same as equations in @xcite where @xmath45 and the channel has no loss i.e. , @xmath104 .
( [ ineqlimit ] ) gives us the possible parameter set @xmath105 for fixed @xmath106 , @xmath107 , @xmath108 , @xmath109 , @xmath39 , and @xmath38 . by taking the optimum ratio between @xmath108 and @xmath109 with respect to @xmath41 for given @xmath35 , @xmath38 and @xmath39 , we obtain an upper bound @xmath110 on the number of phase errors @xmath41 , as a function of the observed values @xmath35 , @xmath38 and @xmath39 , and the chosen parameters @xmath106 and @xmath107 .
in this section , at first we give the examples of the phase error estimation and resulting performance of the b92 protocol .
to illustrate the accuracy of the phase error estimation and security performance of the b92 protocol , we assign values to the observable data as they would arise from a depolarizing quantum channel with loss , i.e. , the state @xmath111 evolves into @xmath112 $ ] , where @xmath113 is the vacuum state , @xmath39 is the loss rate of quantum channel , @xmath114 is the depolarizing rate , and @xmath115 is the pauli matrix . in fig
. [ phase ] , we show the upper bound of the estimated phase error rate , the actual phase error rate , and bit error rate for the case ( a ) @xmath116 , and ( b ) @xmath117 .
these figures are independent of @xmath107 , since we plot the phase error rates normalized by @xmath38 so that the pre - factor @xmath118 is cancelled . in each figure
, it is seen that when @xmath119 becomes smaller , the estimation of the phase errors becomes poorer . on the other hand , larger values of @xmath119
make the signal more vulnerable to noise and loss .
furthermore , due to the loss , the estimation in ( b ) is poorer than that in ( a ) , especially in the region for small @xmath119 .
( dot - dashed line ) , the actual phase error rate ( solid line ) , and bit error rate ( dotted line ) in case of @xmath120 .
diagram ( a ) is for @xmath121 and ( b ) is for @xmath122 .
we assume that quantum channel is a depolarizing channel with loss .
[ phase ] ] since alice and bob apply random permutation in the protocol , the achievable length of the final key is given by @xmath123 $ ] @xcite , when @xmath124 . here
the positivity of @xmath42 means that alice and bob generate a secure key . in fig .
[ keygain ] ( a ) , we have numerically calculated the key generation rate @xmath126 in the case of @xmath45 for the fixed value of @xmath119 , @xmath39 and @xmath114 , and plot @xmath127 optimized over @xmath119 as a function of @xmath114 .
the optimum value of @xmath119 is shown in fig . [ keygain ] ( b ) .
since the optimization of the ratio between @xmath108 and @xmath109 is complicated , we have optimized the ratio numerically . from fig .
[ keygain ] ( a ) , it is seen that the b92 protocol as described here is secure up to @xmath128 ( in the case of @xmath121 ) , @xmath129 ( in the case of @xmath130 ) , and @xmath131 ( in the case of @xmath122 ) . the use of the general b92 measurement , i.e. , @xmath16 yields the higher secret key gain , however this does not change the cutoff of the gain nor the optimum angle in fig .
[ keygain ] ( b ) .
this is again because of the fact that the cutoff is determined by the error rate normalized by @xmath38 . in the bb84 protocol or 6-state protocol @xcite ,
thanks to the randomly chosen conjugate bases , phase error estimation is more precise than that in the b92 protocol . the poor estimation in the b92 makes this protocol more vulnerable to eavesdropping . actually , it is known that the bb84 protocol and 6-state protocol are secure up to @xmath132 @xcite and @xmath133 @xcite corresponding to bit error rates of @xmath134 and @xmath135 , respectively . for the case @xmath45 and ( b ) the optimum value of @xmath119 . here
we assume that the quantum channel is the depolarizing channel with loss .
the dotted line , the dot - dashed , and solid line represent the case where @xmath121 , @xmath130 , and @xmath122 respectively .
[ keygain ] ] in summary , we have shown that the security proof of the b92 over loss - free channel @xcite can be adapted to lossy and noisy channels assuming bob s detector that discriminates between single photon states on one hand and vacuum state or multi - photon states on the other hand .
we have derived the general formula for the estimation of the phase errors from bit errors .
we have also shown that the boundary loss rate at which eve can induce phase errors freely without inducing bit errors coincides with the case where eve employs the optimum unambiguous state discrimination measurement .
finally , we have shown the examples of the security assuming the quantum channel is the depolarizing quantum channel with loss .
it is an interesting problem to compare our results to the case where eve performs only individual attack . for the b92 with loss
, the security is estimated in @xcite .
however in that paper , it is assumed that alice and bob employ error discarding protocol so that we can not compare our result to their result directly .
we leave this problem for future studies .
another interesting problem is to consider the case where bob s detector is not perfect . in this case
, some filtered states may remain vacuum state or multi - photon states so that our proof can not be directly applied .
the security analysis of the b92 with coherent state is another interesting problem .
we hope our study is helpful for solving these problems and we leave them as future studies .
k.t appreciates the warmth and hospitality of the quantum information theory group , universitt erlangen - nrnberg in germany during his stay .
we thank to marcos curty , peter van loock , philippe raynal , and aska dolinska for helpful discussions .
this work has been supported in part by the academic frontiers student exchange promotion program under grant no .
60 by the ministry of education , culture , sports , science and technology , japan , and by the dfg under the emmy - noether programme . | we show that the security proof of the bennett 1992 protocol over loss - free channel in ( k. tamaki , m. koashi , and n. imoto , phys .
rev .
lett . * 90 * , 167904 ( 2003 ) ) can be adapted to accommodate loss .
we assumed that bob s detectors discriminate between single photon states on one hand and vacuum state or multi - photon states on the other hand . |
exploring entanglement has attracted considerable interest in various physical systems , ranging from atoms and photons to solid - state materials @xcite .
entanglement describes a phenomenon in which two or more subsystems are linked through direct or indirect interactions .
this a quantum phenomenon has become an essential ingredient in quantum information processing @xcite . among various architectures ,
a bipartite system is a fundamental unit to exploit the nature of entanglement .
fascinating properties have been extracted from a variety of bipartite systems .
the bell state is a representative example , which plays an important role in dense coding protocol @xcite , quantum teleportation @xcite , and entanglement swapping @xcite . as a result
, generating entanglement becomes an indispensable step for follow - up operations and further applications .
two polar molecules can form a platform for generating entanglement directly through dipole - dipole interaction .
this molecular platform provides an advantage of controllable rotational properties via the coupling of external fields with the dipole moment @xcite .
field sources include static electric fields @xcite , laser pulses charron07 , and optimized complex pulses @xcite .
the rotational properties , such as energy and wave function , are modified to affect the entanglement as well as the dipole - dipole interaction . in most studies , polar molecules
are considered to rotate in three dimensions .
however , only minor efforts have been made that focus on the entanglement generated by two - dimensional rotation .
the energies , level degeneracy , and wave functions of a two - dimensional rotor are different from those of a three - dimensional rotor @xcite .
two - dimensional rotational manner has been observed in the experiments of hydrogen on a surface svensson99,bengtsson00,billy02 .
the planar rotation in diverse configurations further leads to interesting consequences for energy spectrum @xcite , molecular orientation @xcite , and specific heat @xcite . in this work
we study the entanglement in a system of two polar molecules confined to a plane .
both molecules rotate in two dimensions , which are modified by applying an in - plane electric field .
we use concurrence to qualify the degree of entanglement by orienting the electric field . for specific field directions ,
all bell - like states are jointly displayed , corresponding to a situation of overlapping concurrences .
specifically , bell - like states are robustly maintained when the dipole - field and dipole - dipole interactions are tuned . on the contrary ,
these pronounced properties are destroyed in an electric field of another direction .
the obtained results are different from those in the three - dimensional cases .
we further analyze the thermal effect on the entanglement for the bell - like states . .
the left and right molecules are confined to two - dimensional motion , corresponding to the unit vectors of the dipole moments @xmath0 and @xmath1 , respectively .
, width=321 ]
we consider two polar diatomic molecules of rotational constant @xmath2 and dipole moment @xmath3 confined to a plane as depicted in fig .
both molecules rotate in two dimensions , which are spatially separated by an interval @xmath4 .
a static electric field of strength @xmath5 and tilt angle @xmath6 is applied to couple with the dipole moments of the molecules .
the modified two - dimensional rotation leads to specific dipole - dipole interaction .
the hamiltonian describing the system is @xmath7with@xmath8@xmath9 , \label{ddinteraction}\]]where @xmath10 is the strength of dipole - field interaction , @xmath11 is the strength of dipole - dipole interaction , @xmath12 is the unit vector of the dipole moment , @xmath13 is the unit vector , @xmath14 is the permittivity of free space , and @xmath15 is the polar angle between the molecular axis and the @xmath16 axis for the left @xmath17 and right @xmath18 molecules . from the coordinates of the molecules ,
the dipole - dipole interaction @xmath19 is given by @xmath20 the magnitudes of @xmath21 and @xmath22 in eqs .
( [ h0 ] ) and ( dipole ) are dependent on the dipole moment , field strength , and intermolecular distance . for a single molecule in an electric field
, the field - dependent energy @xmath23 and its corresponding wave function @xmath24 are obtained from the equation @xmath25where the index @xmath26 is omitted and the positive integer @xmath27 is defined from @xmath28 .
the wave function @xmath24 is described by @xmath29where @xmath30 is the expanded coefficient , @xmath31 is the field - free eigenfunction , corresponding to the energy @xmath32 for @xmath33 @xcite .
the effect of the dipole - dipole interaction on the system is captured by employing the lowest two states of the molecules .
the chosen states are the ground state @xmath34 and the first excited state @xmath35 , corresponding to the energies @xmath36 and @xmath37 , respectively . here
the index @xmath26 is restored . to solve the hamiltonian in eq .
( [ hamiltonian ] ) , a composite state is taken by the form@xmath38where @xmath39 is the coefficient , and @xmath40 @xmath41 , @xmath42 @xmath43 , @xmath44 @xmath45 , and @xmath46 @xmath47 are the basis states .
by using @xmath48 , the energy @xmath49 and the coefficient @xmath39 can be derived from the matrix @xmath50with @xmath51 and @xmath52 for @xmath53 and @xmath54 being @xmath55 , @xmath56 , @xmath57 , and @xmath58 .
the factors are expressed as @xmath59 , @xmath60 , @xmath61 , and @xmath62 .
the density matrix method is used to examine the entanglement properties in the system @xcite . based on @xmath63 and @xmath64
, the density matrix can be described as @xmath65with the boltzmann constant @xmath66 and temperature @xmath67 .
the degree of entanglement can be measured by the concurrence @xmath68where the quantities @xmath69 are the eigenvalues of the matrix @xmath70 in decreasing order .
the operator @xmath71 is the pauli matrix and @xmath72 is the complex conjugate of @xmath73 .
the value of the concurrence ranges from zero to one , corresponding to an unentangled state and a maximally entangled state , respectively .
@xmath74 as a function of the parameter @xmath75 .
the energy difference between the ground and first excited states is shown in ( b ) .
( c ) probability densities @xmath76 of the ground and first excited states as a function of the angle @xmath77 for @xmath78 .
the tilt angle is @xmath79.,width=321 ]
in this work the systemic properties are explored by controlling the strength of the dipole - field interaction @xmath21 and the dipole - dipole interaction @xmath22 .
the energy - related quantities are expressed in units of @xmath2 throughout the calculations .
figure [ fig2](a ) shows the energy spectra of a single molecule in an electric field of @xmath80 , i.e. , the field direction lies along the @xmath16 axis .
the energy properties are influenced by varying the parameter @xmath81 .
when the field is absent , all excited states are two - fold degenerate except for the ground state at @xmath82 . due to the presence of dipole - field interaction ,
the degenerate levels are split off at @xmath83 .
the energy structure differs from that of a three - dimensional rotor @xcite . for the lowest two states , the energies @xmath36 and @xmath37 display decreasing behavior with increasing the parameter @xmath81 .
the energy difference between the two levels is gradually increased from @xmath2 [ see fig . [ fig2](b ) ] .
for instance , a difference of @xmath84 is determined at @xmath85 . correspondingly , in fig .
[ fig2](c ) the probability densities of the two states are modulated to spread along the @xmath16 axis . as a whole ,
the energies are the same at arbitrary @xmath6 , whereas the spatial distributions of the wave functions depend on the field direction .
it is worth noting that , as defined in eq .
( [ wavefunction ] ) , a two - level subsystem is built from the two individual states @xmath86 and @xmath87 .
however , the two states @xmath87 and @xmath88 are degenerate at @xmath82 , and thus have the same energies @xmath89 @xcite .
the corresponding wave function is either @xmath90 or @xmath91 . at @xmath82 , the transition between the states @xmath92 and @xmath88 can occur as well as the transition between the states @xmath93 and @xmath87 .
the two - level subsystem fails in this situation . as a result , to generate entanglement based on coupled two - level subsystems , the tuning of @xmath94
is excluded throughout the calculation .
( a ) and @xmath95 ( b ) of the factors as a function of the tilt angle @xmath96 for @xmath78 .
square values @xmath97 ( c ) and @xmath98 ( d ) of the contributive factors as a function of the parameter @xmath99 for @xmath79 and @xmath100 . , width=321 ] in eq . ( [ matrix ] ) the energy @xmath63 and the state @xmath101 depend on these factors .
the properties of the wave function plays a determinant role in the contribution of the factors . in figs . [
fig3](a ) and ( b ) an electric field of fixed strength is oriented to modulate the spatial distribution of the wave function so that the absolute values of the factors become angle - dependent . by increasing @xmath102 ,
the value @xmath103 inversely varies against @xmath104 and @xmath105 .
the features of @xmath106 are similar to those of @xmath107 .
if the tilt angle is @xmath108 or @xmath100 , the values of the factors will be either real or purely imaginary . for @xmath109 , only @xmath110 , @xmath105 , and @xmath111 are contributive and maximal , whereas @xmath112 , @xmath113 , and @xmath114 are zero .
a reverse situation occurs for @xmath115 where the contributive roles are interchanged to be @xmath112 , @xmath113 , and @xmath114 . to probe the two cases ,
the square values of the contributive factors are shown in figs . [ fig3](c ) and ( d ) .
we simply set the initial value of @xmath81 as @xmath116 to avoid the three - level situation mentioned above .
the relations @xmath117 , @xmath118 , and @xmath119 hold at arbitrary @xmath120 .
the magnitudes of @xmath121 and @xmath122 , as well as those of @xmath123 and @xmath124 , are raised by increasing the parameter @xmath120 . on the contrary , @xmath125 and @xmath126
are close to zero at large @xmath81 .
( a ) and concurrences @xmath127 ( b ) as a function of the parameter @xmath75 for @xmath79 .
( c ) and ( d ) correspond to @xmath128 .
the insets show the difference between the energies @xmath129 and @xmath130 .
the fixed parameter is @xmath131.,width=321 ] the entanglement property is analyzed in an electric field of a specific tilt angle .
figure [ fig4](a ) shows the four energy levels for @xmath132 and @xmath131 . in this situation
the field direction is parallel to the intermolecular direction .
the energies are gradually diminished by increasing @xmath81 .
the corresponding concurrences reveal pronounced features as illustrated in fig . [
fig4](b ) .
the concurrence for @xmath133 is equal to that for @xmath134 .
both values are decreased by increasing @xmath81 . by contrast , the states @xmath135 and @xmath136 have the maximum concurrences @xmath137 all the time . when the electric field is applied perpendicular to the intermolecular direction , the decreasing energies in fig .
[ fig4](c ) are similar to those for @xmath109 .
however , the energy structures are different in the two cases . for example
, the difference between @xmath129 and @xmath130 for @xmath109 is greater than that for @xmath138 .
different from the case of @xmath109 , a local minimum for @xmath139 is observed at @xmath140 , i.e. , an anticrossing occurs [ see insets of fig .
[ fig4 ] ] . in fig .
[ fig4](d ) the overlapping features of concurrences also occur in the case of @xmath139 .
the concurrences for @xmath141 and @xmath142 are still @xmath137 , while the same concurrences for @xmath133 and @xmath143 are higher than those in fig .
[ fig4](b ) .
( a ) and concurrences @xmath127 ( b ) as a function of the parameter @xmath144 for @xmath79 .
( c ) and ( d ) correspond to @xmath128 .
the fixed parameter is @xmath78.,width=321 ] next , we turn to examining the effect of dipole - dipole interaction on the energies and concurrences by fixing the strength of the dipole - field interaction . the fixed parameter is set to be @xmath85 , corresponding to @xmath145 and @xmath146 in fig .
[ fig2 ] . by increasing the parameter @xmath144
, the four levels for @xmath132 reveal decreasing features [ see fig .
[ fig5](a ) ] .
the energy differences between the adjacent levels increase at large @xmath144 .
however , in fig .
[ fig5](c ) a diverse energy structure is obtained for @xmath147 . the levels for @xmath129 and @xmath130 have a crossing at @xmath148 .
increasing behavior is observed for @xmath149 and @xmath150 . as shown in figs .
[ fig5](b ) and ( d ) , these concurrences can be classified into two types , which are similar to those in fig .
such a result indicates that the overlapping property is robust against the modulation of dipole - dipole interaction . for a system of three - dimensional rotors ,
similar concurrences can be observed only when the parameter is @xmath82 @xcite .
furthermore , the stronger dipole - dipole interaction obviously enhances the concurrences for @xmath151 and @xmath134 .
the degree of enhancement on the concurrence is higher in the case of @xmath147 . to gain insight into the distinctive entanglement properties , the matrix in eq .
( [ matrix ] ) can be simplified according to the results in fig .
fig3 : @xmath152where the parameters @xmath153 , @xmath154 , @xmath155 , @xmath156 , and @xmath157 are for @xmath109 while @xmath158 , @xmath159 , @xmath160 , @xmath161 , and @xmath162 are for @xmath147 . after calculating the matrix ,
one obtains the bell - like states @xmath163@xmath164@xmath165@xmath166and their energies@xmath167@xmath168@xmath169@xmath170with @xmath171 , @xmath172 , and @xmath173 . for @xmath132 , the subscripts @xmath174
are used in eqs .
( [ bell2s ] ) and ( [ bell2e ] ) , corresponding to @xmath175 used in eqs .
( [ bell3s ] ) and ( bell3e ) . the subscripts interchange for @xmath138 .
the concurrence is also derived from the definition @xmath176for the state @xmath177 @xcite .
the concurrences for @xmath141 and @xmath178 are straightforwardly @xmath137 .
the states @xmath179 and @xmath134 have the same concurrence @xmath180with @xmath181 ^{2 } , \label{r0}\]]for @xmath109 and @xmath182 ^{2 } , \label{r90}\]]for @xmath138 .
the analytic results obtained agree exactly with the energies and concurrences in figs .
[ fig4 ] and [ fig5 ] .
once the particular field directions are selected , the four bell - like states can jointly occur as shown in eqs .
( [ bell1s]-[bell4s ] ) .
this situation corresponds to two pairs of identical concurrences .
more importantly , the modulation of the dipole - field and dipole - dipole interactions does not destroy the coexistent characteristic of bell - like states .
( a ) and concurrences @xmath127 ( b ) as a function of the tilt angle @xmath96 for @xmath78 and @xmath183.,width=321 ] when the tilt angle is not limited to @xmath109 and @xmath184 , all factors become contributive to energy and concurrence .
figure [ fig6](a ) shows the angle - dependent energies for @xmath85 and @xmath131 . here
the levels are labeled based on the case of @xmath185 ; this label is consistently retained throughout the tuning of @xmath6 .
the energies @xmath129 and @xmath150 vary monotonically with increasing @xmath6 .
by contrast , the energies @xmath186 and @xmath187 are modulated to intersect at @xmath188 . for @xmath189 and @xmath100 , an explicit signature of entanglement
is that the concurrences no longer coincide together [ see fig .
[ fig6](b ) ] . only the state @xmath135 retains the maximum concurrence @xmath137 , corresponding to @xmath190 throughout the modulation of @xmath6 .
the other states are composed of the four basis states in eq .
( [ wavefunction ] ) , whose concurrences are smaller than @xmath56 .
the result implies that the four bell - like states do not jointly exist at @xmath189 and @xmath100 .
specifically , due to the angle - dependency of the factors , the concurrences for @xmath133 and @xmath143 are modulated to be greatly reduced at @xmath191 and @xmath192 , respectively . for @xmath79 and @xmath184 as a function of the parameters @xmath75 and @xmath193 at different temperatures .
( a ) and ( b ) correspond to @xmath131 while @xmath78 is used in ( c ) and ( d).,width=321 ] the thermal influence on the entanglement is analyzed , specifically with respect to the bell - like states .
figures [ fig7](a ) and ( b ) show the temperature - related concurrences by varying the parameter @xmath81 .
since the temperature induces the contribution of the four bell - like states , the mixing effect leads to a suppressed concurrence [ see eq .
( [ density ] ) ] .
when the temperature is raised , the degree of reduction of concurrence is enlarged .
although the concurrences for @xmath109 and @xmath184 are similar , two dissimilar suppressed features occur , especially in the region of small @xmath81 .
this result originates from the two different energy structures , as illustrated in figs .
[ fig4](a ) and ( c ) .
furthermore , the properties of the energy structure strongly affect the concurrences that are dependent on the parameter @xmath144 [ see figs .
fig7(c ) and ( d ) ] . different from smooth features for @xmath194 , greatly suppressed concurrences can be obtained around the point where an energy crossing between @xmath129 and @xmath130 occurs for @xmath115 , as depicted in figs .
[ fig5](a ) and ( c ) .
we have investigated the entanglement properties in a system of two - dimensional polar molecules by applying an electric field .
the electric field parameters modulate the two - dimensional rotation , leading to distinctive energy spectra and concurrence . when the field direction is perpendicular or parallel to the intermolecular direction , the concurrences become overlapping .
the four bell - like states are collectively generated in the two situations .
the coexistence of bell - like states steadily holds against the modulation of the dipole - field and dipole - dipole interactions .
once other field directions are selected , these significant properties will be destroyed . instead
, the concurrences can be greatly reduced in certain field directions .
the concurrences for the bell - like states are further analyzed at finite temperatures .
the different energy structures cause diverse suppressed features in the two situations . | entanglement generated from polar molecules of two - dimensional rotation is investigated in a static electric field .
the electric field modulates the rotational properties of molecules , leading to distinctive entanglement .
the concurrence is used to estimate the degree of entanglement . when the electric field is applied parallel or perpendicular to the intermolecular direction ,
the concurrences reveal two overlapping features .
such a pronounced signature corresponds to the coexistence of all bell - like states .
the characteristics of bell - like states and overlapping concurrences are kept independent of the modulation of dipole - field and dipole - dipole interactions . on the contrary ,
the bell - like states fail to coexist in other field directions , reflecting nonoverlapping concurrences .
furthermore , the thermal effect on the entanglement is analyzed for the bell - like states .
dissimilar suppressed concurrences occur due to different energy structures for the two specific field directions . |
cassels @xcite was challenged to determine when the sum of three consecutive cubes equals a square .
he @xcite reduced the problem to finding integral points on the elliptic curve @xmath0 . using the arithmetic of certain quartic number fields
, he obtained that the integral points on the above elliptic curve were @xmath1 , @xmath2 , @xmath3 , and @xmath4 . using the classical work of ljunggren @xcite and its generalizations ( see @xcite , @xcite , @xcite , and @xcite ) , luca and walsh @xcite considered the problem of finding the number of positive integer solutions to the diophantine equation @xmath5 , where @xmath6 is a positive integer .
they proved that the number of positive integer solutions to @xmath5 is at most @xmath7 , where @xmath8 is the number of distinct prime factors of @xmath9 . in @xcite ,
chen considered the case where @xmath9 is a prime number greater than @xmath10 .
he proved , in particular , that the diophantine equation @xmath5 has at most two positive integer solutions . recently , togb @xcite considered the more general diophantine equation @xmath11 where @xmath12 is a prime number and @xmath13 is an odd integer greater than @xmath14 .
he proved the following theorem .
[ to1 ] for any prime @xmath12 and any odd positive integer @xmath15 , the diophantine equation has at most seven positive integer solutions @xmath16 . using results obtained through magma
, he then made the following conjecture on sharp bounds for the number of solutions to equation .
[ to2 ] let @xmath12 be a prime and @xmath15 any odd positive integer . 1 .
if @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 , or @xmath27 , then diophantine equation has at most one positive integer solution @xmath16 .
2 . if @xmath28 or @xmath29 , then diophantine equation has at most two positive integer solutions @xmath16 .
3 . if @xmath30 or @xmath31 , then diophantine equation has at most three positive integer solutions @xmath16 .
the aim of this paper is to improve the bound on the number of solutions to the diophantine equation provided in theorem [ to1 ] , and to prove conjecture [ to2 ] in some cases .
the main result of this paper is the following theorem .
[ th1 ] let @xmath12 be a prime and let @xmath15 be an odd integer . 1 . if @xmath32 , then diophantine equation has at most one positive integer solution @xmath16 .
2 . suppose that @xmath33 or @xmath34 , where @xmath12 is odd . 1 .
if @xmath35 or @xmath36 , then diophantine equation has at most three positive integer solutions @xmath16 .
diophantine equation has at most one positive integer solution @xmath16 otherwise .
3 . suppose that @xmath37 , where @xmath12 is odd . 1 . if @xmath38 , @xmath19 , @xmath21 , @xmath24 , @xmath26 , or @xmath27 , then diophantine equation has at most one positive integer solution @xmath16 .
2 . if @xmath17 , @xmath20 , @xmath22 , @xmath23 , @xmath39 , or @xmath40 , then diophantine equation has at most two positive integer solutions @xmath16 .
3 . if @xmath28 or @xmath41 , then diophantine equation has at most three positive integer solutions @xmath16 .
4 . if @xmath42 , then diophantine equation has at most four positive integer solutions @xmath16 .
5 . if @xmath43 , then diophantine equation has at most six positive integer solutions @xmath16
. we will also prove the following result .
[ th5 ] let @xmath12 be a prime and let @xmath15 be an even integer . 1 . if @xmath32 , then diophantine equation has at most two positive integer solutions @xmath16 . moreover , if @xmath44 and @xmath45 , then diophantine equation has at most one positive integer solution @xmath16 .
suppose that @xmath46 or @xmath47 , where @xmath12 is odd . 1 . if @xmath44 , then diophantine equation has at most one positive integer solution @xmath16 .
2 . if @xmath48 , then diophantine equation has at most two positive integer solutions @xmath16 .
3 . suppose that @xmath49 , where @xmath12 is odd . 1 . if @xmath50 , then diophantine equation has at most one positive integer solution @xmath16 .
2 . if @xmath51 , then diophantine equation has at most two positive integer solutions @xmath16 .
3 . if @xmath52 , then diophantine equation has at most three positive integer solutions @xmath16 .
if @xmath53 , then diophantine equation has at most four positive integer solutions @xmath16 .
we present the results required to prove theorem [ th1 ] and theorem [ th5 ] . recall that if @xmath54 is a prime number , @xmath55 denotes the @xmath54-adic valuation of @xmath56 .
let @xmath57 and @xmath58 be odd positive integers for which the equation @xmath59 has a solution in positive integers @xmath60 .
let @xmath61 be the minimal positive solution to this equation and define @xmath62 for an odd integer @xmath63 , define @xmath64 and @xmath65 by @xmath66 luca and walsh proved the following result in @xcite regarding the solutions to the equation @xmath67 [ th2 ] 1 .
if @xmath68 is not a square , then equation has no solution .
if @xmath68 is a square and @xmath69 is not a square , then @xmath70 is the only solution to equation .
if @xmath68 and @xmath69 are both squares , then @xmath70 and @xmath71 are the only solutions to equation .
ljunggren proved the following result in @xcite .
[ th3 ] let @xmath72 and @xmath58 be two positive integers .
the equation @xmath73 has at most one solution in positive integers @xmath60 .
let @xmath74 be a positive non - square integer , and let @xmath75 denote the minimal unit greater than 1 , of norm 1 , in @xmath76 $ ] .
define @xmath77 for @xmath78 .
togb , voutier , and walsh proved the following result in @xcite .
[ th4 ] let @xmath74 be a positive non - square integer .
there are at most two positive integer solutions @xmath60 to the equation @xmath79 . 1 . if two solutions such that @xmath80 exist , then @xmath81 and @xmath82 , except only if @xmath83 or @xmath84 , in which case @xmath81 and @xmath85 .
2 . if only one positive integer solution @xmath60 to the equation @xmath79 exists , then @xmath86 where @xmath87 for some square - free integer @xmath88 , and either @xmath89 , @xmath90 , or @xmath91 for some prime @xmath92 . we will make the theorem [ th1 ] more precise when @xmath74 is even .
[ lem7 ] let @xmath74 be a positive non - square integer .
suppose that @xmath93 where @xmath94 is a positive integer different from @xmath95 . then the equation @xmath79 has at most one positive solution @xmath60 .
suppose that there exist two solutions to the equation @xmath79 .
then there exist positive integer solutions @xmath96 and @xmath97 such that @xmath80 .
it follows from theorem [ th4 ] that @xmath81 , @xmath82 , and @xmath98 , so @xmath99 .
then @xmath100 since @xmath75 is a unit of norm @xmath14 in @xmath76 $ ] and @xmath93 , we obtain @xmath101 , so that @xmath102 is odd .
then @xmath103 , which is a contradiction with .
_ proof of theorem [ th1 ] .
_ let @xmath32 , and let @xmath13 be an odd positive integer .
let @xmath104 be positive integers such that @xmath105 .
it is not difficult to see that @xmath106 divides @xmath107 and @xmath108 .
let @xmath109 and @xmath110 .
then we obtain @xmath111 since @xmath112 , there exist positive integers @xmath113 and @xmath114 such that @xmath115 , @xmath116 , and @xmath117 by lemma [ lem7 ] , this equation has at most one positive integer solution @xmath118 .
let @xmath12 be an odd prime , and let @xmath13 be an odd positive integer .
let @xmath104 be positive integers such that @xmath119 .
we remark that @xmath120 or @xmath121 , so we consider two cases depending on the parity of @xmath107 , with each case yielding two equations .
suppose first that @xmath107 is even , so we let @xmath122 .
since @xmath12 is prime , we let @xmath123 .
then we obtain @xmath124 since @xmath125 , there exist positive integers @xmath113 and @xmath114 such that either @xmath126 , @xmath127 , and @xmath128 or @xmath115 , @xmath129 , and @xmath130 suppose next that @xmath107 is odd . since @xmath12 is prime , we let @xmath131 .
then we obtain @xmath132 since @xmath133 , there exist odd integers @xmath113 and @xmath114 such that either @xmath134 , @xmath135 , and @xmath136 or @xmath137 , @xmath138 , and @xmath139 we consider each of the above four equations separately to determine the number of positive integer solutions to equation .
we next consider equation , which has at most one positive integer solution by theorem [ th3 ] .
it follows from this equation that @xmath114 is odd and that @xmath113 is even if and only if @xmath141 . if @xmath142 or @xmath143 , then @xmath113 is odd , and we obtain @xmath144 .
then equation has a solution only if @xmath17 , @xmath20 , @xmath23 , @xmath29 , @xmath2 , @xmath39 , @xmath22 , or @xmath36 .
furthermore , equation has a solution only if @xmath49 .
equation has at most two positive integer solutions by theorem [ th2 ] .
since @xmath113 and @xmath114 are both odd , we have @xmath145 so @xmath146 and @xmath147 so @xmath148
. then @xmath149 or @xmath143 , and equation has at least one solution only if @xmath35 or @xmath36 .
equation has at most two positive integer solutions by theorem [ th2 ] .
since @xmath113 and @xmath114 are odd , we have @xmath150 so that equation has a solution only if @xmath28 , @xmath151 , @xmath25 , or @xmath152 .
in particular , suppose that equation has two solutions , and let @xmath61 be the minimal positive solution of @xmath153 , so @xmath154 let @xmath155 and compute @xmath156 to obtain @xmath157 since we assume that two solutions exist to equation , @xmath68 and @xmath69 must both be squares by theorem [ th2 ] .
it follows that there exist two positive integers @xmath158 and @xmath159 such that @xmath160 , @xmath161 , and @xmath162 this yields @xmath163 . since @xmath164
, we obtain @xmath165 so @xmath166 .
it follows that @xmath167 , so @xmath168 or @xmath169 .
therefore equation has at most two positive integer solutions only if @xmath30 or @xmath152 , and it has at most one positive integer solution only if @xmath28 or @xmath40 .
furthermore , equation has a solution only if @xmath49 .
+ since the number of solutions to equations and depends on the value of @xmath170 , we first suppose that @xmath171 or @xmath34 , then equations and have no integer solution , equation has at most one solution , and has at most two positive integer solutions only if @xmath35 , or @xmath36 .
therefore when @xmath171 or @xmath172 , equation has at most three positive integer solutions if @xmath35 , or @xmath36 , and it has at most one positive integer solution in all other cases .
+ we next suppose that @xmath37
. then equation has at most one positive integer solution .
+ if @xmath173 , then equation has at most one solution , has at most one solution and only if @xmath168 or @xmath174 , has no solution , and has at most one solution and only if @xmath175 .
+ if @xmath176 , then equation has at most one solution , has at most one solution and only if @xmath168 or @xmath143 , has no solution , and has at most two solutions and only if @xmath177 .
+ if @xmath178 , then equation has at most one solution , has at most one solution and only if @xmath168 or @xmath174 , has no solution , and has at most one solution and only if @xmath179 . +
if @xmath146 , then equation has at most one solution , has at most one solution and only if @xmath168 or @xmath143 , has at most two solutions and only if @xmath168 or @xmath143 , and has at most one solution and only if @xmath180 .
+ _ proof of theorem [ th5 ] .
_ if @xmath13 is even and @xmath12 is odd , we let @xmath181 .
then @xmath182we let @xmath123 , and we obtain @xmath183since @xmath184 there exist positive integers @xmath113 and @xmath114 such that either @xmath185 , @xmath186 , and @xmath187or @xmath188 @xmath189 and @xmath190or @xmath191 @xmath192 and @xmath193or @xmath194 @xmath195 and @xmath196 if @xmath197 is a perfect square , then equation has no positive integer solution , otherwise it has at most one positive integer solution by lemma [ lem7].by theorem [ th3 ] , each of equations , , and has at most one solution .
equation has a solution only if @xmath198 and @xmath199 , equation has a solution only if @xmath197 is odd and @xmath200 and equation has a solution only if @xmath197 is odd .
since the number of solutions to equations and depends on the value of @xmath201 , we first suppose that @xmath171 or @xmath172 .
then equations and have no integer solution . + if @xmath44 , then equation has at most one solution , has no solution , has no solution , and has no solution .
if @xmath48 , then equation has at most one solution , has no solution , has no solution , and has at most one solution .
+ we now suppose that @xmath49 .
then equations and have at most one positive integer solution . if @xmath44 , then equation has at most one solution , has at most one solution only if @xmath198 , has no solution , and has no solution .
+ if @xmath48 , then equation has at most one solution , has at most one solution only if @xmath198 , has at most one solution , and has at most one solution . +
if @xmath13 is even and @xmath32 , we let @xmath181 .
then @xmath202we let @xmath203 , and we obtain @xmath204since @xmath205 , there exist positive integers @xmath113 and @xmath114 such that @xmath206 , @xmath186 , and @xmath207which has no solution if @xmath197 is a perfect square and at most two solutions by theorem [ th4 ] .
moreover , if @xmath197 is even and @xmath208 then by lemma [ lem7 ] equation has at most one solution .
when we had finished writing the paper , we noticed that a proof of the result stated in lemma [ lem7 ] already existed within the proof of theorem 1 by luca and walsh in @xcite .
our proof of lemma [ lem7 ] seems to be different than the proof of the result in @xcite . | in this paper , we find a bound for the number of the positive solutions to the titled equation , improving a result of togb . as a consequence , we prove a conjecture of togb in a few cases . |
this continuation of a previous paper [ 1 ] on the derivation of general beam - referenced stage - two spin - correlation functions is for the analysis of top - antitop pair - production [ 2,3 ] at the tevatron and at the large hadron collider [ 4 ] .
each second at the large hadron collider there will be a top - antitop pair produced when the planned @xmath7 is reached .
this should provide an almost ideal laboratory " for both investigating top - quark physics itself , and for simultaneously improved empirical mastery of reaction backgrounds and detector systematics / performance . as in the previous paper , which we denote as i " , the helicity formalism [ 5 ]
is used for a simple and transparent treatment of all relative phase effects .
we use this formalism for investigating the large effects of w - boson longitudinal - transverse interference in top - antitop pair - production for the charged - lepton plus jets channel , the di - lepton plus jets channel , and the all - jets channel . in i " , only the quark - production contribution was included ; it is the dominant contribution at the tevatron . on the other hand at the large hadron collider , the gluon - production contribution dominates . to leading order in @xmath8 ,
both contributions are included in the analysis in the present paper .
the modular property of the helicity - formalism with respect to incorporation of the various density matrices and symmetries again remains manifest throughout this analysis .
this modularity should be easy to exploit in understanding and checking large interference effects in applications of these results . using the spinor outer - product formulas @xmath9 obtained in appendix a ,
the associated gluon production density - matrix - elements are derived in appendix b in the helicity - amplitude phase convention of jacob & wick ( jw ) .
the analogous quark production density - matrix - elements were obtained in i " .
they also follow from these spinor outer - products . besides their dependence on @xmath10 ,
these gluon and quark density - matrix - elements @xmath11 exhibit several non - trivial overall minus signs and an explicit dependence on the azimuthal - angle @xmath12 .
these important spherical angles @xmath13 and @xmath12 are defined in figs . 1 and 2 .
otherwise , to avoid repetition , we assume that the reader has
i " available with its figures , discussions , and the sequential - decay density matrices @xmath14 and @xmath15 for the @xmath16-conjugate process . for the sequential - decay with the first - stage @xmath17 followed by the second - stage @xmath18 ,
the spherical angles @xmath19 , @xmath20 specify the @xmath21 momentum in the @xmath22 rest frame when there is first a boost from the @xmath23 frame to the @xmath24 rest frame , and then a second boost from the @xmath24 rest frame to the @xmath22 rest frame ( see i " ) .
the @xmath25 direction for the azimuthal - angle @xmath26 is defined by the projection of the @xmath27 momentum direction .
analogously , the spherical angles @xmath28 , @xmath29 specify the @xmath30 momentum in the @xmath27 rest frame .
as in i " , the emphasis is on tests for the charged - lepton - plus - jets reaction @xmath31 or @xmath1 .
however , in contrast with the analysis in i " , because of the differences in dependence on @xmath10 among the 5 sets of gluon - production density - matrix - elements , for clarity @xmath10 is not integrated out in the present paper .
consequently , versus the 3-angle stage - two spin - correlation function @xmath32 in i " which only included the quark - production contribution , there is instead a simple 4-angle beam - referenced stage - two spin - correlation function @xmath33 with both gluon ( 2,3 ) and quark ( 9,10 ) production contributions .
this br - s2sc function can be used for determination of the relative sign of or for measurement of a possible non - trivial phase between the two dominant @xmath2 helicity amplitudes for the @xmath3 decay mode . both in the sm and in the case of an additional large @xmath34 chiral weak - transition moment [ 6 ] , the @xmath35 and @xmath36 @xmath37 amplitudes are more than @xmath38 times larger than the @xmath39 and @xmath40 @xmath41 amplitudes . for the @xmath16-conjugate case ,
there are analogous tests for @xmath4 decay based on the analogous function , ( 21,22 ) and ( 24,25 ) . as in i " , this important signature from w - boson longitudinal - transverse interference requires use of the @xmath42 energy of the hadronically decaying w - boson , or the kinematically equivalent cosine of the polar - angle of @xmath6 emission in the anti - top ( top ) decay frame , @xmath43 . in application , for instance to @xmath44 ,
parton - level top - quark spin - correlation functions [ 7,8 ] need to be smeared with the appropriate parton - distribution functions with integrations over @xmath10 and the @xmath42 energy , @xmath45 .
the color factors have been included in these br - s2sc functions . for the gluon - production - decay sequence @xmath46
we assume that the @xmath35 and @xmath47 @xmath37 amplitudes dominate in @xmath24 and @xmath48 decay .
the simple four - angle distribution @xmath49 for @xmath50 involves the angles @xmath51 ; @xmath52 , @xmath53 , @xmath54 , where @xmath55 is shown in fig.1 .
the other angles are displayed in figs . 1 and 3 of i " .
the gluon contribution " is @xmath56\end{aligned}\ ] ] @xmath57{\mathcal{r}}\end{aligned}\ ] ] with two gluon - beam - referencing factors @xmath58 the tildes denote the fact that these two factors first appear in the slightly more general five - angle br - s2sc functions in sec .
2 in which the @xmath59 energy of the leptonically decaying @xmath60 has not been integrated out , i.e. in which @xmath61 dependence has not been integrated out . the overall pole - factor @xmath62\ ] ]
depends on the @xmath59 center - of - momentum energy @xmath45 and @xmath63 , and includes the gluon color factor .
the signal " contribution is suppressed by @xmath64 } { [ { \overline{\gamma } } ( 0,0)+{\overline{\gamma } } ( 1,1)]}\ ] ] which is associated with the stage - one part of the sequential decay @xmath65 .
there is similarly a suppression factor @xmath66 , see ( 23 ) , for the cp - conjugate channel @xmath67 .
numerically , @xmath68 in both the standard model and in the case of an additional large @xmath69 chiral weak - transition moment . for @xmath70 , in terms of the particles external momenta ,
useful kinematic formulas are : @xmath71 and @xmath72 where @xmath73 with @xmath74 the top - quark mass , energy , and magnitude of 3-momentum in the @xmath75 frame . for the quark - production - decay sequence @xmath76
we also assume that the @xmath35 and @xmath47 @xmath37 amplitudes dominate in @xmath24 and @xmath48 decay . including the color factor , the four - angle distribution @xmath77 for @xmath50 or
quark contribution " is @xmath78\end{aligned}\ ] ] @xmath79{\mathcal{r}}\end{aligned}\ ] ] with two quark - beam - referencing factors @xmath80 in terms of the @xmath3 helicity amplitudes @xmath81 in the @xmath24 rest frame , the decay density matrix element @xmath82 where @xmath83 in terms of the @xmath22 and @xmath84-quark helicities .
the dominant polarized - partial - widths are @xmath85 and w - boson - lt - interference - widths are @xmath86 \nonumber \\ & \equiv & |a(0,-1/2)||a(-1,-1/2)|\cos \beta _ { l } \\ \gamma _ { \mathit{i}}(0,-1 ) & = & -\gamma _ { \mathit{i}}(-1,0 ) \equiv { im}[a(0,-1/2)a(-1,-1/2)^{\ast } ] \nonumber \\ & \equiv & -|a(0,-1/2)||a(-1,-1/2)|\sin \beta _ { l}\end{aligned}\ ] ] where @xmath87 . for the @xmath16-conjugate process , @xmath88 , in the @xmath89 rest frame
, the helicity amplitudes @xmath90 are defined by @xmath91 with @xmath92 , and @xmath93 \nonumber \\ & \equiv & |b(0,1/2)||b(1,1/2)|\cos \overline{\beta } _ { r } \\ \overline{\gamma } _ { i}(0,1 ) & = & - \overline{\gamma } _ { i}(1,0 ) \equiv { { i m } [ b(0,1/2)b(1,1/2)^{\ast } } ] \nonumber \\ & \equiv & -|b(0,1/2)||b(1,1/2)|\sin \overline{\beta } _ { r}\end{aligned}\ ] ] where @xmath94 . for the @xmath16-conjugate channel @xmath95 ,
the analogous four - angle br - s2sc function @xmath96 is a distribution versus @xmath51 ; @xmath97 , @xmath98 , @xmath99 , where the latter three angles are displayed in in figs . 2 and 3 of i " .
the gluon contribution is @xmath100\end{aligned}\ ] ] @xmath101\overline{{\mathcal{r}}}\end{aligned}\ ] ] where @xmath102 } { [ { \gamma } ( 0,0)+{\gamma } ( -1,-1)]}\ ] ] the quark contribution is @xmath103\end{aligned}\ ] ] @xmath104\overline{{\mathcal{r}}}\end{aligned}\ ] ] in general , the @xmath105 decay - structure of the above four - angle br - s2sc functions is exactly analogous to that of the quark - production three - angle non - beam - referenced s2sc functions in i " .
the significant difference is the additional dependence on @xmath63 due to the beam - referencing .
therefore , for gluon - production there are the two gluon - beam - referencing factors @xmath106 , @xmath107 of ( 4,5 ) , and for quark - production the two quark - beam - referencing factors @xmath108 , @xmath109 of ( 11,12 ) .
there is a common final - state interference structure in these br - s2sc functions for the charged - lepton plus jets reaction @xmath0 or @xmath110 .
the final - state relative phase effects do not depend on whether the final @xmath111 system has been produced by gluon or by quark production . in these four - angle expressions , the signal contributions are again suppressed by the factor @xmath112 as a consequence of the dynamical assumption that the @xmath35 and @xmath47 @xmath37 amplitudes dominate . from the @xmath113 dependence of the integrated diagonal - elements of the sequential - decay density matrices
@xmath114 , @xmath115 for @xmath116 , @xmath114 and @xmath115 [ see ( 95 - 96 ) of i " ] , it follows that @xmath117 s numerator appears in @xmath118 multiplied by @xmath119 and that @xmath117 s denominator appears in @xmath120 multiplied by one .
the off - diagonal @xmath121 elements which describe @xmath48-helicity interference do not contribute due to the integration over the opening - angle @xmath122 between the @xmath24 and @xmath48 decay planes .
there are the analogous structures in the four - angle @xmath123 functions for the cp - conjugate tests
. for the production - decay sequence @xmath124 , section 2 of this paper contains the derivation in the @xmath23 frame of the gluon - production br - s2sc functions in the home " or starting coordinate system @xmath125 , which is defined by figs . 1 and 2 with the top - quark @xmath24 is moving along the positive @xmath126 direction .
2.1 lists the gluon - production @xmath127 density - matrix - elements in this coordinate system with the @xmath128 gluon - momentum beam " direction specified by the spherical angles @xmath129 . in sec .
2.2 , the gluon contributions to the general br - s2sc functions @xmath130 are obtained , and in sec .
2.3 applied to the lepton - plus - jets channel of the @xmath131 system , assuming that the @xmath132 and @xmath133 amplitudes dominate .
the above four - angle gluon - production br - s2sc function @xmath134 is obtained , along with addition - angle generalizations which might be useful empirically .
analogously , in section 3 , the quark - production @xmath127 density - matrix - elements for figs .
1 and 2 are listed with @xmath135 quark - momentum beam " direction specified by the spherical angles @xmath129 .
the analogous br - s2sc functions are obtained for the case of quark - production @xmath136 .
section 4 contains a summary and a discussion .
the appendices respectively contain ( a ) the dirac spinors and their outer - products @xmath137 in the jw phase convention including the c - p - t discrete symmetry properties of the spinors , ( b ) the derivation of the gluon - production density - matrix - elements in the jw phase convention for figs . 1 and 2 , and ( c ) the alternative @xmath138 , @xmath139 production density - matrix - elements for the alternative beam - reference system defined by fig .
3 , in which the incident parton beam " specifies the positive @xmath140 axis .
the general beam - referenced angular distribution in the @xmath141 is @xmath142 where the summations are over the @xmath24 and @xmath143 helicities .
the composite decay - density - matrices @xmath144 for @xmath145 and @xmath146 for @xmath147 are given in sec .
2.1 of i " .
this formula holds for any of the above @xmath148 production channels and for either the lepton - plus - jets , the dilepton - plus - jets , or the all - jets @xmath131 decay channels . in the @xmath75 frame , figs . 1 and 2 ,
the angles @xmath149 specify the direction of the incident beam , the @xmath128 or @xmath135 momentum . the @xmath24 momentum is chosen to lie along the positive @xmath126 axis .
the positive @xmath150 direction is an arbitrary , fixed perpendicular direction .
because the incident beam is assumed to be unpolarized , there is no dependence on the associated @xmath151 angle after the observable azimuthal - angles are specified ( see below ) . with respect to the normalization of the various br - s2sc functions ,
the @xmath151 integration is not explicitly performed in this paper . with ( 26 )
there is an associated differential counting rate @xmath152 where , for full phase space , the cosine of each polar - angle ranges from -1 to 1 , and each azimuthal - angle ranges over @xmath153 .
we use the helicity indices to label the successive terms in the gluon contributions to the sum in ( 26 ) .
each term is denoted @xmath154 with @xmath155 the signs of the @xmath24,@xmath48 helicities . unlike in i " , in this paper we do not use superscripts m " and m2 " to emphasize the mixed - helicity and helicity - flip contributions versus the non - superscripted helicity - conserving ones , since a quick glance at the patterns in these subscripted helicity indices provides these distinctions .
the charged lepton s azimuthal - angle @xmath26 , or @xmath156 is always referenced by the opposite @xmath6-boson momentum , so these charged - lepton azimuthal - angles are denoted without tilded accents " .
for the alternative @xmath157 and @xmath158 referencing by the opposite @xmath48 , @xmath24 momentum directions , see the discussion in 2nd and 3rd paragraphs of section 2 of i " .
the @xmath159 helicity - conserving gluon - production density - matrix - elements in the @xmath160 system are @xmath161 @xmath162 where the asterisk denotes complex - conjugation , and the common pole - factor @xmath163 is given in ( 6 ) in the introduction .
note the @xmath164 factor in ( 29 ) .
the mixed helicity - properties gluon - production density - matrix - elements are @xmath165 with an overall minus sign and @xmath166 factor .
the helicity - flip gluon - production density - matrix - elements are @xmath167)\ ] ] @xmath168)\ ] ] the helicity - conserving contribution is @xmath169 where @xmath170 , @xmath171 , @xmath172 , @xmath173 , are the diagonal elements , and @xmath174 , @xmath175 the off - diagonal elements of the sequential - decay density matrices @xmath14 and @xmath15 . as in sec .
2.2.1 of i " , in the @xmath176 and @xmath177 contributions in this paper , the starting angles @xmath178 and @xmath12 have been replaced by the angles @xmath179 and @xmath180 . the first part of the mixed helicity - properties contribution is @xmath181 @xmath182 @xmath183 @xmath184 the second part of the mixed helicity - properties contribution is @xmath185 @xmath186 @xmath187 @xmath188 the helicity - flip contribution is @xmath189 \\ i_{--,--}^{g } = \frac{4m^2}{s } \,\ , c(s,\theta _ { b } ) \,\ , { \mathbf{r}_{--}\overline{\mathbf{r}}_{-- } } [ 1 + \frac{4p^2}{s } ( 1 + \sin ^{4}\theta _ { b } ) ] \\ i_{++,--}^{g } = \frac{4 m^2}{s } \,\ , c(s,\theta _ { b } ) \,\ , e^{\imath \phi } { \mathbf{r}_{+- } } \overline{{\mathbf{r}}}_{+- } [ 1 - \frac{4p^2}{s } ( 1 + \sin ^{4}\theta _ { b } ) ] \\ i_{--,++}^{g } = \frac{4 m^2}{s } \,\ , c(s,\theta _ { b } ) \,\ , e^{-\imath \phi } { \mathbf{r}_{-+ } } \overline{{\mathbf{r}}}_{-+ } [ 1 - \frac{4p^2}{s } ( 1 + \sin ^{4}\theta _ { b } ) ] \end{aligned}\ ] ] from the perspective of specific @xmath131 decay channels and/or specific helicity amplitude tests , one can use the above results to investigate various br - s2sc functions .
for instance , in this paper , we are most interested in the lepton - plus - jets channel " and in tests for the relative sign of or for measurement of a possible non - trivial phase between the @xmath35 helicity amplitudes for @xmath192 .
we assume that the @xmath35 and @xmath47 @xmath37 contributions dominate . for the case
@xmath194 , with @xmath195 decaying into hadronic jets , we separate the intensity contributions into two parts : `` signal terms '' @xmath196 which depend on @xmath197 and @xmath198 , and `` background terms '' @xmath199 which depend on @xmath200 and @xmath201 . in this section , we use a tilde accent on @xmath202 to denote the integration over the @xmath203 , @xmath204 variables . by eqs(95 - 97 ) of i " , this integration directly projects out the various @xmath205 dependencies .
we find for the helicity - conserving contribution @xmath206 @xmath207 \\ + \gamma ( -1,-1)\sin ^{4}\frac{\theta _ { a}}{2}[\overline{\gamma } ( 0,0)(1-\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t})+\overline{\gamma } ( 1,1)(1+\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t } ) ] \nonumber \end{array } \right\}\end{aligned}\ ] ] @xmath208\end{aligned}\ ] ] @xmath209\nonumber\end{aligned}\ ] ] @xmath210 @xmath211 \nonumber\end{aligned}\ ] ] we collect the mixed - helicity contributions in real sums : @xmath212 @xmath213\nonumber\end{aligned}\ ] ] @xmath214 @xmath215\nonumber\end{aligned}\ ] ] @xmath216 @xmath217\nonumber\end{aligned}\ ] ] @xmath218 @xmath219\nonumber\end{aligned}\ ] ] the helicity - flip contributions are @xmath220\end{aligned}\ ] ] @xmath221
\\ + \gamma ( -1,-1)\sin ^{4}\frac{\theta _ { a}}{2}[\overline{\gamma } ( 0,0)(1+\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t})+\overline{\gamma } ( 1,1)(1-\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t } ) ] \nonumber \end{array } \right\}\end{aligned}\ ] ] @xmath222\end{aligned}\ ] ] @xmath223\nonumber\end{aligned}\ ] ] @xmath224\end{aligned}\ ] ] @xmath225\nonumber\end{aligned}\ ] ] @xmath226 \sin \theta _ { 2}^{t}\sin \theta _ { a}\end{aligned}\ ] ] @xmath227\nonumber\end{aligned}\ ] ] for the @xmath16-conjugate process @xmath228 , with @xmath229 decaying into hadronic jets , we similarly separate the contributions : `` signal terms '' @xmath230depending on @xmath231 and @xmath232 , and `` background terms '' @xmath233 depending on @xmath234 and @xmath235 . by integration over @xmath236 , @xmath237 , we find for the helicity - conserving contribution @xmath238 @xmath239 \\
+ \overline{\gamma } ( 1,1)\sin ^{4}\frac{\theta _ { b}}{2}[\gamma ( 0,0)(1-\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t})+\gamma ( -1,-1)(1+\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t } ) ] \end{array } \right\ } \nonumber\end{aligned}\ ] ] @xmath240\end{aligned}\ ] ] @xmath241\nonumber\end{aligned}\ ] ] @xmath242 @xmath243 \nonumber\end{aligned}\ ] ] the mixed - helicity contributions are @xmath244 @xmath245\nonumber\end{aligned}\ ] ] @xmath246 @xmath247\nonumber\end{aligned}\ ] ] @xmath248 @xmath249\nonumber\end{aligned}\ ] ] @xmath250 @xmath251\nonumber\end{aligned}\ ] ] the helicity - flip contributions are @xmath252\end{aligned}\ ] ] @xmath253 \\ + \overline{\gamma } ( 1,1)\sin ^{4 } \frac{\theta _ { b}}{2 } [ \gamma ( 0,0)(1+\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t})+\gamma ( -1,-1)(1-\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t } ) ] \end{array } \right\ } \nonumber\end{aligned}\ ] ] @xmath254\end{aligned}\ ] ] @xmath255 \nonumber\end{aligned}\ ] ] @xmath256\end{aligned}\ ] ] @xmath257\nonumber\end{aligned}\ ] ] @xmath258 \sin \theta _ { 1}^{t}\sin \theta _ { b}\end{aligned}\ ] ] @xmath259\nonumber\end{aligned}\ ] ] in summary , with beam - referencing , for the @xmath261 case there are six `` background terms '' depending on @xmath200 and @xmath262 , and also six `` signal terms '' depending on @xmath263 . as a consequence of lorentz invariance , there are associated kinematic factors with simple angular dependence which can be used to isolate and measure these four @xmath264s
. the following patterns , ( i ) and ( ii ) , for these br - s2sc contributions @xmath265 occur identically in gluon - production and in quark - production : \(i ) @xmath236 polar - angle dependence : the coefficients of @xmath266 vary relatively as the @xmath267-decay @xmath268-squared - intensity - ratios @xmath269 \big / \left\ { \frac{1}{\sqrt{2}}\sin \theta _ { a}\sin ^{2}\frac{\theta _ { a}}{2}\right\ } = \nonumber \\
2(1+\cos \theta _ { a } ) \big / \left [ 1-\cos \theta _ { a}\right ] \big / \left\ { \sqrt{2(1+\cos \theta _ { a})(1-\cos \theta _ { a})}=\sqrt{2}\sin \theta _ { a}\right\}\end{aligned}\ ] ] \(ii ) @xmath270 azimuthal - angle dependence in the `` signal terms '' : the coefficients of @xmath271 vary as @xmath272 in each of the signal terms , i.e. as in the four - angle gluon - signal term ( 3 ) there is a factor @xmath273
. however , in three terms there are also @xmath274 s with the opposite association of these @xmath275 , @xmath276 factors , i.e. in a factor @xmath277 .
this opposite association occurs in half the signal " contributions : the helicity conserving @xmath278 of ( 54 ) , the mixed - helicity @xmath279 of ( 56 ) , and in the helicity - flip @xmath280 of ( 62 ) , along with a different @xmath281 and/or @xmath282 dependence .
this different angular dependence might be used in an empirical separation of these terms from the terms with the normal @xmath283 association .
this different @xmath281 and/or @xmath282 dependence is the reason that only the @xmath284 factor appears in the signal terms in the four - angle and five - angle br - s2sc functions .
there are analogous cp - conjugate patterns in @xmath285 and @xmath286 for the cp - conjugate case @xmath287 .
there is only the factor @xmath288 in the four - angle and five - angle br - s2sc functions . in the cp - conjugate case
the patterns are also identical in gluon - production and in quark - production . to reduce the number of angles to obtain six - angle br - s2sc functions , we integrate out either the beam - referencing azimuthal - angle @xmath281 or the angle @xmath282 between the @xmath24 and @xmath48 decay planes .
we only list the non - vanishing contributions . to clearly label the successive contributions
, we again use the index @xmath290 .
we first consider the @xmath291 channel with the @xmath292 decaying hadronically .
for @xmath293 the non - vanishing six - angle contributions are proportional to the `` background and signal parts '' of the above expressions with @xmath294 for the helicity - conserving , @xmath295 with no @xmath282 dependence . for the helicity - flip @xmath296 with no @xmath282 dependence ; and @xmath297 with a @xmath298 dependence in the background part , and both @xmath298 and @xmath299 dependence in the signal part .
the mixed - helicity contributions vanish .
if instead the @xmath282 dependence is integrated out @xmath302 the non - vanishing six - angle contributions are also proportional to the above expressions : @xmath303 for the helicity - conserving , @xmath295 with no @xmath281 dependence . for the mixed - helicity , @xmath304 with a @xmath305 dependence in the background part , and both @xmath305 and @xmath306 dependence in the signal part . for the helicity - flip @xmath296 with no @xmath281 dependence . if both the @xmath281 and @xmath282 dependence are integrated out , the five - angle distribution @xmath308 is @xmath309 the terms in these expressions only arise from the helicity - conserving ( 51,52 ) , and from the helicity - flip ( 59 , 60 ) : @xmath310\ ] ] @xmath311\ ] ] we obtain @xmath312 \nonumber \\ & & \left\ { \widetilde{g}_{1}^{g}(s,\theta _ { b})+{\mathcal{r\,\,}}\widetilde{g}_{2}^{g}(s,\theta _ { b})\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t}\right\}\end{aligned}\ ] ] @xmath313\,{\mathcal{r\,\,}}\widetilde{g}_{2}^{g}(s,\theta _ { b})\sin \theta _ { 1}^{t}\cos \theta _ { 2}^{t}\end{aligned}\ ] ] where the two gluon - beam - referencing factors @xmath314 are listed in ( 4 , 5 )
. note that versus the four - angle distributions listed in the introduction , in this five - angle distribution @xmath315 the gluon - beam - referencing factor @xmath316 appears in both the background and the signal contributions . for the quark - production contribution to the five - angle distribution @xmath317 of ( 111,112 ) , the analogous situation occurs in the quark - beam - referencing factors @xmath318 versus the four - angle distribution ( 9 , 10 ) .
likewise for the respective cp - conjugate spin - correlation functions .
if the @xmath320 dependence is integrated out , there is a four - angle distribution in @xmath321 with the same @xmath322 " values as in ( 82,83 ) , @xmath323 which is listed in the introduction section in ( 2,3 ) . for the @xmath16-conjugate case in terms of @xmath324 , @xmath97 , @xmath203 , @xmath325 ,
the successively fewer angle distributions similarly follow .
the non - vanishing terms are proportional to the preceding `` background and signal '' expressions ( 63 - 74 ) .
the six - angle expression is @xmath326 for the helicity - conserving , @xmath295 with no @xmath282 dependence . for the helicity - flip , @xmath296 with no @xmath282 dependence ; and @xmath297 with a @xmath298 dependence in the background part , and both @xmath298 and @xmath299 dependence in the signal part .
if instead the @xmath282 dependence is integrated out , there is the six - angle expression @xmath327 for the helicity - conserving , @xmath295 with no @xmath281 dependence . for the mixed - helicity , @xmath304 with a @xmath305 dependence in the background part , and both @xmath305 and @xmath306 dependence in the signal part . for the helicity - flip ,
@xmath296 with no @xmath281 dependence .
if both the @xmath281 and @xmath282 dependence are integrated out , the five - angle distribution @xmath328 is @xmath329 and the terms in these expressions only arise from the helicity - conserving ( 63,64 ) , and from the helicity - flip ( 71,72 ) : @xmath330\ ] ] @xmath331\ ] ] we obtain @xmath332 \nonumber \\ & & \left\ { \widetilde{g}_{1}^{g}(s,\theta _ { b})+\overline{{\mathcal{r}}}{\mathcal{\,\,}}\widetilde{g}_{2}^{g}(s,\theta _ { b})\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t}\right\}\end{aligned}\ ] ] @xmath333\,\overline{{\mathcal{r}}}{\mathcal{\,\,}}\widetilde{g}_{2}^{g}(s,\theta _ { b})\cos \theta _ { 1}^{t}\sin \theta _ { 2}^{t}\end{aligned}\ ] ] where @xmath334 are given in ( 4 , 5 ) . finally ,
if the @xmath335 dependence is integrated out , we obtain the four - angle distribution in @xmath336@xmath337 which is listed in the introduction section in ( 21,22 ) .
unlike the treatment of the quark - production contribution in i " , we do not integrate out the @xmath338 dependence in this paper , as was discussed in the introduction . for comparison with the analogous gluon - density - matrix elements listed in the previous section , including the color factor , the @xmath159 helicity - conserving quark - production density - matrix - elements in the @xmath160 system are @xmath339 @xmath340 with an @xmath341 factor in ( 96 ) .
these and the following density - matrix - elements agree with ( 56 ) of i " . for @xmath342 , @xmath343 , @xmath344 where @xmath345 .
the mixed - helicity properties quark - production density - matrix - elements are @xmath346 with an overall minus sign and @xmath166 factor .
the helicity - flip quark - production density - matrix - elements are @xmath347 these give the quark contributions to the br - s2sc functions listed in section 2 of i " , when each expression in i " is multiplied by the color factor @xmath348 . with the @xmath350 labelling of the present paper , the quark contributions are proportional to the ones in ( 98 - 109 ) of `` i '' as follows . for the helicity - conserving contribution , @xmath351 @xmath352 @xmath353 @xmath354 for the mixed - helicity contribution ,
@xmath355 @xmath356 @xmath357 @xmath358 the helicity - flip contributions are @xmath359 @xmath360 @xmath361 @xmath362 by this @xmath350 labelling , the above general formulas and remarks ( 77 - 83 ) for the successively fewer - angle gluon - distributions apply for the quark - distributions by simply changing the superscript @xmath363 .
thereby , one obtains the six - angle @xmath364 , @xmath365 , and the five - angle @xmath366 .
explicitly , in terms of @xmath367 , the five - angle distribution is @xmath368 \nonumber \\ & & \left\ { \widetilde{g}_{1}^{q}(s,\theta _ { b})+{\mathcal{r\,\,}}\widetilde{g}_{2}^{q}(s,\theta _ { b})\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t}\right\}\end{aligned}\ ] ] @xmath369\,{\mathcal{r\,\,}}\widetilde{g}_{2}^{q}(s,\theta _ { b})\sin \theta _ { 1}^{t}\cos \theta _ { 2}^{t}\end{aligned}\ ] ] where the two quark - beam - referencing factors @xmath318 are listed in ( 11 , 12 ) .
the simpler four - angle distribution @xmath370 is given in the introduction in ( 9 , 10 ) . for this cp - conjugate case ,
the quark contributions are proportional to ( 113 - 124 ) of `` i '' . for the helicity - conserving contribution , @xmath372 @xmath373 @xmath374 @xmath375 for the mixed - helicity contribution , @xmath376 @xmath377 @xmath378 @xmath379 for the helicity - flip contribution , @xmath380 @xmath381 @xmath382 @xmath383 the above general formulas and remarks in ( 87 - 93 ) for the successively fewer - angle gluon - distributions apply for the quark - distributions by simply changing the superscript @xmath384 .
thereby , one obtains the six - angle @xmath385 , @xmath386 , and the five - angle @xmath387 . explicitly , in terms of @xmath388 , the five - angle distribution is @xmath389 \nonumber \\ & & \left\ { \widetilde{g}_{1}^{q}(s,\theta _ { b})+\overline{{\mathcal{r}}}{\mathcal{\,\,\,}}\widetilde{g}_{2}^{q}(s,\theta _ { b})\cos \theta _ { 1}^{t}\cos \theta _ { 2}^{t}\right\}\end{aligned}\ ] ] @xmath390\,\,\overline{{\mathcal{r}}}{\mathcal{\,\,\,}}\widetilde{g}_{2}^{q}(s,\theta _ { b})\cos \theta _ { 1}^{t}\sin \theta _ { 2}^{t}\end{aligned}\ ] ] the simpler four - angle distribution @xmath391 is given in the introduction in ( 24,25 ) .
from the top - quark beam - referenced spin - correlation function @xmath33 with both gluon ( 2,3 ) and quark ( 9,10 ) production contributions , the tests for @xmath392 decay are : \(i ) by measurement of @xmath393 , the relative sign of the two dominant @xmath394 helicity - amplitudes can be determined if their relative phase is @xmath395 or @xmath396 . versus the partial - decay - width @xmath397 , w - boson longitudinal - transverse interference is a large effect because in the standard model @xmath398 without / with a large @xmath399 chiral weak - transition - moment .
[ @xmath400 partial width for @xmath401 ] \(ii ) by measurement of both @xmath393 and @xmath402 via the @xmath26 dependence , limits can be set on a possible non - trivial phase @xmath403 , see ( 14 - 16 ) .
non - trivial relative phases can occur in top - quark decays if @xmath404 invariance is violated [ 7,1 ] .
such a violation will occur if either ( a ) there is a fundamental violation of canonical time - reversal invariance , and/or ( b ) there are absorptive final - state interactions .
explicit expressions for the @xmath405 helicity amplitudes in terms of the most general lorentz coupling @xmath406 where @xmath407 , are given in the nklm paper in [ 7 ] . the canonical decomposition of @xmath408 is @xmath409 @xmath410 where the parameters @xmath411 effective - mass scale of new physics associated with the @xmath322th type additional lorentz structure . " for a general treatment of additional lorentz structures to pure @xmath412 , the @xmath413 or @xmath414 must be considered as complex phenomenological parameters .
details , such as lorentz - structure - equivalence theorems @xmath415 , @xmath416 ; the matrix elements of the divergences of these couplings ; and the definitions of the chiral couplings @xmath417 are in the nklm , nc , na papers in [ 7 ] . for @xmath418 decay , from the gluon ( 21,22 ) and quark ( 24,25 )
production contributions to the cp - conjugate br - s2sc function for @xmath419 , there are two analogous tests for @xmath420 decay . by measurement of @xmath231 ,
the relative sign of the two dominant helicity amplitudes @xmath421 and @xmath422 for @xmath418 can be determined if their relative phase is @xmath395 or @xmath396 ( as in the case of large @xmath423 chiral weak - transition - moment ) . by measurement of both @xmath424 and @xmath232 via the @xmath156 dependence , limits can be set on a non - trivial phase @xmath94 , see ( 18 - 20 ) . in all these br - s2sc functions for top - quark decay tests , the polarized - partial - widths and w - boson - lt - interference - widths ( 15,16 ) appear multiplied by the @xmath425 , @xmath26 angular factors , which are expected in the helicity - formalism for the decay chain @xmath426 .
the spherical angles @xmath19 , @xmath20 specify the @xmath21 momentum in the @xmath22 rest frame when there is first a boost from the @xmath23 frame to the @xmath24 rest frame , and then a second boost from the @xmath24 rest frame to the @xmath22 rest frame .
so , @xmath427 and @xmath428 appear in the background terms @xmath429 , see ( 2 , 9 ) .
similarly , @xmath430 and @xmath431 appear in the signal terms @xmath432 , see ( 3 , 10 ) .
the situation is analogous for the @xmath28 , @xmath29 variables in the cp - conjugate br - s2sc functions .
the above summary is for the leading - order in qcd and the leading - order in electroweak interactions considered in this paper , assuming that the @xmath394 and @xmath433 helicity - amplitudes dominate in @xmath401 decay .
an important consequence of such dominance is that the w - boson longitudinal - transverse - interference effects and br - s2sc signatures treated in this paper are both large versus the non - dominant contributions and versus higher - order qcd and/or higher - order electroweak contributions .
however , for later measurement , there are the two @xmath434 non - dominant amplitudes in @xmath392 decay with their two moduli and two additional relative phases . for a clear and simple visual display " of these measurable phases ,
see @xmath435"-relative - phases in fig . 1 in the nc paper in [ 7 ] .
in this context , next - to - leading - order qcd , next - to - leading - order electroweak , and also w - boson and t - quark finite - width corrections require further theoretical investigation [ 8 ] .
other polarimetry techniques such as @xmath436 polarimetry [ 9 ] , and excellent understandings of detector - systematics and of reaction - backgrounds will be required for a complete measurement of the four @xmath401 decay amplitudes and of the four decay - amplitudes for the cp - conjugate mode .
one learns a number of things from this derivation of the production density matrices and these associated br - s2sc functions .
first , the non - trivial overall minus signs and @xmath437 , @xmath438 factors appear in the same patterns in the gluon and quark production density matrices .
this has the important and empirically useful consequence that most relative phase effects in these br - s2sc functions do not depend upon whether the final @xmath439 system has been produced by gluon or by quark production .
therefore , as discussed in sec .
1.1 , there is a common final - state interference structure of the four - angle br - s2sc functions , and likewise for the other additional - angle generalizations which are listed in sections 2 and 3 .
this is a common interference structure for w - boson longitudinal - transverse - interference .
when the @xmath122-dependence is included , there is also a common interference - structure for @xmath24-quark left - right helicity interference and for @xmath143-antiquark left - right helicity interference .
this left - right spin-@xmath440 interference is in the gluon contribution , see ( 61 - 2 ) , and in the quark contribution , see ( 109 - 110 ) in this paper and in ( 108 - 9 ) in i " .
second , for the @xmath3 decay mode , one learns that for a spin - correlation measurement using the lepton + jets decay channel " to determine the relative sign of or to obtain a constraint on a possible non - trivial phase between the two dominant @xmath2 helicity amplitudes requires use of the @xmath5 energy of the hadronically decaying w - boson , or the kinematically equivalent cosine of the polar - angle of @xmath6 emission in the anti - top ( top ) decay frame .
both this @xmath441 factor and the @xmath442 suppression factor appear in all the four - angle and five - angle signal terms , but not in any of the background terms .
for this application of w - boson longitudinal - transverse - interference , it is fortunate that in the standard model the probabilities for the presence of longitudinal / transverse w - bosons are both large . because of this @xmath117 factor ,
it is also indeed fortunate that these probabilities are unequal , @xmath443 and @xmath444 irrespective of whether there exists a large @xmath69 chiral weak - transition moment .
* acknowledgments * this work was partially supported by u.s .
dept . of energy contract
de - fg 02 - 86er40291 .
in the jw phase - convention , for application to @xmath445 pair production with respective helicities @xmath446 and @xmath447 , the first particle @xmath448 spinor outer - products @xmath449 [ 10 - 12 ] are @xmath450 with @xmath451 with @xmath452 and where the asterisk denotes complex conjugation .
the signs of @xmath446 and @xmath447 are used in labelling the spinors . note that @xmath453 and @xmath454 . for completeness , if the first particle were an anti - particle , then @xmath455 the second anti - particle @xmath456 spinor outer - products @xmath457 are @xmath458 with @xmath459 so @xmath460 and @xmath461 . in the @xmath462 system , where the @xmath463 is back - to - back with the @xmath464 , @xmath465 . in this system ,
@xmath466 and @xmath467 are identical but in calculating , it is sometimes useful to keep them distinct . for completeness ,
if @xmath468 were a particle then @xmath469 these outer - products follow from the following spinors @xmath470 , @xmath471 , @xmath472 constructed following the procedure of [ 12 ] .
the sub - label @xmath122 on these spinors denotes the non - zero value of the third - euler angle in the wigner @xmath473-function , see discussion below .
this sub - label is normally suppressed because it is apparent from the context .
for the first particle @xmath474 , @xmath475 = \sqrt{e+m}\left [ \begin{array}{c } \cos \frac{\theta } { 2 } \\
e^{i\phi } \sin \frac{\theta } { 2 } \\ \frac{p}{e+m}\left
[ \begin{array}{c } \cos \frac{\theta } { 2 } \\
e^{i\phi } \sin \frac{\theta } { 2 } \end{array } \right ] \end{array } \right ] \nonumber\end{aligned}\ ] ] @xmath476 = \sqrt{e+m}\left [ \begin{array}{c } -e^{-i\phi } \sin \frac{\theta } { 2 } \\ \cos \frac{\theta } { 2 } \\ \frac{p}{e+m}\left [ \begin{array}{c } e^{-i\phi } \sin \frac{\theta } { 2 } \\ -\cos \frac{\theta } { 2 } \end{array } \right ] \end{array } \right]\end{aligned}\ ] ] @xmath477 = -\sqrt{e+m}\left [ \begin{array}{c } \frac{p}{e+m}\left [ \begin{array}{c } -\sin \frac{\theta } { 2 } \\ e^{i\phi } \cos \frac{\theta } { 2 } \end{array } \right ] \\
\sin \frac{\theta } { 2 } \\
-e^{i\phi } \cos \frac{\theta } { 2 } \end{array } \right]\end{aligned}\ ] ] @xmath478 = -\sqrt{e+m}\left [ \begin{array}{c } \frac{p}{e+m}\left [ \begin{array}{c } e^{-i\phi } \cos \frac{\theta } { 2 } \\ \sin \frac{\theta } { 2 } \end{array } \right ]
\\ e^{-i\phi } \cos \frac{\theta } { 2 } \\
\sin \frac{\theta } { 2 } \end{array } \right ] \hskip 5 mm { ( a.7 ) } \nonumber\end{aligned}\ ] ] for the second particle specified by @xmath468 of ( a.5 ) , the spinors are @xmath479 = \sqrt{e+m}\left [ \begin{array}{c } \cos \frac{\theta } { 2 } \\
e^{i\phi } \sin \frac{\theta } { 2 } \\ -\frac{l}{e+m}\left
[ \begin{array}{c } \cos \frac{\theta } { 2 } \\
e^{i\phi } \sin \frac{\theta } { 2 } \end{array } \right ] \end{array } \right]\end{aligned}\ ] ] @xmath480 = \sqrt{e+m}\left [ \begin{array}{c }
-e^{-i\phi } \sin \frac{\theta } { 2 } \\
\cos \frac{\theta } { 2 } \\ \frac{l}{e+m}\left [ \begin{array}{c }
-e^{-i\phi } \sin \frac{\theta } { 2 } \\ \cos \frac{\theta } { 2 } \end{array } \right ] \end{array } \right]\end{aligned}\ ] ] @xmath481 = -\sqrt{e+m}\left [ \begin{array}{c } \frac{l}{e+m}\left [ \begin{array}{c } \sin \frac{\theta } { 2 } \\
-e^{i\phi } \cos \frac{\theta } { 2 } \end{array } \right ] \\
\sin \frac{\theta } { 2 } \\
-e^{i\phi } \cos \frac{\theta } { 2 } \end{array } \right]\end{aligned}\ ] ] @xmath482 = -\sqrt{e+m}\left [ \begin{array}{c } -\frac{l}{e+m}\left [ \begin{array}{c } e^{-i\phi } \cos \frac{\theta } { 2 } \\ \sin \frac{\theta } { 2 } \end{array } \right ] \\
e^{-i\phi } \cos \frac{\theta } { 2 } \\
\sin \frac{\theta } { 2 } \end{array } \right ] \hskip 5 mm { ( a.8 ) } \nonumber\end{aligned}\ ] ] a useful alternative spinor - construction is to set the third - euler angle to zero .
when needed for clarity , we denote this alternate spinor - construction by the sub - label 0 " .
the formulas for these @xmath483 and @xmath484 spinors are proportional to those listed above .
explicitly , for @xmath448 , @xmath485 and for @xmath456 , @xmath486 consequently , the @xmath487 factors are absent in their associated outer - products , so in place of the above outer - products : @xmath488 @xmath472 .
similarly , in place of the spin - one particle @xmath489 polarization vectors in [ 11 ] , @xmath490 the @xmath491 parity and @xmath492 time - reversal discrete symmetry properties for the above spinors do not depend on the choice of the third euler angle . with @xmath493 and @xmath494@xmath495 where @xmath496 . in these and the following relations ,
the @xmath497 denotes the @xmath498 spinor which is listed in ( a.7 ) and the @xmath499 denotes the @xmath500 spinor which is listed in ( a.8 ) , etc . .
the gamma matrices are in the dirac representation , which is sometimes named the standard " or
dirac - pauli " [ 12 ] representation . with @xmath501 @xmath322 @xmath502 @xmath503 @xmath504 where @xmath496 , and the asterisk denotes complex conjugation .
the @xmath505 charge - conjugation discrete symmetry properties do depend on the choice of the third euler angle . with @xmath506 @xmath322 @xmath507 , for the first particle @xmath508 spinors ( a.7 ) @xmath509 and for the second particle @xmath468 spinors ( a.8 )
@xmath510 the @xmath505 relations for the @xmath511and @xmath512 spinors are obtained by omitting the @xmath487 factors in these equations . in the gluon ( quark )
production density matrices used in the text , we follow the @xmath513 , @xmath514 choice which corresponds to the choice made in `` i '' to use a non - zero third - euler angle in describing @xmath439 pair production in the helicity - formalism . following the jacob - wick papers [ 5 ] , this non - zero choice of third - argument in the various wigner @xmath473-functions is more often chosen in the literature on applications of the helicity - formalism .
an important exception occurs in treating sequential decay chains .
for instance , in i " for describing the sequential - decay matrix @xmath515 for @xmath516 in `` i '' , we do set the third - euler angle equal to zero . for specifying the second - stage axes orientation for @xmath517 versus the first - stage axes from @xmath3 where the @xmath518-momentum is in the @xmath473-functions ,
it is conceptually simpler not to make an unnecessary additional rotation about the @xmath60-axis in the first - stage @xmath473-functions , see sec .
ii of [ 13 ] . for @xmath519 decay , unlike in completely orienting a rigid body in classical mechanics , it is only necessary to correctly orient the @xmath520 direction of the @xmath60-momentum .
this only involves the first two euler angles . unlike for specifying the @xmath60-momentum , in the case of orienting a rigid body ,
a third euler angle rotation is required for one must also specify the amount of angular rotation of the non - cylindrically - symmetric rigid body about its @xmath521 axis .
obviously , some care is needed to insure that the choice of spinors corresponds to the choice of third argument used in the @xmath473-functions in the helicity - formalism .
this occasionally needs emphasis because it is common to use the @xmath483 and @xmath522 type spinors , and also common to use non - zero third - argument @xmath473-functions in the helicity - formalism .
if used simultaneously to describe the same elementary particle reaction , this would be an inconsistent treatment of relative - phase effects .
for both helicity choices , these spinors were checked by using them to calculate decay matrix elements and comparing their overall signs and @xmath523 , @xmath122 dependence with those of the helicity - formalism : using the spinors for the final neutrinos / anti - neutrinos , the first particle @xmath448 spinors were checked for @xmath524 , @xmath525 and the second particle @xmath456 spinors were checked for @xmath526 , @xmath527 .
they were checked for @xmath528 and @xmath529 .
the quark - production density - matrix elements for @xmath530 were calculated as in appendix b and agree with eq.(56 ) of i " .
for the gluon - production sequence @xmath531 , the density - matrix - elements are @xmath532 for the qcd lorentz - invariant amplitude @xmath533 to the leading - order in @xmath534 .
the initial gluon spin - averaging is over the massless @xmath535 and @xmath536 helicities @xmath537 , @xmath538 .
the @xmath539 and @xmath540 helicites in the amplitude are @xmath446 , @xmath541 . in the complex - conjugate amplitude ,
@xmath542 , @xmath543 are their helicities . for simplicity , the initial color - averaging , the final color - summation , and the associated color - indices are not displayed explicitly in ( b.1 ) . with the spinor outer - products of appendix
a , from ( b.1 ) by normal tracing techniques , the expressions given in the text follow for @xmath544 and analogously for @xmath545 , compare appendix a of [ 14 ] . similarly , for the @xmath546 beam coordinate system of fig .
3 , the alternative @xmath547 gluon production density - matrix - elements of appendix c follow . for either beam coordinate system , there are 5 distinct gluon - production density - matrix - elements and 5 distinct quark - production density - matrix - elements .
the other density - matrix - elements follow either by hermiticity or by the @xmath548 and @xmath549 discrete symmetries which are present in the outer - products and in the qcd feynman - rules .
as shown in fig .
3 , the alternative @xmath552 beam - referenced production density - matrices @xmath553 are for a beam coordinate system @xmath554 in the @xmath23 frame in which the @xmath128 gluon - momentum or @xmath135 quark - momentum beam " direction defines the positive @xmath140 axis .
the final @xmath24 momentum is specified by the spherical angles @xmath555 . in this appendix
, we denote these alternative density - matrix - elements by a tilde " .
we do not use these alternative density - matrix - elements in the derivation of the br - s2sc s because the azimuthal - angle @xmath139 is specified versus the initial - beam direction , whereas in @xmath556 derived in appendix b and used in the text , the azimuthal - angle @xmath12 is defined versus the final @xmath24 momentum direction in the @xmath75 . in using the helicity - formalism for br - s2sc functions for two - body pair production such as @xmath557 , the final @xmath24 momentum direction is the natural / most - convenient axis [ 14 ] for analyzing the two sequential - decay - chain - processes and , therefore , for simultaneously incorporating the beam - referencing . in the jw phase convention ,
the gluon - production density - matrix - elements are @xmath558 @xmath559 @xmath560 @xmath561 ) \hskip 25 mm { ( c.6 ) } \nonumber\end{aligned}\ ] ] @xmath562 ) \hskip 25 mm { ( c.7 ) } \nonumber\end{aligned}\ ] ] where @xmath563 \hskip 28 mm { ( c.8 ) } \nonumber\end{aligned}\ ] ] note that there is not an overall minus sign in ( c.5 ) versus ( 32 ) in the text .
these @xmath564 were derived using the @xmath565 spinor outer - products of appendix a , see remarks after ( a.2 ) .
these gluon - production density - matrix - elements ( c.1-c.7 ) agree in magnitude with @xmath566 constructed from eq(a2 ) of reference [ 15 ] .
the @xmath567 , @xmath568 factors are missing in ( c.9 ) and some overall signs differ between ( c.1-c.7 ) and ( c.9 ) . from the derivation of @xmath569 , one obtains a simple substitution rule " for obtaining these @xmath550 , @xmath570 density - matrix - elements from the @xmath571 , @xmath572 ones in the text : @xmath573 @xmath574 ; @xmath575 ; @xmath576 in @xmath577,@xmath578 " and @xmath579 , @xmath580 @xmath581 " , with complex - conjugation of this rule for @xmath582 " @xmath583 @xmath584 " cases @xmath585 this latter step in the substitution rule , which involves a minus sign , occurs due to the sine functions which arise from @xmath586 factors in the spinor outer - products .
this latter step needs to be performed twice in some of the helicity - conserving density - matrix - elements , and once in the mixed - helicity density - matrix - elements .
the quark - production density - matrix - elements are @xmath587 @xmath588 @xmath589 @xmath590 there is not an overall minus sign in ( c.12 ) versus ( 97 ) in the text .
the above substitution rule also yields these @xmath591 density - matrix - elements from the @xmath592 ones in the text .
nelson , e.g. barbagiovanni , j.j .
berger , e.k .
pueschel , j.r .
wickman , eur .
j. * c45 * , 121(2006 ) ; hep - ph/0506240 .
this paper is referred to as i " . in the present paper ,
g " and q " superscripts are used to distinguish the gluon and quark production contributions to physical quantities .
subscripts @xmath593 , with @xmath446@xmath447,@xmath542@xmath543 the signs of the @xmath24,@xmath48 helicities , are used to label @xmath594 production density - matrix - elements @xmath595 and their associated contributions to beam - referenced stage - two spin - correlation functions ( br - s2sc ) .
f. abe , et.al .
( cdf collaboration ) , phys .
* 74 * , 2626(1995 ) ; and s. abachi , et .
( d0 collaboration ) , _ ibid . _ * 74 * , 2632(1995 ) .
d. acosta , et.al .
( cdf - collaboration ) , phys . rev . *
d71 * , 031101(2005 ) ; v.m .
abazov , et.al .
( d0-collaboration ) , phys . lett . *
b617 * , 1(2005 ) ; f. fiedler , hep - ex/0506005 .
atlas technical proposal , cern / lhcc/94 - 43 , lhcc / p2 ( 1994 ) ; cms technical design report , cern - lhcc- 97 - 32 ; cms - tdr-3 ( 1997 ) . for @xmath596 to tape
, there will be @xmath597 events for @xmath598 , see f. gianotti , ggi inaugural conf .
, arcetri ( 2005 ) .
m. jacob , g. wick , ann .
( ny ) * 7 * , 209(1959 ) ; k .- c .
chou , jetp * 36 * , 909(1959 ) ; m.i .
shirokov , _ ibid . _ * 39 * , 633(1960 ) .
the jacob - wick phase convention includes the phase convention of rose for the @xmath599 functions ; m.e
. rose , _ elementary theory of angular momentum _ , ( wiley , new york , 1957 ) . in the standard model , for the @xmath401 decay mode
, the relative phase is @xmath25 between the dominant @xmath600 and @xmath601 helicity amplitudes .
however , as a consequence of lorentz - invariance , the four intensity - ratios , @xmath602 for the sm s ( @xmath603 ) coupling are identical to those which occur in the case of an additional chiral - tensorial - coupling of relative strength @xmath604 in @xmath605 units ; @xmath606 where @xmath607 , @xmath608 . in the case of such an additional large @xmath609 chiral weak - transition moment
, there is instead a @xmath610 relative phase between the @xmath600 and @xmath601 helicity amplitudes ; see c.a .
nelson , phys . rev . *
d65 * , 074033(2002 ) ; in _ physics at extreme energies _ , nguyen van hieu and jean tran thanh van ( eds ) ( gioi publishers , vietnam , 2001 ) , p. 369 ; in _ results and perspectives in particle physics _ , m. greco ( ed ) ( infn , frascatti , roma , italy , 2003 ) , p. 319
; hep - ph/0411072 and references therein .
nelson , phys . rev . *
d41 * , 2805(1990 ) ; in _ results and perspectives in particle physics _ , m. greco ( ed ) ( editions frontierses , gif - sur - yvette , france , 1994 ) , p. 259 . ; c.a .
nelson , b.t .
kress , m. lopes , t.p .
mccauley , phys .
rev . * d56 * , 5928(1997 ) ; _ ibid . _ * d57 * , 5923(1998 ) ; c.a .
nelson , a.m. cohen , eur .
j. * c8 * , 393(1999 ) ; and c.a .
nelson , l.j .
adler , eur .
j. * c17 * , 399(2000 ) .
next - to - leading - order qcd corrections : w. bernreuther , a. brandenburg , z.g .
si , p. uwer , nucl . phys . * 690 * , 81(2004 ) ; hep - ph/0410197 ; w. bernreuther , m. fuecker , z.g .
si , hep - ph/0508091 ; and references therein .
electroweak corrections : j.h .
kuhn , a. scharf , p. uwer , hep - ph/0508092 .
finite decay - width effects : n. kauer , d. zeppenfeld , phys . rev . *
d65 * , 014021(2002 ) ; h.s .
do , s. groote , j.g .
korner , m.c .
mauser , _ ibid . _ * d67 * , 091501(2003 ) ; and n. kauer , _ ibid . _ * d67 * , 054031(2003 ) .
nelson , eur .
j. * c19 * , 323(2001 ) .
v. barger , j. ohnemus , r.j.n .
phillips , phys . rev . * d35 * , 166(1987 ) .
auvil , j.j .
brehm , phys . rev . * 145 * , 1152(1966 ) .
sakurai , _ advanced quantum mechanics _
, p. 194 - 199
( addison - wesley pub .
co. , reading , ma ; 1967 ) .
nelson , et.al .
, phys . rev .
* d50 * , 4544(1994 ) .
nelson , phys . rev . *
d43 * , 1465(1991 ) .
kane , g.a .
landinsky , c.p .
yuan , phys .
rev . * d45 * , 124(1992 ) .
1 : the derivation of the general beam referenced stage - two - spin - correlation " function begins in the home " or starting coordinate system @xmath125 in the @xmath611 frame .
the top - quark @xmath24 is moving in the positive @xmath126 direction , and @xmath612 specify the @xmath22 momentum direction . the @xmath128 gluon - momentum or @xmath135 quark - momentum beam
" direction is specified by the spherical angles @xmath129 .
note that @xmath613 .
2 : supplement to previous figure to show @xmath614 which specify the @xmath27 momentum direction .
the azimuthal - angles @xmath151 and @xmath178 are lorentz invariant under boosts along the @xmath615 axis .
note that the sum @xmath616 is the angle between the @xmath24 and @xmath48 decay planes , and that the @xmath48 momentum is in the negative @xmath126 direction .
the three angles @xmath617 , @xmath618 and @xmath619 describe the first stage in the sequential - decays of the @xmath620 system in which @xmath621 and @xmath622 . the angles @xmath623 , @xmath618 are defined respectively in the @xmath24 , @xmath143 rest frames , see [ 1 ] .
3 : for derivation of the alternative @xmath552 beam - referenced production density - matrix - elements , a supplement to fig . 1 to show the @xmath138 and @xmath139 angles in the beam coordinate system @xmath624 in the @xmath625 frame . the @xmath128 gluon - momentum or @xmath135 quark - momentum beam "
direction is in the positive @xmath140 direction .
the final @xmath24 momentum is specified by the spherical angles @xmath626 , with the @xmath48 momentum , not shown , back to back with the @xmath24 . | this continuation of the derivation of general beam - referenced stage - two spin - correlation functions is for the analysis of top - antitop pair - production at the tevatron and at the large hadron collider .
both the gluon - production and the quark - production contributions are included for the charged - lepton - plus - jets reaction @xmath0 or @xmath1 .
there is a simple 4-angle beam - referenced spin - correlation function for determination of the relative sign of or for measurement of a possible non - trivial phase between the two dominant @xmath2 helicity amplitudes for the @xmath3 decay mode .
there is an analogous function and tests for @xmath4 decay .
this signature requires use of the @xmath5 energy of the hadronically decaying w - boson , or the kinematically equivalent cosine of the polar - angle of @xmath6 emission in the anti - top ( top ) decay frame .
spinors and their outer - products are constructed so that the helicity - amplitude phase convention of jacob & wick can be used throughout for the fixing of the signs associated with this large w - boson longitudinal - transverse interference effect .
= -5 mm * use of w - boson longitudinal - transverse interference in top quark spin - correlation functions : ii * + charles a. nelson , jeffrey j. berger , and joshua r. wickman + _ department of physics , state university of new york at binghamton + binghamton , n.y . 13902 _ + |
observations have it that magnetic fields of order @xmath3 g are present in clusters of galaxies ( see , for example , kronberg 1994 ; zweibel & heiles 1997 ; carilli & taylor 2002 ; widrow 2002 ; govoni & feretti 2004 ; vogt & enlin 2005 ; ferrari et al . 2008 ; ryu et al .
2012 ) . no physical effect ,
e.g. a battery effect , seems to produce such strong magnetic fields directly .
therefore , it is natural to assume that weak seed magnetic fields were produced at some time and they have been amplified later to current strengths . according to this point of view ,
the origin of cosmic magnetism now involves two separate issues : generation of the initial seed magnetic fields and their amplification . in this paper
, we try to find constraints on the seed magnetic fields by studying their amplification .
theoretically , it is possible to generate weak magnetic fields in the early universe before the recombination ( see widrow et al . 2012 and references therein ) .
however , it is highly unclear whether or not they have survived in the early universe and have become the seeds for present - day magnetic fields in the intracluster medium ( icm ) .
probably , one of the best - known mechanisms for producing cosmic seed magnetic fields is the biermann battery effect ( biermann 1950 ) , which should operate during the formation of the large - scale structure and can produce a seed field of order @xmath4 g .
if seed magnetic fields were produced either in the early universe or during the large - scale structure formation , it is likely that the seed fields were spatially homogeneous on the scale of galaxy clusters .
magnetic fields expelled from astrophysical bodies can also seed the icm .
first stars ( pudritz & silk 1989 ; schleicher et al .
2010 ) or active galaxies ( hoyle 1969 ; rees 1987 ; daly & loeb 1990 ; kronberg et al .
2001 ) can inject magnetized materials into the intergalactic space .
galactic winds , supernova explosions , ram pressure stripping can also provide magnetized materials to the intergalactic space ( see , for example , rephaeli 1988 ; kronberg 1994 ; carilli & taylor 2002 ; donnert et al .
2009 ; arieli et al .
if a seed magnetic field was provided by an astrophysical object , it is likely that the seed magnetic field was spatially localized at the time of injection . as time goes on ,
the seed magnetic field can be dispersed away from the source and ultimately fill the whole system more or less homogeneously .
if the homogenization timescale is greater than the hubble time , it is possible to distinguish the primordial and the astrophysical origins of cosmic magnetism by observing intermittency of magnetic field distribution .
turbulence is commonly observed in clusters and filaments in simulations of the large - scale structure formation ( kulsrud et al .
1997 ; ryu et al .
2008 ; xu et al . 2010 ; vazza et al . 2011 ; miniati 2014 ) .
turbulence can provide an efficient mechanism for amplification of weak seed magnetic fields . in usual astrophysical environments ,
the magnetic diffusivity is very small , which enforces fluid elements and magnetic field lines move together .
therefore , chaotic turbulent motions can stretch magnetic field lines , making the magnetic field s
strength and energy increase ( batchelor 1950 ; zeldovich et al . 1984 ; childress & gilbert 1995 ) .
this process is called small - scale turbulence dynamo . study of small - scale turbulence dynamo has a long history ( e.g. batchelor 1950 ; kazantsev 1968 ; vainstein & ruzmaikin 1972 ; pouquet et al . 1976 ; meneguzzi et al . 1981 ; kulsrud & anderson 1992 ; cho & vishniac 2000 ; haugen et al . 2004
; brandenburg & subramanian 2005 ; schekochihin & cowley 2007 ; cho et al 2009 ; beresnyak 2012 ; brandenburg et al . 2012 ; schober et al . 2012a ; bovino et al . 2013 ; yoo & cho 2014 ) .
most of the studies on the turbulence dynamo deal with growth of a uniform or a homogeneous seed magnetic field in a fluid with unit magnetic prandtl number @xmath5 , which is defined as @xmath6 , where @xmath7 is the viscosity and @xmath8 is the magnetic diffusivity . nevertheless , there are also studies for growth of localized seed magnetic fields ( molchanov et al .
1985 ; ruzmaikin et al .
1989 ; cho & yoo 2012 ; cho 2013 ) in unit magnetic prandtl number turbulence and for growth of a uniform seed magnetic field in high magnetic prandtl number turbulence ( schekochihin et al . 2004
; subramanian et al .
2006 ; schekochihin & cowley 2007 ; bovino et al .
2013 ) . in this paper
, we consider growth of both uniform and localized seed magnetic fields in high magnetic prandtl number turbulence .
we will argue that , when we adopt turbulence dynamo models for high magnetic prandtl number fluids , astrophysical seed magnetic fields are more likely sources of magnetic fields in the icm . in section 2 ,
we discuss turbulence dynamo in the icm .
in section 3 , we describe numerical methods and , in section 4 , we present the results . in sections 5 and 6 ,
we give discussions and summary , respectively .
if the hydrodynamic reynolds number , @xmath9 ( @xmath10 , where @xmath11 is the outer scale of turbulence and @xmath12 is the r.m.s . velocity ) is very large , turbulence dynamo is so efficient that any seed magnetic fields follow virtually the same evolutionary track after a very short initial exponential growth stage ( see appendix ) .
therefore , in this case , it is difficult to find a constraint on the strength of the initial seed magnetic field .
however , it is believed that @xmath9 in the icm does not exceed @xmath0 . in this case
, turbulence dynamo may not be so efficient and the strengths of the seed field may matter . in this section ,
we discuss turbulence dynamo in fluids with @xmath9 @xmath13 .
note that , because the magnetic diffusivity is expected to be still very small in the icm , the magnetic prandtl number , @xmath5 , can be much larger than unity ( i.e. @xmath14 ) .
due to low density and high temperature , the mean free path between collisions , @xmath15 , is very large and , therefore , the viscosity , which is approximately @xmath15 times the particle velocity , may not be negligibly small in the icm .
estimates for @xmath9 in cool cores and hot icm , do not exceed @xmath0 ( robinson 2004 ; ruszkowski et al .
2004 ; reynolds et al . 2005
; brandenburg & subramanian 2005 ; schekochihin & cowley 2006 ; brunetti & lazarian 2007 ) if we use the spitzer ( 1962 ) formula for the viscosity . in fact , the hydrodynamic reynolds number for the central @xmath16400 kpc region of a cluster can be @xmath17 ( modified from equation ( 2 ) in brunetti & lazarian 2007 ) , where @xmath18 is the number density , @xmath19 is the boltzmann constant , @xmath20 is the temperature , and @xmath21 is the coulomb logarithm . in this paper , we mainly assume that @xmath22 in the icm .
if @xmath23 , the viscous - cutoff wavenumber @xmath24 , at which the velocity spectrum drops quickly due to viscous damping , is not much larger than the driving wavenumber . therefore , the eddy turnover time at the viscous - cutoff scale is not much different from the large - scale eddy turnover time @xmath25 , which makes turbulence dynamo inefficient . because of inefficient turbulent dynamo
, the strength of the initial seed magnetic field becomes important . in this section
we consider turbulence dynamo in fluids with high magnetic prandtl numbers and the magnetic reynolds number is much larger than the hydrodynamic reynolds number . ] . at the exponential growth stage ,
the magnetic field lines are stretched by turbulent motions at the viscous - cutoff scale , also known as the kolmogorov scale . since the viscous - cutoff scale is close to the outer scale of turbulence in high magnetic prandtl number fluids ,
the kinetic energy density at the viscous - cutoff scale is not much smaller than that of the outer scale .
therefore , the exponential growth stage lasts for a long time until the magnetic energy density becomes comparable to the kinetic energy density at the outer scale of turbulence . in other words ,
the system is in exponential growth stage most of the time in high magnetic prandtl number turbulence . during the exponential growth stage ,
the magnetic energy density follows @xmath26 where @xmath27 is the strength of the mean magnetic field and @xmath28 is proportional to the eddy turnover time at the viscous - cutoff scale : @xmath29 .
the duration of the exponential growth stage is @xmath30 where we use @xmath31 and @xmath32 . for a fixed @xmath9
, we get @xmath33 where @xmath34 and @xmath35 are constants that depend on @xmath9 .
it is obvious from this equation that a weaker seed magnetic field takes more time to reach saturation .
therefore , if a seed magnetic field is so weak that the exponential growth stage takes more than the hubble time , it can not be the origin of cosmic magnetic fields .
if @xmath36 , we get @xmath37 where @xmath38 is the equipartition magnetic field ( i.e. @xmath39 ) , which is of order @xmath40 g in typical clusters . in this case , a seed magnetic field whose strength is weaker than @xmath41 where @xmath42 is the age of the universe , can not explain present - day magnetic fields in the icm .
in a cluster with @xmath43 years and @xmath44 g , the critical strength would be @xmath45 note , however , that this estimate is highly uncertain because we ignored constants of order unity in many places while deriving equation ( [ eq : t1 ] ) . in actual icm ,
it is more complicated to obtain @xmath46 because @xmath9 can be greater than @xmath47 .
therefore , we need numerical simulations to get better estimates for @xmath46 .
we use a pseudospectral code to solve the incompressible magnetohydrodynamic ( mhd ) equations in a periodic box of size @xmath48 ( @xmath49 ) : @xmath50 where @xmath51 @xmath52 is random driving force , @xmath53 , @xmath54 is the velocity , and @xmath55 is the magnetic field divided by @xmath56 .
we use 22 forcing components in the wavenumber range @xmath57 , which means the driving scale @xmath11 is @xmath58 . in our simulations , @xmath591 before saturation .
therefore , the large - scale eddy turnover time , @xmath60 , is approximately @xmath162.5 before saturation .
other variables have their usual meaning .
we use @xmath61 grid points .
we use normal viscosity for the velocity dissipation term , @xmath62 , and hyperdiffusion for the magnetic dissipation term .
the power of hyperdiffusion is set to 3 , such that the magnetic dissipation term in equation ( [ beq ] ) is replaced with @xmath63 therefore , the magnetic prandtl number ( @xmath64 ) is larger than one .
since we use hyperdiffusion , dissipation of magnetic field is negligible for small wavenumbers and it abruptly increases near @xmath65 , where @xmath66=256 . at @xmath67 , turbulence is fully developed and either a uniform or a localized seed magnetic field gets `` switched on '' . for localized seed magnetic fields , we consider two shapes : tube - like and doughnut - like seed magnetic fields . for tube - shaped seed magnetic fields , we use the following expression for the magnetic field at t=0 : @xmath68 where @xmath69 , @xmath70 , and @xmath71 and @xmath72 are the distances measured from the tube in grid units .
the unit vector @xmath73 is parallel to the x - axis .
the maximum strength of the magnetic field at t=0 is @xmath74 . for doughnut - shaped seed magnetic fields , we use the following expression for the magnetic field at t=0 : @xmath75 where @xmath76 , @xmath77 , @xmath70 , and @xmath78 and @xmath72 are distances measured from the center of the numerical box in grid units .
the unit vector @xmath79 is perpendicular to @xmath80 and @xmath81 .
the maximum strength of the magnetic field at t=0 is @xmath74 .
since @xmath77 in equation ( [ eq : b_doughnut ] ) , the size of the magnetized region at t=0 is @xmath1616 in grid units , which is @xmath161/32 of the simulation box size .
therefore , in a cluster of size @xmath161 mpc , the size of the initially magnetized region corresponds to @xmath1630 kpc .
figure [ fig : unif1 ] shows the time evolution of @xmath82 and @xmath83 . in the simulations ( runs 512-r1-u s in table 1 ) , the viscosity is 0.015 and the driving scale is @xmath84 and the velocity before saturation is @xmath85 .
therefore , the hydrodynamic reynolds number @xmath9 based on the velocity before saturation is roughly 160 .
both the upper and the lower panels of figure [ fig : unif1 ] show the time evolution of the same quantities , but the scales for the vertical axes are different ; linear in the former and logarithmic in the latter .
initially only the uniform component of the magnetic field exists and , as times goes on , turbulent motions stretch magnetic field lines and make the magnetic field grow . as the magnetic energy density increases , the kinetic energy density of turbulence decreases . during the saturation stage ,
magnetic energy density is a bit higher than the kinetic energy density in all runs . in figure [
fig : unif1 ] we can clearly see that the exponential growth stage takes up most of the growth time ( see the lower panel ) .
it is possible that the exponential growth stage is followed by a linear growth stage ( see discussions in cho & yoo 2012 ) .
however , because our @xmath9 is too small and , therefore , we do not fully resolve the inertial range , the duration of the linear growth stage , if any , should be very short . in figure
[ fig : unif1 ] , we can see that a weaker seed magnetic field takes more time to reach the saturation stage .
when the maximum strength of the seed magnetic field is @xmath86 , which means that the maximum strength of the initial magnetic field is roughly @xmath86 times the equipartition magnetic field strength , the system reaches saturation within @xmath87 .
however , if the maximum strength of the seed magnetic field is @xmath88 , the system reaches saturation within @xmath89 .
consider a cluster with @xmath90400 kpc , @xmath59 400 km / s and equipartition magnetic field of @xmath91 g . in this case , the large - scale eddy turnover time is @xmath92 if the strength of the initial seed magnetic field is @xmath93 ng , which is @xmath88 times the equipartition magnetic field strength , the system reaches saturation @xmath1620 billion years after the big bang .
this simple argument implies that any primordial seed magnetic field weaker than @xmath94ng can not be the direct origin of cosmic magnetism .
note that the large - scale eddy turnover time in equation ( [ eq : eddytt ] ) is in rough agreement with the values in some numerical simulations of the large - scale structure formation ( see , for example , ryu et al .
2008 ) . as aforementioned
, we expect that @xmath23 in the icm ( see equation ( [ eq : reicm ] ) ) .
nevertheless , it could be possible that @xmath95 in some clusters .
therefore , it is worth investigating turbulence dynamo for a higher @xmath9 . in figure
[ fig : unif2 ] we present results of simulations for a higher @xmath9 . in the simulations
( runs 512-r2-u s in table 1 ) , the viscosity is 0.003 , the driving scale is @xmath84 , and the velocity before saturation is @xmath96 , so that @xmath9 based on the velocity before saturation is roughly 830 . as in figure
[ fig : unif1 ] , we can clearly see the exponential growth stage and , now , we can see that a short , but clearly visible , linear growth stage follows the exponential growth stage . as in figure
[ fig : unif1 ] , a weaker seed magnetic field takes more time to reach saturation .
but now the growth rate of magnetic field is higher because the eddy turnover time at the viscous - cutoff scale is shorter .
therefore , compared with runs in figure [ fig : unif1 ] , somewhat weaker seed magnetic fields can be the direct origin of magnetic field in the icm .
figure [ fig : unif2 ] shows that a seed magnetic field as weak as @xmath97 times the equipartition strength can reach saturation within @xmath98 . if we consider the same cluster as before , a seed magnetic field as weak as @xmath99ng can reach saturation @xmath100 billion years after the big bang . and @xmath83 , but their vertical scales are different : linear in the upper panel and logarithmic in the middle panel .
the lower panel shows the time evolution of the standard deviation of the magnetic field distribution , @xmath101 . in the lower panel , the two curves for @xmath102 and @xmath99 almost coincide . from runs 512-r1-tb@xmath103 and 512-r1-tb@xmath104 . ] and @xmath83 , but their vertical scales are different : linear in the upper panel and logarithmic in the middle panel .
the lower panel shows the time evolution of the standard deviation of the magnetic field distribution , @xmath101 . in the lower panel , the two curves for @xmath102 and @xmath99 almost coincide . from runs 512-r1-tb@xmath103 and 512-r1-tb@xmath104 . ] in this subsection , we investigate how localized seed magnetic fields grow in high magnetic prandtl number turbulence . in all simulations in this subsection
( runs 512-r1-t s and 512-r1-d s in table 1 ) , the viscosity is 0.015 and the driving scale is @xmath84 and the velocity before saturation is @xmath85 , so that @xmath9 based on the velocity before saturation is roughly 160 . if the seed magnetic field is localized in space , turbulent motions stretch and disperse magnetic field lines , which makes the magnetic field grow and , at the same time , the magnetized region expand in high magnetic prandtl number turbulence as in unit magnetic prandtl number turbulence .
figure [ fig : tubejpg ] shows that the homogenization of a tube - shaped seed magnetic field is very fast .
the figure shows that the whole numerical box becomes magnetized within @xmath105 ( see more quantitative results in the lower panel of figure [ fig : tubet ] ) . in our simulations ,
the size of the computational box is @xmath106 .
therefore , the observed rate of homogenization implies that the speed at which the magnetized region expands is of order @xmath12 in high magnetic prandtl number turbulence as in unit magnetic prandtl number turbulence .
figure [ fig : tubet ] shows the time evolution of @xmath82 , @xmath83 , and the standard deviation of the magnetic field distribution , which can be regarded as an approximate size of the magnetized region .
both the upper and the middle panels of figure [ fig : tubet ] show the time evolution of @xmath82 and @xmath83 , but the scales for the vertical axes are different .
initially only the localized seed magnetic field exists and , as time goes on , turbulent motions make the magnetic field grow . as the magnetic energy density increases
, the kinetic energy density of turbulence decreases .
the values of @xmath82 and @xmath83 at the saturation stage are very similar to those for uniform seed field cases ( compare figures [ fig : tubet ] and [ fig : unif1 ] ) . in the lower panel of figure [ fig : tubet ] , we plot the time evolution of the standard deviation , @xmath101 , of magnetic field distribution : @xmath107 where @xmath108=y and z. initially the standard deviation rises very quickly and the growth rate decreases after @xmath109 . within @xmath110 ,
the standard deviation almost reaches the value for homogeneous distribution .
the behavior of @xmath101 is not sensitive to the value of b@xmath111 : in fact , the curves for @xmath112 and @xmath113 virtually coincide .
since the seed magnetic field in run 512-r1-t@xmath112 ( dashed curve ) is stronger than that in run 512-r1-t@xmath113 ( solid curve ) , run 512-r1-t@xmath112 reaches saturation earlier .
note that run 512-r1-t@xmath113 reaches saturation within @xmath114 .
therefore , for a cluster of galaxies with @xmath115 years , any tube - like seed magnetic field weaker than @xmath116 times the equipartition strength can not produce a strong enough magnetic field within the hubble time .
for example , if we consider the same cluster as before ( @xmath1171 mpc , @xmath118400 kpc , @xmath119 km / s , @xmath120 , and the equipartition magnetic field strength of @xmath91 g ) , it is likely that any localized tube - shaped seed magnetic field that is weaker than @xmath161 ng can not be the origin of magnetic field in the icm . of course , if @xmath9 is larger , the minimum magnetic field strength will decrease . and @xmath83 .
the lower panel shows the time evolution of the standard deviation of the magnetic field distribution @xmath101 . in the lower panel , the two curves for @xmath102 and @xmath99 almost coincide . from runs 512-r1-db@xmath103 and 512-r1-db@xmath104 . ] and
the lower panel shows the time evolution of the standard deviation of the magnetic field distribution @xmath101 . in the lower panel , the two curves for @xmath102 and @xmath99 almost coincide . from runs 512-r1-db@xmath103 and 512-r1-db@xmath104 . ]
figure [ fig : bsprjpg ] shows that the homogenization time for a doughnut - shaped seed magnetic field is also very fast .
the figure shows that the magnetic field spreads out and fills the whole numerical box within @xmath105 .
actually , if we plot the time evolution of the standard deviation , @xmath101 , of magnetic field distribution , @xmath121 where @xmath122 , @xmath123x , y , and z , is defined as in equation ( [ eq : sigma_i ] ) , it reaches the value for homogeneous distribution after @xmath124 ( the lower panel of figure [ fig : bsprt ] ) . as in the tube - like seed magnetic field cases ,
the behavior of @xmath101 is not sensitive to the value of b@xmath111 : the curves for @xmath112 and @xmath113 almost coincide . as we discussed above , the observed rate of homogenization is consistent with the fact that the speed at which the magnetized region expands is of order @xmath12 .
if we compare homogenization of a tube - like seed magnetic field ( figure [ fig : tubejpg ] ) and a doughnut - like seed magnetic field ( figure [ fig : bsprjpg ] ) , the latter produces more intermittent magnetic field distribution during and at the end of the homogenization process .
figure [ fig : bsprt ] shows the time evolution of @xmath82 , @xmath83 , and @xmath101 .
the run 512-r1-d@xmath112 reaches saturation after @xmath125 and the run 512-r1-d@xmath113 reaches saturation after @xmath126 .
therefore , for a cluster of galaxies with @xmath115 years , any doughnut - like seed magnetic field weaker than @xmath127 times the equipartition strength can not produce a strong enough magnetic field within the hubble time .
for example , if we consider the cluster of galaxies mentioned earlier ( @xmath1171 mpc , @xmath118400 kpc , @xmath119 km / s , @xmath120 , and equipartition magnetic field strength of @xmath91 g ) , it is likely that any localized doughnut - shaped seed magnetic field whose maximum strength is weaker than @xmath1610ng can not be the origin of magnetic field in the icm . and @xmath83 . the lower panel shows the time evolution of the standard deviation of the magnetic field ( the dashed line ) and the passive scalar field ( the solid line ) distributions . from run 512-ref . ] and
the lower panel shows the time evolution of the standard deviation of the magnetic field ( the dashed line ) and the passive scalar field ( the solid line ) distributions . from run 512-ref . ]
our simulations show that the strength of a uniform seed magnetic field should be larger than @xmath1 g if @xmath128 in the icm is smaller than @xmath129 , the minimum strength of the seed magnetic field will be larger than @xmath1 g . if @xmath130 as in equation ( [ eq : reicm ] ) , we expect that the minimum strength will be at least @xmath131 g . ] .
note that observations of the cosmic microwave background ( cmb ) place an upper limit of @xmath2 g on the strength of the uniform component of a primordial seed magnetic field ( barrow , ferreira & silk 1997 ; see also widrow et al .
2012 ; durrer & neronov 2013 ) .
therefore , it is unlikely that primordial magnetic fields are the direct origin of present - day magnetic fields in the icm .
if the high magnetic prandtl number model is correct for fluids in the icm , it is also unlikely that a magnetic field generated by the biermann battery ( @xmath132 g ) or aperiodic turbulent fluctuations ( schlickheiser 2012 ; @xmath133 g ) are the direct source of magnetic fields in the icm .
however , it is still possible that pre - galactic magnetic fields generated by the battery effect or the aperiodic turbulent fluctuations became amplified in galaxies ( see kulsrud & zweibel 2008 and references therein ; see also discussions in beck & wielebinski 2013 ) or first stars ( sur et al . 2010 ; schleicher et al .
2010 ; schober et al .
2012b ; schleicher et al .
2013 ) and then ejected into the intergalactic space later .
such ejected magnetic fields can easily be the origin of cosmic magnetic fields even in high magnetic prandtl number turbulence .
all in all , our simulations favor astrophysical origin of the icm magnetic fields .
cho & yoo ( 2012 ) showed that expansion of the magnetized region is fast in unit magnetic prandtl number turbulence : homogenization of a doughnut - shaped seed magnetic field in unit magnetic prandtl number turbulence occurs within @xmath163(l / v ) .
the numerical setup in cho & yoo ( 2012 ) is almost identical to our current numerical setup for the doughnut - shaped seed magnetic fields , excepts the magnetic prandtl numbers ( i.e. the values of the viscosity ) .
our current simulations show that expansion of magnetized region is also fast in high prandtl number turbulence : homogenization of a tube - like seed field occurs within @xmath110 and homogenization of a doughnut - like seed field occurs within @xmath124 . in our current simulations ,
the large viscosity damps small - scale velocity .
nevertheless , we obtain fast homogenization in our current simulations .
why is it so ?
it seems that expansion of the magnetized region is governed by eddy motions at the outer scale of turbulence . inside an outer - scale eddy , magnetization of
the whole eddy happens roughly within one large - scale eddy turnover time if the size of the initially magnetized region is not much smaller than the viscous - cutoff scale .
after full magnetization of the outer - scale eddy , the magnetic field is transported over uncorrelated outer - scale eddies . because any outer - scale eddy which is partially magnetized can be fully magnetized within roughly one large - scale eddy turnover time , the rate of expansion of the magnetized region should be proportional to @xmath12 ( cho 2013 ) .
therefore , full magnetization of the system occurs within @xmath134 , which we actually observe in our simulations .
can we numerically confirm that the expansion speed of the magnetized region is indeed of order @xmath12 in high magnetic prandtl number turbulence ? to measure the expansion rate on scales larger than the outer - scale eddies , we perform a numerical simulation with a small - scale driving ( 512-ref in table 1 ) .
simulation setup for 512-ref is identical to that for 512-r1-db@xmath104 except the the driving scale : the driving wave - numbers for 512-ref are between 15.4 and 26
. the hydrodynamic reynolds number @xmath9 for 512-ref is an order of magnitude smaller than that for 512-r1-db@xmath104 because the driving scale for 512-ref is an order of magnitude smaller .
we plot the simulation results in figures [ fig : k20jpg ] and [ fig : k20 t ] .
figure [ fig : k20jpg ] shows the expansion of the magnetized region and figure [ fig : k20 t ] shows the time evolution of @xmath82 , @xmath83 , and the standard deviations .
if we compare the rates of expansion for @xmath135 and for @xmath136 in figure [ fig : k20jpg ] , the rates look similar , which implies that the expansion rate is linear in time .
indeed the lower panel of figure [ fig : k20 t ] shows that the standard deviation of magnetic field distribution ( the dashed curve ; see equation ( [ eq : sig3d ] ) for definition ) shows a linear growth after a certain time , before which its behavior is very similar to that of a passive scalar field ( the solid curve ) .
this behavior is very similar to that of a localized seed magnetic field in unit magnetic prandtl number turbulence ( see cho 2013 for numerical methods and detailed discussions ) .
if the initially magnetized region is smaller than the viscous - cutoff scale @xmath137 , the expansion speed of the magnetized region can be slower than @xmath138 .
let the size of the initially magnetized region be @xmath139 .
if @xmath140 , then it takes @xmath141 for the magnetic field to fill the ` host ' eddy at the viscous - cutoff scale . after filling the viscous - cutoff - scale eddy , it would take @xmath60 for the magnetic field to fill the whole outer - scale eddy .
therefore , in this case , magnetization of the outer - scale eddy takes @xmath142 the expansion rate of the magnetized region on scales larger than the outer scale of turbulence will be @xmath143 for a cluster with @xmath90400 kpc and @xmath14430 kpc , we will have @xmath145 if @xmath146 .
but , for this @xmath9 , slowdown of the expansion rate will be negligible ( see equation ( [ eq : slowdown ] ) ) .
slowdown of the expansion rate will be important when @xmath14713 in the cluster .
note , however , that , if the magnetized materials in jets or stripped gases provide seeding , the initially magnetized region can be up to hundreds - of - kpc - long and , therefore , @xmath9 can be smaller than this . in the ism of galaxies , strength of the magnetic fields ranges from @xmath148 g ( diffuse ism ) to @xmath149 mg ( molecular clouds )
when these gases are expelled from galaxies , magnetic fields in them will be attenuated due to expansion of the media .
this makes the strength of the localized seed magnetic fields weaker than the galactic values .
if a linear size of the medium expands by a factor of @xmath150 during the ejection , magnetic field becomes weaker by a factor of @xmath151 .
the magnetic attenuation factor @xmath151 is uncertain and may depend on the ejection mechanism .
if a galactic material is ejected via jets or explosions , the expansion factor can be very large .
for example , if a jet is launched from a sub - pc scale and ultimately expands to kpc - scales , the expansion factor can be larger than @xmath152 ( see related discussions in brandenburg & subramanian 2005 ; widrow et al .
2012 ) . however , if a magnetized gas is stripped from a galaxy , the expansion factor can be smaller than @xmath152 .
another factor that can affect the strength of the seed fields is existence of turbulence inside ejecta .
if turbulence exists inside the ejected media , turbulence dynamo can mitigate attenuation of the magnetic fields .
if we use the turbulence dynamo models , we can test whether or not the high magnetic prandtl number turbulence is a correct model for the icm or the intergalactic medium .
suppose that we can observe distribution of magnetic fields in filaments .
if distribution of magnetic fields in filaments is more or less homogeneous , it is likely that the high magnetic prandtl number turbulence model is incorrect .
if we find magnetic fields only near the astrophysical sources , we may be able to conclude that magnetic fields in filaments have astrophysical origins , rather than primordial .
the use of an incompressible code has the advantage that we can easily control the viscosity and the magnetic diffusivity . however
, it is not possible to study the effects of compression with the incompressible code . in galaxy clusters , shocks and mergers
can compress fluids and have a pronounced impact on the growth of magnetic field ( roettiger et al 1999 ; dolag et al .
1999 , 2002 ; iapichino & brggen 2012 ) . for example , a simulation by dolag et al .
( 2002 ) shows that the mean magnetic field within a cluster of comoving radius 1 mpc increases by a factor of @xmath153 from redshift z=0.8 to z=0 .
if this is a typical growth factor for compression amplification , the minimum strength of a uniform magnetic field for @xmath9=160 should be decreased by the same factor : a new minimum strength will be a few times @xmath154 g .
note , however , that @xmath9 in actual clusters is likely to be much smaller than @xmath129 ( see equation ( [ eq : reicm ] ) ) .
therefore , if we take into account both the compression amplification and a realistic @xmath9 , the minimum strength may not be much smaller than @xmath1 g .
it is likely that the central region of a galaxy cluster has formed at a much earlier stage . during these early stages ,
many of the physical properties , including the hydrodynamic reynolds number @xmath9 , could be different .
it is therefore conceivable that some magnetization has occurred already at earlier stages , which can alleviate the constraints found in this paper .
however , unless such central regions were substantially cooler and/or denser than present - day central regions , the effect may not be significant .
note that , even in present - day cool cores , @xmath9 is expected to be less than @xmath155 ( schekochihin & cowley 2006 ) .
in this paper , we have considered turbulence dynamo models in unit and high magnetic prandtl number fluids .
we have found the following results : 1 . in high magnetic prandtl number turbulence ( with a hydrodynamic reynolds number , @xmath9 , less than @xmath0 ) ,
turbulence dynamo is not so efficient that a primordial seed magnetic field weaker than a certain critical value can not be a direct origin of magnetic fields in the icm . for a cluster with driving scale of @xmath16400 kpc and turbulence velocity of @xmath16400 km / s ,
the critical strength is @xmath156 g for @xmath120 , which is very close to an upper limit of @xmath2 g placed by the cmb anisotropy observations .
the critical strengths for seed magnetic fields ejected from galaxies or first stars are about two orders of magnitude higher .
but , since the strengths of magnetic fields in those astrophysical objects are likely to be larger than @xmath3 g , it may not be so difficult for the seed magnetic fields ejected from the astrophysical bodies to have strengths larger than the critical values .
therefore , our calculations favor astrophysical origin of magnetic fields in the icm .
2 . if the high magnetic prandtl number model is correct for clusters and filaments , pre - galactic magnetic fields generated by a battery effect , amplified in galaxies or first stars , and ejected from them later into intergalactic space would be plausible sources of magnetic fields in the large - scale structure of the universe .
we thank the referee for useful suggestions and comments .
this work is supported by the national r & d program through the national research foundation of korea , funded by the ministry of education ( nrf-2013r1a1a2064475 ) .
this work is also supported by the supercomputing center / korea institute of science and technology information with supercomputing resources including technical supports ( ksc-2013-c1 - 034 ) .
99 arieli , y. , rephaeli , y. , & norman , m. l. 2011 , , 738 , 15 banerjee , r. , & jedamzik , k. 2004 , , 70 , 123003 barrow , j. , ferreira , p. , & silk , j. 1997 , phys .
lett . , 78 , 3616 batchelor , g. 1950 , proc .
london a , 201 , 405 beck , r. , & wielebinski , r. 2013 , planets , stars and stellar systems .
volume 5 : galactic structure and stellar populations , 641 beresnyak , a. 2012 , , 86 , 058201 biermann , l. 1950 , zeitschrift naturforschung teil a , 5 , 65 bovino , s. , schleicher , d. r. g. , & schober , j. 2013 , new journal of physics , 15 , 013055 brandenburg , a. , sokoloff , d. , & subramanian , k. 2012 , , 169 , 123 brandenburg , a. , & subramanian , k. 2005 , phys . reports , 417 , 1 brunetti , g. , & lazarian , a. 2007 , , 378 , 245 carilli , c. l. , & taylor , g. b. 2002 , , 40 , 319 childress , s. , & gilbert , a. 1995 , stretch , twist , fold : the fast dynamo ( berlin : springer ) cho , j. 2013 , , 87 , 043008 cho , j. , & vishniac , e. t. 2000 , , 538 , 217 cho , j. , vishniac , e. t. , beresnyak , a. , lazarian , a. , & ryu , d. 2009 , , 693 , 1449 cho , j. , & yoo , h. 2012 , , 759 , 91 daly , r. a. , & loeb , a. 1990 , , 364 , 451 donnert , j. , dolag , k. , lesch , h. , m'uller , e. 2009 , , 392 , 1008 dolag , k. , bartelmann , m. , & lesch , h. 1999 , , 348 , 351 dolag , k. , bartelmann , m. , & lesch , h. 2002 , , 387 , 383 durrer , r. , & neronov , a. 2013 , , 21 , 62 ferrari , c. , govoni , f. , schindler , s. , bykov , a. m. , & rephaeli , y. 2008 , , 134 , 93 haugen , n. e. , brandenburg , a. , & dobler , w. 2004 , , 70 , 016308 hoyle , f. 1969 , , 223 , 936 iapichino , l. , & brggen , m. 2012 , , 423 , 2781 govoni , f. , & feretti , l. 2004 , international journal of modern physics d , 13 , 1549 kazantsev , a. p. 1968
, soviet phys .- jetp lett . , 26 , 1031 kronberg , p. p.
1994 , reports on progress in physics , 57 , 325 kronberg , p. p. , dufton , q. w. , li , h. , & colgate , s. a. 2001 , , 560 , 178 kulsrud , r. m. , & anderson , s. w. 1992 , , 396 , 606 kulsrud , r. m. , cen , r. , ostriker , j. p. , & ryu , d. 1997 , , 480 , 481 kulsrud , r. m. , & zweibel , e. g. 2008 reports on progress in physics , 71 , 046901 meneguzzi , m. , frisch , u. , & pouquet , a. 1981 , physical review letters , 47 , 1060 miniati , f. 2014 , , 782 , 21 molchanov , s. , ruzmaikin , a. , & sokolov , d. 1985 , sov .
, 28 , 307 pouquet , a. , frisch , u. , & leorat , j. 1976 , journal of fluid mechanics , 77 , 321 pudritz , r. e. , & silk , j. 1989 , , 342 , 650 rephaeli , y. 1988 , comments on astrophysics , 12 , 265 rees , m. j. 1987 royal astronomical society , quarterly journal , 28 , 197 reynolds , c. , mckernan , b. , fabian , a. , stone , j. , & vernaleo , j. 2005 , , 357 , 242 robinson , k. , dursi , l. j. , ricker , p. m. , et al .
2004 , , 601 , 621 roettiger , k. , stone , j. m. , & burns , j. o. 1999 , , 518 , 594 ruszkowski , m. , brggen , m. , & begelman , m. 2004 , , 611 , 158 ruzmaikin , a. , sokoloff , d. , & shukurov , a. 1989 , , 241 , 1 ryu , d. , kang , h. , cho , j. , & das , s. 2008 , science , 320 , 909 ryu , d. , schleicher , d. r. g. , treumann , r. a. , tsagas , c. g. , & widrow , l. m. 2012 , , 166 , 1 schekochihin , a. a. , & cowley , s. c. 2006 , physics of plasmas , 13 , 056501 schekochihin , a. , & cowley , s. 2007 , in _ magnetohydrodynamics - historical evolution and trends _ , eds . by molokov , s. , moreau , r. , & moffatt , h. ( berlin ; springer ) , p. 85
( astro - ph/0507686 ) schekochihin , a. a. , cowley , s. c. , taylor , s. f. , maron , j. l. , & mcwilliams , j. c. 2004 , , 612 , 276 schleicher , d. r. g. , banerjee , r. , sur , s. , et al .
2010 , , 522 , a115 schleicher , d. r. g. , latif , m. , schober , j. , et al .
2013 , astronomische nachrichten , 334 , 531 schlickeiser , r. 2012 , physical review letters , 109 , 261101 schlter , a. , & biermann , i. 1950 , zeitschrift naturforschung teil a , 5 , 237 schober , j. , schleicher , d. , federrath , c. , klessen , r. , & banerjee , r. 2012a , , 85 , 026303 schober , j. , schleicher , d. , federrath , c. , et al . 2012b , , 754 , 99 spitzer , l. 1962 , physics of fully ionized gases ( new york : wiley ) subramanian , k. , shukurov , a. , & haugen , n. 2006 , , 366 , 1437 sur , s. , schleicher , d. r. g. , banerjee , r. , federrath , c. , & klessen , r. s. 2010 , , 721 , l134 vainshtein , s. i. , & ruzmaikin , a. a. 1972 , , 15 , 714 vazza , f. , brunetti , g. , gheller , c. , brunino , r. , & brggen , m. 2011 , , 529 , a17 vogt , c. , & enlin , t. a. 2005 , , 434 , 67 wagstaff , j. m. , banerjee , r. , schleicher , d. , & sigl , g. 2014 , , 89 , 103001 widrow , l. m. 2002 , reviews of modern physics , 74 , 775 widrow , l. m. , ryu , d. , schleicher , d. r. g. , et al .
2012 , , 166 , 37 xu , h. , li , h. , collins , d. c. , li , s. , & norman , m. l. 2010 , , 725 , 2152 yoo , h. , & cho , j. 2014 , , 780 , 99 zeldovich , ya . b. , ruzmaikin , a. a. , molchanov , s. a. , & sokoloff , d. d.1984 , j. fluid mech . , 144 , 1 zweibel , e. g. , & heiles , c. 1997 , , 385 , 131 lccccrrr 512-r1-ub@xmath157 & @xmath61 & @xmath16160 & - & @xmath86 & high & uniform & @xmath158 + 512-r1-ub@xmath159 & @xmath61 & @xmath16160 & - & @xmath99 & high & uniform & @xmath158 + 512-r1-ub@xmath160 & @xmath61 & @xmath16160 & - & @xmath88 & high & uniform & @xmath158 + 512-r2-ub@xmath159 & @xmath61 & @xmath16830 & - & @xmath99 & high & uniform & @xmath158 + 512-r2-ub@xmath160 & @xmath61 & @xmath16830 & - & @xmath88 & high & uniform & @xmath158 + 512-r2-ub@xmath161 & @xmath61 & @xmath16830 & - & @xmath97 & high & uniform & @xmath158 + 512-r1-tb@xmath103 & @xmath61 & @xmath16160 & @xmath86 & - & high & tube - shaped & @xmath158 + 512-r1-tb@xmath104 & @xmath61 & @xmath16160 & @xmath99 & - & high & tube - shaped & @xmath158 + 512-r1-db@xmath103 & @xmath61 & @xmath16160 & @xmath86 & - & high & doughnut - shaped & @xmath158 + 512-r1-db@xmath104 & @xmath61 & @xmath16160 & @xmath99 & - & high & doughnut - shaped & @xmath158 + 512-ref & @xmath61 & @xmath1610 & @xmath99 & - & high & doughnut - shaped & @xmath162 [ table_1 ]
in this appendix , we briefly review turbulence dynamo models in unit magnetic prandtl number turbulence and discuss their implications .
we assume that the fluid is incompressible and that both the viscosity and the magnetic diffusivity are very small . when the magnetic prandtl number is unity , the velocity and the magnetic dissipation scales coincide .
if a seed magnetic field has a primordial origin , the coherence length of the seed field can be very large and , therefore , it can be regarded as spatially uniform on the scale of galaxy clusters .
even if the primordial seed magnetic field has a small coherence length ( see , for example , banerjee & jedamzik 2004 ; wagstaff et al .
2014 ) , turbulence dynamo for the seed field will be similar to that for a uniform seed field , as long as the seed field is spatially homogeneous . on the other hand , if a seed magnetic field was provided by an astrophysical object , it is likely that the seed magnetic field was spatially localized at the time of injection . this topic has been studied extensively in literature . in summary , growth of a uniform weak seed magnetic field in unit magnetic prandtl number turbulence
follows the following three stages ( see schlter & bierman 1950 ; schekochihin & cowley 2007 ; cho et al .
2009 ; see also cho & vishniac 2000 ; schober 2012a ) .
( b ) when we consider two points whose separation is @xmath163 , the distance between the two point along the field line doubles after each eddy turnover time , @xmath164 , where @xmath165 is velocity at the velocity dissipation scale @xmath166 .
therefore , the distance grow exponentially , so does magnetic energy density .
( c ) since the characteristic scale in ( b ) is @xmath166 , the magnetic energy spectrum peaks near the dissipation wavenumber @xmath24 .
as time goes on , the magnetic spectrum goes up without changing the shape much .
( d ) in summary , the magnetic field grows in 3 steps : exponential growth , linear growth , and saturation . when the hydrodynamic reynolds number ( @xmath9 ) approaches infinity , the duration of the exponential growth stage becomes zero .
, scaledwidth=48.0% ] \(1 ) the stretching of the magnetic field lines occurs most actively near the velocity dissipation scale first .
this is because eddy turnover time is shortest , hence the rate of field - line stretching is highest , at the dissipation scale .
suppose that we introduce a weak uniform seed magnetic field in a turbulent medium ( see figure [ fig:1](a ) ) .
let us select two points on a magnetic field line whose separation is of order @xmath166 , where @xmath166 is the velocity dissipation scale ( figure [ fig:1](b ) ) .
after approximately one eddy turnover time , @xmath167 , where @xmath165 is the velocity at the dissipation scale , the distance between the two points along the field line will be doubled . after approximately another eddy turnover time @xmath28 , the distance between the two points along the field line will be doubled again .
therefore , magnetic field lines are stretched exponentially and , as a result , the magnetic energy density grows exponentially .
note that the typical scale of the magnetic field - line variation , which is of order @xmath166 , does not change during the process .
therefore , the magnetic energy spectrum peaks near @xmath168 and , as time goes on , it goes upward without changing its shape much ( figure [ fig:1](c ) ) .
\(2 ) as the magnetic energy spectrum moves upward , it will finally touch up the kinetic energy spectrum at the dissipation scale .
then what will happen ?
when the magnetic energy spectrum becomes comparable to the kinetic energy spectrum at the dissipation scale , magnetic back - reaction suppresses stretching of magnetic field lines at the dissipation scale . due to the suppression , the exponential growth stage ends .
note , however , that stretching at scales larger than @xmath166 is still efficient because the kinetic energy spectrum is still higher than the magnetic energy spectrum at the scales .
therefore , now the scale slightly larger than the dissipation scale becomes the most active scale for stretching .
this way , the subsequent stage is characterized by a slower growth of magnetic energy and a gradual shift of the peak of the magnetic energy spectrum to larger scales ( cho & vishniac 2000 ) .
the growth rate of the magnetic energy density at this stage turns out to be linear ( schlter & biermann 1950 ; schekochihin & cowley 2007 ) .
\(3 ) the amplification of magnetic field stops when the magnetic energy density becomes comparable to the kinetic energy density at the outer - scale of turbulence and a final , statistically steady , saturation stage begins . during the saturation stage ,
the peak of the magnetic energy spectrum occurs at a wavenumber a few times larger than that of the kinetic energy spectrum ( cho & vishniac 2000 ) . in the vanishing viscosity ( and diffusivity )
limit , the duration of the exponential growth stage will be very short .
nevertheless , since the eddy turnover time at the dissipation scale is arbitrarily small in the limit , any weak seed field can grow exponentially , reach the equipartition strength at the dissipation scale , and enter the linear growth stage within arbitrarily small amount of time .
therefore , in the vanishing viscosity limit , 1 ) the exponential growth stage shown in figure [ fig:1](d ) becomes invisible , 2 ) @xmath169 shows linear growth most of the time before saturation , and 3 ) regardless of the strength of the initial seed field , @xmath169 follows virtually the same curve of growth when @xmath82 is plotted in a linear scale as in figure [ fig:1](d ) . according to simulations ( e.g. cho et al .
2009 ; cho & yoo 2012 ) , the system reaches the saturation stage in @xmath125 .
amplification of a localized seed magnetic field was studied in detail by cho & yoo ( 2012 ) and cho ( 2013 ) .
if a seed magnetic field is localized in space , turbulent motions make the seed magnetic field transported out from the region of initial injection ( figure [ fig:2 ] ) .
cho ( 2013 ) found that the rate at which the magnetized region expands is @xmath138 .
therefore , the _ homogenization timescale _ ( or the magnetization timescale ) , the time required for the magnetic field to fill the whole system , is given by @xmath170 if the driving scale is comparable to the size of the system , @xmath171 , then homogenization happens fast , i.e. within @xmath172 times the large - eddy turnover time @xmath25 , and the subsequent evolution should be similar to that of a uniform / homogeneous seed magnetic field case ( cho & yoo 2012 ) .
the whole system reaches the saturation stage within @xmath173 . on the other hand , if the driving scale is less than @xmath174 , i.e. if @xmath175 , the magnetic field near the initial injection cite reaches the equipartition strength @xmath38 before the magnetic field fills the whole system . in this case , it takes more than @xmath176 for the whole system to reach the saturation stage . as in uniform / homogeneous seed field cases ,
as long as the strength of the seed magnetic field is sufficiently weaker than the equipartition strength , its absolute value does not matter much in the vanishing viscosity limit .
first , because weak magnetic fields are passively transported and stretched by turbulent motions , the homogenization timescale @xmath177 , which is mainly determined by large - scale turbulent motions , is similar for all weak seed magnetic fields .
second , as in the uniform / homogeneous seed field cases , the strength of the weak seed fields does not affect the saturation time much .
therefore , the turbulence dynamo process should be similar for sufficiently weak localized seed magnetic fields . since turbulence dynamo is so efficient in the vanishing viscosity limit that it is difficult to find a constraint on the strength of the initial seed magnetic field .
this is true for both the primordial and the astrophysical seed magnetic fields .
however , it might be possible to tell the origin of the cosmic magnetism by observing the distribution of magnetic field in some systems .
the key quantity is the homogenization timescale , @xmath177 .
suppose that a localized seed magnetic field is injected into a system .
if @xmath177 is short compared with the age of the universe , the system becomes homogenized quickly and the subsequent evolution will be very similar to that of a uniform seed magnetic field . therefore , in this case , it will be very difficult to distinguish the astrophysical and the primordial seed magnetic fields .
on the other hand , if @xmath177 is longer than the hubble time , the system has not been yet completely magnetized .
therefore , in this case , the magnetic field in the system will be very inhomogeneous .
consider a cluster of galaxies with @xmath1781 mpc . in this case , if @xmath179 km / s , which is very likely , the homogenization timescale is shorter than the hubble time
. therefore , it will be very difficult to distinguish the primordial and the astrophysical origins of magnetic fields in clusters .
on the other hand , if we consider a filament of width @xmath164 mpc , the homogenization timescale is longer than the hubble time , if @xmath180300 km / s .
therefore , it is possible that filaments have not been fully homogenized . if this is the case , the distribution of magnetic fields in filaments is more or less inhomogeneous . | the origin of magnetic fields in clusters of galaxies is still an unsolved problem , which is largely due to our poor understanding of initial seed magnetic fields .
if the seed magnetic fields have primordial origins , it is likely that large - scale pervasive magnetic fields were present before the formation of the large - scale structure . on the other hand ,
if they were ejected from astrophysical bodies , they were highly localized in space at the time of injection . in this paper , using turbulence dynamo models for high magnetic prandtl number fluids , we find constraints on the seed magnetic fields .
the hydrodynamic reynolds number based on the spitzer viscosity in the intracluster medium ( icm ) is believed to be less than @xmath0 , while the magnetic reynolds number can be much larger than that . in this case , if the seed magnetic fields have primordial origins , they should be stronger than @xmath1 g , which is very close to the upper limit of @xmath2 g set by the cosmic microwave background ( cmb ) observations .
on the other hand , if the seed magnetic fields were ejected from astrophysical bodies , any seed magnetic fields stronger than @xmath2 g can safely magnetize the intracluster medium .
therefore , it is less likely that primordial magnetic fields are the direct origin of present - day magnetic fields in the icm . |
this paper is motivated by a recent consulting case in which a company wants to evaluate the performance of its @xmath14 offices at different locations . the performance of office no .
@xmath15 is quantified by a certain measure @xmath16 of costs per unit , but it is clear that it is influenced by various covariables @xmath17 describing , for instance , regional factors which can not be altered by the offices .
the idea is to eliminate these effects via a ( linear ) regression model , assuming that @xmath18 for some unknown handycap function @xmath19 of the covariable vectors @xmath20 , and @xmath8 are the corrected costs per unit of the @xmath15-th office .
now the proposal is to use a linear model for the regression function @xmath19 and to estimate it via least squares or least absolute deviations .
that means , we determine a regression function @xmath21 within a given model @xmath22 such that @xmath23 becomes minimal , where @xmath24 is the residual of the fitted regression function @xmath21 .
based on these residuals one computes the ranks @xmath25 as a surrogate for @xmath26 this procedure is simpler than an established method of benchmarking , called data envelopement analysis ( dea ) , initiated by farrel ( 1957 ) and charnes et al .
very roughly saying , in that approach one assumes that the @xmath8 are non - negative , and @xmath27 describes the minimally achievable costs per unit .
the corrected cost measures @xmath8 are estimated via a linear optimization method .
the main reasons for the company to use the regression approach rather than dea were the higher complexity of dea , which made it difficult to communicate it to employees , and the known sensitivity of dea to errors in the data .
moreover , normal quantile - quantile plots of the residuals showed no serious violation of a gaussian distribution , whereas the dea paradigm would predict a non - symmetric , right - skewed distribution .
this is certainly a non - standard application of regression methods in the sense that the regression function is treated as a nuisance parameter while the `` errors '' @xmath8 are of primary interest .
the problem with that approach is that these `` errors '' may fail to satisfy common assumptions such as independence , mean or median zero and homoscedasticity .
indeed , it may happen that the numbers @xmath8 , as a measure of the offices individual performance ( motivation , efficiency etc . ) , are correlated with the covariable vectors @xmath28 .
but then the residuals @xmath10 are systematically different from the numbers @xmath8 . by the way
, dea may suffer from the same problem , in particular , when the performance of most offices is still far from optimal .
even if the assumed model @xmath22 is correct and if the `` errors '' @xmath8 satisfy the standard assumptions of being independent and following all the same distribution @xmath29 for some @xmath30 , the average absolute difference between the ranks @xmath9 and @xmath7 may be substantial . in the present paper
we derive an explicit expression for the `` rank distortions '' @xmath31 i.e. upper bounds for @xmath32 , in case of traditional least - squares regression .
section [ sec : results ] contains the main results , an exact formula and approximations .
section [ sec : heuristics ] provides a heuristical derivation of an approximation of the rank distortions which also indicates what may happen in non - gaussian settings .
presumably these arguments could be made rigorous by applying similar techniques and arguments as koul ( 1969 , 1992 ) , loynes ( 1980 ) and mammen ( 1996 ) .
section [ sec : proofs ] contains rigorous proofs which do rely on the errors @xmath8 being independent with the same gaussian distribution .
the advantage of that is that minimal assumptions are imposed on the underlying design matrix .
we consider a linear regression model with a random vector @xmath33 here @xmath34 is a given design matrix with @xmath35 , @xmath3 is an unknown parameter vector in @xmath36 , and @xmath5 is an unobserved random vector with distribution @xmath37 with unknown @xmath30 . in our specific setting ,
@xmath38 with the observed covariable vectors @xmath39 and a given basis @xmath40 of the model @xmath22 . let us recall some well - known facts from linear models ( cf .
ryan 1997 ) .
the least - squares estimator of @xmath3 is given by @xmath41 , and the fitted vector @xmath42 may be written as @xmath43 with the `` hat matrix '' @xmath44 this matrix describes the orthogonal projection onto the column space of @xmath1 and satisfies @xmath45 .
moreover , since @xmath46 for any unit vector @xmath47 , one can easily verify that all `` leverages '' @xmath48 satisfy @xmath49 and @xmath50 .
the residual vector @xmath51 may be written as @xmath52 under a mild regularity condition on the hat matrix @xmath53 , the residuals @xmath10 are pairwise different : [ lemma : hat matrix ] for arbitrary indices @xmath54 , @xmath55 the condition @xmath56 implies that @xmath57 .
this lemma remains valid if the errors @xmath58 are only assumed to be independent with continuous distributions .
an immediate consequence of lemma [ lemma : hat matrix ] is that the residuals @xmath10 are pairwise different almost surely , whenever @xmath57 for at most one index @xmath15 .
now we are ready to state our first main result about the ranks @xmath7 and @xmath9 : [ theorem : exact ] suppose that @xmath57 for at most one index @xmath59 . then for arbitrary indices @xmath60 , + & = & _ k,=1^n ( ( ) + ( ) + & & - ( ) - ( ) ) , where @xmath61 , @xmath62 and @xmath63 . in particular , ( _ i - r_i)^2 & = & _ k=1^n ( - 2 ( ) ) + & & + _ 1 k < n ( + ( ) + & & - ( ) - ( ) ) . here and throughout @xmath64 denotes kronecker s symbol , i.e. @xmath64 equals one if @xmath65 and zero otherwise .
theorem [ theorem : exact ] is useful for exact numerical calculations .
it was used in the aforementioned consulting case to show that rank distortions may be substantial .
numerical experiments revealed also that the rank distortions are closely related to the leverages @xmath48 . recall that @xmath66 here is a theoretical result about the rank distortions in case of small maximal leverage : [ theorem : approximate ] suppose that the column space of @xmath1 contains the constant vectors , i.e. @xmath67 . then , as @xmath68 , @xmath69 uniformly in @xmath60 .
[ [ a - numerical - example . ] ] a numerical example .
+ + + + + + + + + + + + + + + + + + + + suppose that @xmath70 , and let @xmath1 be equal to @xmath71 the design matrix for simple linear or quadratic regression , where @xmath72 are equispaced numbers .
figure [ fig : example ] shows the pairs @xmath73 for both cases . in addition the approximations @xmath74 and @xmath75 are shown as lines .
equispaced @xmath76-values.,scaledwidth=100.0% ]
asymptotic statements in this section are meant as @xmath77 .
we assume that the errors @xmath58 are independent and identically distributed with finite standard deviation @xmath78 and c.d.f .
@xmath79 with bounded and uniformly continuous density @xmath19 .
one can easily deduce from @xmath80 that @xmath81 whereas the cauchy - schwarz inequality implies that @xmath82 hence @xmath83 pretending that the empirical c.d.f .
@xmath84 of the errors @xmath8 and the empirical c.d.f .
@xmath85 of the residuals @xmath10 are sufficiently close to @xmath79 , we write r_i = n ( _ i ) & & n f(_i ) , + _
i = n ( _ i ) & & n f(_i ) = n f ( _ i - ( ) _ i ) . but
@xmath86 is quite small , precisely , @xmath87 by .
hence we write @xmath88 moreover , for @xmath60 and @xmath89 , @xmath90 because @xmath91 is very small in the sense that @xmath92 by .
thus we pretend that the random pairs @xmath93 and @xmath94 are stochastically independent and conjecture that @xmath95 now consider the special case of @xmath96 and @xmath97 with the standard gaussian c.d.f .
@xmath98 and density @xmath99 . for @xmath100 , ^2 f(_i )
f(_j ) & = & ( f(_i ) ) ^2 + & = & ( ^-1 ( ^-1 x)^2 dx ) ^2 + & = & ( ( 2)^-1/2 ( x ) dx ) ^2 + & = & ( 4)^-1 , and ^2 f(_i)^2 & = & ^-1 ( ^-1 x)^3 dx + & = & ( 2)^-1 ( x ) dx + & = & ( 2 ) ^-1 .
hence the conjectured approximation equals @xmath101
we write @xmath102 with the companion hat matrix @xmath103 describing the orthogonal projection on the orthogonal complement of the column space of @xmath1 . since @xmath104
we may conclude that @xmath105 has a continuous distribution unless @xmath106 in the latter case , @xmath107 almost surely .
but condition is equivalent to @xmath108 where we utilized @xmath109 .
note further that entails that @xmath110 hence implies that @xmath111 obviously the latter condition yields .
consequently , the three conditions , and are equivalent . since @xmath112 , one may reformulate as @xmath113 since the probability in question does not change when we replace @xmath16 with @xmath121 , we may assume without loss of generality that @xmath122 and @xmath123 .
if @xmath124 denotes a random vector with standard gaussian distribution on @xmath125 , then @xmath117 has the same distribution as @xmath126^\top$ ] , where @xmath127 .
now we write @xmath128 and @xmath129 with @xmath130 $ ] , and @xmath131^\top$ ] , where @xmath132 almost surely , and @xmath133 is uniformly distributed on @xmath134 $ ]
. then ( y_1 0 y_2 0 ) & = & ( ( ) 0 ( ) ( ) + ( ) ( ) 0 ) + & = & ( ( ) 0 ( + ) 0 ) + & = & ( + + 2 ) + & = & ( ) + & = & .
+ according to lemma [ lemma : hat matrix ] , r_i & = & 1 + _ k i 1\{_k _ i } = 1 + _ k i 1\ { _ ki^ 0 } + _ i & = & 1 + _ k i 1\{_k _ i } = 1 + _ k i 1\ { _ ki^ 0 } almost surely , where @xmath135 and @xmath136 with the standard basis @xmath137 of @xmath138 .
consequently it follows from lemma [ lemma : arcsin ] that + & = & _ k i , j ( 1\ { _ ki^ 0 } - 1\ { _ ki^ 0 } ) ( 1\ { _ j^ 0 } - 1\ { _ j^ 0 } ) + & = & _ k i , j ( ( _ ki^ 0 , _
j^ 0 ) + ( _ ki^ 0 , _ j^ 0 ) + & & - ( _ ki^ 0 , _ j^ 0 ) - ( _ ki^ 0 , _ j^ 0 ) ) + & = & _ k i , j ( ( ( _ ki , _ j ) ) + ( ( _ ki , _ j ) ) + & & - ( ( _ ki , _ j ) ) - ( ( _ ki,_j ) ) ) , where @xmath139 note that @xmath140 and with @xmath141 we may write _ ki^_j^ & = & ( _ k - _ i)^ ( _ - _ j ) + & = & g_k + g_ij - g_kj - g_i + & = & _ k , ij - h_k , ij , + _ ki^_j^ & = & ( _ k - _ i)^^ ( _ - _ j ) = ( _ k - _ i)^ ( _ - _ j ) + & = & _ k , ij - h_k , ij .
hence we obtain the formula + & = & _ k i , j ( ( ) + ( ) + & & - ( ) - ( ) ) . but
the restriction to indices @xmath142 and @xmath143 is superfluous , because @xmath144 whenever @xmath145 or @xmath146 .
this yields the first asserted formula .
in the special case of @xmath147 , note that @xmath148 if @xmath149 .
if we replace @xmath150 with @xmath151 in our formula for @xmath152 , we end up with the expression & & _ k,= 1^n ( ( ) + ( ) + & & - ( ) - ( ) ) . but for @xmath145 or @xmath153 the corresponding summands are equal to zero ,
because @xmath145 implies that @xmath154 , and @xmath153 implies that @xmath155 . distinguishing the cases @xmath156 and @xmath157 yields ( _ i - r_i)^2 & = & _ k = 1^n ( - 2 ( ) ) + & & + _ 1 k < n ( + ( ) + & & - ( ) - ( ) ) . finally the assertion follows from the well - known fact that @xmath158 for @xmath159 .
first recall that @xmath160 , whence @xmath161 .
furthermore , @xmath162 whenever @xmath163 and @xmath157 , i.e. @xmath164 for at most @xmath165 index pairs @xmath166 .
elementary calculus shows that @xmath167 for some constant @xmath168 , the optimal one being @xmath169 .
hence + & & - ( ) - ( ) | + & = & o(^1/2 ) uniformly in @xmath170 and @xmath171 .
consequently , + & = & _ k,=1^n ( ( ) + ( ) + & & - ( ) - ( ) ) + o(^1/2 ) uniformly in @xmath60 .
but for @xmath172 $ ] and @xmath173 $ ] , & & ( ) + ( ) - ( ) - ( ) + & & = ( ) - ( ) + & & + ( ) - ( ) + & & + ( ) - ( ) + & & = ( ) - ( ) + & & + ( ) - ( ( 1 + + o(^2 ) ) ) + & & + ( ( 1 + + o(^2 ) ) ) - ( ) + & & =
( + o ( ) ) + & & + ( + o ( ) ) ( + o(^2 ) ) + ( + o ( ) ) ( + o(^2 ) ) + & & = + o(^2 ) .
consequently , @xmath174 but it follows from @xmath175 that @xmath176 + 9 a. charnes , w.w .
cooper and e. rhodes ( 1978 ) . measuring the efficiency of decision making units .
_ europ .
j. oper .
res . * 2 * _ , 429 - 444 .
farrell ( 1957 ) .
the measurement of productive efficiency .
_ j. royal statist .
a * 120 * _ , 253 - 281 .
h. koul ( 1969 ) .
asymptotic behavior of wilcoxon type confidence regions in multiple linear regression .
statist .
* 40 * _ , 1950 - 1979 .
h. koul ( 1992 ) .
_ weighted empiricals and linear models .
_ ims , hayword , ca .
r.m . loynes ( 1980 ) . the empirical distribution function of residuals from generalised regression .
. statist .
* 8 * _ , 285 - 298 .
e. mammen ( 1996 ) .
empiricla process of residuals for high - dimensional linear models .
. statist . * 24 * _ , 307 - 335 .
ryan ( 1997 ) .
_ modern regression methods
. _ wiley , new york . | consider the standard linear regression model @xmath0 with given design matrix @xmath1 ( @xmath2 ) , unknown parameter @xmath3 ( @xmath4 ) and unobserved error vector @xmath5 ( @xmath6 ) with i.i.d
. centered gaussian components .
motivated by an application in economics , we compare the ranks @xmath7 of the errors @xmath8 with the ranks @xmath9 of the residuals @xmath10 , where @xmath11 with the least squares estimator @xmath12 .
exact and approximate formulae are given for the rank distortions @xmath13 .
[ [ key - words ] ] key words : + + + + + + + + + + leverage , rank distortion . |
as a major resource for quantum information processing , entangled states have been intensively studied in a wide variety of quantum systems . from the mathematical point of view ,
when pure states are considered , they correspond to state vectors of a combined system that can not be factorized as a tensor product of vectors belonging to the hilbert spaces of the individual subsystems . in the general case ,
mixed entangled states can not be written as a convex combination of projectors on factorized vectors . from the physical point of view
, these states exhibit strong correlations when the individual systems are locally measured in different ( non orthogonal ) bases , as in the well known setting for the clauser - horn - shimony - holt ( chsh ) inequality @xcite .
for example , in photonic systems it has been tested with polarization entangled photon pairs @xcite .
one can also encode a pair of qubits on the spin and orbital angular momentum of a single photon @xcite .
combination of the two degrees of freedom leads to interesting applications on quantum information protocols including polarization controlled quantum imaging @xcite , algorithms @xcite , gates @xcite , cryptography @xcite , teleportation @xcite , and topological phases @xcite .
of course , the chsh inequality can be naturally applied to the spin - orbit degrees of freedom as well @xcite . despite being frequently addressed in the quantum domain ,
tensor product structures are also present in the classical description of the electromagnetic field .
we may quote the aforementioned example of the spin and orbital angular momentum carried by the optical field .
while spin is related to the polarization , orbital angular momentum ( oam ) has its origin in the wavefront structure @xcite . in a paraxial beam ,
these two degrees of freedom can be addressed independently and a combined mode basis can be built from a tensor product .
of course , one can also conceive non - separable spin - orbit modes and build a bridge for an interesting analogy with entangled states in quantum mechanics .
many concepts can be borrowed from quantum information theory , with a special role for the chsh inequality as a non separability witness .
this classical analogy of the chsh inequality for optical degrees of freedom was suggested in ref.@xcite .
the spin - orbit version of the experiment was published in ref.@xcite together with a theoretical description , including a brief quantum optical approach . since then , other experimental @xcite and theoretical @xcite works have addressed this issue .
spin - orbit entanglement has also been discussed in the quantum optical domain .
the connection between classical inseparability and quantum entanglement was addressed in ref.@xcite , where the fundamental properties of cylindrically polarized modes were thoroughly investigated .
experimental tools for squeezing and entangling spin and orbital degrees of freedom were developed in refs .
@xcite . in this work ,
we discuss quantum optical aspects of the spin - orbit separability , with a special attention to the roles played by coherence and sub poissonian photon number distribution @xcite in connection with the spin - orbit bell measurements performed in ref.@xcite .
we will consider different input states , covering the main attributes directly related to the spin - orbit separability , especially coherence and quantum entanglement .
the manuscript is organized as follows : in section [ polvortices ] we review the basic description of polarization vortices as non - separable spin - orbit modes .
these modes are then quantized in section [ qmodes ] where different mode partitions are envisaged .
the measurement scheme on which we base our theory is described in section [ bellmeasure ] and the corresponding average intensities and quantum noise are derived in section [ qnoise ] for different input states . finally , we summarize our conclusions in section [ conclusion ] .
in this section we discuss the notion of mode entanglement and its subtle distinction from quantum state entanglement . both rely on the concept of non separability in a tensor product vector space .
however , the vector structure of electromagnetic modes is not exclusive to quantum theory , it already occurs in classical electrodynamics .
the polarization of a light beam is directly related to the vector nature of the electromagnetic field . at the same time , a vector space structure can be assigned to the spatial functions used to describe a paraxial beam .
therefore , the complete mode structure requires a combination of these two degrees of freedom in the form of a tensor product between the two vector spaces .
for example , we can combine linear polarization vectors with hermite - gaussian spatial functions , which are discrete solutions of the paraxial wave equation in rectangular coordinates .
we shall label the linear polarization directions with capital letters @xmath0 and @xmath1 , associated with the unit vectors @xmath2 and @xmath3 .
the first - order hermite - gaussian modes oriented along the horizontal and vertical directions will be labeled with small case letters @xmath4 and @xmath5 , respectively , associated with the spatial functions @xmath6\ ; , \nonumber\\ \psi_v(x , y , z ) & = & \mathcal{n}\,y\,\exp \left [ -\frac{(x^2+y^2)}{2w^2(z)}+i\phi(x , y , z ) \right]\ ; , \label{psihv}\end{aligned}\ ] ] where @xmath7 is a normalization constant and @xmath8 is the phase distribution .
the most general first - order vector mode can be written as @xmath9 where @xmath10 .
we can borrow the definition of concurrence from quantum information theory to characterize the spin - orbit separability @xmath11 so that @xmath12 for product modes and @xmath13 for non factorisable ones .
a first - order spin - orbit mode basis can also be built with maximally non - separable modes analogous to bell states in quantum mechanics @xmath14 they correspond to non uniform polarizations on the wavefront .
two simple examples of maximally non separable modes are the well known radially ( @xmath15 ) and azimuthally ( @xmath16 ) polarized modes , characterized by the polarization patterns shown in fig.([fig : modes ] ) .
numerous methods for their production are available in the literature .
they have attracted a considerable attention for their special focusing properties .
it is easy to check from eq.([concurrence ] ) that @xmath17 for all elements of the basis given by eq.([bellbasis ] ) .
first order polarization vortices analogous to the four bell states . ]
let us consider the heisenberg operator describing the positive frequency component of the electric field associated with a first - order paraxial beam propagating along @xmath18 . in the separable mode basis @xmath19 , with @xmath20 and @xmath21
, it reads @xmath22 where @xmath23 is the operator that annihilates a photon on mode @xmath24 . as usual
, we can also define new annihilation operators when a different mode basis is used to decompose the field operator . for the bell modes defined in section [ polvortices ] , we obtain the following transformation equations for annihilation operators @xmath25 and the corresponding conjugate transformations for the creation operators . the electric field operator can then be written as @xmath26 now it is interesting to investigate the transformations for different kinds of states .
a fock basis of the hilbert space can be built for either mode decompositions by acting on the vacuum state with the corresponding creation operators .
let us consider , for example , an @xmath27photon fock state on mode @xmath28 , with all other modes empty . for simplicity , we shall omit the empty modes in the state vector , unless they are directly involved in the transformations that we are interested , however we must keep in mind that we will be always dealing with a multimode hilbert space .
thus , for the mode in question we have @xmath29 if we use the conjugate transformations ( [ annihibell ] ) and write the creation operator for @xmath30 in terms of those for @xmath31 and @xmath32 , we obtain @xmath33 from the analogous definitions of the fock states on modes @xmath31 and @xmath34 , we readily identify them in the right hand side of eq.([fockpsi+2 ] ) , so that @xmath35 for example , a single photon fock state on mode @xmath36 can be written @xmath37 note that the left hand side of eqs .
( [ fockpsi+3 ] ) and ( [ fockpsi+1photon ] ) has a product state while the right hand side has an entangled one .
of course , entanglement is independent of the basis chosen for the hilbert space , however these equations are _ not _ a change of basis , they correspond to different mode partitions of the electromagnetic field .
therefore , the mode basis must not be confused with the quantum state basis . in this sense , we prefer to employ the term _ mode partition _ in order to avoid this kind of confusion . entanglement does depend on the way the subsystems are defined and , although this partition is usually evident for material systems , it is not so clear for the electromagnetic field since one often has more than one mode decomposition adapted to the same boundary conditions
. therefore , we must be careful when comparing entanglement in the two cases . in this respect ,
pure coherent states play a very special role because they are _ always _ a product state in whatever mode decomposition .
indeed , let @xmath38 be the displacement operator that produces the coherent state @xmath39 when acting on the vacuum state .
since @xmath40 and @xmath41 commute , it can be factorized as a product between the corresponding displacement operators of modes @xmath31 and @xmath34 . from eqs.([annihibell ] ) one easily obtains @xmath42 so that the coherent states in the two mode decompositions are related by @xmath43 which is a product state in any case . from this discussion
we develop some intuition about the role played by photon number noise in the interplay between mode separability and quantum entanglement .
we next discuss this relationship more closely in the context of the experiment done in ref.@xcite .
let us consider the sketch shown in fig.([fig : interferometer ] ) for a spin - orbit bell measurement as the one done in @xcite .
an input mode initially prepared in some quantum state @xmath44 enters a mach - zehnder interferometer with an extra mirror in one arm .
this type of interferometer will be designated by the acronym mzim @xcite . when the two arms are equilibrated , the interferometer sorts even and odd modes at different output ports , the components @xmath31 and @xmath32 exit from the even port while @xmath45 and @xmath46 exit from the odd port .
the input beams are described by the electric field operators @xmath47 where @xmath20 and @xmath48 refer to polarization and transverse mode orientation respectively , @xmath49 refers to the input port , and @xmath50 is the corresponding annihilation operator .
the input field coming through port 1 will be transmitted through unitary transformation elements acting on the transverse mode and polarization degrees of freedom .
these elements are used to set the bases of the bell measurement .
they can be a dove prism ( dp ) for transverse mode and a half wave plate ( hwp ) for polarization transformation .
when the hwp is oriented at an angle @xmath51 and the dp at @xmath52 , the annihilation operators are transformed according to @xmath53 where @xmath54 is the unitary transformation describing the combined action of the two elements .
the hwp and the dp both act in a similar manner , they produce a reflection operation along their orientation axis .
therefore , their combined action can be written as the tensor product @xmath55 , where @xmath56 in matrix notation we have @xmath57 the same decomposition also applies to the output field operators @xmath58 where @xmath59 is the annihilation operator corresponding to mode @xmath60 at output @xmath61 .
scheme for the spin - orbit bell measurement . ]
one of the output ports of the interferometer will combine the odd modes from the vacuum input with the even modes from the excited input and the opposite combination is performed in the other output port . the quantized mode amplitudes @xmath62
are then determined by the input - output relations for the interferometer @xmath63 we now define the intensity operators integrated over the detectors area @xmath64 from the input - output relations of the measurement apparatus ( transformation elements plus interferometer ) one finds @xmath65 where @xmath66 .
note that modes @xmath45 and @xmath46 on output port 1 are not excited , as well as modes @xmath31 and @xmath32 on output port 2 .
therefore , they do not contribute to intensity measurements and the relevant operators at the output of the interferometer are @xmath67 the total input - output intensity does not depend on the measurement settings @xmath68 . in analogy with the parameter used in the clauser - horn - shimony - holt ( chsh ) inequality
, we define @xmath69 where the averages are calculated with the initial quantum state @xmath70 .
for single photon states , the normalized detected intensities correspond to detection probabilities and one finds the usual scenario for bell measurement where @xmath71 for single photon spin - orbit entangled states .
as discussed in ref.@xcite , this violation can be achieved with a polarization vortice coherent state , which is a product state in the @xmath31 and @xmath32 decomposition , as shown in eq.([coherentpsi+ ] ) .
therefore , in order to evidence quantum behavior , it will be necessary to monitor the intensity difference noise @xmath72 and compare it with the corresponding shot noise limit .
this figure of merit will be used in the next section to set up the comparison between different input states .
let us assume that only modes @xmath73 and @xmath74 are initially occupied with some quantum state @xmath70 , while all other modes are empty .
we will proceed as follows , first all output intensity operators are calculated in terms of the input modes ( before hwp and dp ) using the input - output relations outlined above .
then the mean values giving the average intensity and the corresponding quantum fluctuation are calculated with the specified initial state .
we will discuss different input states , their similarities and differences .
let us first consider the single photon entangled state given by eq.([fockpsi+1photon ] ) .
in this case , we have @xmath75\ ; , \label{fockm } \end{aligned}\ ] ] where @xmath76 . for the measurement settings @xmath77 , @xmath78 , @xmath79 , @xmath80 , we obtain the limiting value @xmath81 expected for a maximally spin - orbit entangled state .
we also evaluate the corresponding intensity difference fluctuations for any setting @xmath82 @xmath83\;. \label{fockdeltam } \end{aligned}\ ] ] these results can be easily generalized for the arbitrary @xmath27photon entangled state @xmath84 given by eq.([fockpsi+3 ] ) for which @xmath85 . for the bell measurement settings we obtain @xmath86 intensity squeezing . when @xmath87 , all photons exit through the same output port and the intensity difference is perfectly squeezed .
note that inter - mode coherence is a key ingredient for violation of the chsh inequality , since the summations on the right hand side of eqs .
( [ fockpsi+3 ] ) and ( [ fockpsi+1photon ] ) involve a coherent superposition with well defined relative phases .
when those are randomized , one gets statistical mixtures of fock states which no longer violate the inequality , as we will see next
. it will be interesting to evaluate the spin - orbit chsh inequality for a mixed fock state exhibiting strong photon number correlations but no entanglement .
this state can be represented by the density matrix @xmath88 with a well defined total photon number @xmath89 , randomly distributed between modes @xmath31 and @xmath34 . for this state
we obtain @xmath90 what results in @xmath91 for the measurement settings . the intensity difference noise is @xmath92 which can exhibit perfect squeezing for specific settings @xmath82 , but in general scales with the photon number .
for example , if we set @xmath93 , expression ( [ fockmixeddeltam ] ) gives @xmath94 while ( [ fockdeltam ] ) predicts perfect squeezing . in order to capture the role played by entanglement in the quantum noise properties of fock states ,
let us consider a partially entangled state of the kind @xmath95 where @xmath96 is given by eq.([fockmixed ] ) .
it is analogous to the werner states frequently used in quantum information science to discuss correlations present in partially entangled states @xcite .
the average intensity difference for setting @xmath82 is @xmath97 + ( 1-p)\,\cos\,2\alpha\,\cos\,2\beta\ ; , \nonumber\\ \label{wernerfockm}\end{aligned}\ ] ] which gives @xmath98 , corresponding to a violation threshold at @xmath99 .
the intensity noise can be readily calculated as a combination of the results given by eqs.([fockdeltam ] ) and ( [ fockmixeddeltam ] ) .
in fact , one can easily verify the relation @xmath100^{\,\,2}\ , , \label{wernerdm } \end{aligned}\ ] ] where @xmath101 is given by eq.([fockdeltam ] ) , which scales as @xmath102 while @xmath103 is given by eq.([fockmixeddeltam ] ) , which scales as @xmath104 .
note that both @xmath105 and @xmath106 are proportional to @xmath107 , so that @xmath108 scales as @xmath107 .
for a two - mode coherent state @xmath39 given by eq.([coherentpsi+ ] ) we obtain @xmath75\ ; , \label{purecoherentm}\end{aligned}\ ] ] with @xmath109 , giving the same limiting value @xmath81 as for the entangled fock state , however , the input state is now factorized . in order to distinguish the two cases ,
we calculate the intensity difference noise to find @xmath110 so that shot noise is expected for any @xmath82 settings , which is radically distinguished from the corresponding result for entangled fock states .
we now investigate the role played by coherence in the violation of the spin - orbit chsh inequality .
suppose a laser beam is initially prepared in a coherent state @xmath111 .
this beam is then split in two equal parts , one of them is transformed to mode @xmath32 and sent to a beam splitter where it gets _ contaminated _ with an independent laser also prepared in @xmath34 .
finally , the mixed @xmath32 mode is recombined with @xmath31 on a polarizing beam splitter ( pbs ) as sketched in fig.([fig : interferometer2 ] ) . since the two lasers have random relative phase , the quantum state describing the final beam can be written as the following statistical mixture @xmath112 where @xmath113 , with @xmath114 and @xmath115 being the reflection and transmission coefficients of the contaminating coupler .
let @xmath116 , where @xmath117 is the intensity reflectivity and @xmath118 is the phase acquired on reflection , then we obtain @xmath119 where @xmath120 , giving @xmath121 for the bell measurement settings . a violation threshold is predicted at @xmath122 . for @xmath123 and @xmath124 one
recovers the coherent state result given by eq.([purecoherentm ] ) . proposed scheme to produce an adjustable mixed coherent state . ] for a completely incoherent superposition ( @xmath125 ) we obtain @xmath126 giving @xmath91 for the bell measurement settings .
this result suggests that coherence is an essential ingredient for violation of the inequality for the intensity average .
meanwhile , the intensity noise for a setting @xmath82 is @xmath127 which scales as @xmath102 and is above shot noise for most settings .
it is also interesting to inspect the results obtained for the two - mode squeezed vacuum state , which presents quadrature entanglement .
it can be constructed from the action of the two - mode squeezing operator on the vacuum state as given by @xmath128 where @xmath129 is the squeezing parameter and the total intensity is @xmath130 . in this case
we find @xmath126 giving @xmath91 for the measurement settings , the same result as for the mixed coherent state .
for the intensity noise we get @xmath131 which also scales as @xmath102 for most settings . therefore , despite its quadrature entanglement , the quantum properties of squeezed vacuum do not show up in this kind of spin - orbit bell measurement .
in summary , we have discussed spin - orbit bell measurements in paraxial polarization vortices from a quantum optical perspective . the vortices are treated as non - separable modes and their quantum descriptions in different mode decompositions are compared .
the mode separability is evaluated with an inequality for the average intensities analogous to the chsh inequality .
it is violated for a single - photon spin - orbit entangled state , as expected , but also by a factorized coherent state .
these two limiting cases were experimentally investigated in refs .
therefore , a natural question arises about the interplay between entanglement and coherence in the violation . in this article
we compared different input states in order to develop some intuition .
our results indicate that the roles played by coherence and photon number fluctuations can be summarized by the following examples : * without coherence we expect no violation of the intensity based chsh inequality , as illustrated by the examples with mixed coherent or fock states ( sections [ qnoisemixfock ] and [ qnoisemixcoherent ] ) . * with coherence and poissonian distribution for the total photon number ,
we expect violation but no entanglement , as illustrated by the example with product coherent states ( section [ qnoisepurecoherent ] ) .
we remark that pure coherent states are factorized in any mode decomposition but exhibit maximal violation .
* with coherence and reduced ( below shot - noise ) photon number fluctuations , we expect violation and entanglement , as illustrated by the entangled fock states ( section [ qnoisefock ] ) . it is important to stress that in all cases the relevant coherence being considered is the one characterized by inter - mode relative phase , not the absolute one .
we also present examples that interpolate between pure and mixed states .
finally , section [ qnoisesqueezed ] illustrates that quadrature entanglement does not imply in violation of the intensity based chsh inequality .
experimentally , structural inseparability and entanglement can be simultaneously certified by measuring intensity averages and the corresponding fluctuations .
violation of the chsh inequality with the averages certify structural inseparability while intensity difference squeezing evidences entanglement .
coherent states should exhibit shot noise level for any measurement settings while photon number squeezed states should present 50% squeezing for the settings used in the inequality .
these results can be experimentally tested with polarization vortices produced on intensity squeezed sources like optical parametric oscillators @xcite or pump - noise - supressed diode lasers @xcite .
funding was provided by coordenao de aperfeioamento de pessoal de nvel superior ( capes ) , fundao de amparo pesquisa do estado do rio de janeiro ( faperj - br ) , and instituto nacional de cincia e tecnologia de informao quntica ( inct - cnpq ) .
99 j. f. clauser , m. a. horne , a. shimony , and r. a. holt , * 23 * , 880 ( 1969 ) .
a. aspect , p. grangier , and g. roger , * 49 * , 91 ( 1982 ) .
m. j. padgett and j. courtial , opt
* 24 * , 430 ( 1999 ) .
m. frana santos , p. milman , a. z. khoury , and p. h. souto ribeiro , , 023804 ( 2001 ) .
d. p. caetano , p. h. souto ribeiro , j. t. c. pardal , and a. z. khoury , , 023805 ( 2003 ) .
a. n. de oliveira , s. p.walborn , and c. h. monken , j. opt .
b : quantum semiclassic . opt .
* 7 * , 288 ( 2005 ) .
l. marrucci , c. manzo , and d. paparo , * 96 * , 163905 ( 2006 ) .
e. nagali , f. sciarrino , f. de martini , l. marrucci , b. piccirillo , e. karimi , and e. santamato , * 103 * , 013601 ( 2009 ) .
e. nagali , f. sciarrino , f. de martini , b. piccirillo , e. karimi , l. marrucci , and e. santamato , opt .
express * 17 * , 18745 ( 2009 ) . c. e. r. souza and a. z. khoury , opt .
exp . * 18 * , 9207 - 9212 ( 2010 ) .
a. r. vieira , m. hor - meyll , and a. z. khoury , j. opt .
b * 30 * , 1623?1626 ( 2013 ) .
l. aolita and s. p. walborn .
* 98 * , 100501 ( 2007 ) . c. e. r. souza ,
c. v. s. borges , a. z. khoury , j. a. o. huguenin , l. aolita , and s. p. walborn . * 77 * , 032345 ( 2008 ) .
v. dambrosio , e. nagali , s. p. walborn , l. aolita , s. slussarenko , l. marrucci , and f. sciarrino , nat . comm . * 3 * , 961 ( 2012 ) .
l. chen and w. she , * 80 * , 063831 ( 2009 ) .
j. t. barreiro , t .- c .
wei , and p. g. kwiat , * 105 * , 030407 ( 2010 ) .
a. z. khoury and p. milman , * 83 * , 060301(r ) ( 2011 ) .
m. erhard , h. qassim , h. mand , e. karimi , and r. w. boyd , arxiv:1404.7573 [ quant - ph ] ( 2014 ) . c. e. r. souza , j. a. o. huguenin , p. milman , and a. z. khoury , , 160401 ( 2007 ) . c. e. r. souza , j. a. o. huguenin , and a. z. khoury , j. opt .
a * 31 * , 1007 - 1012 ( 2014 ) .
l. chen and w. she , j. opt .
b * 27 * , a7-a10 ( 2010 ) . c. v. s. borges , m. hor - meyll , j. a. o. huguenin , and a. z. khoury ,
* 82 * , 033833 ( 2010 ) .
e. karimi , j. leach , s. slussarenko , b. piccirillo , l. marrucci , l. chen , w. she , s. franke - arnold , m. j. padgett , and e. santamato , * 82 * , 022115 ( 2010 ) .
m. padgett and l. allen , contemp .
phys . * 41 * , 275 - 285 ( 2000 ) .
r. j. c. spreeuw , found
* 28 * , 361 ( 1998 ) .
k. h. kagalwala , g. di giuseppe , a. f. abouraddy , b. e. a. saleh , nature photonics * 7 * , 72 ( 2013 ) .
a. valls , v. dambrosio , m. hendrych , m. micuda , l. marrucci , f. sciarrino , j. p. torres , arxiv:1407.0144v1 ( 2014 ) . b. n. simon , s. simon , f. gori , m. santarsiero , r. borghi , n. mukunda , and r. simon , * 104 * , 023901 ( 2010 ) . x. f. qian , j. h. eberly , opt . lett . * 36 * , 4110 ( 2011 ) .
p. chowdhury , a. s. majumdar , and g. s. agarwal , * 88 * , 013830 ( 2013 ) .
f. de zela , * 89 * , 013845 ( 2014 ) .
a. holleczek , a. aiello , c. gabriel , ch .
marquardt and g. leuchs , opt . exp . * 19 * , 9714 ( 2011 ) . c. gabriel , a. aiello , w. zhong , t. g. euser , n. y. joly , p. banzer , m. fortsch , d. elser , u. l. andersen , ch .
marquardt , p. st .
j. russell and g. leuchs * 106 * , 060502 ( 2011 ) . c. gabriel , a. aiello , s. berg - johansen , ch .
marquardt , g. leuchs eur .
j. d * 66 * , 172 ( 2012 ) .
e. s. guerra , a. z. khoury , l. davidovich , and n. zagury , * 44 * , 7785 ( 1991 ) .
l. davidovich , s. y. zhu , a. z. khoury , and c. su , * 46 * , 1630 ( 1992 ) .
m. k. olsen , l. i. plimak , a. z. khoury , opt
. comm . * 201 * , 373 ( 2002 ) .
h. sasada and m. okamoto , * 68 * , 012323 ( 2003 ) .
r. f. werner , * 40 * , 4277 ( 1989 ) .
a. heidmann , r. j. horowicz , s. reynaud , e. giacobino , c. fabre , and g. camy , * 59 * , 2555 ( 1987 ) .
b. coutinho dos santos , k. dechoum , and a. z. khoury , * 103 * , 230503 ( 2009 ) .
y. yamamoto , s. machida , and o. nilsson , * 34 * , 4025 ( 1986 ) . | in this work we investigate the quantum noise properties of polarization vortices in connection with an intensity based clauser - horne - shimony - holt inequality for their spin - orbit separability .
we evaluate the inequality for different input quantum states and the corresponding intensity fluctuations .
the roles played by coherence and photon number squeezing provide a suitable framework for characterizing pure state spin - orbit entanglement .
structural inseparability of the spin - orbit mode requires coherence , an issue concerning either classical or quantum descriptions . in both cases
, it can be witnessed by violation of this intensity based chsh inequality . however , in the quantum domain , entanglement requires both coherence and reduced photon number fluctuations . |
the possibility of getting information about the thermodynamics of nuclear matter from the analysis of intermediate energies heavy ion collisions has triggered a lot of interest in this field .
starting from the pioneering work of the purdue group , when for the first time a power law was used to fit the mass spectra resulting from highly excited nuclear systems and thus suggesting that a critical phenomena was taking place , it has open the study of a rather new branch of thermodynamics i.e the study of phase transitions ( liquid - vapor ) in finite systems@xcite .
one of the most challenging features that nuclear multi - fragmentation phenomena presents , is that signals of phase transitions of different orders can be extracted from experimental data . on one hand
, the analogy between the nuclear force and a van der waals interaction suggests that the nuclear equation of state should reproduce the main features that characterize a liquid - gas phase transition .
different indications of this kind of transition have been reported @xcite . on the other , signatures of a second order phase transition or critical behavior
have been obtained assuming fisher - like scaling relations for fragment distributions @xcite . moreover , in recent works @xcite is suggested that the observed critical behavior is compatible with a first order phase transition , and it is due exclusively to finite size effects .
an important point to be kept in mind is that the above mentioned approaches to the fragmentation problem are based on descriptions where not only the concept of thermodynamical equilibrium , but also the macroscopic constraints imposed to the system play a relevant role .
for example , several statistical descriptions of the nuclear multi - fragmentation process , e.g. the statistical multifragmentation model ( smm ) @xcite and the microcanonical metropolis monte carlo model ( mmmc ) @xcite , employ the concept of a _ freeze - out volume _ inside of which the existence of a thermodynamical equilibrated ensemble of fragments is assumed .
these statistical models have been widespreadly used by the nuclear community in a successful manner to described some experimental observations . however , from the experience gained in numerical simulations , such a concept as _ freeze - out volume _
, even as an approximation , does not seem to be completely correct .
it is then important to study which kind of differences arise as a consequence of not assumming a finite volume scenario , as it is not _ a priori _ evident that the evolution of a fragmenting system confined in a finite volume would produced the same macroscopic observable when compared with a non confined one , in which an expansive motion is present as an extra collective degree of freedom . in previous works @xcite
we have studied the fragmentation of a simple classical system where the dynamics is governed by a hamiltonian with a two body interaction lennard - jones term .
a microscopic description , employing molecular dynamics techniques , was used in order to adequately handle the possible presence of a non equilibrium behavior .
it was shown that a fragmentation time can be defined , after which a certain degree of local equilibrium is achieved in the system .
this fact allowed us to calculate a caloric curve for our expanding - fragmenting system , which is characterized by the absence of a vapor branch .
the aim of the present communication is to study how the restriction of a finite volume , and then the imposition of equilibration , affects some of the results obtained in the unconstrained case .
we will show that one of the main effects is seen in the behavior of the caloric curve ( cc ) .
for the constrained system it clearly shows a vapor - branch , which is absent in the free expanding case . this behavior is a direct consequence of the presence of a confining volume , which destroys the velocity correlations that , in the case of the free expanding system , build up an expansive flux that acts as a heat sink .
nevertheless for both cases ( at rather low densities for the constrained system ) a local maximum and a loop in the cc can be seen .
this feature can be associated with a negative branch in the corresponding thermal response function ( trf ) , that might signal a phase transition of first order .
a gradual smoothing of the mentioned indicators is observed for the confined system at higher densities .
differences in fragment mass distributions are also reported and related to the microscopic origin of such behavior .
this paper is organized as follows . in section [ numsim ]
we will describe the model used in our simulations .
a brief review of the results already obtained for the unconstrained expanding system is included . section [ seccc ] is devoted to the study of the caloric curves . in section [ secvfo ]
we study possible choices for the freeze out volume for the constrained case . in section [ sectrf ]
we calculate the thermal response functions and the kinetic energy fluctuations in order to signal the presence of possible phase transitions .
a microscopic correlations study is performed in section [ secspectra ] analyzing the results of two different clusterization algorithms .
finally , in section [ secconclu ] , conclusions are drawn .
the system under study is composed by excited drops made up of particles interacting via a 6 - 12 lennard jones potential , which reads : @xmath0 r \le r_{c } & \\ 0 r$>$r_{c } & \end{array } \right.\ ] ] we took the cut - off radius as @xmath1 .
energies and distances are measured in units of the potential well ( @xmath2 ) and the distance at which the potential changes sign ( @xmath3 ) , respectively while the unit of time used is : @xmath4 .
we integrated the set of classical equations of motion using the well known verlet algorithm @xcite , taking @xmath5 as the integration time step .
initial conditions were constructed using the already presented method of cutting spherical drops composed of 147 particles out of equilibrated and periodic 512 particles per cell l.j .
system @xcite .
a broad energy range was considered such that the asymptotic mass spectra of the fragmented drops for the unconstrained system changes from a u shaped
pattern to an exponentially decaying one .
somewhere in between this two extremes a power law like spectra can be obtained . for the sake of completeness we summarize in this section the main results obtained in previous work for the non - constrained fragmenting system ( see @xcite for details ) .
due to the fact that we are dealing with a non - stationary process the determination of the time at which fragments are formed becomes one of the key ingredients in the analysis . to accomplish this task one can use a simple and intuitive cluster definition , that is based on correlations in configuration space :
a particle @xmath6 belongs to a cluster @xmath7 if there is another particle @xmath8 that belongs to @xmath7 and @xmath9 , where @xmath10 is a parameter called clusterization radius ( in this work we used @xmath11 ) .
the recognition algorithm introduced by this definition is known as minimum spanning tree ( mst ) fragment recognition method . in @xcite
it was shown that a different recognition method , the early cluster formation model and its companion practical realization the so called early cluster recognition algorithm ( ecra ) , outperforms the mst clusterization algorithm in the sense that it finds the asymptotic fragmentation pattern in phase space at early stages in the evolution , when fragments are still not observables in configuration space . at that time
the system still looks like a rather compact piece of excited matter . instead of being defined through a proximity criteria ,
the ecra fragments are associated with the set of clusters @xmath12 for which the sum of the fragment internal energies attains its minimum value : @xmath13 } } \nonumber \\
e_{int}^{c_{i } } & = & \sum_{i}[\sum_{j\in c_{i}}k_{j}^{c.m.}+\sum_{{j , k\in c_{i}% } { j\le k}}v_{j , k } ] \label{eq : eecra}\end{aligned}\ ] ] where the first sum in ( [ eq : eecra ] ) is over the clusters of the partition , @xmath14 is the kinetic energy of particle @xmath8 measured in the center of mass frame of the cluster which contains particle @xmath8 , and @xmath15 stands for the inter - particle potential .
we dub the partition found by ecra as the most bound density fluctuation in phase space ( mbdf ) and we define as time of fragment formation ( @xmath16 ) the time at which the mbdf attain microscopic stability ( see appendix for details ) . in this way
@xmath17 is related to the time at which the systems switches from a regime dominated by fragmentation to one in which the dominant decay mode is evaporation of light aggregates ( mostly single particles ) by the excited fragments . in this way
the ecfm - ecra outperforms the mst not only in terms of the time at which fragments are detected but also in it s capability of unveiling the nature of fragmentation process .
it shows that fragments are formed in phase space as a consequence of correlations in both q - p space .
once the time of fragment formation is determined it is possible to calculate several properties of the system at fragmentation time . for this purpose
the expanding system is decomposed in concentric shells and the mean radial velocity is calculated .
it can be seen that the expansion is almost linear with the distance to the center of mass of the system , and that a local temperature can be defined as the fluctuations of the velocity around the local expansive collective motion .
moreover , the isotropical character of those fluctuations supports the idea that local equilibrium is achieved @xcite .
it is worth noticing that , on one hand at @xmath16 most of the system is still interacting , and on the other the local temperature of the inner shells attain a rather constant value that can be consistently considered as the temperature of the system at fragmentation time .
this can be seen in figure [ figtloc ] where we show , for different energies , the dependence of the local temperature , @xmath18 , and the local density , @xmath19 , as a function of the distance to the center of mass at fragmentation time .
in order to study the consequences of imposing a finite volume constraint to our system we used a spherical confining ` wall ' .
the considered external potential behaves like @xmath20 with a cut off distance @xmath21 , where it smoothly became zero along with its first derivative .
a rather broad range of values for @xmath22 were used . inside this potential ,
a highly excited drop was initialized in the way already described above and the corresponding equations of motion were integrated .
once the transient behavior was over we performed a microcanonical sampling of particle configurations every @xmath23 up to a final time of @xmath24 .
in this case the standard prescription for temperature calculation in the microcanonical ensemble was used , i.e. the kinetic energy ( @xmath25 ) was related to the temperature of our @xmath26-particles system using : @xmath27
one of the thermodynamics measurements that remains useful in the small system limit ( hundred of particles ) is the caloric curve , i.e. the functional relationship of the system temperature with its excitation energy . the resulting caloric curve ( cc ) for the unconstrained system
is displayed in fig .
[ figcc ] a ) . in this figure circles
denote the temperature of the system , already defined in section [ sfes ] , as a function of the energy . on the same figure
a curve denoted by squares is also present , for this one the total kinetic energy , i.e. including both the fluctuations around the collective motion and the collective motion itself , is used instead .
of course this is not a temperature , just a fraction of the total kinetic energy , but it is then obvious that collective motion begins to be noticeable around a value of @xmath28 , and it becomes dominant at around @xmath29 .
two main features are to be noticed , in first place the cc develops a maximum , and second the cc has no vapor branch but develops a rather constant behavior .
these features are in contrast with the standard expectations inherited from the thermodynamics of infinite systems . in fig.[figccv3d ] ) we show the behavior of @xmath30 for the constrained system ( see also fig.[figccv]a ) , where for the sake of clarity , we include just three different values of @xmath31 ) . for all density values a vapor branch is clearly observed .
moreover , comparing fig.[figcc]a ) with fig.[figccv3d ] ) and fig.[figccv]a ) , it is clear that the vapor branch is due to the presence of the confining volume which inhibits the formation of radial collective motion that behaves like a heat sink in the non constrained case . inside the volume
the emitted aggregates can do nothing but interact among themselves , and with the constraining wall , until thermal equilibrium is attained .
the effect of the boundaries on the cc is not important as long as the total energy in the system is small enough , and the expansive collective motion ( in the unconstrained case ) can be neglected .
but this is not the case for highly excited finite systems , for which the presence of the constraining wall prohibits the formation of the expansive radial motion . for that range of energies the local equilibrium features that can be found in the fragmentation of the unconstrained system
are replaced by global equilibrium ones when the dynamics is confined to finite volumes .
this means that one has to be at least extremely careful if one intends to analyze expanding systems using confining volumes or any other model that rely on a global equilibrium hypothesis .
another interesting feature to be notice in fig.[figccv3d ] ) and fig.[figccv]a ) is the presence of a loop at the beginning of the gas rise in the cc for diluted enough systems .
this back - bending behavior is gradually smeared out as the system density is increased . in section [ sectrf ] we will relate this to the behavior of the thermal response function of the system and in section [ secspectra ] we will find that this can also be understood with the aid of the mass spectrum resulting from the mst and ecra analysis .
as was already shown , due to the intrinsic non - equilibrium character of the fragmentation process in the non - confined case , an assumption of local equilibrium ( ` mounted ' over an expansive radial flux ) instead of a global one has been found to be more appropriate .
accordingly , distributions of density and temperature , and not unique values , are necessary to described the system at fragmentation time in a rigorous way(see fig.[figtloc ] ) .
it is then not expected to find an exact mapping in the ( @xmath32 ) space between the constrained and non - constrained dynamics . nevertheless if one insists on establishing such a comparison ,
two different criteria can be adopted in order to choose the appropriate volume for the constrained case ( which we will call _ freeze - out volume _ , @xmath33 ) .
on one hand it can be seen from fig.[figtloc ] ) that at @xmath16 the local density value for the inner cells ( where most of the mass is present at the time of fragment formation ) remains rather constant ( @xmath34 ) for the broad energy range presented in the figure .
accordingly , one can choose @xmath33 in order to attain such density value in the constrained case .
this election , that could be considered as a ` density - guided ' choice , corresponds to @xmath35 and a _ freeze out _ density value of @xmath36 .
for the other hand a different approach can also be adopted looking at fig.[figcc]a ) and fig.[figccv]a ) .
it can be verified that for energy values where the onset of the fragmentation process occurs ( @xmath37 ) the thermal energy in the expanding system is clearly lower than for the @xmath38 case in the constrained case .
in fact , similar temperatures can be found in the constrained case only for much diluted situations , i.e. for @xmath39 . in this sense , from a ` temperature point of view ' , a different _ freeze out _ density @xmath40 can be established . as we mentioned above this ambiguity in the definition of a _
freeze out _ volume ( or density ) is a consequence of trying to reproduce the behavior of an expanding fragmenting system using a global equilibrium scenario .
in recent works @xcite a lot of attention has been paid to the role played by the behavior of the specific heat ( or more generally speaking : the thermal response function , @xmath41 ) as a signal of the occurrence of a phase transition in finite systems .
moreover , it has been shown ( see @xcite ) that the presence of a negative branch in the @xmath41 can be related to a first order phase transition taking place in an isolated finite system .
this kind of analysis can straightforwardly be performed over the systems under the current study , using the already calculated cc and taking into account that the respective trf s can be calculated as : @xmath42 in fig.[figcc]b ) we show the trf for the unconstrained system .
two poles can be seen .
the first one ( @xmath43 ) signals the entrance of the system into the multifragmentation regime , while the second one is related to the leveling off in the corresponding cc , and can be related to the increasing limit imposed by the strong flux to the thermalization of the total available energy .
fig.[figccv]b ) shows the corresponding trf for the constrained case .
we include the curves for the two limiting cases discussed in section [ secvfo ] as possible freeze out choices , i.e. @xmath44 and @xmath45 , and an intermediate @xmath46 value ( @xmath47 ) .
we notice that two poles and a negative branch can be observed for the @xmath48 trf curve . in this case
we can relate the first pole with the onset of the transition , and the second one with the entrance in the gas phase , the energy distance between them being related to the associated latent heat . for @xmath49
it can be seen that the curve exhibits a qualitatively similar behavior , the distance between the two poles is just reduced .
nevertheless , for higher density cases , such as @xmath44 , the two poles merge into a single ` singularity ' limited by finite - size effects . in that case
, the trf remains positive for all energies , showing a peak as a signature of the transition , as a consequence of the fact that the corresponding caloric curve does not display a loop but instead a simple change of slope ( see next section for an analysis of this change of behavior ) . before leaving this section we would like to discuss , using an approach introduced in @xcite , the behavior of the fluctuations in the kinetic energy as an indicator of the occurrence of a phase transition in our finite system . in
@xcite it is shown that for an isolated and equilibrated system in which the total energy can be decomposed as : @xmath50 the heat capacity can be calculated as : @xmath51 where @xmath52 is the heat capacity associated with the subsystem-@xmath53 , @xmath54 is the fluctuation of the partial energy @xmath55 , and @xmath56 the temperature of the system . in this context , the presence of poles and negative values in trf s can be associated to abnormally large fluctuations of the partial energy stored in subsystem-@xmath53 , i.e. @xmath57 , during the phase transition . in fig.[figflucv ] ) we show , for the constrained case , the relative fluctuation , @xmath58 , for the kinetic energy defined as : @xmath59 where @xmath26 is the number of particles in the system , and @xmath60 stands for an average over the microcanonical sampling of configurations .
we show the direct calculations using equation [ ae ] for both , @xmath44 and @xmath61 , and we also include the estimation of @xmath58 derived for the respective cc s in fig.[figflucv ] ) .
a rather good agreement can be seen between both ways of calculating this magnitude . in the same graph
a reference level marks the canonical value for @xmath58 .
as expected , for the appropriate range of energies , ` unusually ' large relative fluctuations for the @xmath48 case can be observed , while for the @xmath44 case a local maximum , that does not exceed the canonical value can be seen .
we postpone an analysis of this behavior until the next section , where we will study the properties of the system clusters .
we now try to extend the above performed analysis for the unconstrained system . in order to do that , we considered the central region ( or core ) defined by @xmath62 , and we calculated the total energy , @xmath63 , and the number of particles @xmath64 that remains inside the core at fragmentation time .
binning the whole set of events according to these variables we can sample unconstrained fragmenting events in a _ pseudo - microcanonical _ way in what concerns core - quantities at @xmath16 . in fig.[figfluc ] ) we show the behavior of the relative fluctuation of the core kinetic energy , @xmath65 , calculated using eq .
[ ae ] with @xmath66 and @xmath67 . in the same figure
we include the canonical reference level . even with a severe reduction of statistics imposed by the binning , a region of large fluctuations in @xmath68 , for @xmath69 and @xmath63 fixed , can be easily recognized .
this finding suggest that the large kinetic energy fluctuation signal , counterpart of the behavior observed for the corresponding cc and trf , can still be found for the expanding case .
in previous sections we have analyzed our system studying thermodynamical features like the behavior of the cc and the trf . in this part of the work
we will focus our attention to a different aspect of the process , more directly related to microscopic correlations , that can be studied using the clusterization algorithms already introduced in section [ sfes ] .
let us begin with the simple mst analysis .
as this cluster recognition method is exclusively based on spatial correlations one expects for high density situations a big mst - cluster to be present . on the other hand , for lower densities a deviation from the u - shaped behavior in the mass spectra can be expected . as was already stressed , mst underestimates the number of clusters because it does not take into account the relative velocities of the cluster constituent particles . for unconstrained systems this effect was apparent when we considered the time of fragment formation , at this time the system was already fragmented in phase space , but in configuration space , a big cluster was still present . due to the fact that the system was free to expand the asymptotic stage was such that mst converged to ecra .
but in the constrained case such an expansion is prohibited and then mst will not converge to ecra . nevertheless , an interesting point to be notice is that the mst algorithm can still provide useful information about the limit imposed by the constraining finite volume to the formation of well defined fragments in configurational space because it reflects the size of the interacting subsystems present at a given time . to further explore this we studied the behavior of the mst clusterization when applied to constrained systems at high energies .
we observed that once we enter the ` vapor - branch ' , the spectra remain unchanged . in fig.[figmstgas ] ) we show the obtained mst cluster distribution for four different densities at energies that situate the systems in the vapor - branch in the respective cc .
it can be seen that the mst - spectra go from a u - shape up to an exponential decaying behavior as the system density decreases .
notice that for densities around @xmath70 a power - law like shape can be recognized .
that means that from the spatial point of view the system admits , at least in principle , fragment configurations formed by well separated clusters of almost all possible sizes . in fig.[figmstortiz ] ) we show the mean value of the size of the maximum mst - fragment as a function of the energy for the two densities taken as a reference .
it can be seen that for the dense case , @xmath71 , the maximum fragment comprises almost all of the mass of the system irrespective of the value of the energy deposited in it . on the other hand for @xmath72 the size of the biggest fragment decreases with increasing energy .
this results support the idea that for the high density constrained case , no surfaces can be built in order to get small mst - fragments at high densities , whereas at low densities mst - fragments are well defined structures in configuration space and the surfaces appear .
we will now discuss the results obtained within the ecfm model . in fig.[figecramst ] ) we show the ecra and mst mass distribution for the unconstrained system , both calculated at @xmath16 .
we also show the same quantities calculated for the constrained case at densities @xmath73 and @xmath72 .
first we can notice that for the dense constrained case mst always gives essentially the same u shaped curve , while the ecra spectra show the usual transition from u shaped to exponentially decaying behavior .
on the other hand , for the low density constrained case and the unconstrained case both the mst and ecra results show such a transition .
it is interesting to note that when comparing the most energetic case the ecra mass distributions corresponding to the constrained cases present a steeper slope than the corresponding to unconstrained situation ( figures [ figecramst ] ( d ) , ( h ) , and ( l ) ) .
this happens as a direct consequence of the existence of boundaries .
for the constrained cases the violent inter particle collisions , that are present at all times due to the reflections of particles with the constraining walls , destroy almost every correlation in phase space . on the other hand , for the unconstrained case the system develops a collective expansive motion which limits the ` thermal ' component of the kinetic energy , allowing for the build up of well correlated density fluctuations in phase space , i.e.ecra-clusters ( note also that while the ecra clusters are microscopically stable for the unconstrained case , the ones for the constrained system are not ) .
it is in this regime where the differences between a local equilibrium picture and a global one become important .
as long as the collective expansive energy in the free case can be neglected ( @xmath74 ) the ecra spectra present a rather similar shape for the three shown cases . moreover , a clear signature of a first order phase transition can be recognized in the shape of the ecra spectra noticing the simultaneous presence of a ` liquid - phase ' ( mass fragments @xmath75 ) and a ` gas - phase ' ( mass fragments @xmath76 ) for @xmath77 this findings clarify the shapes of the corresponding cc . for the dense constrained system , @xmath44 , we have a big configurational mst - cluster all the time , inside of which an ( unstable ) ecra partition can be defined that exhibits features expected for a first order phase transition .
the fact that the whole system is strongly interacting ( there are no spatial surfaces dividing ecra - phases ) is responsible of the reduction of the back - bending behavior in the corresponding cc and the lack of strong kinetic energy fluctuations . for the more dilute case , @xmath45
, the system is capable of ` reproduce ' phase space surfaces ( at least at some degree ) in configurational space and the mentioned signals start to be noticeable . an extra support to the idea that fragmentation do take place in phase space not only for the free expanding system ( see @xcite ) but also for the constrained case can be gained if we remove the external ` wall ' potential and let the system expand . in fig.[figfreevol ] ) we compare the asymptotic spectra resulting from such a process of removal of the constraints , with the ecra one obtained with boundaries for @xmath48 and @xmath44 cases at @xmath78 and @xmath79 .
a quite remarkable agreement can be found .
the fact that this agreement is better for lower densities reflects that different degree of spatial - fragmentation is achieved at different densities . in this way , for @xmath80 , the differences in both spectra ( slight suppression of intermediate mass fragments favoring bigger clusters ) can be understood as consequence of an aggregation - like process occurring during the early expansive stage due to the strong interacting nature of the system . in section [ sectrf ]
we showed for the confined system that the diverging behavior of the trf is related to abnormally large fluctuations of the kinetic energy . in order to understand the origin of those fluctuations we show in fig.[figcorrel ] ) the system kinetic energy @xmath25 and the mass of the biggest ecra - cluster as a function of time for @xmath72 at @xmath78 .
also the mass of the biggest mst - cluster is shown .
it is clearly seen that the first two quantities are well correlated while the mst biggest cluster is not .
this indicates that from a microscopic point of view the systems is sometimes mainly liquid ( big ecra biggest cluster ) and others it is mainly vapor ( small ecra biggest cluster ) .
this phase alternation is the kind of coexistence expected for finite system ( see @xcite ) .
in this work we have presented a complete analysis of caloric curves , thermal response functions and mst and ecra fragment distributions for constrained and unconstrained excited liquid drops .
we have shown that the effect of constraints is quite important , and that its main result is to allow the system to reach thermal equilibrium .
in particular the caloric curves for constrained cases show a vapor branch which is clearly an ` equilibrium effect ' that is not present for unconstrained systems .
we have also verified , studying the behavior of ecra - cluster mass distributions , that fragmentation does occur in phase space in all the studied cases .
moreover , phase coexistence can be found in all ecra spectra at appropriate energies ( we have shown that the ecra fragments in constrained systems correspond closely to the mst asymptotic clusters when the constraints are removed .
this reinforces the validity of the ecfm - ecra approach to study fragmentation phenomena occurring in finite volumes ) .
it is interesting to notice that for the constrained case , as the system density is increased , the expected signals of a first order phase transition in the cc , trf and kinetic energy fluctuations are gradually smoothed , and eventually disappear .
this can be associated to the fact that at low densities
internal surfaces can be developed in the constrained system allowing the transition to be traced in configurational space ( i.e. mst clusters are formed reflecting the fact that well separated aggregates appear in coordinate space ) .
such a feature is not possible in the dense case , all of the time the system is composed by a big configurational cluster that comprises more than 95% of the total mass .
moreover for the constrained case , a relation between the fluctuations in size of the maximum ecra - fragment and the kinetic energy fluctuations was established . as a consequence , the presence of abnormally large fluctuations in the system kinetic energy
could directly be linked with a phase coexistence phenomenon taking place in a finite system . as a final remark
we want to emphasize that , even some similarities can be found between the behavior of low density constrained systems and unconstrained ones at low energies , such a comparison fails at and above the energy that corresponds to the onset of the fragmentation process , i.e. when the collective radial modes begin to drive the evolution of the system .
as we mentioned , the role of this collective motion is to behave as a heat sink , precluding the system from developing a vapor branch , and freezing the most bound density fluctuations in phase space allowing them to become the asymptotic clusters .
this is why a ` local ' equilibrium picture , and not a ` global ' one , is necessary to describe the fragmentation process correctly , i.e. taking into account the effects of the radial flux .
this teaches us that when dealing with fragmentation phenomena of the kind appearing in nuclear multifragmentation , caloric curves which display vapor branches should be taken with caution , and consequently the results obtained from models that display such a feature should be at least critically reexamined .
this work was done under partial financial support from the university of buenos aires via grant no.tw98 , and conicet via grant no .
pip 4436/96 .
a.chernomoretz acknowledges conicet for financial support and the warm hospitality of the laboratoire de physique nuclaire , dpartement de physique , universit laval ( quebec , canada ) .
m.ison aknowledges finantial support from uba via a student scholarship . once the clusters have been calculated using the ecfm we have to determine the time at which the fragmentation is over and the systems enters an evaporation regime ( i.e the fragmentation pattern is formed and the fragments undergo a simple evaporation process ) .
we will then define the time of fragment formation to that time at which the mbdf attain microscopic stability .
this property of the clusters has been calculated using what we call the short time persistence , in which we calculate the stability of mbdf against evaporation and coalescence .
we define the time of stabilization in the following way : given a configuration resulting from the ecra analysis at a given time @xmath81 we analyze the microscopic stability of each fragment @xmath82 of size @xmath83 by searching on all the fragments @xmath84 present at time @xmath85 for the biggest subset @xmath86_{i}^{t+dt}$ ] of particles that belonged to @xmath87 we then we assign to this fragment a value @xmath88 @xmath89_{i}^{t+dt}}{n_{i}^{t}}% .
$ ] in this way we are taking into account what we call the evaporation process .
we also have to take into account that there can be some realization for which cases the @xmath90_{i}^{t+dt}$ ] does not constitute a subset of the original cluster @xmath91 but is embedded in a bigger fragment of mass @xmath92 .
we include this effect by defining @xmath93_{i}^{t+dt}}{n_{i}^{t+dt}}$ ] ( with @xmath6 standing for @xmath94 .
finally the short time persistence reads : a reference value for @xmath99 can be obtained considering that the fragments undergo only a simple evaporative process . in this way
we say that when @xmath100 reaches this reference value the system goes from fragmentation to evaporation .
we call this time @xmath16 . | we study , using molecular dynamics techniques , how boundary conditions affect the process of fragmentation of finite , highly excited , lennard - jones systems .
we analyze the behavior of the caloric curves ( cc ) , the associated thermal response functions ( trf ) and cluster mass distributions for constrained and unconstrained hot drops .
it is shown that the resulting cc s for the constrained case differ from the one in the unconstrained case , mainly in the presence of a `` vapor branch '' .
this branch is absent in the free expanding case even at high energies .
this effect is traced to the role played by the collective expansion motion . on the other hand
, we found that the recently proposed characteristic features of a first order phase transition taking place in a finite isolated system , i.e. abnormally large kinetic energy fluctuations and a negative branch in the trf , are present for the constrained ( dilute ) as well the unconstrained case .
the microscopic origin of this behavior is also analyzed . |
lead magnesium niobate pbmg@xmath0nb@xmath1o@xmath2 ( pmn ) is a canonical relaxor material characterized by an extremely large magnitude of dielectric permittivity over a broad temperature range and having logarithmically - wide frequency dispersion @xcite that is strongly demanded in many valuable applications .
for example , this is important to produce small - size capacitors robust to the temperature change .
dilution of pmn with ferroelectrics led to the discovery of one of the best piezoelectrics ever found , which have now extremely wide applications , in all brunches of life , science and technology @xcite . at the same time , the understanding of the main property of relaxors , which is the colossal dielectric response , is still debated @xcite .
the source for the main properties of relaxors is usually related to the temperature growth of the polar nano regions ( pnr s ) on cooling @xcite .
it is believed that pnr s start their development , at a temperature called the burns temperature of @xmath3 , and increase their strength on cooling @xcite .
chu , setter and tagantsev @xcite suggested that @xmath4 might manifest a local phase transformation and heterophase fluctuations . according to ref .
[ ] , this phase is characterized by a critical change of the lower boundary of the pb off - centering magnitude .
specifically , this lower boundary , above @xmath4 , is zero , but critically increases on cooling below @xmath4 .
we will call the phase below this temperature as spherical glass , because the pb - off - centerings were shown to be spherical @xcite .
the spherical model was recently confirmed also in ref .
[ ] , and it is also in line with the suggestion that this phase corresponds to the heisenberg - like universality class , while at lower temperatures reduces to the ising universality class @xcite . in ref .
[ ] the emergence of the relaxor properties at @xmath5 was related to the crossover between the lattice and relaxational dynamics , leading to a specific phonon confinement called also breathers .
we assume that this phase freezes only the magnitude , but not the direction of the pb displacement .
this assumption implies that the corresponding order parameter is scalar .
the correlation of these scalar ( looking like a higgs boson @xcite ) off - centerings can happen via the lattice volume modulation , which is scalar too .
it is interesting to notice also that some of such off - centerings can have intense rotational dynamics , and the transverse instability of such rotators might result in specific nontrivial relaxations @xcite and contributions to dielectric permittivity @xcite that , actually , has been observed as a non - landau contribution @xcite .
vogel - fulcher freezing temperature @xmath6 is another famous temperature in relaxors @xcite .
below this temperature , pmn reaches a glass - like state associated with critical slowing down of the local dipoles on cooling together with gradual broadening of the relaxation frequency spectrum @xcite . nevertheless ,
in contrast to the glass - like idea @xcite , @xmath6 can be alternatively considered as the temperature of the emergence of the polarization nanodomains , controlled by the quenched random electric fields @xcite arising from the @xmath7 and @xmath8 random distribution over the crystal lattice .
a combined concept has been introduced by glinchuk , according to which the low temperature phase in pmn can be called as mixed ferroglass @xcite .
this author defines two quantities , @xmath9 is the most probable magnitude of the long - range dipolar field and @xmath10 is the half width of the random field distribution function .
when @xmath10 is much larger than @xmath9 , a glass order stabilizes .
if @xmath10 is much smaller then ferroelectricity shows up . in between of these two limits ,
the ferroglass is the stable phase , which is a mixture of the glass and ferroelectric phases .
the formation of the ferroglass on cooling has been thoroughly studied directly by neutron diffraction @xcite and dielectric spectroscopy @xcite methods .
knowing the observations at both @xmath4 and @xmath6 peculiar temperatures , several elegant theoretical models have been proposed .
they include , for instance , the quenched - random - field - model @xcite , the random - site - electric - dipole - system theory @xcite or the most recent and accepted spherical - random - bond random - field model @xcite .
this latter model considers a new type of glass by introducing an analog of the edvards - anderson parameter in the theory of relaxors .
it is worth also mentioning first - principles calculations of relaxors @xcite , which exploited the idea @xcite that the ferroelectric relaxors possess nanoscale chemically ordered regions facilitating the emergence of pnr s .
in contrast to this idea , some other calculations have shown a possibility to calculate some relaxor properties without the use of the idea about the chemically ordered regions @xcite .
the idea of the frustration of ferroelectricity has been also considered in the frame of recent first - principles calculations @xcite .
in addition to the relaxor - like features related to @xmath4 and @xmath6 , viehland _ et al _ emphasized the fundamental importance of temperature @xmath11 , which could manifest a new phase having glass features @xcite .
interestingly , the authors got this temperature by extrapolating the curie - weiss temperature , from the high - temperature part of dielectric permittivity .
et al _ @xcite found a critical temperature @xmath12 , on the basis of the lattice parameter temperature behavior and x - ray and neutron integrated diffraction intensity . in qualitative agreement with these findings ,
svitelskiy _ et al _ @xcite found that e - raman - lines split up , critically , at @xmath13 , on cooling , with the ratio of the components of @xmath14 .
the authors explained this splitting by a ( local ) tetragonal distortion .
recent measurements @xcite of the thermally stimulated acoustic emission revealed another temperature , @xmath15 , which is significantly higher than 400 k. at this temperature , the acoustic emission had a pulse , on cooling .
these authors related this pulse to the formation of static pnr s .
most recently , new measurements @xcite of diffuse neutron scattering , with a higher than earlier precision @xcite , revealed the former temperature , @xmath16 , again . according to the authors
, this temperature marks the start of the diffuse neutron scattering , having a so - called butterfly shape , on cooling .
less precise earlier measurements exhibited a strong tail in this integral intensity @xcite .
the authors @xcite concluded that pnr s develop on cooling starting from @xmath16 rather than from @xmath4 .
to the best of our knowledge , all these facts have not got any reasonable explanation , in the frame of a microscopic model of relaxors , so far .
actually , this transformation has not been taken into account in the previous theoretical models @xcite . in the present paper , we will focus on the significance of this issue .
( notice that the notations @xmath17 and @xmath18 are ours .
the authors of refs .
[ ] both used the notation @xmath19 as in ref .
. we will relate all these @xmath19 s , @xmath20 , and @xmath21 to different stages of the freezing of the random local tetragonal distortions , on cooling , and this is why we will call this phase as quadrupole glass ( see also studies of other quadrupole glasses in refs .
this phase differs from the mixed ferroglass phase by the fact that the local dipoles are still dynamic , but being compared with the spherical glass , they are now strongly cooperative , not only in the magnitude but also in the direction .
we will argue , in the manuscript , that the quadrupole glass phase transition is diffuse , and this makes the dielectric permittivity also diffuse .
let us remind that this diffusness has been considered so far as one of the main features of the relaxors @xcite .
now , we are going to give more details , derivations , and experiments , supporting this picture of pmn .
the plan of the paper is the following .
section 2 discusses the temperature dependence of the lattice parameter in pmn .
section 3 suggests the idea of the quadrupole glass .
section 4 reviews the diffuse scattering of neutrons in pmn .
section 5 addresses the main point of the paper , the giant dielectric response in pmn .
section 6 gives some scenarios of the possible local tetragonal distortion , in the temperature interval of interest .
section 7 presents our data of the photoelectric current .
finally , section 8 summarizes our results and discusses possible implications .
in order to illustrate an experimental evidence of the quadrupole phase , first , we present the temperature dependence of the lattice parameter @xmath22 in pmn @xcite .
these data manifest two different anomalies , at @xmath17 and @xmath18 ( the anomaly at @xmath18 was denoted earlier also as @xmath20 @xcite ) . at @xmath17 ,
the lattice parameter temperature dependence has the strongest curvature .
contrary to this , at @xmath18 , the temperature change of the lattice parameter vanishes , at all .
let us discuss , first , the change at @xmath23*. according to general thermodynamics @xcite , the largest curvature of @xmath24 results in the largest thermally stimulated strain .
indeed , @xmath25\delta t\ ] ] @xmath26 is the temperature lowering , @xmath27 is the volume of the sample , and @xmath28 is the change of the strain on temperature cooling . the former term in eq .
( [ eq1 ] ) corresponds to the straight dependence of the volume on temperature .
the latter one originates from the dependence of the heat capacity on pressure @xcite : @xmath29 we believe @xcite that this last term in equation ( 1 ) is responsible for the strong thermally stimulated acoustic emission at @xmath30*. indeed , the largest curvature of the lattice parameter temperature change implies the largest second derivative of the volume with respect to temperature , and this results , according to equation ( [ eq1 ] ) , in the largest strain , which can trigger the thermally stimulated acoustic emission ( via , for example , appearance of cracks ) .
the significance of temperature t@xmath31 * deserves special attention . at this temperature , the thermal change of the lattice parameter disappears and , correspondingly , the lattice anharmonicity vanishes .
this kind of transformation has been found also in other relaxors @xcite .
for the sake of illustration , we want to make some bridge between this phenomenon and the temperature dependency of the density of @xmath32he , at so called @xmath33 point .
@xmath32he has a discontinuity at this point , above which the density is rapidly growing on cooling , while , below , it saturates after showing some small spike @xcite .
this fact was explained by a phase transformation , at the @xmath33 point , from a classic liquid regime to quantum one . in the quantum he ii phase , below the @xmath33 point ,
the @xmath32he bosons cooperatively and synchronically rotate , in a vortex .
thus , the additional cooperative quantum permanent movement of the bosons makes the density of the matter to saturate .
quantum paraelectrics are other examples , which possess extremely high values of dielectric susceptibility in a wide temperature interval owing to quantum beatings @xcite . the dielectric permittivity in the quantum paraelectrics saturates , below the so called saturation temperature ( which is typically of 50 k ) .
we emphasize here this kinetic aspect , because dielectric permittivity in the temperature interval below @xmath34 and above @xmath35 , in pmn , possesses giant magnitudes , which can be explained by the extreme softness of the vibration or / and relaxation modes .
being based on these analogies , one can expect that , in pmn , at @xmath36 * , there also appears a cooperative atomic movement ( not necessarily totally synchronized quantum mechanically ) , which breaks the phonon anharmonicity .
it is worth noting that the saturation of the dielectric permittivity in relaxors , below some temperature , has been already discussed @xcite .
for example , in ref .
[ ] , the dielectric permittivity is represented in the form @xmath37 where @xmath38 is constant , @xmath39 with the relaxation time given by the vogel - fulcher formula @xmath40 here @xmath41 and @xmath42 are constants .
@xmath43 in equation ( [ epsilon ] ) is given by the landau - type expansion @xmath44 where @xmath45 is the spontaneous polarization , and the coefficient @xmath46 is given by a barrett - like quantum formula @xmath47 here @xmath48 is a characteristic temperature , which can be called also the saturation temperature ( in the spirit of the barrett formula @xcite ) .
the theory under discussion does not give a clue about the physical origin of this temperature .
notice that only dc dielectric susceptibility saturates , at low t , below @xmath48 , while , at any finite frequency , formula ( [ epsilon ] ) gives a one - peak function , with the peak position dependent on the frequency . in the same year
, experimental data , obtained for pmn , confirmed this kind of saturation of the dc dielectric susceptibility @xcite .
later on @xcite , it was realised that the dielectric susceptibility can be presented in the form of the sherrington - kirkpatrick formula @xcite @xmath49 where @xmath50 , @xmath51 and @xmath52 are constants , and @xmath53 is the ferroglass edwards - anderson parameter @xcite .
thus , the deviation of the temperature evolution of the dielectric permittivity from the curie - weiss law was related to the emergence of the glass order parameter at @xmath35 @xcite . in the next sections
, we will show that this understanding is not complete and requires to be supplemented by another temperature , namely , @xmath19 .
besides the lattice parameter measurements and thermally stimulated acoustic emission described above , there are other evidences of the peciliar behavior of pmn in the close vicinity of @xmath54 , which we have already mentioned in introduction .
these are dielectric permittivity anomaly @xcite , which was described as the emergence of a glass behavior , around 400 k , and the x - ray and neutron integrated diffraction intensity @xcite , which show clear appearance of the integral intensity around 400 k. we will treat both these evidences below , with the help of a model describing the quadrupole glass formation . at last , but not least , raman spectra in many studies ( see e.g. ref . [ ] ) confirmed to have the tetragonal - like splitting found , first , in ref .
[ ] . in this section ,
we suggest a model of the transformation of the spherical glass into the quadrupole glass .
then , we will fit this model to the experimental data .
raman spectra will be discussed in sec .
let us focus on the lead atomic displacements in pmn . according to experimental findings , below @xmath5 ,
the lead ions get off - center , and the strength of these off - centering s increases on cooling @xcite .
this conclusion follows from the fact that the radial distribution of the lead displacements , below @xmath5 , gets split into two sets , with a gap in between .
the gap appears , first , at those lead positions , which were not disturbed by the random fields ( high - temperature lead displacement is absent , for such positions ) .
these off - centerings being combined into polar nano regions ( pnr s ) increase their strength on cooling , and , thus , increase the strength of the pnr s . however
, this increase has some limit , because , when the random pnr s start touching each other , they organize larger superclusters ( not necesarily having global dipole moments , because of the disorder of the dipole directions in each of the pnr s ) , and , finally , these superclusters organize a net through the whole crystal . for
the sake of illustration , let us consider a simple model free energy describing two interacting pnr s : @xmath55 where @xmath56 is the distance between the lead pairs , @xmath57 is the correlation radius , the power @xmath58 is supposed to be large compared to 1 , @xmath59 is the relative correlation volume ( or density of the spherical off - centerings ) .
@xmath60 is the inverse local ( spherical ) susceptibility , which decreases on cooling and acquires negative values below @xmath4 , resulting in the appearance of finite off - center spherical displacements . the first term in equation ( [ eq2 ] )
represents the repulsion term which increases the free energy on approaching the correlation spheres of the nearest pnr s , each to the other , and has , originally , elastic nature @xcite .
the next two terms ( @xmath61 is supposed to be positive , for simplicity ) define the temperature dependence of the volume @xmath62 , in the case , when the spherical centers are far each from the other .
finally , by minimizing the free energy ( [ eq2 ] ) , one can find the equilibrium radii of the nearest pnr s .
the average radius can be given by the cubic root of the inverse concentration of the lead pairs .
the emergence of the quadrupole glass can be determined by the magnitude of the corresponding quadrupole glass order parameter , which can be constructed in the same way as the edwards - anderson parameter @xcite . for this purpose
, one should take into account that this parameter is conjugate to the corresponding random strain variance , which makes the phase transition spread out over a significant temperature range , as we shall see later on .
the development of the quadrupole glass in pmn is followed by the transformation of the phonon dynamics to the order - disorder one .
indeed , we assume that the dipoles in the pnr s , due to the strong dipole - dipole interaction inside each of the pnr s , are relaxators .
the growth of the correlation radius of the pnr s can be associated with the growth of the relaxation contribution to the dielectric permittivity on cooling .
simultaneously , when these dynamical pnr s touch each other , i.e. , in other words , start to acquire a static component of strain ( this , we believe , happens below @xmath17 @xcite ) , they organize a new phase with the quadrupole order , which , fully or partially replaces the spherical glass state below @xmath18 .
thus , we associate the temperature interval [ @xmath18,@xmath17 ] with the one , where the inhomogeneous strain organizes a steady state , and the relaxation contribution to the dielectric permittivity becomes critically important and practically substitutes the phonon contribution .
this consideration correlates well with the development of the dielectric spectrum of pmn recently found experimentally by bovtun _
et al _ @xcite .
the quadrupole glass can be described phenomenologycally by employing the following simple free energy @xmath64 where @xmath65 is polarization , q is the quadrupole glass order parameter , @xmath66 is the critical temperature , which , in the first approximation , equals @xmath18 , @xmath67 is the random strain variance , which becomes significant below @xmath17 .
the equilibrium condition for @xmath68 now reads @xmath69 solutions of eq .
( [ q ] ) were studied in ref .
[ ] . here , we consider the case of large @xmath67 : @xmath70 this order parameter appears well above @xmath66 and nearly linearly increases on cooling below @xmath71 .
the emergence of the quardrupole glass order parameter is also behind the suppression of the ferroelectric phase transition .
the soft ferroelectric mode behavior can be described by the following simple formula @xmath72 if we define temperature @xmath19 such that @xmath73 then , in the closest vicinity of @xmath19 , @xmath74 equation ( [ diffuse ] ) with @xmath75 being representative of the diffuseness of the transition , is one the main great features describing the dielectric behavior of relaxors , and now , we see that it might arise owing to the diffuse phase transition from the spherical glass to the quadrupole glass phase .
figure 2 presents the result of our extraction of the @xmath76 temperature dependence from the experimentally measured dielectric permittivity at 100 khz , in the same manner as it was done in ref .
the main result of this extraction is the following .
@xmath76 is zero , above @xmath4 .
it appears at @xmath5 , and gradually increases on cooling , down to @xmath77 .
this latter temperature is the curie - weiss extrapolated critical temperature , from @xmath78 part .
the smooth temperature behavior of @xmath76 is characteristic of diffuse phase transitions .
below @xmath66 , the dependence of @xmath76 becomes more and more linear that is inherent to normal behavior of a scalar order parameter like the quadrupole glass order parameter , in the landau theory .
we tried fitting eq .
( [ formula ] ) to the extracted from experiment @xmath76 curve , at a fixed value of @xmath79 .
one can see from figure 3 ( curve 1 ) that the fitting curve well fits the linear portion of experiment , but lies well above experimental @xmath76 , at high temperatures .
this deflection naturally stems from the fact that the random strain variance @xmath67 in formula ( [ formula ] ) is temperature independent .
this makes @xmath76 finite at @xmath5 and above , in contrast to the experimental result . in order to fix this problem
, one can follow the assumption about a temperature dependence of the random strain variance .
let @xmath80 , at temperatures above @xmath5 , and , at lower temperatures , @xmath67 gradually increases , until a maximal ( saturated ) magnitude of this variance .
one of the ways to check this assumption is using a hyperbolic tangent : @xmath81 for @xmath82 , and @xmath80 for @xmath78 . here
@xmath48 is the temperature interval , where @xmath67 saturates .
figure 3 ( curve 2 ) shows that this assumption well suits experiment .
thus , our data witness that the random field variance is zero , above @xmath5 , starts from zero at burns temperature and increases on cooling until some magnitude , and then saturates .
very probably , the saturation happens in the quadrupole phase .
figure 4 shows the integral intensity of the ( 001 ) bragg peak extracted from neutron pmn powder diffraction pattern measured at institut laue - langevin ( grenoble , france ) between 100k and 850k .
the temperature dependence of this intensity is representative for all the bragg peaks ( not shown here ) .
notice the changes of the slopes .
the first change happens at a temperature located in the interval [ @xmath18,@xmath17 ] .
the second does at the so - called freezing temperature @xmath6 ( see arrows ) .
these changes are in good agreement with those observed on the temperature dependence of the lattice parameter ( figure 1 ) .
we assume that these abrupt changes of the integral intensity stem from the change of the diffuse scattering intensity .
indeed , it was shown earlier @xcite that the diffuse scattering of neutrons with butterfly shape ( see the inset in figure 4 ) strongly increases below @xmath17 on cooling , while some diffuse scattering exists even above @xmath4 @xcite ( the latter is known as huang scattering on the quenched lattice parameter fluctuations , which can be described by chemical disorder of the nb and mg ions in the same sublattice ) .
notice that the nature of the butterfly diffuse scattering has been intensively discussed in literature ( see the most resent discussion in ref .
[ ] and references therein ) .
this scattering , as it was understood , arises due to the thermal excitations of the combined optical and acoustic atomic vibrations , having strong anisotropy .
the nature of this anisotropy is still debating . thus , this scattering manifests , in some sense , the transformation of the spherical glass into an anisotropic glass ( like the transformation of the heisenberg spins into ising ones @xcite ) .
figure 5 presents fitting of the same function as in fig . 4 ( curve 2 ) , differing by only a factor , to the integral intensity of the neutron data , after subtraction of the high - temperature value of this intensity @xcite .
one can see that this function nicely fits these data in the temperature interval above @xmath83 .
this finding gives us significant information that the integral intensity of the diffuse scattering can be considered as a general order parameter of relaxors , in the same way as we got this parameter , @xmath76 , above , from the dielectric permittivity data .
as it follows from equation ( [ eq6 ] ) , the soft mode hardens , because of the appearance of @xmath68 . on the other hand
, the appearance of @xmath68 leads to a new contribution to the dielectric permittivity , due to the relaxations .
the relaxation polarization , due to similarity in symmetry , may couple to the phonon displacements , and finally , can get cooperative .
the combined phonon - relaxation response can be described analytically , in the mean - field approximation @xcite , and the final formula is given by equation ( [ epsilon ] ) . in the case if one considers zero - frequency , i.e. the dc ( field - cooled ) response , the trend of the total dielectric susceptibility , which includes both the relaxation and phonon parts is such that , below @xmath84 the dielectric permittivity keeps growing @xcite , with temperature decrease , and saturates , below @xmath6 .
this makes the quadrupole phase having an extremely large value of the dielectric permittivity , in a wide temperature interval .
however , at finite frequencies , this permittivity strongly decreases , below a frequency - related temperature , owing to the vogel - fulcher temperature dependence of the relaxation time .
thus , the colossal dielectric response in pmn arises from the combination of the minimum of the soft mode and critically increasing , on cooling , edwards - anderson parameter @xmath68 corresponding to the quadrupole phase .
these anomalies arise , in our opinion , due to three consequent diffuse phase transitions , at @xmath4 , @xmath19 , and @xmath6 , first , owing to phonon softening , towards @xmath19 , on cooling , and , then , due to the conversion of the phonon dynamics to the relaxation one , at lower temperatures , down to @xmath6 .
it is important to stress the cooperative character of the polarization dynamics , due to the coupling of the phonons with the relaxations .
recent experimental data @xcite showed that , in pmn , there are actually two soft modes , @xmath85 and @xmath86 , with quite different frequencies .
this finding has not got yet a reasonable explanation .
let us suggest that these two modes correspond to the regions covered by the pnr s correlation spheres , in which the soft mode frequency @xmath85 is supposed to be high , due to the existence of the local @xmath68 order parameter , and to the regions away from the pnr s volume , where the soft mode @xmath87 is not disturbed by @xmath68 . according to the experimental data @xcite , the temperature dependence of the mode @xmath87 frequency can be described by the curie - weiss law , with @xmath88 .
this fact well corresponds to our assumption that @xmath89 , and that this temperature corresponds to the phase transition from the spherical glass to quadrupole glass phase .
the temperature dependence of @xmath85 is much weaker @xcite , and this also corresponds to our assumption that the softening of this mode is responsible for the phase transition from the quadrupole to ferroglass phase . in order to check our assumption
, one can measure the temperature dependency of the relative weight of these two modes .
we suppose that @xmath85 becomes more populated than @xmath85 , with temperature decrease , because of the development of the quadrupole phase on cooling .
one more remark concerns our assumption about the closeness of @xmath20 and @xmath90 ( see above ) .
this conception assumes that , above @xmath4 , pmn presents a well seen development of the ferroelectric frequency softening , ornstein - cernice ( landau ) correlations , and curie - weiss behaviour of the dielectric permittivity .
however , at lower temperatures , another tendency competes with the ferroelectric scenario .
we believe that the appearance of @xmath68 order parameter can be described in a manner similar to the one suggested in ref . [ ] ( for the first general mathematical exploration if this competition see ref .
[ ] , which introduced a new , trigger type , phase transition ) .
these authors suggested that , in some of the relaxors , there can exist a scenario of a local phase transition and heterophase fluctuations .
more specifically , the nonpolar local phase transition can happen in , say , pmn , at @xmath4 , because of the competition between the ferroelectric and , in our terminology , spherical glass order parameter . such a transition must be of first order , but , in our opinion , might be diffuse @xcite .
so , similarly , we suggest that such kind of the phase transformation may happen also at a lower temperature , which we denoted as @xmath90 . here
the phase transformation occurs between the spherical glass and quadrupole glass phases , and this transition , as we showed above ( see also ref . [ ] ) to be strongly diffuse .
interestingly , at a lower temperature , @xmath35 , we again meet absolutely the same situation , but , now , between the quadrupole- and ferro- glass phases @xcite .
thus , in our opinion , all three transformations follow the chu - setter - tagantsev scenario @xcite suggested earlier also as a general trigger mechanism [ ] .
in this section , we will discuss possible scenarios of the tetragonal deformation of the lattice of pmn revealed by the raman studies @xcite .
first of all , as soon as the shape of the crystal , in the whole temperature range , remains cubic , these tetragonal distortions must be local . in principle , the source for this distortion can be whether the strong dipole - dipole interactions between nearest pb displacements or the elastic ( or antiferroelectric ) interaction between next - to - nearest - neighbors . in the former case , the pb pairs , having in pmn [ 001 ] directions , can serve as tetragonal seeds for the lattice . in the latter case , the disordered incommensurate patterns , for example , in [ 110 ] directions , can result in the local tetragonal distortions .
in general , the competition between the ferroelectric and antiferroelectric order can be behind the emergence of the stable pnr s with low local symmetry @xcite . as a consequence , these nano - objects can break the macroscopic polarization and help the emergence of the quadrupole glass phase . as an example of such glass formation , the li pairs in the solid solution ktao@xmath91:li ( klt ) are tetragonal centers and organize an orthogonal ( 90-degree ) patterns made of these pairs , owing to the next - to - nearest elastic interactions @xcite .
the situation in pmn can be different , because the pb sublattice does not contain any other ions but lead ( klt contains k and li in the same sublattice ) .
however , the random fields , created by the random distribution of the mg and nb , in pmn , can make the lead ions different , and this difference is characterised by the magnitude ( and , generally , direction ) of the local electric field .
for example , it was realized earlier that those lead ions , which possess the smallest local electric field , stemming from the random distribution of the mg and nb ions , are more ferroelectrically active , while those , which are embedded into a strong local field , are not ferroelectrically active @xcite .
it is worth citing the recent first - principles calculations of the pmn relaxors diluted with pbtio@xmath91 revealing the tetragonal prevalence in the correlations between the atomic displacements @xcite .
weak crystallization theory might be an adequate model for the pnr s growth in pmn @xcite . according to this theory
, the nuclei of the new phase grow in such a manner that the magnitude of the correlation radius is an order parameter , but the direction is not ( the symmetry of the nucleus is spherical ) .
a similar situation was recently shown to have some relevance to higgs boson @xcite .
another example is the labyrinth structures observed in ultrathin films of lead titanate @xcite . to answer the question about the reasons resulting in the tetragonal splitting of the raman lines in pmn @xcite
, one needs performing more additional experiments and derivations .
we have also considered the effect of the dynamic polarization , inherent to the quadrupole phase , on the dc photoconductivity @xmath92 of pmn . we studied a flux - grown pmn crystal having a form of cube ( with edges of 1.8 mm ) and with the faces parallel to the ( 001 ) planes of the prototype perovskite lattice .
we put electrodes made of aquadag on two opposite faces of the crystal parallel to the incident light beam .
we illuminated the crystal with incandescent light having the intensity of 100 mw .
the absence of substantial heating of the crystal by the incident light was proved by close proximity of the permittivity - temperature curves measured at 1 khz , in the darkness and under continuous illumination .
electric current was measured during slow ( 2 - 3 k / min ) heating rate by means of a v7 - 30 electrometer .
figure 6 presents the results obtained . at low temperatures , below 210
k i.e. @xmath6 , we found the logarithm of the photoelectric current increasing approximately linearly with respect to inverse temperature , like in semiconductors , with an activation energy of 0.08 ev . between 210 k and approximately 240 k ,
the photoelectric current heavily decreased manifesting a catastrophic decrease of the mobility ( this finding is in line with the earlier measurements of the electron mobility in pmn @xcite ) , which we relate to the scattering of the electrons by the fluctuations of the crystal structure at the border between the ferro- and quadrupole glasses . then above 240 k
, our experiment again shows a linear dependence of the logarithm of the electric current on the inverse temperature , but , this time , with a much larger activation energy than at low temperatures i.e. 0.35 ev against 0.08 ev at low temperature .
this very strong increase of the activation energy is in line with our assumption of the existence in the lattice of a strong ( relaxation ) dynamic of the polarization .
this dynamic makes the dielectric permittivity having extremely large magnitudes , and is reflected in the large magnitudes of the electric conductivity activation energies inherent to polarons @xcite .
it is worth noting that similar decrease of @xmath93 on heating was observed previously in the region of phase transitions between the ferroelectric and antiferroelectric phases in nanbo@xmath2 @xcite and between ferroelectric and paraelectric phases in cd@xmath94nb@xmath94o@xmath95 @xcite .
it should be stressed that the measuring field strength was rather small ( @xmath96 along the [ 001 ] direction ) and could not induce the ferroelectric phase at low temperatures .
finally , we have developed a model of pmn relaxor involving the recently discovered @xmath19 temperature .
our model considers @xmath19 as a quasicritical temperature which manifests the transformation of the lead displacement pattern from the spherical glass to a quadrupole glass order .
one of the evidences of the anisotropic order is the intense butterfly - like diffuse scattering which develops in this temperature interval , the tetragonal splitting of the raman lines , specific features in the neutron scattering @xcite , and the viehland _
et al _ @xcite evidence of the glass order parameter .
we believe that , on further temperature decrease , the quadrupole glass transforms into a ferroglass phase .
we argued that all three diffuse phase transitions ( transformations ) ( at @xmath4 , @xmath66 , and @xmath6 ) follow the chu - setter - tagantsev scenario known generally also as a trigger mechanism @xcite .
we have emphasized that the quadrupole glass possesses a cooperative relaxation - phonon dynamics , looking as cooperative beatings of the local polarization , which have some similarities with the boson vortex dynamics in the quantum phase of @xmath32he and with the zero - point atomic vibrations , in quantum paraelectrics . these beatings ( cooperative relaxations )
suppress the development of the lattice parameter , form a colossal dielectric permittivity response , and result in a polaronic - like effect for the electrons .
we hope that our model will find further experimental evidences , theoretical exploration and will be utilized to take advantage of the discovered transformations in the applications desiring giant dielectric responses , of non - electron nature , in wide temperature intervals around room temperature .
sp and bd appreciate discussions of the topic with sergey vakhrushev , alexander tagantsev , and peter gehring .
this study is partially supported by russian foundation for basic research grant no .
12 - 08 - 00887-a .
sp also appreciates office of naval research grants n00014 - 12 - 1 - 1034 and n00014 - 11 - 1 - 0384 .
1900 g. a. smolensky and a. i. agranovskaya , sov .
phys . * 3 * , 1380 ( 1958 ) . l. e. cross , ferroelectrics * 151 * , 305 ( 1994 ) .
d. viehland , s. j. jang , l. e. cross , and m. wuttig , philos . mag .
b * 64 * , 335 ( 1991 ) .
s. zhang and f. li j. appl .
111 * , 031301 ( 2012 ) .
a. e. glazounov , a. j. bell , and a. k. tagantsev , j. phys . : condens .
matter * 7 * , 4145 ( 1995 ) .
v. a. isupov .
( b ) * 213 * , 211 ( 1999 ) ; v. a. isupov , sov .
. solid state * 28 * ( 1986 ) 1253 .
e. v. colla , e. yu .
koroleva , n. m. okuneva , and s. b. vakhrushev , phys .
lett . * 74 * , 1681 ( 1995 ) .
b. e. vugmeister and m. d. glinchuk , rev .
* 82 * 993 ( 1990 ) .
r. pirc and r. blinc , phys .
b * 60 * , 13470 ( 1999 ) .
r. blinc , j. dolinek , a. gregorovi , b. zalar , c. filipi , z. kutnjak , a. levstik , and r. pirc , phys .
lett . * 83 * 424 ( 1999 ) .
r. pirc , r. blinc , and v. bobnar , phys .
b * 63 * , 054203 ( 2001 )
. a. l. kholkin , a. n. morozovska , d. a. kiselev , i. k. bdikin , b. j. rodriguez , p. wu , a. a. bokov , z .- g .
ye , b. dkhil , l .- q .
chen , m. kosec , s. v. kalinin , adv .
func . mat . *
11 * , 1977 ( 2011 ) .
r. a. cowley , s. n. gvasaliya , s. g. lushnikov , b. roessli , and g. m. rotaru , advances in physics * 60 * , 229 ( 2011 ) .
a. a. bokov and z. g. ye , phys . rev .
b * 66 * , 064103 ( 2002 ) .
g. burns and f. h. dacol , solid state commun . * 48 * , 853 ( 1983 ) .
f. chu , n. setter , and a. k. tagantsev , j. appl .
phys . * 93 * , 5129 ( 1993 ) .
vakhrushev , and n.m .
okuneva , aip conf . proc .
* 626 * , 117 ( 2002 ) .
s. b. vakhrushev , a. a. naberezhnov , n. m. okuneva , and b. n. savenko , phys .
solid state * 37 * , 3621 ( 1995 ) ; s. b. vakhrushev , b. e. kvyatkovsky , a. a. naberezhnov , n. m. okuneva , and b. toperverg , ferroelectrics * 90 * , 173 ( 1989 ) .
m. paciak , t. r. welberry , j. kulda , m. kempa , and j. hlinka , phys .
b * 85 * , 224109 ( 2012 ) . c. stock , guangyong xu , p. m. gehring , h. luo , x. zhao , h. cao , j. f. li , d. viehland , and g. shirane , phys . rev .
b * 76 * , 064122 ( 2007 )
. a. r. bishop , a. bussmann - holder , s. kamba , and m. maglione , phys .
b * 81 * , 064106 ( 2010 ) .
d. podolsky , a. auerbach , and d. p. arovas , phys .
b * 84 * , 174522 ( 2011 ) .
s. a. prosandeev , phys .
state * 45 * , 1774 ( 2003 ) .
s. a. prosandeev , i. p. raevski , a. s. emelyanov , _ et al _ , j. appl . phys . * 98 * , 014103 ( 2005 ) .
a. k. tagantsev , phys .
lett . * 72 * 1100 ( 1994 ) .
v. westphal , w. kleemann , and m. d. glinchuk , phys .
* 68 * , 847 ( 1992 )
. s. a. prosandeev , u. waghmare , i. p. raevski and l. jastrabik , integrated ferroelectrics * 58 * , 1359 ( 2003 ) ; s. a. prosandeev , i. p. raevski and u. v. waghmare , aip conference proceedings,677 , 41 ( 2003 ) ; arxiv : cond - mat/0302496 . m. d. glinchuk , british ceramic transactions * 103 * , 76 ( 2004 ) . w. dmowski , s. b. vakhrushev , i .-
jeong , m. p. hehlen , f. trouw , and t. egami , phys .
lett . * 100 * , 137602 ( 2008 ) .
jeong , t .
w. darling , j. k. lee , th .
proffen , r. h. heffner , j. s. park , k. s. hong , w. dmowski , and t. egami , phys .
lett . * 94 * , 147602 ( 2005 ) .
m. d. glinchuk and v. a. stephanovich , j. phys . :
condensed matter * 6 * , 6317 ( 1994 ) .
s. tinte , b. p. burton , e. cockayne , and u.v .
waghmare , phys .
lett . * 97 * , 137601 ( 2006 ) .
p. k. davies and m. a. akbas , j. phys .
solids * 61 * , 159 ( 2000 ) .
a. r. akbarzadeh , s. prosandeev , eric j. walter , a. al - barakaty , and l. bellaiche , phys .
* 108 * , 257601 ( 2012 ) .
jorge iguez and l. bellaiche , physical review b * 73 * , 144109 ( 2006 ) .
hiroyuki takenaka , ilya grinberg , and andrew m. rappe , phys . rev . lett . * 110 * , 147602 ( 2013 ) .
s. m. nakhmanson and i. naumov , phys .
lett . * 104 * , 097601 ( 2010 ) .
n. choudhury , l. walizer , s. lisenkov and l. bellaiche , nature * 470 * , 513 ( 2011 ) .
d. viehland , s. j. jang , l. e. cross , and m. wuttig , phys .
b * 46 * , 8003(1992 ) .
b. dkhil , j. m. kiat , g. calvarin , g. baldinozzi , s. b. vakhrushev , and e. suard , phys . rev .
b * 65 * , 024104 ( 2001 ) .
o. svitelskiy , j. toulouse , g. yong , and z .- g . ye , phys . rev .
b * 68 * , 104107 ( 2003 ) .
b. dkhil , p. gemeiner , a. al - barakaty , l. bellaiche , e. dulkin , e. mojaev , and m. roth , phys .
b * 80 * , 064103 ( 2009 ) . c. stock , l. van eijck , p. fouquet , m. maccarini , p. m. gehring , guangyong xu , h. luo , x. zhao , j .- f . li , and d. viehland , phys .
b * 81 * , 144127 ( 2010 ) .
the higher precision means here smaller wave vectors , and , correspondingly , larger affected regions .
ivliev , v.p .
sakhnenko , soviet phys .
solid state phys .
* 28 * , 356 ( 1986 ) .
h. vollmayr , r. kree , and a. zippelius , phys .
b * 44 * , 12238 ( 1991 ) .
r. m. lynden - bell and k. h. michel , rev .
66 * , 721 ( 1994 ) .
b. tadic , r. pirc , and r. blinc , phys .
b * 55 * , 816 ( 1997 ) . see similar results , but in a smaller temperature interval , in ref .
[ ] . l. d. landau and e. m. lifshitz , statistical physics ( butterworth - heinemann , london , 1980 ) , vol .
5 , part 1 .
e. dulkin , e. mojaev , m. roth , i. p. raevski , and s. a. prosandeev , appl .
lett . * 94 * , 252904 ( 2009 ) .
sciau and a - m .
castagnos , ferroelectrics * 270 * , 259 ( 2001 ) .
f. london , superfluids ( new york : dover publications , inc . , 1964 ) .
a. tipler and r. a. llewellyn , _ modern physics _ , 6-th eddition , ( w.h . freeman and co , new york 2012 ) . k. a. mller and h. burkard , phys . rev .
b * 19 * , 3593 ( 1979 ) .
b. e. vugmeister and h. rabitz , phys .
b * 57 * 7581 ( 1998 )
. a. levstik , z. kutnjak , c. filipi , and r. pirc , phys .
b * 57 * , 11204 ( 1998 ) .
r. pirc , z. kutnjak , r. blinc , and q. m. zhang , j. appl . phys . * 110 * , 074113 ( 2011 ) .
b. e. vugmeister and h. rabitz , j. phys .
* 61 * , 261 ( 2000 ) .
r. pirc , r. blinc , v. bobnar , and a. gregorovi , phys .
b * 72 * , 014202 ( 2005 ) .
k. binder and a. p. young , rev .
* 58 * , 801 ( 1986 )
. s. f. edwards and p.w .
anderson , j. phys .
f * 5 * , 965 ( 1975 ) . note that , in the model of ref .
[ ] , the glass order parameter does not have formal meaning of the glass order , but rather reflects the fact that the relaxors contain the relaxation processes , which might be longer than the inverse measuring frequency .
s. a. prosandeev , e. cockayne , b. p. burton , et al .
b * 70 * , 134110 ( 2004 ) .
s. a. prosandeev , e. cockayne , and b. p. burton , phys .
b * 68 * , 014120 ( 2003 ) .
v. bovtun , s , kamba , a. pashkin , p. samoukhina , j. petzelt , i. p. bykov , and m. d. glinchuk , ferroelectrics * 298 * , 23 ( 2004 ) .
v. bovtun , j. petzelt , v. porokhonskyy , s.kamba , and y. yakimenko . j. europ .
* 21 * , 1307(2001 ) .
s. prosandeev , m. panchelyuga , s. raevskaya , and i. raevski , appl .
lett . * 91 * , 242904 ( 2007 ) .
a. bosak , d. chernyshov , sergey vakhrushev , and m. krisch , acta cryst .
a * 68 * , 117 ( 2012 ) .
r. g. burkovsky , a. v. filimonov , a. i. rudskoy , k. hirota and m. matsuura , s. b. vakhrushev , phys .
b * 85 * , 094108 ( 2012 ) .
we subtracted a constant from the experimental data , so , to have zero at high temperatures .
this zero is the origin for our fit .
one can see from figure 4 , where we present the original ( untreated ) data , that the subtracted constant is of about half of the total maximal intensity . in principle , bragg intensity may vary with temperature by mild release of neutron extinction due to small local distortion , in case of relaxors .
thus , the tail in figure 5 may contain the noncritical scattering , in particular , huang scattering , by the imperfections .
this part we describe by the random strain , and we show that the temperature dependent part of this scattering ( we are not interested in the temperature independent part here ) does not originate from the defects , but rather it comes from the pnr s emerging at @xmath5 ( otherwise , the tail would go father than @xmath5 ) .
in addition , diffuse scattering also may vary by spectrometer condition .
this part is also taken into account by the random strain tensor .
indeed , by changing the precision of the experiment , one simply changes the important wave vectors , that is the important scale , which , obviously , corresponds to the scale of the random strain tensor .
s. a. prosandeev , and v. a. trepakov , jetp , * 94 * , 419 ( 2002 ) .
b. e. vugmeister , phys .
b * 73 * 174117 ( 2006 ) .
s. b. vakhrushev , p. g. burkovskiy , s. shapiro , and a. ivanov , phys . solid state * 52 * , 889 ( 2010 ) .
a. al - zein , j. hlinka , j. rouquette , and b. hehlen , phys .
* 105 * , 017601 ( 2010 ) .
j. holakovsky , phys .
status solidi b * 56 * , 615 ( 1973 ) .
i. p. raevski , s. i. raevskaya , s. a. prosandeev , v. a. shuvaeva , a. m. glazer , and m. s. prosandeeva , j. phys . : condens .
matter * 16 * , l221 ( 2004 ) .
s. a. prosandeev , m. s. panchelyuga , s. i. raevskaya , and i. p. raevskii .
solid state , * 53 * , 147 ( 2011 ) .
a. tkachuk and h chen , aip conf .
* 677 * , 55 ( 2003 ) . v. m. ishchuk , v. n. baumer , and v. l. sobolev , j. phys .
matter * 17 * , l177 ( 2005 ) .
i. p. swainson , c. stock , p. m. gehring , guangyong xu , k. hirota , y. qiu , h. luo , x. zhao , j .- f .
li , and d. viehland , phys .
b * 79 * , 224301 ( 2009 ) .
s. miao , j. zhu , x. zhang , and z .- y .
cheng , phys .
b * 65 * , 052101 ( 2001 ) . e. i. kats , v. v. lebedev , and a. r. muratov , physics reports * 228 * , 1 ( 1993 ) .
d. d. fong , g. b. stephenson , s. k. streiffer , j. a. eastman , o. auciello , p. h. fuoss , c. thompson , science * 304 * , 1650 ( 2004 ) .
trepakov , a.v .
babinskii , n.n .
krainik , g.a .
smolenskii , and a.n .
samukhin , jetp lett .
* 26*,341(1977 ) .
i. p. raevski , s. m. maksimov , a. v. fisenko , s. a. prosandeev , j .
phys . : condens .
matter * 10 * , 8015 ( 1998 ) .
i. p. raevskii , m. a. malitskaya , o. i. prokopalo , v. g. smotrakov , and e. g. fesenko , soviet physics - solid state * 21 * , 715 ( 1979 ) .
i. p. raevski , m. a. malitskaya , p. f. tarasenko , o. i. prokopalo , ya .
e. cherner , and v. e. zakapko , soviet physics - solid state * 25 * , 1621 ( 1983 ) . | we present neutron diffraction , dielectric permittivity and photoconductivity measurements , evidencing that lead - magnesium niobate experiences a diffuse phase transformation between the spherical glass and quadrupole glass phases , in the temperature interval between 400 k and 500 k , with the quadrupole phase possessing extremely high magnitudes of dielectric permittivity .
our analysis shows that the integral diffuse scattering intensity may serve as an order parameter for this transformation .
our experimental dielectric permittivity data support this choice .
these data are important for the aplications desiring giant dielectric responses , in a wide temperature intervals and not related to electron s excitations . |
the performance of communication systems that rely on electromagnetic wave propagation are ultimately limited by noise of quantum - mechanical origin . moreover ,
high - sensitivity photodetection systems have long been close to this noise limit . hence determining the ultimate capacities of lasercom channels
is of immediate relevance .
the most famous channel capacity formula is shannon s result for the classical additive white gaussian noise channel . for a complex - valued channel model in which we transmit @xmath0 and receive @xmath1 , where @xmath2 is the channel s transmissivity and @xmath3 is a zero - mean , isotropic , complex - valued gaussian random variable that is independent of @xmath0 , shannon s capacity is @xmath4\,\mbox { nats / use } , \label{shannon}\ ] ] with @xmath5 and @xmath6 . in the quantum version of this channel model
, we control the state of an electromagnetic mode with photon annihilation operator @xmath7 at the transmitter , and receive another mode with photon annihilation operator @xmath8 , where @xmath9 is the annihilation operator of a noise mode that is in a zero - mean , isotropic , complex - valued gaussian state . for lasercom , if quantum measurements corresponding to ideal optical homodyne or heterodyne detection are employed at the receiver , this quantum channel reduces to a real - valued ( homodyne ) or complex - valued ( heterodyne ) additive gaussian noise channel , from which the following capacity formulas ( in nats / use ) follow : @xmath10 \label{homodyne}\\[.06 in ] c_{\rm heterodyne } & = & \ln[1 + \eta\bar{n}/((1-\eta)n+1 ) ] .
\label{heterodyne}\end{aligned}\ ] ] the + 1 terms in the noise denominators are quantum contributions , so that even when the noise mode @xmath9 is unexcited these capacities remain finite , unlike the situation in eq .
( [ shannon ] ) .
the classical capacity of the pure - loss bosonic channel in which the @xmath9 mode is unexcited ( @xmath11)was shown in @xcite to be @xmath12 nats / use , where @xmath13 is the shannon entropy of the bose - einstein probability distribution with mean @xmath14 .
this capacity exceeds the @xmath11 versions of eqs .
( [ homodyne ] ) and ( [ heterodyne ] ) , as well as the best known bound on the capacity of ideal optical direct detection . the ultimate capacity of the thermal - noise ( @xmath15 ) version of this channel is bounded below as follows , @xmath16 , and this bound was shown to be the capacity if the thermal channel obeyed a certain minimum output entropy conjecture @xcite .
this conjecture states that the von neumann entropy at the output of the thermal channel is minimized when the @xmath7 mode is in its vacuum state .
considerable evidence in support of this conjecture has been accumulated @xcite , but it has yet to be proven . nevertheless , the preceding lower bound already exceeds eqs .
( [ homodyne ] ) and ( [ heterodyne ] ) as well as the best known bounds on the capacity of direct detection .
more recently , a capacity analysis of the bosonic broadcast channel led to an inner bound on the capacity region , which was shown to be the capacity region under the presumption of a second minimum output entropy conjecture @xcite .
both conjectures have been proven if the input states are restricted to be gaussian , and they have been shown to be equivalent under this input - state restriction . in this paper , we show that the second conjecture will establish the privacy capacity of the lossy bosonic channel , as well as its ultimate quantum information carrying capacity .
the entropy power inequality ( epi ) from classical information theory is widely used in coding theorem converse proofs for gaussian channels . by analogy with the epi ,
we conjecture its quantum version , viz . , the entropy photon - number inequality ( epni ) . in this paper
we show that the two minimum output entropy conjectures cited above are simple corollaries of the epni .
hence , proving the epni would immediately establish key results for the capacities of bosonic communication channels .
the term wiretap channel " was coined by wyner @xcite to describe a communication system , in which alice wishes to communicate classical information to bob , over a point - to - point discrete memoryless channel that is subjected to a wiretap by an eavesdropper eve .
alice s goal is to reliably and securely communicate classical data to bob , in such a way that eve gets no information whatsoever about the message .
wyner used the conditional entropy rate of the signal received by eve , given alice s transmitted message , to measure the secrecy level guaranteed by the system .
he gave a single letter characterization of the rate - equivocation region under a limiting assumption , that the signal received by eve is a degraded version of the one received by bob .
csiszr and krner later generalized wyner s results to the case in which the signal received by eve is not a degraded version of the one received by bob @xcite .
these classical - channel results were later extended by devetak @xcite to encompass classical transmission over a quantum wiretap channel .
a quantum channel @xmath17 from alice to bob is a trace - preserving completely positive map that transforms alice s single - use density operator @xmath18 to bob s , @xmath19 .
the quantum wiretap channel @xmath20 is a quantum channel from alice to an intended receiver bob and an eavesdropper eve .
the quantum channel from alice to bob is obtained by tracing out @xmath21 from the channel map , i.e. , @xmath22 , and similarly for @xmath23 .
a quantum wiretap channel is degraded if there exists a degrading channel @xmath24 such that @xmath25 the wiretap channel describes a physical scenario in which for each successive @xmath26 uses of @xmath20 alice communicates a randomly generated classical message @xmath27 to bob , where @xmath28 is a classical index that is uniformly distributed over the set , @xmath29 , of @xmath30 possibilities . to encode and transmit @xmath28
, alice generates an instantiation @xmath31 of a discrete random variable , and then prepares @xmath26-channel - use states that after transmission through the channel , result in bipartite conditional density operators @xmath32 .
a @xmath33 code for this channel consists of an encoder , @xmath34 , and a positive operator - valued measure ( povm ) @xmath35 on @xmath36 such that the following conditions are satisfied for every @xmath27 . 1
. bob s probability of decoding error is at most @xmath37 , i.e. , @xmath38 2 . for any povm @xmath39 on @xmath40 ,
no more than @xmath37 bits of information is revealed about the secret message @xmath28 .
using @xmath41 , this condition can be expressed , in terms of the holevo information @xcite , as follows , @xmath42 here , @xmath43 is the holevo information , where @xmath44 is a probability distribution associated with the density operators @xmath45 , and @xmath46 is the von neumann entropy of the density operator @xmath47 .
because holevo information may not be additive , the classical privacy capacity @xmath48 of the quantum wiretap channel must be computed by maximizing over successive uses of the channel , i.e. , for @xmath26 being the number of uses of the channel , @xmath49 .
\label{eq : privcap } \end{aligned}\ ] ] the probabilities @xmath50 form a distribution over an auxiliary classical alphabet @xmath51 , of size @xmath52 .
the ultimate privacy capacity is computed by maximizing the expression specified in over @xmath53 , @xmath54 , @xmath55 , and @xmath26 , subject to a cardinality constraint on @xmath52 . for a degraded wiretap channel ,
the auxiliary random variable is unnecessary , and eq . reduces to @xmath56 .
\label{eq : privcap_degraded}\ ] ]
the noiseless bosonic wiretap channel consists of a collection of spatial and temporal bosonic modes at the transmitter that interact with a minimal - quantum - noise environment and split into two sets of spatio - temporal modes en route to two independent receivers , one being the intended receiver and the other being the eavesdropper .
the multi - mode bosonic wiretap channel is given by @xmath57 , where @xmath58 is the wiretap - channel map for the @xmath59th mode , which can be obtained from the heisenberg evolutions @xmath60 where the @xmath61 are alice s modal annihilation operators , and @xmath62 , @xmath63 are the corresponding modal annihilation operators for bob and eve , respectively .
the modal transmissivities @xmath64 satisfy @xmath65 , and the environment modes @xmath66 are in their vacuum states .
we will limit our treatment here to the single - mode bosonic wiretap channel , as the privacy capacity of the multi - mode channel can in principle be obtained by summing up capacities of all spatio - temporal modes and maximizing the sum capacity subject to an overall input - power budget using lagrange multipliers , cf .
@xcite , where this was done for the multi - mode single - user lossy bosonic channel .
* theorem * assuming the truth of minimum output entropy conjecture 2 ( see sec .
v ) , the ultimate privacy capacity of the single - mode noiseless bosonic wiretap channel ( see fig . [
fig : boswiretap ] ) with mean input photon - number constraint @xmath67 is @xmath68 for @xmath69 and @xmath70 for @xmath71 .
this capacity is additive and achievable with single - channel - use coherent - state encoding with a zero - mean isotropic gaussian prior distribution @xmath72 .
+ ) encodes her messages to bob ( @xmath73 ) in a classical index @xmath74 , and over @xmath26 successive uses of the channel , thus preparing a bipartite state @xmath75 where @xmath76 represents @xmath26 channel uses of an eavesdropper eve ( @xmath21 ) . for @xmath69 , this channel
is degraded , as eve s state can be recreated by passing bob s state through a beamsplitter of transmissivity @xmath77.,width=226 ] * proof * devetak s result for the privacy capacity of the degraded quantum wiretap channel in eq .
requires finite - dimensional hilbert spaces .
nevertheless , we will use this result for the bosonic wiretap channel , which has an infinite - dimensional state space , by extending it to infinite - dimensional state spaces through a limiting argument .
furthermore , it was recently shown that the privacy capacity of a degraded wiretap channel is additive , and equal to the single - letter quantum capacity of the channel from alice to bob @xcite , i.e. , @xmath78 where the superscript @xmath79 denotes single - letter capacity .
it is straightforward to show that if @xmath69 , the bosonic wiretap channel is a degraded channel , in which bob s is the less - noisy receiver and eve s is the more - noisy receiver .
the degraded nature of the bosonic wiretap channel has been depicted in fig .
[ fig : boswiretap ] , where the quantum states @xmath80 of the constructed system @xmath81 are identical to the quantum states @xmath82 for a given input quantum state @xmath83 . using eq . for the bosonic wiretap channel
, we have @xmath84 } \nonumber \\[.06 in ] & = & \max_{\langle \hat{a}^\dagger\hat{a}\rangle \le \bar{n } } [ s({\hat \rho}^b ) - s({\hat \rho}^{e^\prime } ) ] \nonumber \\[.06 in ] & = & \max_{0 \le k \le g(\eta{\bar n } ) } \{\mbox{$\max _ { \langle \hat{a}^\dagger\hat{a}\rangle \le \bar{n } , s({\hat \rho}^b ) = k}$ } [ s({\hat \rho}^b )
- s({\hat \rho}^{e^\prime})]\ } \nonumber \\[.06 in ] & = & \max_{0 \le k \le g(\eta{\bar n})}\{k - \mbox{$\min _ { \langle \hat{a}^\dagger\hat{a}\rangle \le \bar{n } , s({\hat \rho}^b ) = k}$ } [ s({\hat \rho}^{e^\prime})]\ } \nonumber \\[.06 in ] & = & \max_{0 \le k \le g(\eta{\bar n})}\{k - g[(1-\eta)g^{-1}(k)/\eta]\ } \nonumber \\[.06 in ] & = & g(\eta{\bar n})-g((1-\eta){\bar n})\,\mbox { nats / use } \nonumber \\[.06 in ] & = & q^{(1)}({\cal n}_{a\mbox{-}b}).\end{aligned}\ ] ] the first equality above follows from lemma 3 of @xcite .
the second equality follows from @xmath20 being a degraded channel .
the restriction to @xmath85 in the third equality is permissible because @xmath86 .
the fifth equality follows from minimum output entropy conjecture 2 ( see sec .
v ) . the @xmath87 that achieves this equality is a thermal state , which is realized when alice employs coherent - state encoding with a zero - mean isotropic gaussian prior distribution @xmath88 .
the sixth equality now follows from @xmath89 being a monotonically increasing function of @xmath90 , for @xmath91 a constant satisfying @xmath92 , and the equality to the single - letter quantum capacity follows from eq . .
note that the privacy capacity of this channel is zero when @xmath71 .
it is straightforward to show that in the limit of high input photon number @xmath93 , @xmath94 a result that wolf et .
@xcite independently derived by a different approach without use of an unproven output entropy conjecture .
let @xmath95 and @xmath96 be statistically independent , @xmath26-dimensional , real - valued random vectors that possess differential ( shannon ) entropies @xmath97 and @xmath98 respectively . because a real - valued , zero - mean gaussian random variable @xmath99 has differential entropy given by @xmath100 , where the mean - squared value , @xmath101 , is considered to be the _ power of @xmath99 , the entropy powers of @xmath102 and @xmath103 are taken to be @xmath104 in this way , an @xmath26-dimensional , real - valued , random vector @xmath105 comprised of independent , identically distributed ( i.i.d . ) , real - valued , zero - mean , variance-@xmath106 , gaussian random variables has differential entropy @xmath107 .
we can similarly define an i.i.d .
gaussian random vector @xmath108 with differential entropy @xmath109 .
we define a new random vector by the convex combination @xmath110 where @xmath111 .
this random vector has differential entropy @xmath112 and entropy power @xmath113 .
furthermore , let @xmath114 .
three equivalent forms of the entropy power inequality ( epi ) , see , e.g. , @xcite , are then : @xmath115 h({\bf z } ) & \ge & h(\tilde{\bf z } ) \label{epi2}\\[.12 in ] h({\bf z } ) & \ge & \eta h({\bf x } ) + ( 1-\eta ) h({\bf y } ) .
\label{epi3}\end{aligned}\ ] ] _ let @xmath116 $ ] and @xmath117 $ ] be vectors of photon annihilation operators for a collection of 2@xmath26 different electromagnetic field modes of frequency @xmath118 @xcite .
the joint state of the modes associated with @xmath119 and @xmath120 is given by the product - state density operator @xmath121 , where @xmath122 and @xmath123 are the density operators associated with the @xmath119 and @xmath120 modes , respectively .
the von neumann entropies of the @xmath119 and @xmath120 modes are @xmath124 $ ] and @xmath125 $ ] .
the thermal state of a mode with annihilation operator @xmath7 has two equivalent definitions : @xmath126 and @xmath127 where @xmath128 is the average photon number . in eq .
( [ cohstateform ] ) , @xmath129 is the coherent state of amplitude @xmath130 , i.e. , it satisfies @xmath131 , for @xmath130 a complex number . in eq .
( [ numberstateform ] ) , @xmath132 is the @xmath133-photon state , i.e. , it satisfies @xmath134 , for @xmath135 , with @xmath136 being the photon number operator .
physically , eq . ( [ cohstateform ] ) says that the thermal state is an isotropic gaussian mixture of coherent states .
equation ( [ numberstateform ] ) , on the other hand , says that the thermal state is a bose - einstein mixture of number states . from eq .
( [ numberstateform ] ) we immediately have that @xmath137 , because the photon - number states are orthonormal .
the entropy photon - numbers of the density operators @xmath122 and @xmath123 are defined as follows : @xmath138 thus , if @xmath139 and @xmath140 , where @xmath141 is the thermal state of average photon number @xmath142 for the @xmath143 mode and @xmath144 is the thermal state of average photon number @xmath145 for the @xmath146 mode , then we have @xmath147 and @xmath148 .
we define a new vector of photon annihilation operators , @xmath149 $ ] , by the convex combination @xmath150 and use @xmath151 to denote its density operator .
this is equivalent to saying that @xmath152 is the output of a lossless beam splitter whose inputs , @xmath143 and @xmath146 , couple to that output with transmissivity @xmath153 and reflectivity @xmath154 , respectively .
we can now state two equivalent forms of our conjectured entropy photon - number inequality ( epni ) @xcite : @xmath155 s(\hat{\rho}_{\boldsymbol c } ) & \ge & s(\hat{\rho}_{\tilde{\boldsymbol c } } ) , \label{epni2}\end{aligned}\ ] ] where @xmath156 with @xmath157 being the thermal state of average photon number @xmath158 for @xmath152 .
by analogy with the classical epi , we might expect there to be a third equivalent form of the quantum epni , viz .
, @xmath159 it is easily shown that ( [ epni1 ] ) implies ( [ epni3 ] ) @xcite , but we have not been able to prove the converse .
indeed , we suspect that the converse might be false . more important than whether or not ( [ epni3 ] ) is equivalent to ( [ epni1 ] ) and ( [ epni2 ] ) , is the role of the epni in proving classical information capacity results for bosonic channels .
in particular , the epni provides simple proofs of the following two minimum output entropy conjectures .
these conjectures are important because proving minimum output entropy conjecture 1 also proves the conjectured capacity of the thermal - noise channel @xcite , and proving minimum output entropy conjecture 2 also proves the conjectured capacity region of the bosonic broadcast channel @xcite .
furthermore , as we have shown above , proving minimum output entropy conjecture 2 also establishes the privacy capacity of the bosonic wiretap channel and the single - letter quantum capacity of the lossy bosonic channel . * minimum output entropy conjecture 1 *
let @xmath160 and @xmath161 be @xmath26-dimensional vectors of annihilation operators , with joint density operator @xmath162 , where @xmath163 is an arbitrary zero - mean - field pure state of the @xmath160 modes and @xmath164 with @xmath144 being the @xmath146 mode s thermal state of average photon number @xmath165 .
define a new vector of photon annihilation operators , @xmath149 $ ] , by the convex combination and use @xmath151 to denote its density operator and @xmath166 to denote its von neumann entropy . then choosing @xmath163 to be the @xmath26-mode vacuum state minimizes @xmath166 . * minimum output entropy conjecture 2 *
let @xmath160 and @xmath161 be @xmath26-dimensional vectors of annihilation operators with joint density operator @xmath162 , where @xmath167 is the @xmath26-mode vacuum state and @xmath123 has von neumann entropy @xmath168 for some @xmath169 .
define a new vector of photon annihilation operators , @xmath149 $ ] , by the convex combination and use @xmath151 to denote its density operator and @xmath166 to denote its von neumann entropy . then choosing @xmath164 with @xmath144 being the @xmath146 mode s thermal state of average photon number @xmath165 minimizes @xmath166 . to see that the epni encompasses
both of the preceding minimum output entropy conjectures is our final task in this paper .
we begin by using the premise of conjecture 1 in ( [ epni1 ] ) .
because the @xmath119 modes are in a pure state , we get @xmath170 and hence the epni tells us that @xmath171 taking @xmath172 on both sides of this inequality , we get @xmath173 $ ] .
but , if @xmath163 is the @xmath26-mode vacuum state , we can easily show that @xmath174 , with @xmath157 being the @xmath152 mode s thermal state of average photon number @xmath175 .
thus , when @xmath163 is the @xmath26-mode vacuum state we get @xmath176 $ ] , which completes the proof .
next , we apply the premise of conjecture 2 in ( [ epni1 ] ) . once again , the @xmath119 modes are in a pure state , so we get @xmath177 and hence @xmath173 $ ] .
but , taking @xmath164 , with @xmath144 being the @xmath146 mode s thermal state of average photon number @xmath165 , satisfies the premise of minimum output entropy conjecture 2 and implies that @xmath174 , with @xmath157 being the @xmath152 mode s thermal state of average photon number @xmath175 .
in this case we have @xmath176 $ ] , which completes the proof .
we conjectured a quantum version of the classical entropy power inequality , which subsumes two minimum output entropy conjectures that prior work has shown to be sufficient to prove the capacity of the point - to - point thermal - noise lossy bosonic channel , and the bosonic broadcast channel respectively @xcite .
even though proving this more general inequality the entropy photon - number inequality ( epni)might seem harder than the two minimum output entropy conjectures , there is a possibility of drawing parallels from the proofs of the classical entropy power inequality @xcite . in this paper
, we have also shown that the epni also implies the proof of the privacy capacity of the bosonic wiretap channel . furthermore , using a result from @xcite , we have that the degraded nature of the bosonic wiretap channel implies that its privacy capacity equals the single - letterquantum capacity of the lossy bosonic channel .
moreover . both of these capacities are achieved by coherent - state encoding using an isotropic gaussian prior .
this research was supported by the w. m. keck foundation center for extreme quantum information theory . 2 v. giovannetti , s. guha , s. lloyd , l. maccone , j. h. shapiro , and h. p. yuen , `` classical capacity of lossy bosonic channels : the exact solution , '' phys . rev . lett .
* 92 , * 027902 ( 2004 ) .
v. giovannetti , s. guha , s. lloyd , l. maccone , j. h. shapiro , b. j. yen , and h. p. yuen , `` classical capacity of free - space optical communication , '' in o. hirota , ed .
, _ quantum information , statistics , probability _ , ( rinton press , new jersey , 2004 ) pp .
v. giovannetti , s. guha , s. lloyd , l. maccone , and j. h. shapiro , `` minimum output entropy of bosonic channels : a conjecture , '' phys .
a * 70 , * 032315 ( 2004 ) .
s. guha , j. h. shapiro , and b. i. erkmen , `` classical capacity of bosonic broadcast communication and a minimum output entropy conjecture , '' phys .
a * 76 , * 032303 ( 2007 ) .
a. d. wyner , `` the wiretap channel , '' bell .
sys . tech .
* 54 , * 13551387 ( 1975 ) .
i. csiszr and j. krner , `` broadcast channels with confidential messages , '' ieee trans .
inform . theory
* 23 , * 339348 , ( 1978 ) .
i. devetak , the private classical capacity and quantum capacity of a quantum channel , " arxiv : quant - ph/0304127v6 .
a. s. holevo , `` the capacity of a quantum channel with general input states , '' ieee trans .
theory * 44 , * 269273 ( 1998 ) ; p. hausladen , r. jozsa , b. schumacher , m. westmoreland , and w. k. wootters , `` classical information capacity of a quantum channel , '' phys .
a * 54 , * 18691876 ( 1996 ) ; b. schumacher and m. d. westmoreland , `` sending classical information via noisy quantum channels , '' phys .
rev . a * 56 , * 131138 ( 1997 ) .
g. smith , `` the private classical capacity with a symmetric side channel and its application to quantum cryptography , '' arxiv : quant - ph/0705.3838 .
m. m. wolf , d. prez - garca , g. giedke , `` quantum capacities of bosonic channels , '' arxiv : quant - ph/0606132 .
o. rioul , `` information theoretic proofs of entropy power inequalities , '' arxiv : cs.it/07041751 .
l. mandel and e. wolf , _ optical coherence and quantum optics , ( cambridge university press , cambridge , 1995 ) . to show that ( [ epni1 ] ) implies ( [ epni2 ] ) , assume ( [ epni1 ] ) is true : @xmath178 & = & \eta n(\hat{\rho}_{\tilde{\boldsymbol a } } ) + ( 1-\eta ) n(\hat{\rho}_{\tilde{\boldsymbol b } } ) \label{deriv12}\end{aligned}\ ] ] now , if @xmath179 is the joint density operator of the @xmath119 and @xmath120 modes , we find that the state of the @xmath180 modes is @xmath181 , where @xmath157 is a thermal state with average photon number given by @xmath182 , so that @xmath183 $ ] .
thus , from ( [ deriv12 ] ) we get @xmath184 .
taking @xmath172 of both sides of this inequality completes the proof . _ to show that ( [ epni2 ] ) implies ( [ epni1 ] ) , assume ( [ epni2 ] ) is true : @xmath185 & \ge & g^{-1}(s(\hat{\rho}_{\tilde{\boldsymbol c}})/n ) = g^{-1}[g(\eta n(\hat{\rho}_{\tilde{\boldsymbol a } } ) + ( 1-\eta ) n(\hat{\rho}_{\tilde{\boldsymbol b } } ) ) ] \nonumber \\[.12 in ] & = & \eta n(\hat{\rho}_{\tilde{\boldsymbol a } } ) + ( 1-\eta ) n(\hat{\rho}_{\tilde{\boldsymbol b } } ) \nonumber \\[.12 in ] & = & \eta n(\hat{\rho}_{\boldsymbol a})+ ( 1-\eta ) n(\hat{\rho}_{\boldsymbol b}),\end{aligned}\ ] ] where the inequality is due to @xmath186 being a monotonically increasing function of @xmath187 , and the proof is complete .
assume that ( [ epni1 ] ) is true .
we then have that @xmath188 , so that @xmath189 \ge
ng[\eta n(\hat{\rho}_{\boldsymbol a } ) + ( 1-\eta ) n(\hat{\rho}_{\boldsymbol b } ) ] \\[.12 in ] & \ge & \eta ng[n(\hat{\rho}_{\boldsymbol a } ) ] + ( 1-\eta)ng[n(\hat{\rho}_{\boldsymbol b } ) ] \\[.12 in ] & = & \eta s(\hat{\rho}_{\boldsymbol a } ) + ( 1-\eta)s(\hat{\rho}_{\boldsymbol b}),\end{aligned}\ ] ] where the second inequality follows from @xmath190 being concave , and
the proof is complete . | determining the ultimate classical information carrying capacity of electromagnetic waves requires quantum - mechanical analysis to properly account for the bosonic nature of these waves .
recent work has established capacity theorems for bosonic single - user and broadcast channels , under the presumption of two minimum output entropy conjectures . despite considerable accumulated evidence that supports the validity of these conjectures , they have yet to be proven . in this paper
, it is shown that the second conjecture suffices to prove the classical capacity of the bosonic wiretap channel , which in turn would also prove the quantum capacity of the lossy bosonic channel .
the preceding minimum output entropy conjectures are then shown to be simple consequences of an entropy photon - number inequality ( epni ) , which is a conjectured quantum - mechanical analog of the entropy power inequality ( epi ) form classical information theory . |
the initial mass function ( imf ) of stars is one of the most fundamental and crucial ingredients in models of galaxy formation and stellar evolution .
it determines several key parameters in stellar populations , such as the yield of heavy elements , the mass - to - light ratio , luminosity evolution over time , and the energy input into the interstellar medium .
the determination of the imf is therefore of great astrophysical importance .
the imf of low - mass stars is of special interest in this context , since they contain a major fraction of the stellar mass , and have been hypothesized to contain a significant fraction of the total mass in the universe ( see , e.g. , fukugita , hogan and peebles , 1998 ) . in this paper
we mainly deal with objects having masses less than 0.5@xmath8 , the lowest mass of the detectable objects being as low as 0.025@xmath8 .
these low - mass stars evolve little over the lifetime of the universe , and hence the observed present day mass function of these stars is likely to be a good representation of their imf .
but , the imf at or below the hbml remains poorly known , mainly for two reasons .
first , such objects are faint and hence difficult to detect .
second , the mass - luminosity relation of these objects is uncertain and model - dependent , and hence their mass determination is imprecise .
significant improvements are being made on both these aspects , as described below .
the difficulty caused by their faintness can be greatly alleviated by concentrating on young , low - mass objects since the low - mass stars at or below the hbml are expected to be warmer and more luminous when young , although they rapidly cool and fade with age ( burrows et al .
1997 , dantona & mazzitelli 1997 , baraffe et al .
hence young and nearby open clusters provide a good opportunity to study the low end of the stellar imf , since the suitable combination of their youth and proximity makes it possible to detect objects well below the hbml in these clusters , particularly at near - infrared wavelengths .
the uncertainty in the mass - luminosity relation has been greatly reduced by the tremendous progress in the theoretical models for the evolution of these cool and dense objects over the past few years .
these models play a crucial role in predicting masses of these low - mass objects from the observable quantities like colors and luminosities .
burrows et al .
( 1997 ) have generated models of spectra , colors and evolution of brown dwarfs using nongrey calculations .
their models span the mass range of 0.3@xmath9 to 70@xmath9 ( where @xmath9 refers to the mass of jupiter ) with effective temperatures varying from @xmath71300 k to 100 k. dantona & mazzitelli ( 1997 ) simulate the evolution of objects in the mass range 20@xmath10 .
they describe the star s evolution from the hydrostatic phases of pre - main sequence contraction to the hydrogen burning main sequence phase through deuterium and lithium burning .
baraffe et al.(1998 ) have developed the evolutionary models in the 0.075 @xmath8 to 1 @xmath8 mass range for solar type metallicities based on the nextgen atmospheric models of allard et al .
( 1996 ) , and chabrier et al .
( 2000 ) have extended this study by including dust formation and opacity .
the recent models for dwarfs by marley et al ( 2002 ) take the extra effect of sedimentation into account , which suggest that some of the colors , particularly the sloan i@xmath11-z@xmath11 , can be greatly affected by sedimentation .
the advent of the red - sensitive ccds and 2-dimensional near - ir detectors in the last decade has made it possible to detect such low - mass objects , and there have been numerous imaging surveys targeted towards open clusters to probe the substellar domain .
surveys by wilking et al .
( 1999 ) and luhman et al .
( 1999 ) for @xmath12 ophiuchi , herbig ( 1998 ) and luhman ( 1999 ) for ic 348 , bjar et al .
( 1999 ) and zapatero osorio et al .
( 1999a ) for @xmath0 orionis , zapatero osorio et al .
( 1996 ) and stauffer et al .
( 1999 ) for @xmath13 persei , zapatero osorio et al .
( 1997;1999b ) , bouvier et al . ( 1998 ) and hambly et al .
( 1999 ) for pleiades , hambly et al .
( 1995 ) , pinfield et al .
( 1997 ) and magazz et al.(1998 ) for praesepe , gizis et al .
( 1999 ) and reid & hawley ( 1999 ) for hyades , and barrado y navascus et al .
( 2001a ) for ic 2391 are to name a few . the recent release of the 2mass catalogue in the near - infrared wavelengths with a limiting magnitude of @xmath1 , and the latest ( development ) version of the guide star catalogue ( gsc ) with a limiting magnitude of @xmath14 form an ideal combination to study low - mass objects in nearby open clusters using a statistical approach .
we have used these two datasets in conjunction with the recent evolutionary models to isolate the low mass members of the clusters .
the cluster membership is not ascertained by follow - up spectroscopy or proper motion studies .
but the background / foreground contamination is accounted for statistically by studying nearby control fields . in 2 , we review some previous work on the derivation of the mass function ; in 3
, we describe the rationale of our sample selection and the details of the 3 individual clusters selected for this study ; in 4 , we describe the data , the procedure adapted in selecting the cluster members and the method used for their mass determination ; in 5 , we describe the specific selection criteria used for the individual clusters and the slope of the resultant mass function ; and we end with a discussion of the results in 6 .
we plan to extend this work to more clusters in the future .
the stellar mass function is defined as the number density of stars per unit mass bin , and is universally represented as @xmath15 the stellar imf for the high - mass stars ( @xmath16 ) has been long established and well studied following the pioneering work of salpeter ( 1955 ) .
salpeter ( 1955 ) derived the imf from the luminosity function of the present day field stars assuming a constant rate of star formation and correcting for the stellar evolution . in linear units the salpeter mass function is given by @xmath17 where @xmath13 = 2.35 for stars in the mass range 110 @xmath8 .
the steep slope of the imf indicates that the low - mass stars greatly outnumber their high - mass counterparts and account for the major fraction of the stellar mass .
miller & scalo ( 1979 ) and scalo ( 1986 ) rederived the stellar imf by extending the study to the subsolar domain .
their derived values of @xmath13 are 1.4 , 2.5 and 3.3 for the mass ranges 0.1@xmath18 , 1.0@xmath19 and 10@xmath20 , respectively .
there have been several studies recently to extend the stellar mass function to lower masses .
kroupa et al ( 1993 ) took the age and metallicity dependence of the mass - luminosity relation into account in determining the mass functions .
their derived values of @xmath13 are 2.7 for stars more massive than 1 m@xmath5 , 2.2 in the mass range @xmath21 and @xmath22 in the range @xmath23 .
chabrier ( 2001 ) has derived the stellar mass function of the galactic disk stars down to the vicinity of the hbml using their luminosities determined from parallax measurements .
his study indicates that the mass function may be slightly better represented by a log - normal or exponential form than the power - law form given by eq .
his results suggest that the mass function flattens out below 1 m@xmath5 but keeps rising down to the bottom of the main sequence which is consistent with the results of kroupa et al ( 1993 ) .
all these studies suggest a flattening of the imf at the low mass end . however , the nature of the mass function below the hbml , and particularly its spatial variation if any , is still poorly known . on the other hand , to understand the galactic structure , it is important to know the temporal and spatial variation of the imf . according to the study by adams & fatuzzo ( 1996 )
, the imf shows a low mass turn over and they attribute this turn over to the suppression of the very formation of low - mass objects by winds .
models by price & podsiadlowski ( 1995 ) suggest that the stellar imf of a cluster should depend on the cluster density because of the stellar interactions during the process of accretion .
whereas , larson ( 1992 ) links the dependence of the stellar imf on the geometrical structure of star forming clouds and in particular , he attributes the power law form of the upper imf to the presence of hierarchical or fractal structure in these clouds .
the fragmentation of these filaments is predicted to yield a minimum stellar mass of @xmath24 .
the searches for low - mass objects in young clusters by various groups mentioned in 1 have also found many hundreds of candidates with estimated masses below the hydrogen - burning limit of @xmath25 .
in addition , some of these searches have also discovered free - floating objects with inferred masses possibly below the deuterium - burning limit of @xmath26 ( see , e.g. lucas & roche 2000 ; zapatero osorio et al .
the scenarios proposed for the formation of such low - mass objects generally involve ejections from a multiple system .
reipurth & clarke ( 2001 ) argue that brown dwarfs are not formed as a result of low mass core collapse but they are ejected stellar embryos from multiple systems for which the star formation process was aborted before the onset of hydrogen burning .
this ejection model could appreciably explain the flattening of the imf at low masses .
boss ( 2001 ) has suggested that , under certain conditions , the same collapse and fragmentation process that produce single and multiple stars can also produce planetary - mass , self - gravitating objects .
the ejection of these fragments from an unstable , protostellar system may explain the formation of the free - floating planetary - mass objects found in some young clusters .
reid et al . ( 1999 ) derive the mass function of the field stars using the 2mass and the denis data within 8 pc of the sun in the mass range 0.1 1 @xmath8 .
their study shows that in the lower mass range the mass function is flatter ( @xmath27 ) , the value of @xmath13 being closer to the lower limit . extrapolating their results to the substellar regime down to 0.01 @xmath8
they predict that the ratio of the number of brown dwarfs to the main sequence stars is 2:1 in the solar neighborhood .
their study implies that brown dwarfs contribute less than 15% of the total mass of the disk and hence they are unlikely to be the major constituent of galactic dark matter .
however , their mass function is necessarily derived from a mixed sample of stars formed at different times and different environments . studying the stellar imf in different clusters
provides a means to check whether the imf in clusters is different from that of the field population and also to check the universality of the imf .
low mass objects can , in principle , escape a cluster due to cluster dynamics and galactic tides .
in addition , the internal velocity dispersion may result in mass segregation in the cluster which is expected to increase with cluster age ( spitzer & mathieu 1980 ; kroupa 1995 ) .
however , clusters younger than one relaxation time should not show these effects .
mass segregation has been observed in very young ( @xmath28 1 myr ) clusters such as the orion trapezium ( hillenbrand & hartman 1997 ) , but these are most likely the result of the formation process rather than the dynamical evolution .
the typical relaxation time of clusters like pleiades is @xmath7 10 myr ( raboud & mermilliod 1998 ) .
hence , clusters with ages younger than @xmath7 10 myr would be ideal to understand the mass function at the low mass regime , as these clusters are not old enough to have lost members due to stellar evolution , or to have suffered mass segregation due to dynamical effects such as evaporation or violent relaxation ( lada & lada 1991 and references therein ) .
moreover , young clusters are ideally suited for the detection of very low mass objects and brown dwarfs and hence for deriving the mass function in the substellar domain .
this fact can be better appreciated in figure 1 which shows the @xmath29 magnitudes as a function of age and mass of the low - mass objects .
the figure , which uses the models of baraffe et al .
a larger set of model isochrones spanning a more extensive grid of ages and masses are available electronically via anonymous ftp from _ ftp.ens - lyon.fr_. ] and chabrier et al .
( 2000 ) , shows that objects with masses less than 0.1 m@xmath5 are more than 5 magnitudes brighter when they are 1 myr old than when they are 1 gyr old , which makes them more easily detectable .
the main objective of the present study is to derive the stellar mass function down to very low masses for a sample of young open clusters , in an attempt to obtain a global view of the mass function of low - mass objects in such young clusters .
low mass stars ( @xmath30 ) have effective temperatures below 3500 k ( berriman & reid 1987 ) and brown dwarfs ( @xmath31 ) have temperatures below 2800 k ( chabrier et al . 2000 ) , which implies that the spectral energy distribution of these objects peaks in the near infrared between 1 3 @xmath32 m .
hence , the near ir bands are ideal for detecting these objects .
we use the data from the 2mass second incremental release and the latest version of the gsc catalogue .
this 2mass release covers @xmath33 47% of the sky in @xmath34 near - ir bands .
fortunately , it covers the major area of the three clusters and their corresponding control fields chosen by us .
the incomplete coverage of our fields by the 2mass data is accounted for by appropriate normalization as explained in 5 . the latest version of the gsc catalog is an all - sky catalog of stars in the @xmath35 and @xmath36 bands with limiting magnitudes of 22 and 21 in each passband respectively .
( currently , this catalog has restricted access , and is in its final stages of preparations for release , expected to be in late 2002 , mclean private communication ) .
the red plates used in making the gsc catalogue had used slightly different emulsions for the northern and the southern surveys . accordingly , the f - passbands are slightly different for the northern and the southern sources , which we have taken into account in our analysis of the three clusters .
the 2mass catalogue has roll over limiting magnitudes of 15 in @xmath37 , 15.5 in @xmath38 and 16.5 in @xmath39 ( the photometric accuracies are shown in fig . 7 of the documentation pages of the 2mass second incremental release ) .
this allows us to probe low - mass objects down to about 0.03 @xmath40 of all nearby clusters ( d @xmath41 200pc ) with ages less than about 100 myr ( see fig .
2 ) . unlike most other previous studies which rely on confirming the candidate low - mass objects through spectroscopic observations
, we use a statististical approach to estimate the number of low - mass objects . in a statistical approach
, it is important to use several control fields close to each cluster to subtract the contribution of foreground and background objects .
the nature of these two extended surveys enables us to use several such control fields and we are not limited by our choice of the field sizes for the clusters or the control fields . to establish the viability of a statistical approach ,
it is important to apply this procedure for a few well - studied clusters .
the clusters we have chosen for this work are pleiades , ic 348 and @xmath0 orionis , all of which are young and nearby .
the masses of the objects corresponding to the faint limiting magnitudes in each passband for the 3 clusters are given in table 1 .
for ic 348 and @xmath0 orionis , a member with mass @xmath7 0.025 @xmath8 would have a @xmath42 magnitude @xmath7 15 and for the relatively older cluster pleiades , @xmath43 15 corresponds to objects with masses @xmath7 0.04 @xmath8 .
pleiades and ic 348 have been well - studied through ground - based optical and near - ir photometry and optical spectroscopy , and hence it would be useful to check the consistency of our results with that of the previous studies .
previous studies of @xmath0 orionis are confined to photometry in optical wavelengths , but our study includes the near ir data . a brief description of the individual clusters is given below .
the details of the derivation of mass functions and comparison with other studies will be discussed in a later section .
ic 348 is a relatively dense , rich and compact cluster at a distance of @xmath7310 pc , with a size @xmath7 20@xmath11 ( luhman 1999 ) and located in one end of the perseus molecular cloud .
herbig ( 1998 ) found a significant spread in the age of the cluster ( 0.7 12 myr ) based on imaging observations in @xmath44 .
ground based @xmath45 imaging by lada & lada ( 1995 ) revealed a central subcluster with a radius of 0.5 pc containing half the cluster members .
luhman et al .
( 1998 ) derived an age spread of 5 10 myr for this subcluster which was confirmed by najita et al .
there have been several surveys to detect and catalog the low mass members of this cluster ( luhman et al .
1998 ; luhman 1999 ; najita et al .
luhman et al .
( 1998 ) obtained infrared ( @xmath29 band ) and optical spectroscopy and @xmath45 photometry of the stellar population within the @xmath46 core of this cluster to study the star formation , disk properties and the mass function . in a continuing program to identify and characterize
the low mass stellar and substellar populations in this cluster , luhman ( 1999 ) carried out a wide and deep photometric survey in @xmath47 and @xmath48 covering a total area of @xmath49 .
low resolution optical spectroscopy of a subset of the candidate substellar objects found from the survey was performed to confirm their membership . in conjunction with the evolutionary models of dantona & mazzitelli ( 1997 ) and baraffe et al .
( 1998 ) , luhman ( 1999 ) concluded that the hbml for this cluster occurs at a spectral type of m6 and that several objects found in the survey fall below the substellar boundary with masses as low as 20 - 30 @xmath9 .
najita et al .
( 2000 ) , using the @xmath50/nicmos narrow band imaging , investigated the low mass population down to the deuterium burning limit .
their study spans a mass range of @xmath7 0.7 @xmath8 to 0.015 @xmath8 .
the orion complex is one of the richest star forming regions in our galaxy .
@xmath51 observations in the orion belt led to the discovery of a large sample of x - ray sources near the bright young multiple star @xmath0 orionis ( walter et al . 1994 ) which belongs to the orion 1b association .
subsequent photometry and spectroscopy of the x - ray sources detected by @xmath51 revealed the existence of a young stellar cluster ( wolk 1996 ) .
this cluster is located at a distance of 352 pc .
the age of this cluster is in the range of 1 5 myr ( bjar et al .
1999 and references therein ) and its estimated size is 25@xmath11 ( lynga 1983 ) .
bjar et al .
( 1999 ) present the ccd observations in the @xmath47 , @xmath48 and @xmath52 bands covering a total area of 870 @xmath53 around @xmath0 orionis , which led to the detection of objects with masses down to @xmath70.02 @xmath8 . by combining results from imaging surveys and
follow up photometric and spectroscopic observations , bjar et al .
( 2001 ) identified 64 very low mass members in this cluster , in the mass range 0.2 @xmath8 to 0.013 @xmath8 .
barrado y navascus et al .
( 2001b ) have recently reported spectroscopic observations of planetary mass candidates in this cluster .
pleiades is by far the best studied open cluster for low - mass stars . with an age of @xmath7120 myr and a distance of 125 pc
, it is an ideal hunting ground for low mass stars and brown dwarfs .
imaging surveys to detect brown dwarfs began almost a decade ago ( jameson & skillen 1989 ; stauffer et al .
1989,1994 ; simons & becklin 1992 ; rebolo et al .
1995 ; cosburn et al . 1997 ; zapatero osorio et al .
1997,1999b ; bouvier et al .
1998 ; festin 1998 ) .
zapatero osorio et al . ( 1997 ) reported the @xmath54 observations of some of the least luminous members of pleiades and proposed a substellar limit at spectral type m7 . in a deep @xmath55 survey covering an area @xmath7 1 @xmath56 in the central region , zapatero osorio et al .
( 1999b ) detected substellar candidates ranging in mass from 0.075 @xmath8 to 0.03@xmath8 .
bouvier et al .
( 1998 ) performed a wide field imaging survey of the pleiades cluster in the @xmath47 and @xmath48 passbands covering a large area of @xmath7 2.5 @xmath56 to a completeness limit of @xmath57 23 and @xmath58 22 .
they found 17 objects satisfying the criteria for brown dwarfs ranging in mass from the hbml down to 0.045@xmath8 .
as mentioned earlier , we have used the data from the 2mass sky survey and the recently made available guide star catalogue in this study . in selecting the 2mass sources ,
we have constrained our sample to include only those sources for which the error in the @xmath42 band is less than or equal to 0.15 mag to avoid the noisy data at the faint end .
this would translate to a signal - to - noise ratio of @xmath597 in @xmath42 .
this limit ensures positive detection in the other two bands as well .
we have also taken into account the different flags so that our sample is not affected by the spikes of nearby bright stars , contamination from extended sources , or saturation in any of the three bands .
care is taken to exclude asteroids and minor planets , identified by the 2mass catalog .
we merged the 2mass data with the gsc sources by taking the 2mass coordinates , and cross - correlating them with the gsc catalogue .
for such a correlation , it is important to take an appropriate search radius which takes into account the uncertainty in the coordinates .
the uncertainty in the coordinates mainly comes from the frame of reference used in the two catalogs .
comparison of hipparcos data with the gsc data reveals that the gsc coordinates can differ by as much as @xmath60 from the coordinates in the hipparcos reference frame that 2mass uses ( bakos , sahu and nemeth , 2002 ) .
a search radius of @xmath60 was found to be an optimum value for this cross - correlation , which was small enough to reject spurious and multiple detections which could be the result of diffraction spikes due to bright stars in the field , and large enough to include any positional uncertainties in the two catalogs . assigning cluster membership to sources
is a difficult task .
the color magnitude diagrams [ cmds ] for various clusters given in the literature show that the density of faint members is usually comparable or even significantly less than the field / background / foreground population ( barrado y navascus et al.(2001a ) for ic 2391 , barrado y navascus et al.(2001c ) for m35 etc . )
the use of cmds can be used as an effective tool to delineate possible members from the field contaminants .
of course , proper motion and/or spectroscopic studies are needed to establish the membership of specific sources .
the cmds best suited for this purpose can be determined by examining the locations of sources with different masses and temperatures in different cmds , and the slope and direction of the low - mass main sequence and the reddening vectors .
an example @xmath42 versus @xmath61 cmd is shown in figure 3 where the dots represent the stars in the direction of the pleiades cluster . to differentiate the line - of - sight contaminants from the cluster members we need a reliable locus for the low - mass main sequence .
these open clusters are located in the spiral arm of the galaxy which comprises of the young disk population . to account for the contaminants ,
we construct a @xmath62 locus for the young - disk objects using the empirical data from table 6 of leggett ( 1992 ) . the solid line in fig .
3 shows such a disk sequence for all stars with spectral types m0 ( corresponding to 0.5 m@xmath5 ) and later as taken from leggett ( 1992 ) . here
, we have appropriately shifted the leggett sequence to take the distance and the extinction of the pleiades cluster into account .
the lowest two points in this line are identified as _ young - old _
( y / o ) by leggett ( 1992 ) .
we derive the reddening vectors from the interstellar extinction laws of rieke & lebofsky ( 1980 ) and bessell & brett ( 1988 ) ; these reddening vectors are shown as dashed lines in the fig .
we have chosen to plot the young - disk population because such a plot is likely to be most useful in rejecting the older non - member stars along the line - of - sight which are expected to fall to the left of this sequence owing to metallicity effects . as seen from the figure , the stars with spectral types m0 and later fall more or less vertical in this diagram , and the @xmath63 color changes by @xmath7 0.35 over this entire spectral range .
the reddening vector runs almost perpendicular to this sequence with a slope of @xmath70.67 .
since the effect of the distance is to move a particular source vertically in this diagram , there is a clear degeneracy between mass and distance for the mass range from 0.6 @xmath8 to @xmath7 0.06@xmath8 ( leggett 1992 ) . for lower masses ,
the extinction vector runs almost parallel to the direction of the leggett sequence creating a degeneracy between mass and extinction in this spectral range .
the inclusion of one optical color in the cmd [ e.g. @xmath36 versus @xmath64 overcomes the aforementioned degeneracies as explained below .
one problem , however , is that the original data of leggett ( 1992 ) give the absolute magnitudes and the intrinsic colors for the standard broad band filters in the cousins , johnsons and the cit systems , and not in the f - passband .
so we need to first convert the r magnitudes to f magnitudes by applying appropriate @xmath65 color corrections . for this purpose , we compared the leggett young disk sequence with model isochrones of different ages in a @xmath36 versus @xmath65 plot and found that the 100 myr reasonably fits the legett young disk sequence as expected ( e.g. allen 1973 ) .
so we used the 100 myr model isochrones in f and r to calculate the values of @xmath65 , and applied this correction to the empirical @xmath47 values derived from leggett ( 1992 ) .
the resulting cmd in @xmath36 versus @xmath65 is shown in fig.4 , where the color increases rapidly with spectral type ( @xmath7 3 mags ) and the reddening vector is steeper with a slope of @xmath71.59 .
use of this plot avoids the degeneracies found in the @xmath62 cmd by minimizing the overlap between the reddened background stars and the low mass members of the cluster , and offers an efficient rejection criterion for non - members .
hence we have used the @xmath36 vs. @xmath66 cmd to delineate field stars from the cluster members .
figure 4 shows a clear separation between the cluster and non - cluster members , which justifies the use of this cmd .
while rejecting the field star population , care has to be taken not to reject genuine candidates from the sample .
the disk population in the spiral arms are the field contaminants in the line - of - sight .
as will be shown later , the main sequence from leggett ( 1992 ) , corresponding to this young disk population , falls to the left of the theoretical isochrones from baraffe et al .
( 1998 ) for the clusters ic 348 and @xmath0 orionis .
hence , this sequence from leggett ( 1992 ) is preferred over the other for the rejection criterion for these two clusters . for the pleiades cluster , which is @xmath7100myr old
, the two tracks overlap for most of the masses but the baraffe model falls to the left for some part of the low mass sequence and hence is preferred over the leggett main sequence to reject non members , and at the same time to ensure that none of the genuine members are rejected . note that throughout the selection process , we have erred on the side of including some contaminants while ensuring that we do not reject _ any _ genuine low - mass objects .
the price we pay in following this conservative approach is that our sample will inevitably include some foreground / background contamination .
this is , however , a small price to pay since the contamination can be corrected through observations of the control fields . on the contrary , if some of the genuine objects are rejected , the effect can be disastrous since the observed number of stars would then be smaller than the actual number , thus making the statistics potentially incorrect .
it is worth noting here that , since we have followed this conservative approach , the sources lying between the leggett main sequence and the theoretical isochrone should be treated with caution . even though the @xmath36 vs. @xmath66 cmd is a better choice to separate the cluster members from the non - members ,
we still see a smooth transition between the field stars and the cluster members in fig .
4 which implies that our cluster sample would still be contaminated . this contamination could be due to the background reddened stars as well as the foreground population .
we need to correct for this contamination before we can derive a reliable mass function .
this becomes especially important in our study since we are not carrying out follow - up spectroscopic or proper motion studies to confirm the membership of individual members .
as we will demonstrate in detail later , the contamination can be satisfactorily taken into account by studying the properties of off - fields or control fields in the vicinity of the clusters .
we have tried to use several control fields , each covering the same area as the cluster , which are located near the cluster and are free from anomalies .
these control fields were selected by visually inspecting the star density from the poss plates . for ic 348 ,
we have used the co maps of bachiller & cernicharo ( 1986 ) to find suitable control fields .
table 2 gives the location and sizes of the cluster and control fields used in our study .
the masses of our selected candidates are determined by comparing the observed magnitudes with those predicted by the evolutionary models .
we have used the evolutionary tracks of baraffe et al .
( 1998 ) for this purpose , as they provide the magnitudes and colors as a function of mass for various ages in the passbands of interest .
the magnitudes in the f - bands were specially calculated and were kindly provided to us by baraffe and allard at our request , which are used in this analysis .
whereas , in order to transform the effective temperatures and luminosities of other models we need to use the bolometric corrections of the baraffe model . additionally , the tracks by baraffe et al .
( 1998 ) have been successful in fitting the mass - luminosity relation in various optical and infrared passbands and predicting coeval ages for members of several young multiple systems ( white et al .
1999 ; luhman 1999 ) .
it is also seen that these models provide good fits to the infrared photometric sequence in the pleiades and @xmath0 orionis clusters ( martn et al .
2000 ; zapatero osorio et al .
it is worth emphasizing here that the mass determinations are model dependent and are subject to possible systematic effects arising from the different parameters used in the models .
in this section we discuss the selection criteria and the resulting mass functions derived for the individual clusters . to sample the entire cluster , we consider all the sources within a radius of @xmath67 around the cluster center of ic 348 .
the wide and deep study of luhman ( 1999 ) also covers almost the entire cluster , and includes optical photometry and follow - up spectroscopy .
the study by luhman et al .
( 1998 ) includes both optical and near - ir observations for the central @xmath68 of the cluster .
since our study includes both optical and near - ir photometry , ic 348 provides an ideal opportunity to compare the results , and to check the consistency between the different approaches . for this cluster ,
we derive the value of interstellar reddening from the sample of 70 confirmed low mass cluster members of luhman ( 1999 ) .
the observed distribution of the extinction values listed in luhman ( 1999 ) ranges from @xmath69 to @xmath70 with a peak around @xmath71=0.3 magnitudes .
we have used the lower value of @xmath72 magnitudes in our selection criteria to ensure that none of the genuine cluster members are rejected .
the fact that some members of ic 348 have zero extinction is also seen in the data of najita et al .
( 2000 ) where there are a few sources with @xmath73 .
5 shows the @xmath36 versus @xmath66 plot for all the sources in this region of the cluster and the two control fields .
the figure also shows the main sequence from leggett ( 1992 ) and the theoretical isochrone for 5 myr from baraffe et al .
( 1998 ) appropriately scaled to a distance of 316 pc ( herbig 1998 ) and an interstellar reddening of @xmath71 of zero .
the number of sources increases to the left of the main sequence locus which can be attributed to the field star contamination .
hence , the objects bluer and fainter than the leggett main sequence are rejected , giving us the first criterion for the choice of candidates , which is derived by fitting a straight line to the data points of leggett ( 1992 ) @xmath74 we then use the model isochrones from baraffe et al.(1998 ) in other cmds appropriately scaled to the distance and extinction of ic 348 , to select the low mass members of the cluster . as
discussed in section 3.1 , this cluster exhibits appreciable spread in the age .
we adopt a mean age of 5 myr and use the magnitudes and colors of the baraffe models to formulate the following criteria for selecting the objects with mass @xmath75 .
@xmath76 @xmath77 @xmath78 @xmath79 the sources satisfying all the four criteria are designated as potential low - mass members of the cluster and the sources in the control fields which satisfy these criteria are used to correct for the possible contaminants .
6 shows the @xmath42 versus @xmath61 plot of the sources satisfying the low mass selection criteria .
the closed circles represent the sources from the cluster and the open triangles are the candidates from the control fields which represent the possible contaminants .
the plot also shows the theoretical isochrone for a 5 myr cluster scaled to the distance and reddening of ic 348 .
as one can see , there are a few sources from the control samples to the right of the theoretical isochrone .
we correct for this contamination while deriving the mass function .
also plotted are the confirmed low mass members from the luhman ( 1999 ) sample ( open circles ) .
the fact that they all fall in the region satisfying all our selection criteria proves the validity of this approach .
this figure also suggests low internal extinction in this cluster . if the internal extinction were high , the stars in the near side of the cluster would undergo less extinction , and the stars on the far side would undergo more extinction . as a result ,
the observed stars would be expected to fall on both sides of this theoretical main sequence .
thus , the fact that very few stars lie to the left of the main sequence implies that most of the extinction must be foreground rather than internal . on the other hand
, we do observe many stars on the right side of the theoretical sequence .
since the internal extinction is not large , this would argue that the most of the observed extinction is not interstellar , but circumstellar . in a recent work muench et al .
( 2001 ) find , from the analysis of the @xmath45 colors , that @xmath750 % of the brown dwarf candidates in the trapezium cluster display significant infrared excess .
this suggests that these sources are extremely young and provides independent confirmation of their cluster membership and low mass nature .
7 shows the @xmath80 versus @xmath81 color - color diagram for the low mass stars of the cluster and the control fields . in this figure , the region to the left of the redenning band
is forbidden for young stellar objects .
hence , the location of sources in this region is most likely due to the combined uncertainties in the models , the derived magnitudes and the extinction characteristics .
the region to the right of the redenning band is occupied by sources with infrared excess ( lada & lada 1995 ) .
this is in agreement with the discussion on the circumstellar extinction . to derive the mass function , we construct mass bins of 0.01 @xmath8 in the mass range 0.02 0.09 @xmath8 .
as the number of candidates decreases for higher masses , the width of the bin is increased to 0.1 @xmath8 for the mass range 0.1 0.5 @xmath8 .
the members of the cluster field and the two control fields satisfying the five selection criteria are grouped into these mass bins .
the grouping is done based on the model - predicted @xmath29 magnitudes for these mass bins taking the distance , age and mean extinction of ic 348 into account .
the number of sources in each mass bin is then normalized with the total number of sources detected in that field .
this is essential in order to correct for the unequal area covered by the cluster and the control fields , often caused by incompleteness of the survey .
the mean percentage of the number of sources present in each mass bin of the control fields quantifies the contamination one expects .
this percentage contamination is then removed from the cluster bins and divided by the bin width to obtain the value of @xmath82 .
this as a function of the mass gives the mass function ( eqn 1 & 2 ) .
8 shows the resultant mass function for ic 348 .
the closed circles are based on the models of baraffe et al .
( 1998 ) and the open triangles are derived using the models of chabrier et al .
( 2000 ) which include dust opacities .
the error bars in @xmath83 show the @xmath84 errors involved in the counting statistics .
the least squares fit to the data points derived from the baraffe et al .
( 1998 ) model yield a slope of @xmath85 ) , with an uncertainty of @xmath86 .
in deriving this slope , we have rejected the lowest mass bin since this is affected by the limiting magnitude of the 2mass survey . as seen from this figure the results from the models of chabrier et al .
( 2000 ) are in agreement with the non - dusty models of baraffe et al .
( 1998 ) within the given scatter .
the mass function derived here is consistent with that derived by najita et al .
( 2000 ) and luhman et al .
( 1998 ) . using the tracks from dantona & mazzitelli ( 1997 ) ,
luhman et al . ( 1998 ) derive a mass function which , in logarithmic units , slowly rises from the hbml to @xmath70.25 @xmath8 with a slope of @xmath87 . converting their results to our system ( eqn . 1 & 2 ) yields a value of @xmath13 = 0.6 .
najita et al .
( 2000 ) derive the mass function using the models of baraffe et al .
( 1998 ) as well the evolutionary tracks of dantona & mazzitelli ( 1997 ) .
they derive mass functions with @xmath13 = 0.5 and @xmath13 = 0.4 respectively for the two models .
we have assumed a mean age of 3 myr from the age estimates of bjar et al .
( 1999 ) for @xmath0 orionis .
lee ( 1968 ) finds a low extinction of @xmath88 = 0.05 for the multiple star @xmath0 orionis which suggests that the reddening of the associated cluster is small .
we also assume this value of @xmath88 in our analysis as the value of the foreground extinction for the cluster .
we have covered an area of @xmath7 0.8 @xmath56 centered around the star @xmath0 orionis .
taking the distance to be 352 pc ( bjar et al .
1999 ) , we derive the following criteria for field star rejection and selection of low mass members @xmath89 @xmath90 @xmath91 @xmath92 @xmath79 fig .
9 - 11 show the different cmds and color - color plots for @xmath0 orionis .
as is evident from the figures , the contamination for this cluster is higher than that of ic 348
. the possible reason might be the location in the orion complex and inadequate correction for variable extinction in the cluster and the control field .
the derived mass function is shown in fig .
12 . as seen in the figure
there is a sudden rise in the mass function at 0.045 @xmath8 , which could be because of the large contamination from the background sample .
the lowest mass bin is affected by the limiting magnitude of the survey .
hence , these two lowest mass bins were excluded in deriving the slope of the mass function .
the resultant mass function has a slope of @xmath13 = [email protected] .
bjar et al .
( 2001 ) have previously derived the mass function for this cluster in the mass range from 0.2 @xmath8 to 0.013 @xmath8 based on deep photometry in i , z , j and k bands .
they obtain a slope of @xmath13 = [email protected] assuming the age to be 5 myr .
our study , which covers a larger area and includes photometry in all the 2mass near - ir bands , is consistent with the results of bjar et al .
( 2001 ) . for pleiades
, we have selected four fields of 1 deg radius each and centers offset by a degree from the cluster center towards the north , south , east and west .
these four fields are then combined together to one field with an effective radius of 1.5 deg centered around the cluster center and hence amounts to a total area of @xmath77 @xmath56 .
in contrast to many of the previous studies , our study thus covers a large area and includes the photometry in optical as well as the near - ir wavelengths .
the control fields are generated in similar fashion .
as compared to ic 348 and @xmath0 orionis , pleiades is an older cluster with an age estimate varying between @xmath93 myr . in our analysis , we assume an age of 100 myr ( bouvier et al .
( 1998 ) , a distance of 125 pc ( bouvier et al . 1998 ) and a redenning of @xmath94=0.12 ( crawford & perry 1976 ) for the cluster .
13 shows the @xmath36 versus the @xmath66 diagram for pleiades . because of the high galactic latitude of this cluster one sees a better separation between the field and the cluster members unlike that in the other two clusters .
as explained in section 4.2 , we have used the locus of the baraffe model to set a criterion for rejection of the field population .
a 4th order polynomial best fits the theoretical isochrone for a 100 myr cluster .
the other criteria are derived analogous to the previous two clusters .
the resultant criteria are as follows .
@xmath95 @xmath96 @xmath91 @xmath97 @xmath98 fig .
13 - 15 show the color magnitude and the color color plots for pleiades .
16 shows the derived mass function .
the linear fit to the data points yields a value of @xmath13=0.5 .
this is in good agreement with the results of bouvier et al .
( 1998 ) who derive a mass function which still rises below the hbml with @xmath13=0.6 .
martn et al .
( 1998 ) , from a compilation of independent photometric survey covering different areas of the cluster , derive a steeper slope in the mass range 0.04 0.25 @xmath8 with @[email protected] .
hambly et al .
( 1999 ) derive a value of @xmath13=0.7 from a @xmath47 and @xmath48 survey covering an area of 6 @xmath99 6 degrees centered on pleiades .
it is worth noting here that pleiades is a relatively older cluster , where the mass segregation and the escape mechanism of low - mass stars described earlier might have played a role .
this may be the reason for the flatter slope compared to the other two clusters , and may also also explain the observed discrepancies in various estimates of the slopes indicating that the mass function may not be uniform over the entire region of the cluster .
recent surveys have found a significant population of low mass stars , brown dwarfs and planetary mass objects in young open clusters . we have adopted a statistical approach to determine the mass spectrum , @xmath100 , of objects in the mass range 0.5@xmath8 to 0.025 - 0.055@xmath8 , using the data from the recently released 2mass and the gsc catalogues .
unlike some of the previous studies , our study makes use of both the optical data as well as the near - ir data .
since these datasets cover a large portion of the sky , they allow us to study the entire area covered by each cluster . as a result ,
the areas covered in our study are generally larger than the areas covered in most of the previous studies . these datasets also allow us to apply a statistical approach to efficiently subtract the background contribution using several control fields close to the cluster .
we carried out a detailed study for ic 348 .
fo this cluster , we derived the mass functions using the solar metalicity model of baraffe et al .
( 1998 ) and the dusty models of chabrier et al .
( 2000 ) , both of which gave very similar results .
we also compared the confirmed low - mass members from luhman ( 1999 ) with our isolated low - mass candidates which further strengthened the validity of this technique .
we then used the same technique to @xmath0 orionis and pleiades .
the resultant slopes of the mass functions for ic 348 , @xmath0 orionis and pleiades are 0.7 , 1.2 and 0.5 respectively , with an estimated error of @xmath60.2 . for ic 348
we have used the mass range from 0.5@xmath8 to 0.025@xmath8 in deriving the slope of the mass function whereas for @xmath101 orionis and pleiades , the lowest mass bins correspond to 0.045 @xmath8 and 0.055@xmath8 , respectively .
taking into consideration the effect of mass segregation and preferential loss of low mass members from the cluster , the mass function derived here for the inner 1.5 deg radius of pleiades cluster could be considered as a lower limit to the true mass function . as discussed under individual sections ,
the mass functions derived here are in good agreement with the value derived by other groups based on studies of confirmed low mass members of these clusters , which demonstrates the consistency of different approaches . within the uncertainties , the values of @xmath13 derived for the ic 348 and @xmath0 orionis
is in agreement with that derived for the local sample of low - mass stars ( reid et al .
the derived slopes imply that the mass spectrum continues to rise well below the hbml , but the mass functions are appreciably flatter in the low - mass regime than the salpeter mass function .
the results are summarized in table 3 .
taking the salpeter exponent of 2.35 in the mass range 1 10 @xmath8 , the chabrier exponent of 1.55 in the mass range 0.5 1 @xmath8 , and the values obtained by us below 0.5 , we calculate the mass contribution to be about 40% for objects below 0.5 m@xmath5 , and about 4% for objects below the hbml of 0.08 m@xmath5 .
( note that the contributions are not sensitive to the choice of the slope in the higher mass regime .
for example , if we use the value of @xmath13 as 2.7 for m@xmath102 as derived by chabrier ( 2001 ) instead of the salpeter value of 2.35 , the mass contributions change only by @xmath103 ) .
our results are consistent with that of the previous studies ( e.g. bjar et al .
2001 ) , and suggest that , although the low mass stars are at least as numerous as their high mass counter parts ( as seen from figures 5,7 and 9 ) , their contribution to the total mass is small .
the contributions of low - mass objects to the total mass in the clusters seem to be marginally smaller than that of the low - mass objects in the local sample ( e.g. reid et al . 1999 ) , but the slope of the mass function is less steep for the relatively older pleiades cluster .
this is not surprising since the high - mass stars are likely to be preferentially lost in an older and mixed population such as the local sample .
follow up spectroscopic observations to confirm the isolated low mass members of the clusters would further strengthen the results derived from this purely statistical approach .
we would like to thank brian mclean and mario lattanzi ( the gsc - ii project scientists ) for access to the development version of gsc - ii in advance of publication and their technical comments .
we would like to thank the anonymous referee for valuable comments and suggestions .
we are grateful to i. baraffe and f. allard for making the electronic versions of the latest models available and generating the model isochrones for the non - standard f passbands .
we would also like to thank neill reid for useful discussions .
the visit of a. tej to stsci was supported by a ddrf grant of stsci awarded to k. sahu .
the digitized sky surveys and the guide star catalogues were produced at the space telescope science institute under us government grant nag w-2166 ; the images of these surveys are based on photographic data obtained using the oschin schmidt telescope on palomar mountain and the uk schmidt telescope at siding spring .
the 2mass project is a collaboration between the university of massachusetts and the infrared processing and analysis center ( jpl/ caltech ) , the funding for which is provided primarily by nasa and the nsf .
research work at prl is funded by dept .
of space , govt of india .
adams , f.c . , & fatuzzo , m. 1996 , apj , 464 , 256 allen , c. w 1973 , astrophysical quantities , univ . of london the athlone press , p242 allard , f. , hauschildt , p.h . , baraffe , i. , & chabrier , g. 1996 , apj , 465 , l123 bachiller , r. , & cernicharo , j. 1986 , a & a , 166 , 283 bakos , g.a . ,
sahu , k.c . , & nmeth , p. 2002 , apj suppl .
( in press ) ( astro - ph/0202164 ) baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p.h . 1998 , a & a , 337 , 403 barrado y navascus , d. , stauffer , j.r . ,
briceo , c . ,
patten , b. , hambly , n.c . , & adams , j.d .
2001a , apjss , 134 , 103 barrado y navascus , d. , zapatero osorio , m. r. , bjar , v. j. s. , rebolo , r. , martn , e. l. , mundt , r. , bailer - jones , c. a. l. 2001b , a & a , 377 , l9 barrado y navascus , d. , stauffer , j.r . , bouvier , j. , & martn , e.l .
2001c , apj , 555 , 546 bjar , v.j.s . ,
zapatero osorio , m.r .
, & rebolo , r. 1999 , apj , 521 , 671 bjar , v.j.s . , martn , e. l. , zapatero osorio , m. r. , rebolo , r. , et al .
2001 , apj , 556 , 830 berriman , g. , & reid , n. 1987 , mnras , 227 , 315 bessell , m. s. , & brett , j. m. 1988 , pasp , 100 , 1134 boss , a. 2001 , apj , 551 , l432 bouvier , j. , stauffer , j.r . ,
martn , e.l .
, barrado y navascus , d. , wallace , b. , & bjar , v.j.s .
1998 , a & a , 336 , 490 burrows , a. , marley , m. , hubbard , w.b . ,
lunine , j.i . ,
guillot , t. , & saumon , d. 1997 , apj , 491 , 856 chabrier , g. 2001 , apj , 554 , 1274 chabrier , g. , baraffe , i. , allard , f. , & hauschildt , p. 2000 , apj , 542 , 464 cosburn , m.r . ,
hodgkin , s.t . ,
jameson , r.f . , & pinfield , d.j .
1997 , mnras , 288 , l23 crawford , d.l . , & perry , c.l .
1976 , aj , 81 , 419 dantona , f. , & mazzitelli , i. 1997 , mem .
soc . astron .
italiana , 68 , 4 festin , l. 1998 , a & a , 333 , 497 fukugita , m. , hogan , c. j. , & peebles , p. j. e. 1998 , apj , 503 , 518 gizis , j.e . , reid , i.n . , &
monet , d.g .
1999 , aj , 118 , 997 hambly , n.c . ,
hodgkin , s.t . ,
cossburn , m.r . , & jameson , r.f .
1999 , mnras , 303 , 835 hambly , n.c . ,
steele , i.a .
, hawkins , m.r.s . , & jameson , r.f .
1995 , mnras , 273 , 505 herbig , g.h .
1998 , apj , 497 , 736 hillenbrand , l.a . , &
hartman , l.w .
1997 , apj , 492 , 540 jameson , r.f . , & skillen , w.j.i . , 1989 , mnras , 239 , 247 koorneef , j. 1983 , a & a , 128 , 84 kroupa , p. 1995 , mnras , 277 , 1522 kroupa , p. , tout , c.a . , & gilmore , g. 1993 , mnras , 262 , 545 lada , c.j . , & lada , e.a .
1991 , the formation and evolution of star clusters , asp conf .
ser 13 , p3 lada , e.a . , & lada , c.j . 1995 ,
aj , 109 , 1682 larson , r.b .
1992 , mnras , 256 , 641 lee , t.a . 1968 , apj , 152 , 913 leggett , s.k .
1992 , apjs , 82 , 351 lucas , p.w .
, & roche , p.f .
2000 , mnras , 314 , 858 luhman , k.l . ,
rieke , g.h . ,
lada , c.j .
, & lada , e.a .
1998 , apj , 508 , 347 luhman , k.l . 1999 ,
apj , 525 , 466 luhman , k.l . , liebert , j. , & rieke , g.h .
1999 , apj , 489 , l165 lynga , g. 1983 , 5th ceskoslovenska akademie conf . , star clusters and associations and their relation to the evolution of the galaxy , p80 magazz , a. , rebolo , r. , zapatero osorio , m.r . ,
martn , e.l . , & hodgkin , s.t .
1998 , apj , 497 , l47 martn et al .
2000 , apj , 543 , 299 miller , g.e . , &
scalo , j.m .
1979 , apjs , 41 , 513 muench , a.a . ,
alves , j. , lada , c.j .
, & lada , e.a .
2001 , apj , 558 , l51 najita , j.r . ,
tiede , g.p . , & carr , j.s .
2000 , apj , 541 , 977 pinfield , d.j . ,
hodgkin , s.t . ,
jameson , r.f . , cossburn , m.r . , &
von hippel , t. 1997 , mnras , 287 , 180 price , n.m . , & podsiadlowski , ph .
1995 , mnras , 273 , 1041 raboud , d. , & mermilliod , j .- c .
1998 , a & a , 333 , 897 rebolo , r. , zapatero osorio , m.r . , &
martn , e.l .
1995 , nature , 377 , 129 reid , i.n . , & hawley , l.s . 1999 ,
aj , 117 , 343 reid , i.n . ,
kirkpatrick , j.d .
, et al . 1999 ,
apj , 521 , 613 reipurth , b. , & clarke , c. 2001 , apj , 122 , 432 rieke , g.h . , & lebofsky , m.j .
1985 , apj , 288 , 618 salpeter , e.e . 1955 , apj , 121 , 161 scalo , j.m .
1986 , fundam . cosmic phys . , 11 , 1 simons , d.a . , & becklin , e.e .
1992 , apj , 390 , 431 spitzer , l.jr . , & mathieu , r.d .
1980 , apj , 241 , 618 stauffer , j.r . , hamilton , d. , probst , r.g . ,
rieke , g. , mateo , m. 1989 , apj , 344 , l21 stauffer , j.r . , hamilton , d. , probst , r.g .
1994 , aj , 108 , 155 stauffer , j.r .
, barrado y navascus , d. , et al .
1999 , aj , 527 , 219 walter , f.m . ,
vrba , f.j . ,
mathieu , r.d .
, brown , a. , & myers , p.c .
1994 , aj , 107 , 692 white , r.j . ,
ghez , a.m. , reid , i.n .
, & schultz , g. 1999 , apj , 520 , 811 wilking , b.a . , greene , t.p .
, & meyer , m.r . 1999 ,
aj , 117 , 469 wolk , s.j .
1996 , ph.d .
thesis , univ .
new york at stony brook zapatero osorio , m.r . ,
bjar , v.j.s . ,
rebolo , r. , martn , e.l . ,
& basri , g. 1999a , apj , 524 , l115 zapatero osorio , m.r .
, rebolo , r. , martn , e.l . ,
hodgkin , s.t . ,
cossburn , m.r . ,
magazz , a. , steele , i.a . , & jameson , r.f .
1999b , a & as , 134 , 537 zapatero osorio , m.r .
e.l . , & rebolo , r. 1997 , a & a , 323 , 105 zapatero osorio , m.r . ,
rebolo , r. , martn , e.l . , & garca lpez , r.j .
1996 , a & a , 305 , 519 zapatero osorio , m.r . ,
bjar , v.j.s . ,
martn , e.l . ,
rebolo , r. , barrado y navascus , d. , bailer - jones , c.a.l . , & mundt , r. 2000 , science , 290 , 103 | in this paper we present the mass functions in the substellar regime of three young open clusters , ic 348 , @xmath0 orionis and pleiades , as derived using the data from the 2 micron all sky survey ( 2mass ) catalogue which has a limiting magnitude of @xmath1 , and the latest version of the guide star catalogue ( gsc ) which has a limiting magnitude of @xmath2 21 . based on recent evolutionary models for low mass stars ,
we have formulated the selection criteria for stars with masses below @xmath3 . using a statistical approach to correct for the background contamination , we derive the mass function of objects with masses ranging from 0.5@xmath4 down to the substellar domain , well below the hydrogen burning mass limit .
the lowest mass bins in our analysis are 0.025 , 0.045 and 0.055 m@xmath5 for ic 348 , @xmath0 orionis and pleiades , respectively .
the resultant slopes of the mass functions are 0.8 @xmath6 0.2 for ic 348 , 1.2 @xmath6 0.2 for @xmath0 orionis and 0.5 @xmath6 0.2 for pleiades , which are consistent with the previous results .
the contribution of objects below 0.5 m@xmath5 to the total mass of the cluster is @xmath740% , and the contribution of objects below 0.08 m@xmath5 to the total mass is @xmath74% . |
young stellar associations ( @xmath11 50 myr ) and open clusters have a crucial importance in advancing our understanding of star formation and the first stages of stellar evolution . since the galactic acceleration does not have a chance to affect the kinematical properties of these young stellar groups , the stellar content of an association is preserved .
consequently one can obtain kinematical , dynamical and chemical properties of these young stellar groups by studying their secure members .
today nearby associations within the solar neighbourhood have been very well identified .
their stellar content of each association has been precisely determined up to a magnitude limit of @xmath1210.5 mag using astrometric data of the hipparcos satellite ( i.e. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in addition the new reduced hipparcos catalogue @xcite gave an opportunity to investigate the astrometric data of the stellar content of a large number of open clusters , associations and moving groups more accurately .
the first application of the new reduced hipparcos astrometric data is applied to stellar groups within the solar neighbourhood by @xcite .
this and forthcoming studies of young stellar groups will improve our understanding of the history of star formation , the initial mass function , chemical and dynamical evolution of the milky way .
recent statistical studies , such as @xcite , @xcite and @xcite show a high ratio of multiplicity in stellar formation regions ( sfrs ) and claim that it is not a coincidence , but a characteristic of star formation .
the detailed study of multiple systems ( especially those with eclipsing components ) in sfrs will reveal the fundamental stellar parameters more directly and with higher precision compared to those obtained from single stars and thus impose more stringent tests on stellar evolution .
critical tests of stellar evolution require masses and radii with a precision better than 3 per cent ( i.e. @xcite ) .
however , methods developed for single stars are not capable of delivering masses with a precision better than 5 per cent and radii remain uncertain by a factor of 1.5 . consequently studying single stars
does not enable us to obtain accurate dimensions and , therefore , to test the most recent evolutionary models .
recent studies on @xmath13 mus by @xcite in lower centaurus - crux association , on v578 mon by @xcite in ngc 2244 , and on ab dor by @xcite in the ab dor association , demonstrate the precision with which age , chemical composition and kinematical properties can be determined by studying such high - mass systems . in the present study , we analyzed the high resolution spectra and _ bvh@xmath0 _ photometric data of i m mon , which is located in the region of ori ob1a association .
i m mon is a bright ( [email protected] mag ) , early - type ( ( _ b - v_)=0.14 mag ) and short orbital period ( [email protected] days ) eclipsing binary system .
its spectroscopic and photometric variations were discovered by @xcite and @xcite respectively .
the eccentricity of the spectroscopic orbit obtained by @xcite was commented to be spurious by @xcite in their study on the determination of photometric elements of 14 detached systems .
@xcite studied early photometric observations of i m mon , which were collected by @xcite in integral light and by @xcite in @xmath15 and @xmath16 filters . however , due to a large scatter in all photometric observations , which is attributed to the intrinsic variability of one of the components by @xcite , none of the authors was able to find a unique and precise solution for the system .
a recent spectroscopic study of @xcite revealed the spectroscopic orbital elements of i m mon and showed that its orbit is circular . in order to reveal more precise absolute dimensions of i m mon and to test its membership to ori ob1a
, we included it into our list of eclipsing binaries in the region of ob associations . using all literature based data the orbital period of i m mon
is revised in 2.3 .
high resolution spectral lines of i m mon are modelled and atmosphere parameters are derived in 3 .
the close binary stellar parameters of the system are determined by the analysis of light and radial velocity ( rv ) curves in 4 . in 5 the absolute parameters of the components are derived together with the age and distance of the system .
this information enabled us to establish the absolute dimensions of the close binary and properties of the ori ob1a association through the kinematical and dynamical properties of i m mon .
finally we summarized our study and present our conclusions in 6 .
we collected as many original individual measurements of i m mon as possible from the literature . in table
[ listsources ] , all available photometric and spectroscopic observations of i m mon are listed .
all the data given in table [ listsources ] are used for ephemeris determination for o c analysis , whereas only relatively more precise photometric data are used for light curve ( lc ) modelling . as a starting orbital period
, we adopted @xmath17 , which is published by @xcite and later used by @xcite for their radial velocity analysis . [ cols="^,^,^,^,^,^",options="header " , ] + ( a ) @xcite , ( b ) @xcite , ( c)@xcite , ( d)@xcite , ( e)@xcite , ( f ) @xcite , ( g ) @xcite
ob associations are young galactic clusters where star formation is ongoing or has just ended .
the study of ob associations yields useful information about the characteristic of star formation such as formation history , binary population and initial mass function .
however , this information can be obtained only if the properties of the ob association such as distance , age , metallicity and kinematics are very well established .
observing single stars in an ob association does not provide information of sufficient precision unless many of them are observed , which requires a lot of observing time . in this case , eclipsing binaries which are the royal road to the stars can yield precise age , metallicity and kinematics of the medium in which they are embedded , and they do not require much observing time provided that stars with relatively short periods are selected . in this work , using sophisticated modelling tools
, we studied the close binary system i m mon together with all its available photometric , spectroscopic , kinematical and dynamical data .
the membership of i m mon to ori ob1a sub - group has been established securely by means of comparing the dynamical galactic orbits of 29 ori ob1a members with the galactic orbit of i m mon .
the absolute dimensions we derived for i m mon in this study lead to a reliable distance determination .
the location of both components of i m mon in the plane of @xmath18 - @xmath19 is fully compatible with the formerly derived ages for this association .
the metallicity ( @xmath20=$]0.20(0.15 ) dex ) of i m mon obtained in this study has a large uncertainty which may explain the disagreement between the average metallicity ( @xmath20=-$]0.01(0.04 ) dex ) of the onc and i m mon . using the information derived in the present work ,
we conclude that ori ob1a is located at a distance of 353(59 ) pc , has an age of 11.5(1.5 ) myr and has a metallicity of 0.20(0.15 ) dex . in summary ,
i m mon is a secure member of the ori ob1a subgroup .
* acknowledgements * + this study is fully supported by the scientific & technological research council of turkey ( tubitak ) with the project code 109t449 .
zdenk mikulek is supported by the grants gaav iaa 301630901 and gar 205/08/0003 .
we thank s. n. de villiers for valuable comments to our text and the anonymous referee who improved the manuscript by his / her very useful comments .
andersen , j. 1991 , a&arv , 3 , 91 baki , v. , baki , h. , eker , z. , & demircan , o. 2007 , mnras , 382 , 609 baki , v. , baki , h. , bilir , s. , soydugan , f. , et al .
2010 , newa . , 15 , 1 bevington , p. r. , robinson , d. k. 2003 , in data reduction and error analysis for the physical sciences , third edition , mcgraw - hill , 2003 , edited by daryl bruflodt and spencer j. cotkin .
blaauw , a. 1991 , the physics of star formation and early stellar evolution , nato advanced science institutes ( asi ) series c , vol . 342 , held in agia pelagia , crete , greece , may 27th - june 8th , dordrecht : kluwer , 1991 , edited by charles j. lada and nikolaos d. kylafis .
, p.125 bouy , h. , martin , e. l. , brandner , w. , zapatero - osorio , m. r. , bjar , v. j. s. , schirmer , m. , hulamo , n. , & ghez , a. m. 2006 , a&a , 451 , 177 bozic , h. , wolf , m. , harmanec p. , et al .
2007 , a&a , 464 , 263 briceno , c. , calvet , n. , hernandez , j. , hartmann , l. , muzerolle , j. , dalessio , p. , & vivas , a. k. 2005 , star formation in the era of three great observatories , meeting abstracts from the conference held july 13 - 15 , 2005 in cambridge , ma .
http://cxc.harvard.edu/stars05/agenda/program.html , p.61 brown , a. g. a. , de geus , e. j. , & de zeeuw , p. t. 1994 , a&a , 289 , 101 brown , a. g. a. , blaauw , a. , hoogerwerf , r. , de bruijne , j. h. j. , & de zeeuw , p. t. 1999 , the origin of stars and planetary systems . edited by charles j. lada and nikolaos d. kylafis .
kluwer academic publishers , 1999 , p.411 brown , a. g. a. 2001 , an , 322 , 43 calvet , n. , briceo , c. , hernndez , r. , et al . 2005 , aj , 129 , 935 castelli f. , & cacciari , c. 2001 , a&a , 380 , 630 cester , b. , fedel b. , giuricin , g. , mardirossian , f. , mezzetti , m. 1978 , a&as , 33 , 91 cousins , a. w. j. 1987 , south african astronomical observatory , circulars , 11 , 93 cramer , k. 1984 , a&a , 132 , 283 cunha , k. & lambert , d. l. 1994 , apj , 426 , 170 de zeeuw , p. t. , hoogerwerf , r. , de bruijne , j. h. j. , brown , a. g. a. , & blaauw , a. c. s. 1999 , aj , 117 , 354 deutschman w. a. , davis r. j. , & schild r. e. 1976 , apjs , 30 , 97 dinescu , d. i. , girardi , t. m. , & van altena , w. f. 1999 , aj , 117 , 1792 dorazi , v. , randich , s. , flaccomio , e. , palla , f. , sacco , g. g. , & pallavicini , r. 2009 a&a , 501 , 973 esa 1997 , in the hipparcos and tycho catalogues , esa sp-1200 , noordwijk flower , p. j. 1996 , apj , 469 , 355 girardi , l. , bressan , a. , bertelli , g. , & chiosi , c. 2000 , a&as 141 , 371 gum , c.s .
1951 , mnras , 111 , 634 hadrava , p. 1995
, a&as , 114 , 393 hoogerwerf , r. 2000 , mnras , 313 , 43 johnson , d. r. h. , & soderblom , d. r. 1987 , aj , 93 , 864 johnson , h. l. , & morgan , w. w. 1953 , apj , 117 , 3131 johnson , h. l. , & morgan , w. w. 1951 , apj , 114 , 522 kharchenko , n. v. , scholz , r. d. , piskunov , a. e. , rser , s. , & schilbach , e. 2007 , an , 328 , 889 kreiner , j. m. , 2004 .
acta astron .
54 , 207 kodaira , k. , & scholz , m. 1970 , a&a , 6 , 93 kouwenhoven , m. b. n. , brown , a. g. a. , portegies zwart , s. f. , & kaper , l. 2007 , a&a , 474 , 77 kurucz r. l. 1993 , cd - rom 13 , 18 , http://kurucz.harward.edu luhman , k. l. , & potter , d. 2006 , apj , 638 , 887 maek , k. , batsch , t. , czyrkowski , h. , cwiok , m. , dabrowski , r. , dominik , w. , kasprowicz , g. , et al . 2010 , advances in astronomy , 2010 , 1 melnik , a. m. , & dambis , a. k. 2009 , mnras , 400 , 518 mihalas , d. , & binney , j. 1981 , galactic astronomy : structure and kinematics , 2nd edn .
( w.h.freeman , san fransisco ) page181 .
mikulek , z. , krtika , j. , henry , g. w. , et al .
2008 , a&a , 485 , 585 nieva , m. f. , & przybilla , n. 2007 , a&a , 467 , 295 nieva , m. f. , & przybilla , n. 2010 , hot and cool : bridging gaps in massive star evolution asp conference series vol .
425 , proceedings of a workshop held at the california institute of technology , pasadena , california , 10 - 12 november 2008 . edited by claus lietherer , philip bennett , pat morris , jacco van loon .
san francisco : astronomical society of the pacific , p.146 omara , b. j. , & simpson , r. w. 1972 , a&a , 19 , 167 pavlovski , k. , & hensberge , h. 2000 , mixing and diffusion in stars : theoretical predictions and observational constraints , 24th meeting of the iau , joint discussion 5 , august 2000 , manchester , england , meeting abstract .
pavlovski , k. , & hensberge , h. 2005 , a&a , 439 , 309 pojmaski , g. 1997 , acta astron . , 47 , 467 pearce , j. a. 1932 , publ .
dominion astrophys .
victoria , 6 , 59 pi of the sky 2010 , http://grb.fuw.edu.pl/ pourbaix d. , tokovinin a. a. , batten a. h. , fekel f. c. , hartkopf w. i. , levato h. , morrell n. i. , torres g. , udry s. , 2004 , a&a , 424 , 727 sanyal , a. , & sinvhal , s. d. 1964 , the observatory , 84 , 211 sanyal , a. , mahra , h. s. , & sanwal , n. b. 1965 , bull .
16 , 209 shobbrook , r. r. 2004 , journal astron . data , 10 , 1 simon , k. p. , sturm , e. 1994 , a&a , 281 , 286 straiys , v. , & kuriliene , g. 1981 , ap&ss , 80 , 353 .
torres , g. , andersen , j. , & gimnez , a. 2010 , a&arv , 18 , 67 van hamme , w. 1993 , aj , 106 , 2096 .
van leeuwen , f. 2007 , a&a , 474 , 653 wilson , r. e. , devinney , e. j , 1971 , apj , 166 , 605 .
wilson , r. e. , 1994 , pasp , 106 , 921 .
von zeipel , h. 1924 , mnras , 84 , 665 , 684 , 702 .
zverko , j. , iovsk , j. , mikulek , z. , et al .
2009 , a&a , 506 , 845 . | all available photometric and spectroscopic observations were collected and used as the basis of a detailed analysis of the close binary i m mon .
the orbital period of the binary was refined to 1.19024249(0.00000014 ) .
the roche equipotentials , fractional luminosities ( in _ b , v _ and _ h@xmath0 _ - bands ) and fractional radii for the component stars in addition to mass ratio @xmath1 , inclination @xmath2 of the orbit and the effective temperature @xmath3 of the secondary cooler less massive component were obtained by the analysis of light curves .
i m mon is classified to be a detached binary system in contrast to the contact configuration estimations in the literature .
the absolute parameters of i m mon were derived by the simultaneous solutions of light and radial velocity curves as m@xmath4 m@xmath5 and 3.32(0.16 ) m@xmath5 , r@xmath6 r@xmath5 and 2.36(0.03 ) r@xmath5 , @xmath7 k and 14500(550 ) k implying spectral types of b4 and b6.5 zams stars for the primary and secondary components respectively .
the modelling of the high resolution spectrum revealed the rotational velocities of the component stars as v@xmath8 kms@xmath9 and v@xmath10 kms@xmath9 .
the photometric distance of 353(59 ) pc was found more precise and reliable than hipparcos distance of 341(85 ) pc .
an evolutionary age of 11.5(1.5 ) myr was obtained for i m mon .
kinematical and dynamical analysis support the membership of the young thin - disk population system i m mon to the ori ob1a association dynamically .
finally , we derived the distance , age and metallicity information of ori ob1a sub - group using the information of i m mon parameters . |
liquid crystals spatially confined on the nanometer scale exhibit structural and thermodynamical properties which differ markedly from their bulk counterparts .
both the collective orientational ( isotropic - to - nematic ) and translational ( smectic ) transitions have turned out to be significantly by finite size , quenched disorder and interfacial ( solid - liquid ) interactions introduced by the confining solid walls or the pore topology @xcite . for example
, the first - order discontinuous isotropic - to - nematic transition of rod - like liquid crystals condensed in pores a few nanometers across goes to a critical point , above which , strictly speaking , it disappears .
then the transition is rendered continuous with a paranematic high - temperature phase , where even at highest temperatures the pore walls ( anchoring fields ) induce a residual collective orientational , paranematic order and no isotropic liquid is observable @xcite .
this sensitivity of liquid crystalline phase transitions stimulated a plethora of experimental studies .
typically they were performed upon complete filling of the confining geometry @xcite and thus revealed integral quantities on the phase behavior .
phenomenological considerations @xcite and computer simulations on liquid crystals in thin film and nanopore confinement @xcite indicate , however , pronounced heterogeneities , in particular interface - induced molecular layering and orientational immobility in the pore wall proximity . despite recent experimental advancements in optical techniques directly probing orientational order parameter profiles in the proximity of planar , solid walls @xcite , achieving the spatial resolution necessary to rigorously explore such inhomogeneities in nanometer - sized capillaries remains still experimentally extremely demanding . here
we present an experimental scheme which allows us to investigate the thermotropic orientational order of a classic , rod - like nematogen ( 7cb ) as a function of gradual filling of parallel - aligned silica capillaries in a monolithic membrane .
we find a gradual evolution of the thermotropic orientational order behavior and can infer four distinct phase transition behaviors depending on the filling fraction , and thus spatial arrangement of the liquid crystalline material in the capillaries .
moreover , we find hints of tensile pressure effects on the orientational behavior of the nano - confined liquid crystal .
in our experiments we employ a monolithic silica membrane of 320 @xmath2 m thickness .
it is traversed by an array of parallel - aligned , non - interconnected capillaries .
the membrane is prepared by thermal oxidation of a free - standing mesoporous silicon membrane @xcite at 810@xmath3c for 12 h. the mean capillary diameter @xmath4 and porosity @xmath5 , as determined by recording a volumetric nitrogen sorption isotherm at @xmath677 k , are @xmath7 @xmath8 nm and @xmath9 @xmath10 % , respectively . in the bulk state 7cb
exhibits an isotropic - to - nematic ordering transition at a temperature @xmath11=315 k. given the low vapor pressure of 7cb at room temperature and the channel geometry with large , inner surfaces we can not fill the matrix via the gas phase .
therefore , we imbibe it with eight binary 7cb / acetone solutions @xcite . after evaporation of the high vapor pressure solvent ,
the matrix is filled with the remaining , low - vapor pressure liquid crystals up to a filling fraction @xmath12 , that depends on the initial solute concentration .
it is instructive to compare this filling procedure with the standard way for the preparation of partial liquid fillings of nanoporous substrates via the gas phase , i.e. vapor sorption isotherms . for nanoporous solids they exhibit distinct regimes typical of characteristic arrangements of
the pore liquid : at small fractional filling @xmath12 , below a certain threshold value @xmath13 , an adsorbed film is homogeneously formed at the pore walls , see sketch in fig .
[ fig1]d , @xmath14 .
for @xmath15 the pores fill via capillary condensation , that is the formation of liquid bridges terminated by concave menisci .
both regimes can be inferred by light adsorption experiments . upon onset of capillary condensation or evaporation typically a drastic decrease in optical transmission , @xmath16 is observable @xcite ( here @xmath17 and @xmath18 are the intensities of incident and transmitted laser light , respectively ) .
the matrix turns milky and the reduction in the transmitted light originates in scattering at the capillary bridges / void structures , which have sizes on the length scale of the wavelength of visible light @xcite .
by contrast , the matrix remains transparent both in the film condensed state ( @xmath14 ) and in the completely filled state ( @xmath191 ) , where such structures are absent .
laser light adsorption experiments on our solute - prepared samples exhibit an analogous behavior as a function of @xmath12 , see fig .
[ fig1]a . a sizeable decrease in light transmission
is observed for @xmath20 .
the gradual disappearance of the bubbles and/or capillary bridges upon reaching a complete filling is indicated for @xmath21 and , eventually the full optical transmission upon vanishing of all empty pore segments is recovered for @xmath22 .
note that the transmission drops by four orders of magnitude in the capillary condensate regime , eventually even below the sensitivity of our setup .
in contrast to nanoporous silica powders or disordered , sponge - like pore structures , our monolithic silica membrane is particularly suitable in order to evaluate the collective orientational order of liquid crystals by optical birefringence measurements in a transmission geometry @xcite . in the bulk nematic
state the optical birefringence changes linearly with the orientational order parameter , @xmath23 , where @xmath24 is the angle between the long axis of a single molecule and a direction of preferred orientation of that axis , the director @xcite .
the brackets denote an averaging over all molecules under consideration .
the simple relation between optical birefringence and @xmath25 the bulk system is also conserved for the liquid crystal / matrix composite material , provided the geometrical birefringence is subtracted @xcite .
it was measured with a self - designed , high - resolution polarimetry setup presented in refs .
@xcite . the high sensitivity of this device , which employs a dual lock - in detection scheme , allows one to obtain data even for low optical transmissions of the matrix .
therefore , it is possible to gain reliable data even upon onset of capillary condensation and upon reaching complete filling , as long as the transmission is larger than 10@xmath26 . in fig .
[ fig2]c @xmath27 is plotted versus temperature for a selected set of @xmath12 s in the film - condensed state .
there exists a final birefringence and thus orientational order at high-@xmath28 in all measurements .
it is typical of a sizable collective ordering of the molecules induced by the interaction with the wall ( interface anchoring ) .
silica surfaces favor a planar anchoring in the case of the family of cyanobiphenyl liquid crystals ( cbs ) , as documented in experiments @xcite and atomistic computer simulations @xcite .
the magnitude and the positive sign of the changes of the optical birefringence for the lowest filling fraction ( @xmath290.12 , corresponding to a partially filled , first adsorbed monolayer ) indicates indeed such an alignment .
it is even compatible with a preferred orientation of the long axis of the liquid crystals along the axial direction of the nano channels - as illustrated in fig .
note that there is a small , gradual increase @xmath30 upon cooling for this filling fraction , but no indication of a distinct change in collective orientational order . if one assumes a perfect cylindrical geometry and an intermolecular distance of 0.4 nm @xcite , the molecules added in the next filling steps correspond to @xmath311.5 and 2 monolayers ( @xmath290.17 and 0.225 ) .
they exhibit sizeable increases in @xmath30 upon cooling , which are characteristic of a rearrangement between partial collective orientational order ( a paranematic state ) to full collective orientational order ( nematic order ) .
the symmetry of those states is identical ; the transition is no phase transition in a classical sense .
it is , however , possible to define @xmath32 as a temperature of most pronounced changes of @xmath33 , in analogy to the true " phase transition temperature .
we determine this quantity by a calculation of the maximum in the @xmath28-derivative of @xmath30 - see insets in fig .
[ fig2 ] . for the nanometric liquid crystal film it yields a @xmath32 of 306 k. this downward - shift @xmath34 of 10k in respect to the bulk system compares surprisingly well with an expected @xmath34 of @xmath31 9 k , which one achieves by an extrapolation of thickness - dependent boundary effects , explored in thin - film geometry on silica ( with film thicknesses of several hundred nanometers ) @xcite , down to the nano metric films investigated here .
the gradual increase in thermotropically induced molecular reorientations upon film - thickness increase , observed here , corroborates a continuous increase in molecular orientational freedom with increasing distance from the pore walls documented in computer simulations for cylindrical nanopores @xcite . in fig .
[ fig1]c the @xmath12-dependence of @xmath35 is shown as determined in the confined nematic phase at @xmath36k for each selected @xmath12 .
it increases linearly with @xmath12 both in the film - condensed and in the capillary - condensed state .
it is striking , however , that the slope in the experimentally accessible capillary - condensed regime ( 0.25@[email protected] ) is by more than one order of magnitude larger than in the film - condensed state - see also inset in fig .
this means that the film - condensed and capillary - condensed state is clearly distinguishable in terms of the evolution of thermotropic orientational order : in agreement with molecular simulation studies @xcite , the freedom in orientational molecular ordering is considerably larger near the pore axis ( i.e. in the core of the pore filling ) than in the proximity of the channel wall , where the molecules are much more strongly influenced both by channel - wall roughness ( quenched disorder ) and the anchoring field of the channel walls @xcite .
the linear increase of @xmath35 with @xmath12 in the film - condensed and capillary - condensed regime suggests that material added in one step does not affect the reorientational behavior of the material adsorbed in the previous filling steps .
we can not exclude , however , that the formation of capillary condensate in the pore center and the corresponding vanishing cylindrical liquid - vapor interfaces affects the reorientation behavior of the molecules in the film - condensed , boundary layer state .
the temperature of the paranematic - to - nematic transition exhibits only a weak increase in the entire region of capillary condensation ( @xmath38 ) and remains at least @xmath31 4 k below the transition point @xmath32 of the fully filled ( @xmath39 ) porous substrate , see fig . [ fig1]b .
note , however , that there is a sizable upward jump of @xmath32 between the slightly underfilled ( @xmath40 ) and the completely filled state ( @xmath39 ) - see fig .
[ fig1]b . hence
, @xmath32 of the continuous transition of the completely filled state is close to the temperature of the discontinuous bulk transition @xmath11 , in agreement with our previous experimental study and phenomenological treatment of this state with a landau - de - gennes model enriched by confinement and quenched disorder effects ( kklz - model ) @xcite .
the abrupt @xmath28-shift in the transition temperature may result from the complete vanishing of any liquid / vapor interface and the resulting reduction of inhomogeneities in the system , which rather favor the disordered paranematic state .
there is another alternative explanation , which is related to the hydrostatics of the confined liquid . in the capillary - condensed state
the liquid experiences a tensile pressure , dictated by the concave curvature of the menisci terminating the liquid bridges .
it is known that this negative hydrostatic pressure causes not only subtle deformations of the rigid , nanoporous matrix @xcite , it also significantly affects density , and thus pressure - dependent first - order transitions , most prominently the liquid - solid transition @xcite .
therefore , it is in principal also expected to affect the isotropic - to - nematic transition @xcite . according to the laplace formula applied to the menisci in the capillaries , while assuming good wetting conditions ( mean curvature radius = - pore radius = -6 nm , surface tension of 7 cb=31.7 mn / m @xcite ) the tensile pressure in the liquid bridges amounts to @xmath31 - 10.5 mpa for @xmath41 .
this pressure is completely released upon reaching @xmath39 ( mean curvature radius = @xmath42 ) .
an extrapolation of the pressure - dependence of @xmath11 of 7cb , reported in the literature @xcite , towards negative pressures yields a corresponding downward shift of 3.8 k between partially and completely filled state , in good quantitative agreement with the observed @xmath28-shift .
it is interesting to note that this @xmath28-shift is much larger than any tensile pressure effect on liquid crystals that has been reported to date .
presumably , this results from the extreme challenges @xcite associated with experiments aimed at reaching tensile pressures comparable to the ones established here in the partially filled state . since the true curvature of the menisci at high @xmath12-values is not accessible in our experiments and since studies with higher f - resolution are experimentally demanding both due to the sizable light scattering in that regime and because of the error bars in the filling fraction ( resulting from @xmath28-dependent volume changes in the liquid ) , this conclusion remains somewhat speculative and more rigorous studies are planned for the future .
in summary our optical studies allowed us to experimentally document an inhomogeneous thermotropic orientational order behavior upon sequential filling of silica nano - capillaries with a rod - like liquid crystal which encompasses four distinct regimes : the monolayer state , the film - condensed state , the capillary - condensed regime and the complete filling . since orientational ordering is key to functional materials with switching capability , we believe that our observations are not only of high fundamental interest , but also of importance for the emerging field of nanophotonics @xcite : our study documents that the thermooptical properties of nanoporous systems ( or single nano - capillaries ) can be tailored over a surprisingly wide range simply by variation of the the filling fraction with liquid crystals . for the future we envision analogous experimental studies on the molecular dynamics , which should show a similar complex behavior as the static properties investigated here .
we also hope that our findings stimulate temperature - dependent simulation studies and model calculations on partially filled nano - capillaries , aimed at an exploration of the remarkably complex phase transition phenomenologies revealed here as a function of material distribution in the capillaries .
this work has been supported by the polish national science centre under the project `` molecular structure and dynamics of liquid crystals based nanocomposites '' ( decision no dec-2012/05/b / st3/02782 ) and by the german research foundation ( dfg ) by grant no .
hu850/3 - 1 . | optical birefringence and light absorption measurements reveal four regimes for the thermotropic behavior of a nematogen liquid ( 7cb ) upon sequential filling of parallel - aligned capillaries of 12 nm diameter in a monolithic , mesoporous silica membrane .
no molecular reorientation is observed for the first adsorbed monolayer . in the film - condensed state ( up to 1 nm thickness ) a weak , continuous paranematic - to - nematic ( p - n ) transition is found , which is shifted by 10 k below the discontinuous bulk transition at t@xmath0305 k. the capillary - condensed state exhibits a more pronounced , albeit still continuous p - n reordering , located 4 k below t@xmath1 . . |
in certain situations , geometrical information of manifolds might be encoded in differential forms . in the presence of symmetries of the manifold via the action of a lie group ,
the behavior of these differential forms under the group action may lead to a better understanding of the manifold itself . in some well known instances of actions of compact lie groups ,
the action is of a particular type whenever the differential form may be extended to an equivariant one in the cartan model of equivariant cohomology @xcite ; this is the case for example in hamiltonian actions on symplectic manifolds @xcite , in hamiltonian actions on exact courant algebroids @xcite or in gauged wzw actions which are anomaly free @xcite and its generalizations @xcite .
when the lie group is not compact , the cartan complex associated to the action in general does not compute the equivariant cohomology of the manifold , and therefore many of the results that hold for compact lie groups may not hold for the non - compact case .
but since the cartan model is very well suited for studying the infinitesimal behavior of the differential forms with respect to the action of the lie algebra , it would be worthwhile knowing more about the relation between the cohomology of the cartan complex and the equivariant cohomology . in this paper
we study this relation and we obtain some interesting results which in particular permit us generalize the conditions for cancellation of anomalies on gauge wzw actions developed by witten @xcite , to the case on which the gauge group is the non - compact group @xmath1 .
the main ingredient of this work is the equivariant de rham complex developed by getzler @xcite whose cohomology calculates the equivariant cohomology independent whether the group is compact or not .
we show that there is an inclusion of complexes of the cartan complex into the equivariant de rham one , and therefore we obtain a homomorphism between the cohomology of the cartan complex to the equivariant de rham cohomology .
this in particular implies that any closed form in the cartan complex defines an equivariant cohomology class , but note that the converse may not be true .
we furthermore show that there is a spectral sequence converging to the equivariant de rham cohomology such that its @xmath2-term is isomorphic to the cohomology of the cartan complex , and in this way an explicit relation between the cohomology of the cartan complex and the equivariant de rham cohomology is obtained .
now , since the first page of this spectral sequence could be understood in terms of the differential cohomology of a lie group , we recall its definition and some of its properties and we reconstruct some calculations for the group @xmath1 .
we conclude this work by applying our results to the gauged wzw actions whenever the gauge group is @xmath1 .
we recall the result of witten which claims that the gauged wzw action on compact lie groups is anomaly free , if and only if the wzw term @xmath3 can be extended to a closed form in the cartan complex , and we generalize it to the case on which the gauge group is @xmath1 ; this is theorem [ equivalence of extension for sl(n , r ) ] .
we use this theorem to construct explicit examples where the equations representing the condition of anomalies cancelation hold and where they do not .
the physical implications of our work will appear elsewhere .
* acknowledgments * the first author was partly supported by the conacyt - mxico through grant number 128761 .
the third author would like to acknowledge and to thank the financial support provided by the alexander von humboldt foundation and also would like to thank the hospitality of prof .
wolfgang lck at the mathematics institute of the university of bonn .
let @xmath4 be a lie group and @xmath5 a manifold on which @xmath4 acts on the left by diffeomorphisms .
the @xmath4-equivariant cohomology of @xmath5 could be defined as the singular cohomology of the homotopy quotient @xmath6 @xmath7 where @xmath8 is the universal @xmath4-principal bundle @xmath9 . the previous definition works for any topological group and any continuous action , but
sometimes it is convenient to have a de rham version with differentiable forms for the equivariant cohomology whenever the group is of lie type and the action is differentiable.when the group @xmath4 is compact , the weil and cartan models provide a framework on which the equivariant cohomology could be obtained via a complex whose ingredients are the local action of the lie algebra @xmath10 , and the differentiable forms @xmath11 .
when the group @xmath4 is not compact there is a more elaborate model for equivariant cohomology that we will describe in the next section .
one way to obtain a de rham model for equivariant cohomology is through the total complex of the double complex @xmath12 that is obtained after applying the differentiable forms functor to the simplicial space @xmath13 which is the nerve of the differentiable groupoid @xmath14 . by the works of bott - shulman - stasheff @xcite and getzler @xcite we know that one way to calculate the cohomology of
the total complex of the double complex of differentiable forms @xmath12 is through the differentiable cohomology groups @xmath15 of the group @xmath4 with values on the differentiable forms of @xmath5 tensor the symmetric algebra of the dual of the lie algebra @xmath10 . in @xcite getzler
has shown that there is a de rham theorem for equivariant cohomology showing that there is an isomorphism of rings @xmath16 between the de rham model for equivariant cohomology and the cohomology of the homotopy quotient .
the de rham model for equivariant cohomology defined by getzler is described as follows .
consider the complex @xmath17 with elements smooth maps @xmath18 which vanish if any of the arguments @xmath19 equals the identity of @xmath4 .
the operators @xmath20 and @xmath21 are defined by the formulas @xmath22 as in the case of the differential in cartan s model for equivariant cohomology @xcite .
recall that the elements in @xmath23 are defined to have degree 2 , and therefore the operator @xmath21 has degree 1 .
denote the generators of the symmetric algebra @xmath24 by @xmath25 where @xmath26 runs over a base of @xmath10 .
the coboundary @xmath27 is defined by the formula @xmath28 and the contraction @xmath29 is defined by the formula @xmath30 where @xmath31 . if the image of the map @xmath32 is a homogeneous polynomial of degree @xmath33 , then the total degree of the map @xmath34 equals @xmath35 .
it follows that the structural maps @xmath36 and @xmath37 are degree 1 maps , and the operator @xmath38 becomes a degree 1 map that squares to zero .
the elements of the complex @xmath39 will be called _ equivariant de rham forms _ and its cohomology @xmath40 will be called the equivariant de rham cohomology . in @xcite
it was shown that the complex @xmath41 together with the cup product @xmath42 for @xmath43 , becomes a differential graded algebra , and moreover that there is a canonical isomorphism of rings @xmath44 with the cohomology of the homotopy quotient .
the cartan model for equivariant cohomology is the differential graded algebra @xmath45 endowed with the differential @xmath46 .
therefore there is a natural homomorphism of differential graded algebras @xmath47 given by the inclusion @xmath48 since the restriction of @xmath49 to @xmath50 is precisely @xmath51 because the operators @xmath52 and @xmath37 act trivially on the invariant elements of @xmath53 .
the induced map on cohomologies @xmath54 is far from being an isomorphism as the case of @xmath55 and @xmath56 shows .
+ in this case @xmath57 $ ] and @xmath58 , hence @xmath59 $ ] where @xmath60 . on the other hand @xmath61 since @xmath62 is contractible .
nevertheless , when the lie group @xmath4 is compact , the map @xmath63 induces an isomorphism in cohomology @xcite .
now , in order to understand in more detail the relation between the cohomology of the cartan model and the equivariant cohomology we will introduce a spectral sequence suited for this purpose .
let us filter the complex @xmath64 by the degree in @xmath65 ; namely , if we consider maps @xmath66 with image homogeneous elements of degree @xmath33 , we will denote @xmath67 and @xmath68
. then we can define the filtration @xmath69 where @xmath70 .
we have that the differentials have the following degrees : @xmath71 and therefore the filtration is compatible with the differentials .
the spectral sequence associated to the filtration @xmath72 has for page 0 : @xmath73 and the 0-th differential is @xmath74 because the other three differentials raise @xmath75 .
therefore the page 1 is : @xmath76 the differentiable cohomology of @xmath4 with coefficients in the representation @xmath65 .
the 0-th row of the first page is precisely the cartan complex @xmath77 and the first differential on this row @xmath78 becomes precisely the cartan differential @xmath51 .
therefore we see that on the second page we get that @xmath79 namely that the 0-th row of the second page is isomorphic to the cohomology of the cartan complex .
therefore we conclude that the composition @xmath80 of the surjective homomorphism @xmath81 with the inclusion @xmath82 is equivalent to the induced map on cohomologies @xmath54 defined previously .
notice that in de rham model for equivariant cohomology the operator @xmath83 is defined in a similar fashion as the differential for group cohomology in the case of discrete groups .
the cohomology groups defined by the differential @xmath83 are called the _ differentiable cohomology groups _ and are defined for any @xmath4-module @xmath84 ; i.e. if @xmath4 is a lie group and @xmath84 is a @xmath4-module then the _ differentiable cohomology of @xmath4 with values in @xmath84 _ is the cohomology of the complex @xmath85 where @xmath86 consists of differentiable maps @xmath87 such that @xmath34 vanishes if any of the arguments @xmath19 equals the identity of @xmath4 , and the differential is @xmath83 is defined by @xmath88 we denote this cohomology by @xmath89 for @xmath84 in the category of topological @xmath4-modules , the cohomology groups @xmath90 can be seen as the relative derived functor associated to the @xmath4 invariant submodule @xmath91 .
in particular we have that @xmath92 and whenever @xmath4 is compact the functor of taking the @xmath4-invariant submodule is exact and therefore in that case @xmath93 .
the first page of the spectral sequence that was defined in section [ subsubsection spectral sequence ] @xmath76 is precisely the differential cohomology defined above @xmath94 for the @xmath4-module @xmath65 .
for @xmath84 a vector space over @xmath95 , by differentiating the functions from @xmath96 to @xmath84 , van est @xcite proved that for @xmath4 connected @xmath97 whenever @xmath98 is the maximal compact subgroup of @xmath4 , @xmath99 and @xmath10 are their corresponding lie algebras , and @xmath100 denotes the lie algebra cohomology defined by chevalley and eilenberg in @xcite . whenever the group @xmath4 has a compact form @xmath101
, i.e. a compact lie group whose complexification is isomorphic to the complexification of @xmath4 @xmath102 then we have that @xmath103 where the second isomorphism was proved by chevalley and eilenberg in @xcite for compact lie groups , and the first isomorphism follows from the isomorphisms @xmath104 which follow from the isomorphism of complex lie algebras @xmath105 .
in this case we have that the differentiable cohomology of @xmath4 can be calculated by topological methods , i.e. @xmath106 let us consider the non - compact group @xmath107 .
in this case we have @xmath108 therefore for @xmath109 we have that @xmath110,\ ] ] where the degree of @xmath111 is @xmath112 and @xmath113 is the largest odd integer which is less or equal than @xmath114
. the equivariant de rham cohomology of this group is @xmath115 . the spectral sequence defined in section [ subsubsection spectral sequence ] has for first page @xmath116 and since the algebra @xmath117 is reductive and @xmath24 is a finite dimensional semi - simple @xmath10-module in each degree @xcite , then we have that @xmath118 and therefore @xmath119 the ideal of @xmath4-invariant polynomials is known to be @xmath120,\ ] ] where the degree of @xmath121 is @xmath122 , and so we get that the first page of the spectral sequence converging to @xmath123 is @xmath124 \otimes { \ensuremath{{\mathbb r}}}[c_2,c_3, ... ,c_n].\ ] ] since we know that for @xmath109 @xmath125}]\ ] ] is the free algebra on the pontrjagin classes , we obtain that the @xmath126-th differential of the spectral sequence maps the class @xmath111 to the class @xmath121 @xmath127 in particular we obtain that @xmath128 and therefore we see that the fourth cohomology group of @xmath62 is generated by the @xmath4-invariant quadratic forms in @xmath129 .
in many instances in geometry , the action of a compact lie group on a manifold being of a certain kind is equivalent to the existence of an equivariant lift of a specific invariant closed differential form on the cartan model .
some examples of this phenomenon are the following : * the action of @xmath4 on a symplectic manifold @xmath130 being hamiltonian is equivalent to the existence of a closed equivariant lift @xmath131 of the symplectic form ; in this case the maps @xmath132 may be assembled into a map @xmath133 , @xmath134 , which becomes the moment map ( see @xcite ) . *
the action of @xmath4 on an exact courant algebroid @xmath135_h)$ ] , with courant - dorfman bracket twisted by the closed three form @xmath136 , is hamiltonian provided there exists a closed equivariant lift @xmath137 ( see @xcite ) .
* let @xmath138 be a connected , simple , simply connected and compact matrix group , i.e. @xmath139 , and denote by @xmath140 the product of two copies of @xmath138 acting on @xmath138 on the left by the action @xmath141 a subgroup @xmath142 is called an _ anomaly - free subgroup _ if there the wzw action could be gauged with respect to the group action given by @xmath4 . in
@xcite witten showed that the anomaly - free subgroups are precisely the subgroups @xmath4 on which the wzw term @xmath143 could be lifted to a closed equivariant 3-form in the cartan complex @xmath144 .
( we will elaborate this construction in the next chapter ) . the previous examples are not exhaustive , but they give the idea of the general principle . in these cases
we have a @xmath4-invariant closed form @xmath145 and we need to find a closed equivariant lift @xmath146 .
note that since the group @xmath4 is compact , the existence of a closed equivariant lift in the cartan model is equivalent to the existence of a lift of the cohomology class @xmath147 \in h^*(m ; { \ensuremath{{\mathbb r}}})$ ] on the equivariant cohomology @xmath148
. therefore the obstructions of the existence of the equivariant lift could be studied via several methods , for instance , with the use of the serre spectral sequence associated to the fibration @xmath149 , or with the use of the spectral sequence associated to a filtration of the cartan complex @xmath150 given by the degree of @xmath151 as it is done in the papers @xcite .
but what happens in the case that the lie group @xmath4 is not compact ?
we speculate that the situation should be similar , in the sense that the action being of certain kind is equivalent to the existence of a lift on the equivariant de rham complex defined in section
[ equivariant de rham complex ] .
this situation has not been explored so far but we believe that the equivariant de rham complex provides a framework in which actions on non - compact lie groups could be better understood .
since in many instances the geometric information is captured by closed forms in the cartan complex , we would like to study the relation between the cohomology of the cartan complex and the equivariant de rham cohomology whenever the group is not compact .
let us consider the diagram of complexes @xmath152 \ar[r]^(.35){i } & c^*(g , s{{\mathfrak}g}^ * \otimes \omega^\bullet m ) \ar[ld ] \\
\omega^\bullet m & } \ ] ] where the horizontal map is the injective map of complexes defined in and the vertical maps are the natural forgetful maps .
let us take a closed @xmath4-invariant form @xmath136 on @xmath5 and let us suppose that we can lift this closed form to a closed form in the cartan model @xmath153 , then equivariant form @xmath154 becomes a closed lift for @xmath136 in the complex of equivariant de rham forms , i.e. @xmath155 .
we have then [ extension in cartan implies equivariant extension ] take a closed @xmath4-invariant form @xmath156 . if @xmath136 can be lifted to a closed form @xmath157 in the cartan model of the equivariant cohomology , then the form @xmath158 is a closed lift for @xmath136 in @xmath159 , the closed forms of the equivariant de rahm complex .
note in particular that lemma [ extension in cartan implies equivariant extension ] implies that if one can extend an invariant closed form to a closed form in the cartan model for equivariant cohomology , then the cohomology class @xmath160 $ ] lies in the image of the canonical forgetful homomorphism @xmath161 and therefore the cohomology class @xmath160 $ ] could be extended to an equivariant cohomology class in any model for the equivariant cohomology of @xmath5 .
the converse of lemma [ extension in cartan implies equivariant extension ] would say that if one knows that an invariant differentiable form @xmath136 could be lifted to a closed form in @xmath159 then a lift could be written as a closed element in the cartan model .
the converse of lemma [ extension in cartan implies equivariant extension ] is indeed true whenever the lie group @xmath4 is compact , but for general group actions it does not hold . from the spectral sequence defined in section
[ subsubsection spectral sequence ] we have seen that in the first page we have @xmath162 with differential @xmath78 equivalent to @xmath51
. therefore on the second page we get that @xmath79 namely that the 0-th row of the second page is isomorphic to the cohomology of the cartan complex .
since we have the surjective homomorphism @xmath81 we can conclude that [ conditions extension in cartan ] for a closed @xmath4-invariant form @xmath156 , it can be lifted to a closed form in the cartan complex if , firstly the cohomology class @xmath160 $ ] could be lifted to an equivariant cohomology class in @xmath163 , and secondly , if the lift lies on the subgroup @xmath164 . the second condition of proposition [ conditions extension in cartan ] is more difficult to check than the first one , since it depends explicitly on the equivariant de rham model for equivariant cohomology ; for the first condition any model for the equivariant cohomology works . in certain specific situations , extensions in the cartan model of closed forms may be obtained , and this is the subject of the next and final chapter .
we note here that a sequence of obstructions for lifting a @xmath4-invariant form to a closed form in the cartan complex can be determined , when studying the spectral sequence associated to appropriate filtrations of the cartan complex , as it is carried out in @xcite .
our approach is different since we are interested in using the fact that an extension in the cartan model can only exist if there is an equivariant extension , i.e. an extension in the homotopy quotient .
in the physics literature ( see @xcite and the references therein ) it has been argued that the condition of anomaly cancelation for the gauged wzw action is given by the equation @xmath165 and this equation is moreover equivalent to the existence of an equivariant extension of the wzw term @xmath143 on the cartan model for equivariant cohomology . in the case that the group that we are gauging is compact , the anomaly cancellation is equivalent to the existence of an equivariant lift of the cohomology class @xmath147 $ ] , and therefore the anomaly cancellation becomes topological and could be checked with topological methods . in this section
we study in detail the case on which the gauge group is @xmath107 ( or any subgroup of it ) and we show that the anomaly cancelation condition for the wzw action is also topological and only depends on the existence of an equivariant lift of the cohomology class @xmath147 $ ] ; in this way we find a large family of @xmath1 actions with anomaly cancellation .
let us start by recalling the explanation given by witten @xcite that asserts that the condition for anomaly cancelation is equivalent to the existence of a lift in the cartan complex of the form @xmath145 .
let @xmath138 be a connected and simple matrix group , i.e. @xmath139 , such that its fundamental group is finite .
denote by @xmath140 the product of two copies of @xmath138 acting on @xmath138 on the left by the action @xmath166 and consider a subgroup @xmath142 acting on the left on @xmath138 by the induced action of @xmath140 .
the embedding @xmath142 determines a map at the level of lie algebras which can be written as @xmath167 and the canonical vector fields @xmath168 on @xmath138 generated by the @xmath4 action could be written as @xmath169 for all @xmath170 .
the matrix 1-forms @xmath171 and @xmath172 satisfy the equations @xmath173 and we can take the differential form @xmath174 which defines the wzw action
. the form @xmath175 is @xmath140 invariant , therefore @xmath176 , it is closed and is a generator of the cohomology group @xmath177 . to find a closed extension of @xmath145 in the cartan complex we need to find 1-forms @xmath178 such that the following equations are satisfied : @xmath179},\end{aligned}\ ] ] where the first two imply that the form @xmath180 is @xmath181-closed , and the third one implies that @xmath146 is @xmath4-invariant .
calculating @xmath182 we obtain @xmath183\end{aligned}\ ] ] and therefore we see that we can define @xmath184 satisfying the equation @xmath185 . the fact that @xmath186}$ ] is satisfied , is a tedious but straightforward computation .
now we compute @xmath187 noting that for @xmath188 the function thus defined become constant , and therefore we have that @xmath189 is a 3-form in the cartan complex @xmath190 and its failure to be closed is the quadratic form @xmath191 where the coefficient of @xmath192 of this quadratic form is precisely @xmath193 since it was known in the literature that the condition for the absence of anomalies was @xmath194 witten concluded that the absence of anomalies was equivalent to the existence of a closed extension of @xmath145 in the cartan complex . whenever the lie group @xmath4 is compact , the existence of such an extension is equivalent to the existence of an equivariant extension of the cohomology class @xmath147 $ ] , and this can be checked with the use of the serre spectral sequence associated to the fibration @xmath195 the second page of the spectral sequence becomes @xmath196 and since @xmath197 ,
the only non - trivial differential that affects @xmath147 \in e_2^{0,3}$ ] is @xmath198 thus defining an element @xmath199 ) \in e_2^{4,0 } \cong h^4(bg;{\ensuremath{{\mathbb r}}});\ ] ] this implies that the only obstruction to lift @xmath147 $ ] to an equivariant class is precisely @xmath200)$ ] . since we assumed that @xmath4 is compact , we know that @xmath201 and therefore we must have that @xmath199 ) = ( d+\iota)\tilde{\omega},\ ] ] namely that the two obstructions are the same .
the previous argument permits to find several cases on which there is anomaly cancellation .
the simplest of all is the adjoint action of @xmath4 on itself @xmath202 since in this case the spectral sequence associated to the fibration @xmath203 always collapses at the second page , and therefore @xmath204 . in this section
we will argue that if the gauge group is @xmath107 then the cancellation of anomalies is topological , and therefore it is equivalent to the existence of an equivariant lift of the cohomology class @xmath147 $ ] .
[ equivalence of extension for sl(n , r ) ] let @xmath138 be a connected , simple matrix group with finite fundamental group .
let @xmath107 and consider an action on @xmath138 defined by an injection @xmath142 .
the existence of an equivariant extension on @xmath205 of the cohomology class @xmath147 $ ] is equivalent to the existence of a closed lift to the cartan model of @xmath145 . we know that if there is an extension @xmath146 in the cartan model , then the cohomology class @xmath206 $ ] represents the lift in the cohomology group @xmath207 . to prove the converse we will make use of the constructions and results of sections
[ section differentiable cohomology of sl(n , r ) ] , [ section extension differential form ] and [ section gauged wzw actions ] . from the equivariant de rham theorem
we know that @xmath208 and therefore we could take the class @xmath200)$ ] , which is the obstruction of extending @xmath147 $ ] to an equivariant class , to be an element in @xmath199 ) \in h^4(g , s { \mathfrak}{g}^*).\ ] ] we already know that @xmath209 and its cohomology class in @xmath210 represents the same obstruction for an equivariant lift , i.e. @xmath199 ) = [ ( d+ \iota ) \tilde{\omega}].\ ] ] in general it may happen that the cohomology class @xmath211 $ ] is zero even though the form @xmath212 is different from zero .
but in the particular case of @xmath213 we have already seen in that the inclusion map of the cartan complex into the equivariant de rham complex @xmath214 induces an isomorphism in degree 4 @xmath215\ ] ] and therefore we have that the vanishing of the class @xmath200)$ ] is equivalent to the vanishing of the quadratic form @xmath216 , i.e. @xmath199)=0 \ \ \mbox{if and only if } \ \
( d+\iota)\tilde{\omega}=0.\ ] ] we see that for the case on which the gauge group is @xmath1 , the equations of cancellation of anomalies , namely that for all @xmath217 @xmath194 are equivalent to the existence of an equivariant extension on @xmath205 of the cohomology class @xmath147 $ ] .
now we are ready to give examples on both the existence and the non existence of equivariant extensions of @xmath145 .
let @xmath218 and consider the diagonal injection @xmath219 which induces the adjoint action of @xmath4 on @xmath220 . in this case
the cohomology of the homotopy quotient @xmath221 is isomorphic to the cohomology of @xmath4 tensor the cohomology of @xmath62 : @xmath222 this isomorphism can be obtained from the serre spectral sequence associated to the fibration @xmath223 which collapses at level 2 because the classes in @xmath224 can be lifted to classes to @xmath225 : take a primitive class in @xmath226 ( namely a class in @xmath226 which is in the image of a primitive element in @xmath227 of one of the differentials of the serre spectral sequence associated to the fibration @xmath228 ) pull it back to @xmath229 via the evaluation map where @xmath230 is the space of free loops of @xmath62 , then integrate over @xmath231 and get a class in @xmath232 of degree 1 less ; this class in @xmath232 once restricted to the based loops @xmath233 is precisely the class in @xmath227 that defined the primitive class in @xmath234 that we started with ; finally recall that @xmath235 .
if we take @xmath236 , we know by theorem [ equivalence of extension for sl(n , r ) ] that the existence of an equivariant extension of @xmath147 $ ] is equivalent to the cancelation of anomalies for this gauged action , and since the class @xmath147 $ ] can be extended to an equivariant one , we conclude that in this case there is an anomaly cancelation .
whenever the action of @xmath107 on @xmath138 is obtained by a left action induced by an inclusion @xmath238 , we have that the @xmath4 action on @xmath138 is free and therefore the homotopy quotient @xmath239 and the quotient @xmath240 are homotopy equivalent . for @xmath109
the serre spectral sequence tells us that @xmath199)=c_2 \in h^4(bsl(n,{\ensuremath{{\mathbb r}}});{\ensuremath{{\mathbb r}}})\ ] ] and therefore the class @xmath147 $ ] does not extend to an equivariant one . by theorem [ equivalence of extension for sl(n ,
r ) ] we know that this implies that there is no cancellation of anomalies . we conclude that for left free actions of the group @xmath1 for @xmath109 there must exist @xmath217 such that @xmath241 | consider a manifold endowed with the action of a lie group .
we study the relation between the cohomology of the cartan complex and the equivariant cohomology by using the equivariant de rham complex developed by getzler , and we show that the cohomology of the cartan complex lies on the @xmath0 row of the second page of a spectral sequence converging to the equivariant cohomology .
we use this result to generalize a result of witten on the equivalence of absence of anomalies in gauged wzw actions on compact lie groups to the existence of equivariant extension of the wzw term , to the case on which the gauge group is the special linear group with real coefficients . |
shot noise in a mesoscopic conductor is a consequence of the stochastic character of electron tunneling and of the discreteness of charge . unlike thermal noise , shot noise describes the non - equilibrium fluctuations of current
; therefore , the study of current fluctuations can provide further understanding of properties related to correlation mechanisms , internal energy scales or the carrier statistics which can not be obtained by measuring the average current @xcite .
a well studied example of physical processes where electron correlations play a dominant role is the phenomenon of coulomb blockade . in a system of small tunneling junctions , due to the large electrostatic energy ( as compared to temperature or voltages ) ,
the electronic charge is transported one by one .
this effect leads to many remarkable features in transport properties and has been a subject of extensive study for the last decades @xcite . as an example
, the strong dependence of the current - voltage characteristics on the gate charge was exploited to use a single electron transistor ( set ) as a highly sensitive charge detector @xcite , and proposed as measuring apparatus of the charge state of a josephson quantum bit @xcite .
since it leads to a strong correlation of consecutive tunneling events , coulomb blockade has turned out to manifest itself in a peculiar way on the current noise .
such an effect has been studied both in the sequential tunneling @xcite and in the cotunneling regime @xcite .
additional interest in studying noise in single electron devices comes from the recently proposed schemes that employ current fluctuation measurements to detect entanglement in solid state systems @xcite .
an even richer scenario occurs when the coherence of charge carriers is maintained over a significant portions of the system .
such a circumstance is encountered quite often when the tunneling junctions are superconducting . in this case , the charge carriers are cooper pairs , and their coherent tunneling across the junctions gives rise to a series of pronounced structures in the @xmath0-@xmath1 characteristics at sub - gap voltages @xcite . furthermore , as analyzed in ref . , the scattering of quasi - particles ( and consequently the shot noise ) in superconducting point contacts
is significantly enhanced in the presence of the supercurrent produced by a coherent flow of cooper pairs . in this paper
, we analyze a superconducting double tunnel junction device , operating in a suitably chosen bias voltage regime , such that one of the junctions of the set is on resonance for cooper pair tunneling ( the case where cooper pair resonance occurs on both junctions has been recently analyzed in ref . ) . the interplay between coherence and interaction
is explored by sweeping the operating point of the device through the cooper pair resonance .
we will show that the fluctuations of the charge on the central island are sensitive to both coulomb blockade and quantum coherence .
more pronounced effects arise in the regime in which the rates of incoherent quasi - particle tunneling matches the frequency of coherent cooper - pair oscillation .
this gives rise to an enhanced fluctuation of charge in the central - island and to a substantial suppression of the current noise . by investigating the statistics of the tunneling events ,
we show that the suppression in the shot noise is related to the deviation of the counting statistics from the poissonian distribution .
the probability distribution of tunneling events exhibits a parity dependence and remains non - poissonian in a wide range of parameter values .
the interplay between coherence and coulomb blockade affects the overall charge transport and is also clearly observed in the finite frequency behavior of the current noise .
its power spectrum displays a sharp resonance peak at the josephson frequency , resulting from coherent oscillations between two quantum states .
the work presented here applies to the setup used in a recent experiment @xcite to probe the coherent evolution of quantum states in a cooper pair box as well as in an earlier experiment @xcite on resonant cooper pair tunneling . in this paper
we extend the results of ref . .
the paper is organized as follows . in section
[ long : sec2 ] , we introduce the model to describe the set transistor and we describe the relevant processes involved in the josephson - quasiparticle cycle . in the same section
we also introduce the master equation , in its general form , used to obtain all the results of this work .
the solution of the master equation is worked out to analyze in detail the fluctuations of the charge on the central electrode and , in particular , the spectrum of the charge fluctuation as measured by a detector coupled to the island ( section [ long : sec3 ] ) , the counting statistics for the tunneled charges ( section [ long : sec4 ] ) , and the shot noise of the current ( section [ long : sec5 ] ) . in section [
long : sec5 ] we discuss in some details the properties of the shot noise at zero frequency , making explicit the results already contained in the counting statistics . in the same section we also discuss the frequency dependence of the shot noise .
section [ long : sec6 ] is devoted to the conclusions .
the system we consider is a superconducting set ( see fig .
[ fig01 ] ) , which consists in a small central electrode ( island ) connected by tunnel junctions to two leads and capacitively coupled to a gate electrode .
the electrostatic energy of the island can be adjusted by controlling the gate voltage @xmath2 .
a transport voltage @xmath1 is applied to the outer leads , determining the current flowing through the device .
letting @xmath3 and @xmath4 be the gate and left and right junction capacitances , respectively , the electrostatic charging energy is given by @xmath5 where @xmath6 is the total capacitance of the island .
the device operates in the coulomb blockade regime , i.e. with the charging energy much larger than both the josephson coupling energy @xmath7 and the thermal energy @xmath8 ( @xmath9 ) , but still much smaller than the superconducting gap @xmath10 which we suppose to be the largest energy scale in the problem .
charge can pass through the tunnel barriers coherently and incoherently .
in addition to quasi - particle tunneling , josephson effect allows to maintain coherence in the cooper - pair tunneling processes
. this coherence does play an important role in transport as long as charge states which differ by one cooper pair are almost degenerate .
the hamiltonian of the system is given by @xmath11 where @xmath12 ( @xmath13 ) is the bcs hamiltonian of the left ( l ) , right ( r ) lead and of the central island ( i ) .
the tunneling hamiltonian @xcite is @xmath14 \,,\ ] ] where @xmath15 is the tunneling amplitude and @xmath16 ( @xmath17 ) creates ( annihilates ) a particle with momentum @xmath18 and energy @xmath19 in electrode @xmath20 .
the variable @xmath21 is the superconducting phase difference at the left ( right ) junction and it is canonically conjugated to the number @xmath22 of electrons that have passed across the left ( right ) junction _ out of _ the central electrode ( @xmath23=i\delta_{jk}$ ] ) .
finally , @xmath24 is the electrostatic energy , @xmath25 where @xmath26 is the number of excess electrons on the central island , while @xmath27 is the offset charge due to the applied voltages . for later convenience
it is useful to define the part of the hamiltonian which accounts for the coherent dynamics of the macroscopic variable @xmath28 .
it includes the charging and the josephson terms . by properly adjusting the bias and gate voltages
, one can put either the right or left junction at such a resonance for cooper - pair tunneling .
we consider the case of resonance across the left junction , and consequently we keep only the corresponding josephson coupling @xmath29 a quasi - particle tunneling _ into _ ( _ out of _ ) the island across the junctions leads to the transition @xmath30 ( @xmath31 ) of the island charge .
the rates of these incoherent processes are given by the relation @xmath32
\frac{{\mathop{\mathrm{im}}\nolimits}i_\mathrm{qp}({{\mathcal e}}_{n,\pm}^{l / r})}{2e } \,,\ ] ] where @xmath33 , @xmath34 , @xmath35 , and @xmath36 is the quasi - particle tunneling current at the bias voltage @xmath37 in the absence of charging effects ( see , e.g. , ref . for an explicit expression of @xmath38 in terms of @xmath15 ) .
since the set transistor operates in the charge regime ( @xmath39 ) , we can take advantage of the strong suppression of charge fluctuations to use the eigenstates of @xmath28 as basis states for the island .
moreover , we focus on the bias regime @xmath40 where only the two charge states with @xmath41 and @xmath42 , are nearly degenerate .
such a condition implies that quasi - particle tunneling only takes place from the central island toward the right electrode , while the left junction allows only for coherent cooper pair tunneling .
furthermore , we suppose that the josephson energy of the right junction is negligible ( the corresponding term has already been omitted from @xmath43 , which is justified within the rotating wave approximation ) .
all these conditions are met in the recent experiment by nakamura _
et al . _
@xcite , designed to probe the state of the island via the detection of the incoherent tunneling current . in this situation one can imagine that the coherent cooper - pair tunneling occurring across the left junction is interrupted from time to time by quasi - particle tunneling across the right junction , as sketched in fig.[fig01 ] ( b ) . due to the strong coulomb blockade
, it suffices to keep the three charge states , @xmath44 , and two tunneling rates , @xmath45 and @xmath46 ; the other tunneling rates are exponentially suppressed . in order to simplify the notation , we assume that @xmath47 , which is a very good approximation in the regime we are interested in .
for example , in the experiment of ref . , @xmath48 and @xmath49 .
the transport properties of the system in the set - up described above can be well described in terms of two variables , either @xmath28 and @xmath50 or @xmath28 and @xmath51 ( @xmath26 ) . however , the quantum dynamics of these system is affected by the quantum noise due to the coupling to the environment provided by the fermionic bath . in order to describe this effect
, we adopt a master equation approach , which has been widely used to describe quantum open systems @xcite .
a master equation for the reduced density matrix @xmath52 is obtained by taking the trace over the fermionic degrees of freedom from the liouville equation ( @xmath53 ) @xmath54\ ] ] for the density matrix @xmath55 of the system plus environment .
the resulting equation can then be written in the lindblad form as @xcite @xmath56 + \frac{1}{2}\sum_{n=1,2}\gamma_n\bigl [ 2l_n\,\rho(t)\,l_n^\dag - l_n^\dag l_n\,\rho(t ) - \rho(t)\,l_n^\dag l_n \bigr ] \,.\ ] ] here @xmath57 is a lindblad operator corresponding to the quantum jump @xmath31 and @xmath58 , i.e. , in the @xmath59-basis @xmath60 .
the first term describes a purely phase - coherent dynamics , while the second one is responsible for both dephasing and relaxation due to the quasi - particle tunneling .
the solution of the eq . behaves in distinct ways in the two limiting cases of strong and weak coupling with the quasi - particle reservoir . in the _ strong dephasing limit _ ( either @xmath61 or @xmath62 , see below ) , the dephasing time @xmath63 , which describes the decay of the off - diagonal elements of @xmath64 to their stationary values , is small compared to the relaxation time @xmath65 which sets the time - scale for the variation of the diagonal elements ( i.e. population of the charge states ) . the relaxation time is given by @xmath66 on the other hand , in the _ weak dephasing limit _ ( @xmath67 )
, there is no such a clear separation of time scales ; both the diagonal and off - diagonal elements vary over the same time scale @xmath68 .
a first insight into the interplay among coherent cooper - pair tunneling , coulomb blockade , and incoherent quasi - particle tunneling can be obtained by examining the fluctuations of the charges on the island as a function of the quasi - particle tunneling rate @xmath69 , the gate voltage , and the josephson coupling energy . in this case
, we only need to keep track of the variable @xmath28 , and thus define a reduced density matrix for the central island charge , @xmath70 , which satisfies an equation identical to eq . , but with @xmath71 now operating on the reduced ` @xmath28 ' space only @xcite . in the stationary state , the master equation has the solution ( @xmath72 ) [ long : sigma ] @xmath73 here @xmath74 measures the energy difference between the state with @xmath41 and @xmath42 charge on the island ; the cooper pair resonance corresponds to @xmath75 . from the stationary - state solution
, we evaluate the characteristic function @xmath76 for the quantum variable @xmath28 .
it is given by @xmath77 from @xmath78 one can evaluate all the statistical moments of the charge on the island ; in particular , we concentrate on the variance @xmath79 ^ 2}\ ] ] shown in fig .
[ fig02 ] . of the charge on the island as a function of @xmath80 and @xmath81.,title="fig : " ] of the charge on the island as a function of @xmath80 and @xmath81.,title="fig : " ] an interesting point about the result in eq .
is that the maximum of the variance is found at @xmath82 for @xmath75 . in other words ,
the fluctuation is enhanced when the decoherence time matches the time for the josephson coherent oscillations .
moreover as @xmath83 increases ( up to @xmath84 ) , the optimal value of @xmath69 decreases as @xmath85 ; meaning that @xmath81 enhances the _ effective _ dephasing rates .
these two features will appear more clearly when we discuss the statistics ( section [ long : sec4 ] ) and noise ( section [ long : sec5 ] ) of the transport across the junctions .
an important quantity to consider is the fluctuation spectrum for the number of electron charge residing on the island .
as already discussed by many authors , it is important to include the back - action of the measuring apparatus @xcite , which could be a set transistor capacitively coupled to the central island .
we include the charge detector coupling via an hamiltonian term of the form @xmath86 where @xmath87 is a detector operator .
assuming the correlation time for the detector to be the fastest time scale of the problem , we write ( here we follow averin s treatment @xcite ): @xmath88 the non - zero value of @xmath89 is the essential cause of the measurement back - action .
indeed , a term proportional to @xmath89 enters the master equation eq.([long : mastereq1 ] ) , thus affecting the time evolution of the system variables . to effectively measure the charge number , we look at an output detector operator @xmath90 , which , in the linear regime , evolves as @xmath91 , ( see ref . ) .
the response coefficient @xmath92 is determined by the imaginary part of the equilibrium correlation function @xmath93 .
furthermore , @xmath92 can be related to @xmath89 so that we can write for the signal to noise ratio @xmath94 where it is assumed that the real part of the @xmath90@xmath87 correlator vanishes ( which is the most favorable case for a measurement ) . here
@xmath95 is the charge number fluctuation spectrum evaluated as @xmath96 where the time evolution is obtained from a modified master equation including the back - action .
note that here we use the symmetric correlation function since the island charge itself is coupled to the quasi - particle bath , so that the detector can also receive energy from the system . ) , with the indicated back - action rate values and with @xmath97 ( left ) and @xmath98 ( right).,title="fig : " ] ) , with the indicated back - action rate values and with @xmath97 ( left ) and @xmath98 ( right).,title="fig : " ] as shown in fig .
( [ fig03 ] ) , the spectrum displays a resonance peak at the josephson frequency which is broadened by the quasi - particle rate @xmath69 ( the peak is only visible in the weak dephasing regime , otherwise it is completely washed out independently of the value of @xmath89 ) .
as the back - action rate @xmath89 increases , a maximum develops at zero frequency , which finally hides the resonance structure .
this enhanced zero frequency noise results from incoherent transition induced by the detector coupling . near the maximum at @xmath99 , and for @xmath100 ,
the spectrum takes the approximate form @xmath101
we turn now to a description of the statistical distribution of the number of charges transmitted through the system during a period @xmath102 , @xcite .
specifically , we will examine the probability @xmath103 that @xmath104 electrons have been transferred across the right junction during the interval @xmath105 $ ] .
we note that @xmath106 where @xmath107 is the joint probability that @xmath51 electrons have passed across the right junction up to the time @xmath108 and @xmath109 electrons up to time @xmath110 . to obtain @xmath103
, we define a characteristic matrix @xmath111 defined so that @xmath112 is the characteristic function for @xmath103 .
namely , @xmath113 where the trace is taken over the states @xmath114 .
following the same procedure that led to eq . , one can show that @xmath115 satisfies the following master equation : @xmath116 + \frac{1}{2}\sum_{n=1,2}\gamma_n \left [ 2 e^{i \theta } l_n g_t l_n^{\dag } - g_t l_n^{\dag } l_n - l_n^{\dag } l_n g_t \right]\ ] ] with the initial condition @xmath117 . here
we will consider two limiting cases for the solution , the strong and the weak dephasing limit ( see the discussion at the end of section [ long : sec2 ] ) .
we find that in the strong dephasing case ( @xmath118 or @xmath119 ) @xmath120 \exp\left[- \frac{\gamma_\mathrm{r}\tau}{2 } ( 1-z^2)\right]\ ] ] while in the weak dephasing limit at resonance ( @xmath121 ) @xmath122}{f(z ) } \sinh\frac{\gamma\tau f(z)}{4 } \right ] \\
+ \frac{\gamma}{e_j}(1-z){\mathop{\mathrm{im}}\nolimits}\sigma_{02}(t ) \left[\cosh\frac{\gamma\tau f(z)}{4 } + \frac{4z+1}{f(z)}\sinh\frac{\gamma\tau f(z)}{4 } \right ] \biggr\ } \exp\left(-\frac{3\gamma\tau}{4}\right ) \\ - \frac{\gamma}{e_j}(1-z)\left\ { { \mathop{\mathrm{im}}\nolimits}\sigma_{02}(t)\cos(e_j\tau ) - \frac{1}{2}\left[\sigma_{00}(t)-\sigma_{22}(t)\right]\sin(e_j\tau ) \right\ } \exp\left(-\frac{\gamma\tau}{2}\right ) \,,\end{gathered}\ ] ] where @xmath123 , and @xmath124 . in figs .
[ fig04 ] and [ fig05 ] , the resulting statistics are shown for the weak and strong dephasing cases , respectively , in the transient state ( i.e. @xmath125 ) .
obtained by numerically solving equation ( [ long : mastereq2 ] ) , in the case of @xmath126 , with initial condition @xmath127 . ] ) and at resonance ( @xmath75 ) . in this regime ,
the counting probability does not depend on the initial condition for @xmath64 ] at @xmath129 in the strong dephasing regime with @xmath130 for the three initial conditions a ) @xmath131 ( dashed line ) , b ) @xmath132 ( dotted line ) c ) @xmath133 ( continuous line ) .,width=226 ] as seen in eqs . and , @xmath115 and hence @xmath103 depend on the charge state of the island at time @xmath108 .
this point is further illustrated in fig .
[ fig06 ] where the counting statistics at short time is shown in the strong dephasing limit for various starting conditions @xmath134 . in the limit of strong dephasing , the counting statistics depends sensitively on the initial state @xmath134 .
this has been exploited in ref . , where the measurement of the quantum state is performed with the system taken far from degeneracy .
on the contrary , the dependence on the initial condition is quickly lost in the weak dephasing regime , since the strong josephson energy can rapidly produce a change in the state , before quasi - particles have any time to be produced . another important limit to consider is the stationary state ( @xmath135 ) , where physical properties do not depend on the initial preparation of the system . in the strong dephasing limit ,
eq . is reduced to the simple form @xmath136 \,.\ ] ] it gives the probability distribution function for the transmitted charges [ countstat ] @xmath137 @xmath138 shows a strong even - odd asymmetry : for even @xmath104 , the distribution is poissonian , but the probability that an odd number of electrons has passed is negligible .
below we will see that this strong parity effect manifests itself as an enhancement of zero - frequency shot noise .
we leave the physical interpretation of the parity effect until we discuss shot noise in section [ long : sec5 ] . in the weak dephasing limit , eq .
is reduced to @xmath139\end{gathered}\ ] ] so that [ p1 ] @xmath140 where @xmath141 this distribution function shows a much weaker ( but still finite ) even - odd asymmetry than the previous case [ cf .
eq . ( [ countstat ] ) ]
. furthermore , the distribution clearly deviates from a poissonian function , indicating that the presence of the strong coherent tunneling of cooper pairs tends to correlate the quasi - particle tunneling events across the right junction .
this is further reflected in the deviations of the current noise from the classical shot noise value ( see discussions in section [ long : sec5 ] ) .
one may expect that for a long waiting time ( @xmath142 ) , implying very large numbers of tunneled charges , the distribution of @xmath104 should approach a gaussian . in particular , this becomes an exact result if the distribution is poissonian . in our case , on the other hand , we have [ long : p2 ] @xmath143 where @xmath144 is a gaussian distribution , @xmath145 \label{normal}\ ] ] with @xmath146 .
the distributions for both even and odd @xmath104 are separately gaussian , but @xmath147 as a whole is not , since even - odd asymmetry is still present . in fig .
[ fig07 ] we compare the stationary - state results for @xmath148 in the weak and strong dephasing limits with the gaussian distribution . at @xmath149 ( a )
@xmath150 and ( b ) @xmath151 . for a comparison , a normal distribution function given in eq .
( [ normal ] ) is also shown ( dashed line).,title="fig:",width=226 ] at @xmath149 ( a ) @xmath150 and ( b ) @xmath151 . for a comparison , a normal distribution function given in eq .
( [ normal ] ) is also shown ( dashed line).,title="fig:",width=226 ] at @xmath152 for @xmath153 ( filled circle ) . for a comparison ,
the poissonian distribution is also plotted ( empty circle ) .
] finally , it is interesting to understand what happens to the stationary counting probability @xmath154 in the intermediate regime , i.e. , when the dephasing rate is comparable to the josephson energy .
unfortunately , an analytic expression is not available in this case ; the numerical results , however , are shown in fig.[fig08 ] , where one can see that the distribution function deviates significantly from a poissonian , being suppressed ( enhanced ) for odd ( even ) @xmath104 .
a deeper insight into the transport process can be obtained in the frequency domain , from a careful analysis of the spectral power of current fluctuations .
the zero - frequency shot noise could be directly determined by the second moment of the counting probability eq .
( [ serve ] ) , see ref . .
here , however , we follow a different route which allows us to get the entire current spectrum . to this end
, we define the noise spectrum as @xmath155 where @xmath156 and @xmath157 .
the total current @xmath158 through the system is related to the _ tunneling _
currents @xmath159 across each junction by @xcite @xmath160 to simplify the evaluation of @xmath161 , it is convenient to introduce the spectral densities of currents flowing across the individual junctions and the cross correlation spectral powers . therefore , in a way analogous to eq .
( [ sdef ] ) , we write ( @xmath162 ) @xmath163 which allows to express the total shot noise spectrum in the form @xmath164 \,.\end{gathered}\ ] ] in the stationary state @xmath165 , so that @xmath166 in the zero - frequency limit . in the opposite limit ( @xmath167 ) , @xmath168 @xcite . in our case ,
the left junction is ( nearly ) at resonance for the cooper pair tunneling and therefore @xmath169 . in order to obtain @xmath170 we have to calculate two - time correlation functions .
we follow the standard procedure based on the quantum regression theorem @xcite and define the auxiliary matrices [ long : chieta ] @xmath171 where the index @xmath172 runs over the left and right junctions ( @xmath173 ) .
these auxiliary operators , @xmath174 and @xmath175 , satisfy a master equation of exactly the same form as eq .
( but with respect to @xmath102 instead of @xmath108 and , of course , with different initial conditions ) .
their relevance can be understood by noticing that the correlation functions can be expressed directly in terms of their average values@xcite : @xmath176 @xmath177 and @xmath178 the problem is now reduced to ( a ) solving a master equations of the form given in eq . to get @xmath179 , @xmath180 , and @xmath181 for respective initial conditions , and ( b ) evaluating eqs . , , and to obtain the correlation functions . following this procedure and performing the fourier transforms of the resulting correlation functions
, we find that in the stationary state @xmath182 @xmath183 and @xmath184 where we have used the bra - ket notations @xmath185 @xmath186 and @xmath187 from eqs .
, , , and , it follows that the zero - frequency noise is given by @xmath188 in the strong dephasing limit ( @xmath189 ) , the second term in eq . becomes negligibly small , as it vanishes as @xmath190 .
therefore , the zero - frequency shot noise is enhanced approximately by a factor @xmath191 compared with its classical value , @xmath192 .
this can be understood in terms of the josephson quasi - particle ( jqp ) cycle @xcite .
because of the fast quasi - particle tunneling across the right junction , each cooper pair that has tunneled into the central island breaks up immediately into quasi - particles , and quickly tunnels out .
the charge is therefore transferred in units of @xmath193 ( compared with @xmath194 in classical charge transfer ) for each jqp cycle .
this was already confirmed in the counting statistics of the transmitted charges . according to eq .
, the probability that an odd number of electrons are transferred is zero and charges are transferred only in pairs . in the weak ( @xmath195 ) and moderate ( @xmath196 ) dephasing limits , the semiclassical jqp
picture breaks down and we do not have shot noise enhancement any longer . in the limit @xmath195 ,
the period of oscillations of cooper pair is very short compared to the typical time for for quasi - particles to tunnel out of the central island .
the system can be regarded as a single - junction circuit , where the quasi - particle tunneling events are independent .
the fano factor @xmath197 , becomes much closer to unity in this limit .
the small deviation from the poisson value is due to the fact that the quasi - particle tunneling events can not be considered as independent because coulomb blockade allows only one cooper pair to oscillate coherently across the left junction .
therefore , the tunneling process corresponding to @xmath198 is likely to be followed by @xmath199 .
it is clear that this behavior is related to the residual even - odd asymmetry we found in the counting statistics , eq .
, even in the long waiting - time limit ( @xmath200 ) , eq . . with moderate dephasing ( @xmath196 ) ,
quasi - particle tunneling events across the right junction are strongly affected by the _
coherent _ oscillation of cooper pairs across the left junction .
indeed , this effect gives rise to the significant deviation from the poissonian distribution of the tunneling statistics , eq . .
most remarkably , it leads to a suppression of the shot noise . the strongest suppression , by a factor of @xmath201 ,
is achieved at resonance ( @xmath75 ) for @xmath153 , see fig .
[ fig09 ] .
this is reminiscent of the shot noise suppression in ( non - superconducting ) double - junction systems@xcite , whose maximal suppression is by factor @xmath202 for the symmetric junctions .
we emphasize , however , that in the latter case , the coherence was not essential . in our case , on the contrary , the role of coherence becomes evident by noticing that the dip in fano factor disappears when moving away from the resonant condition as shown in fig .
[ fig09 ] . .
the dip in the noise is most pronounced at resonance ( @xmath75).,width=226 ] ) for @xmath203 , and ( b ) at a fixed weak dephasing rate ( @xmath204 ) for @xmath205.,title="fig : " ] ) for @xmath203 , and ( b ) at a fixed weak dephasing rate ( @xmath204 ) for @xmath205.,title="fig : " ] as a function of frequency @xmath206 in the ( a ) strong ( @xmath207 ) and ( b ) weak ( @xmath208 ) quasi - particle tunneling limits .
for both plots , @xmath75 and @xmath209 were assumed.,title="fig : " ] as a function of frequency @xmath206 in the ( a ) strong ( @xmath207 ) and ( b ) weak ( @xmath208 ) quasi - particle tunneling limits .
for both plots , @xmath75 and @xmath209 were assumed.,title="fig : " ] in figs .
[ fig10 ] and [ fig11 ] we show the typical behavior of the finite - frequency noise spectrum in the ( a ) strong and ( b ) weak dephasing limits .
it is interesting to notice that ( only ) in the weak dephasing limit ( @xmath195 ) , there is a resonance peak of the form @xmath210 where the resonance frequency is given by @xmath211 \,.\ ] ] clearly , the peak is an effect of coherent quantum oscillations between the two energy levels separated by @xmath212 , induced by the josephson effect across the left junction . as expected , the resonance peak is reduced in its height and broadened in its width with increasing @xmath69 . on the contrary , as @xmath81 increases , the peak gets sharper and the peak height increases quadratically with @xmath81 .
however , this should not be confused with the zero - frequency case , where @xmath81 effectively enhances the decoherence effects .
as @xmath81 increases , the josephson oscillation across the left junction becomes faster , and there are less chances that it is interrupted by the quasi - particle tunneling across the right junction .
this , in turns , implies that the coherent oscillation is better defined and the spectral component ( especially @xmath213 ) at frequency @xmath214 is highly enhanced . for a vanishingly small quasi - particle tunneling rate
, @xmath215 would approximately become a delta like function , centered at @xmath216 .
however , one should not be misled by this result , since the noise is always proportional to the average current , which vanishes in this limit .
it is worth mentioning here on the relation between this result and the description of the noise output from linear detector @xcite . in the setup considered in this work
, the right electrode has the role of the detector of cooper pair oscillations ; since the total current in the circuit is due to quasi - particle tunneling ( i.e. it is a _ dissipative _ current ) , the output signal may be considered as classical .
this has to be compared to the case of a detector measuring the charge on the island .
there , the back - action of the detector was essential to produce an observable result . in the case of the current , instead
, the `` detector '' is intrinsically part of the system and it couples to the observed quantity in an essentially non - linear way .
in this paper we considered properties of the distribution of the transmitted charge in a superconducting set tuned close to a cooper pair resonance .
the dominant process to the transport in the regime considered here , is the jqp cycle , a process in which coherent cooper pair oscillations are accompanied by ( incoherent ) quasi - particle tunneling .
the interplay between the coherence and the strong coulomb blockade manifests itself in various ways both in the counting statistics and in the shot noise .
we found two distinct regimes characterized by different ratios of the time scales for dephasing and relaxation , @xmath217 in the strong dephasing limit or @xmath218 in the opposite case of weak dephasing .
a generic feature of the counting statistics , valid in both the regimes , is its even - odd asymmetry related to the fact that charge transport is mediated by the cooper pair tunneling .
other properties are more pronounced in one of the two regimes .
an example is the dependence of @xmath103 on the initial time @xmath108 .
this is clearly visible in the strong dephasing limit while quickly lost in the weak dephasing regime since , because of the strong josephson energy , the state changes significantly before quasi - particles have any time to be produced .
another important point is that the counting statistics is not poissonian , due to the relevance of correlations between different tunneling events . as a consequence
the fano factor is different from the classical value .
the maximal suppression of the zero - frequency shot noise is observed when the quasi - particle tunneling rate is comparable to the frequency scale of the coherent cooper pair oscillations .
we finally investigated the shot noise at finite frequencies , which shows a resonance peak at the josephson oscillation frequency .
this maximum can be interpreted as an effect of coherent quantum transitions between the two energy levels involved in the transport phenomena in the device .
we thank d.v .
averin , y. blanter , and j. siewert for very useful discussions .
we acknowledge financial support from european community ( ist - fet - squbit ) and infm ( pais - tin ) .
m .- s.c . acknowledges the support from the swiss - korean outstanding research efforts award program .
d. v. averin , in _
`` exploring the quantum classical frontier : recent advances in macroscopic and mesoscopic quantum phenomena '' _ , eds .
j. r. friedman and s. han , to be published ( cond - mat /0004364 ) ; d. v. averin , preprint ( 2002 ) , cond - mat/0202082 .
d. loss and e.v .
sukhorukov , phys .
* 84 * , 1035 ( 2000 ) ; m .- s . choi , c. bruder , and d. loss , phys .
b * 62 * , 13 569 ( 2000 ) ; f. plastina , r. fazio , and g. m. palma , phys . rev .
b * 64 * , 113306 ( 2001 ) .
a microscopic derivation is given in g .-
ingold and y. v. nazarov , in _ single charge tunneling : coulomb blockade phenomena in nanostructures _ , edited by h.
grabert and m. devoret ( plenum press , new york , 1992 ) . | we analyze charge tunneling statistics and current noise in a superconducting single - electron transistor in a regime where the josephson - quasiparticle cycle is the dominant mechanism of transport . due to the interplay between coulomb blockade and josephson coherence ,
the probability distribution for tunneling events strongly deviates from a poissonian and displays a pronounced even
odd asymmetry in the number of transmitted charges .
the interplay between charging and coherence is reflected also in the zero - frequency current noise which is significantly quenched when the quasi - particle tunneling rates are comparable to the coherent cooper - pair oscillation frequency . furthermore the finite frequency spectrum shows a strong enhancement near the resonant transition frequency for josephson tunneling . |
all graphs considered in this paper are simple , connected and undirected .
we follow the terminology and notation of bondy and murty @xcite .
an edge - colored graph @xmath1 is _ rainbow connected _ if any two vertices are connected by a path whose edges have distinct colors .
the _ rainbow connection number _ @xmath14 of @xmath1 , introduced by chartrand et al .
@xcite , is the minimum number of colors that results in a rainbow connected graph @xmath1 .
later , another generalization of rainbow connection number was introduced by chartrand et al.@xcite in 2009 .
a tree @xmath0 is a _ rainbow tree _ if no two edges of @xmath0 are colored the same .
let @xmath2 be a fixed integer with @xmath15 .
an edge coloring of @xmath1 is called a _ @xmath2-rainbow coloring _ if for every set @xmath3 of @xmath2 vertices of @xmath1 , there exists a rainbow tree in @xmath1 containing the vertices of @xmath3 .
@xmath2-rainbow index _
@xmath5 of @xmath1 is the minimum number of colors needed in a @xmath2-rainbow coloring of @xmath1 .
it is obvious that @xmath16 .
let @xmath2 be a positive integer .
a subset @xmath17 is a @xmath2-dominating set of the graph @xmath1 if @xmath18 for every @xmath19 .
the @xmath2-domination number @xmath20 is the minimum cardinality among the @xmath2-dominating sets of @xmath1 .
note that the @xmath21-domination number @xmath22 is the usual domination number @xmath23 .
a subset @xmath3 is a connected @xmath2-dominating set if it is a @xmath2-dominating set and the graph induced by @xmath3 is connected .
the connected @xmath2-domination number @xmath24 represents the cardinalities of a minimum connected @xmath2-dominating set . for @xmath25 , we write @xmath26 instead of @xmath27 .
chakraborty et al .
@xcite showed that computing the rainbow connection number of a graph is np - hard .
so it is also np - hard to compute @xmath2-rainbow index of an arbitrary graph .
chandran et al .
@xcite use a strengthened connected dominating set ( connected @xmath11-way dominating set ) to prove @xmath28)+3 $ ] .
this led us to the investigation of what is strengthening of a connected dominating set which can apply to consider @xmath6-rainbow index of a graph .
recently , for @xmath6-rainbow index , li et al . did some basic results and they obtained the following theorem .
@xcite[thm0 ] let @xmath1 be a @xmath11-connected graph of order @xmath29 @xmath30 .
then @xmath31 , with equality if and only if @xmath32 or @xmath1 is a spanning subgraph of @xmath6-sun or @xmath1 is a spanning subgraph of @xmath33 or @xmath1 is a spanning subgraph of @xmath34 . here , a @xmath6-sun is a graph @xmath1 which is defined from @xmath35 = @xmath36 by adding three edges @xmath37 , @xmath38 and @xmath39 .
chartrand et al .
@xcite obtained that for integers s and t with @xmath40 , @xmath41{t},4\}$ ] .
thus , @xmath42 .
@xcite consider the regular complete bipartite graphs @xmath43 .
they show @xmath44 for integer @xmath45 with @xmath46 . in this paper , we focus on @xmath6-rainbow index . in section @xmath11 , we adopt connected @xmath11-dominating set to study @xmath6-rainbow index .
a coloring strategy is obtained which uses only a constant number of extra colors outside the dominating set .
we prove that @xmath47)+4 $ ] , where @xmath10 is the connected @xmath11-dominating set of @xmath1 . in section @xmath6 , we determine a sharp bound of @xmath6-rainbow index for @xmath48 ( @xmath49 ) and an upper bound for @xmath13-free . in section @xmath50 , we investigate a sharp upper for @xmath7 of general graphs by block decomposition and an upper bound for graphs with @xmath8 by connected @xmath11-dominating set .
[ thm5 ] let @xmath1 be a connected graph with minimal degree @xmath51 .
if @xmath10 is a connected @xmath11-dominating set of @xmath1 , then @xmath47)+4 $ ] and the bound is tight .
we prove the theorem by demonstrating that @xmath1 has a 3-rainbow coloring with @xmath9)+4 $ ] colors . for @xmath52 , its neighbors in @xmath10 will be called foots of @xmath53 , and the corresponding edges will be called legs of @xmath53 .
we give @xmath54 $ ] a 3-rainbow coloring using colors @xmath55))$ ] .
let @xmath56 .
partition @xmath57 into sets @xmath58 as follows .
@xmath59 is the set of all isolated vertices of @xmath60 . in every nonsingleton connected component of @xmath60 ,
choose a spanning tree .
so we construct a forest on @xmath61 and choose @xmath62 and @xmath63 as any one of the bipartitions defined by this forest .
color every @xmath64 edge with @xmath65 or @xmath66 where each of @xmath67 appears at least once , every @xmath68 edge with @xmath65 or @xmath69 where each of @xmath70 appears at least once , every edge between @xmath62 and @xmath63 with @xmath71 .
since @xmath1 has a minimal degree @xmath51 , every vertex in @xmath59 will have at least three neighbors in @xmath10 .
color two of them with @xmath65 and @xmath69 and all the others with @xmath71 .
next , we show that under such an edge coloring for any three vertices in @xmath10 there exists a rainbow tree containing them . for three vertices
@xmath72 , there is already a rainbow tree containing them in @xmath54 $ ] . for three vertices
@xmath73 @xmath74 ( or @xmath75 ) , join any one leg of @xmath76 ( or @xmath65 , @xmath69 ( @xmath66 ) legs of @xmath77 and @xmath76 ) with a rainbow tree containing the corresponding foot ( or two foots ) , @xmath53 and @xmath77 ( or @xmath53 ) in @xmath54 $ ] .
now we consider the case three vertices @xmath73 @xmath78
. for three vertices @xmath79 , join three edges which color @xmath80 and @xmath69 with a rainbow tree containing the corresponding foots @xmath81 in @xmath10 . for two vertices @xmath82 , @xmath83 ,
join a @xmath65 leg of @xmath76 and @xmath84 legs of @xmath85 with a rainbow tree containing the corresponding foots in @xmath54 $ ] .
consider one vertex @xmath86 , two vertices @xmath87 .
if @xmath88 , join a @xmath71 leg of @xmath53 and @xmath67 legs of @xmath77 and @xmath76 with a rainbow tree containing the corresponding foots in @xmath54 $ ] .
if @xmath89 or @xmath90 , join a @xmath71 leg of @xmath53 and @xmath91 legs of @xmath77 and @xmath76 with a rainbow tree containing the corresponding foots in @xmath54 $ ] . then consider three vertices @xmath92 .
if @xmath93 , we know , for @xmath94 , @xmath53 has a neighbor @xmath95 .
@xmath96 edge ( colored @xmath71 ) and @xmath69 leg of @xmath97 , join @xmath65 leg of @xmath77 and @xmath66 leg of @xmath76 with a rainbow tree containing the corresponding foots in @xmath54 $ ] .
similarly , in other cases , we can find a rainbow tree containing them .
hence , @xmath1 has a 3-rainbow coloring with @xmath9)+4 $ ] colors .
the proof of tightness is given in the next section .
in this section , we consider two special graphs : complete bipartite graphs @xmath48 and @xmath13-free graphs .
[ thm6 ] for any complete bipartite graphs @xmath48 with @xmath98 , @xmath99 , and the bound is tight . because @xmath48 with @xmath98 is a 2-connected graph , by theorem [ thm0 ] , we have , @xmath100 .
the equality clearly holds for @xmath101 since @xmath102 .
thus , to complete the proof , it suffices to show @xmath103 , @xmath98 .
let @xmath104 and @xmath105 be the two partite sets of @xmath48 , where @xmath106 and @xmath107 .
suppose @xmath108 .
clearly we can find a connected 2-dominating set @xmath109 of @xmath48 .
in addition , @xmath110 is connected , @xmath111 , by theorem [ thm5 ] , @xmath112)+4=6 $ ] . to prove the sharpness of the above upper bound , we derive the following claim . * claim . * for any @xmath113 , @xmath114 , @xmath115 .
firstly , we consider the graph @xmath116 .
we may assume that there exists a 3-rainbow coloring @xmath117 of @xmath116 with @xmath2 colors .
corresponding to this 3-rainbow coloring , for every vertex @xmath118 in @xmath105 , there is a color code , code(@xmath118 ) , assigned @xmath119 , @xmath120 .
observe that any three vertices have at least three distinct colors appeared in their color codes .
thus , we know that at most two vertices have the common code except possibly when @xmath121 . otherwise , there is no rainbow tree containing these three vertices which have the same code and at most two colors in color code .
therefore , when @xmath122 , there must exist three vertices @xmath123 , @xmath124 , @xmath125 such that code ( @xmath123)=code(@xmath124)=code(@xmath125)=@xmath126 and @xmath121 . if a rainbow tree containing @xmath127 , it must contain @xmath128 and @xmath129 to guarantee its connectivity , where @xmath129 belongs to @xmath105 and code(@xmath129)=@xmath130 , where @xmath131 , @xmath132 are different from each other , @xmath133 .
thus @xmath134 .
so @xmath135 , when @xmath136 .
similarly , we can prove @xmath115 , for @xmath137 , @xmath114 .
thus , this claim also provides the tight proof of the theorem [ thm5 ] . here , we can simply check that the upper bound can not be generalized to the graphs @xmath138 . by the same method used in the above claim
, we may assume that there exists a 3-rainbow coloring @xmath117 of @xmath138 with @xmath2 colors .
corresponding to this 3-rainbow coloring , there is a color code , code(@xmath118 ) , assigned @xmath119 for @xmath139 . observe that at most two vertices have the common code .
it follows , @xmath140 .
thus , @xmath2 is not less than a certain constant when @xmath141 is enough large . to state
next theorem , we need to make more definitions .
a graph @xmath1 is called _ a perfect connected dominant graph _ if @xmath142 , for each connected induced subgraph @xmath62 of @xmath1 . if @xmath1 and @xmath60 are two graphs , we say that @xmath1 is _ @xmath60-free _ if @xmath60 does not appear as an induced subgraph of @xmath1 . furthermore ,
if @xmath1 is @xmath143-free and @xmath144-free , we say that @xmath1 is _
@xmath145-free_. next , we determine the upper bound for @xmath6-rainbow index of @xmath13-free graphs . zverovich has obtained the following result .
@xcite[thm11 ] a graph @xmath1 is a perfect connected - dominant graph if and only if @xmath1 contains no induced path @xmath146 and induced cycle @xmath147 .
as shown in theorem [ thm5 ] , in order to obtain a better bound of @xmath6-rainbow index , we may turn to a smallest possible connected @xmath11-dominating set . for a graph with minimal degree @xmath51 , b.reed proved the following conclusion in @xcite .
@xcite[thm12 ] if @xmath1 is connected graph with @xmath148 , then @xmath149 . for @xmath13-free graphs @xmath148 , we have @xmath150 . inspired by this result , the extension of the idea of connected dominating set to connected @xmath11-dominating set is what gives the following lemma .
[ lem3 ] let @xmath1 be a connected graph of order @xmath29 with minimal degree @xmath151 .
if @xmath10 is a connected dominating set in a graph @xmath1 , then there is a set of vertices @xmath152 such that @xmath153 is a connected 2-dominating set and @xmath154 .
there are two types of the components of @xmath155 : singletons and connected subgraphs .
let @xmath156 be the set of the singletons , and @xmath157 be the set of the connected components of @xmath155 .
note that @xmath158 . since @xmath151 , for any vertex @xmath159 in @xmath156 , it has at least two neighbors in @xmath10 . in every non - singleton connected component of @xmath157 , we choose a spanning tree
this gives a spanning forest on @xmath160 .
choose @xmath62 and @xmath63 as any one of the bipartitions defined by this forest .
without loss of generality , we suppose that @xmath161 . stage @xmath162 while @xmath163 such that @xmath164 @xmath165 if @xmath166 pick a vertex @xmath167 .
let @xmath168 else @xmath169 @xmath170 clearly @xmath153 remains to be connected .
since stage ends only when any vertex in @xmath160 has at least 2 neighbors in @xmath153 .
so the final @xmath153 is a connected @xmath11-dominating set .
let @xmath2 be the number of iterations executed . since we add a vertex in @xmath62 to @xmath153
, @xmath171 reduces by 1 in every iteration , @xmath172 , so latexmath:[$|d'|\leq for a connected @xmath13-free graph @xmath1 with @xmath148 , we can derive the following result by theorem [ thm5 ] , theorem [ thm11 ] , theorem [ thm12 ] and lemma [ lem3 ] .
[ thm4 ] for every connected @xmath13-free graphs @xmath1 with @xmath8 , @xmath174 .
for every connected @xmath13-free graphs @xmath1 with @xmath8 , from the theorem [ thm11 ] , @xmath175 . and by the theorem [ thm12 ] , we have @xmath149 .
thus , @xmath150 . combining this with lemma [ lem3 ] , the graph @xmath1 have a connected @xmath11-dominating set @xmath10 with order less than @xmath176 .
observe that the connected @xmath11-dominating set @xmath10 can get a 3-rainbow coloring using @xmath177 colors by ensuring that every edge of some spanning tree gets distinct color .
so the upper bound follows immediately from theorem [ thm5 ] .
in this section , we derive a sharp bound for @xmath6-rainbow index of general graphs by block decomposition . and
we also show a better bound for @xmath6-rainbow index of general graphs with @xmath8 by connected @xmath11-dominating set .
let @xmath178 be the set of blocks of @xmath1 , whose element is @xmath179 ; let @xmath180 be the set of blocks of @xmath1 , whose element is @xmath181 ; let @xmath182 be the set of blocks of @xmath1 , whose element @xmath62 is a cycle or a block of order @xmath183 ; let @xmath184 be the set of blocks of @xmath1 , whose element @xmath62 is not a cycle and @xmath185 .
let @xmath1 be a connected graph of order @xmath29 with @xmath189 blocks in its block decomposition .
if @xmath190 , then we have done by theorem [ thm0 ] and @xmath191 , which satisfies the above bound .
thus , we suppose @xmath192 . in order to prove that the upper bound is tight
, we construct the graph @xmath1 of order @xmath29 , as shown in figure 1 , consisting of @xmath199 @xmath179 , @xmath45 cycles of order 4 and one 7-length - cycle with a chord .
it is clear that @xmath200=@xmath45 , @xmath201 .
we consider the size of a rainbow tree @xmath0 contain the vertices @xmath202 . @xmath203 and @xmath204 by the above theorem .
we have @xmath205 .
we finish this section with general graphs with minimal degree at least @xmath6 . here
, we denote as @xmath206 the maximum number of components of @xmath207 among all vertices @xmath208 .
the following result is needed in the sequel . | a tree @xmath0 , in an edge - colored graph @xmath1 , is called _ a rainbow tree _ if no two edges of @xmath0 are assigned the same color .
a _
@xmath2-rainbow coloring _ of @xmath1 is an edge coloring of @xmath1 having the property that for every set @xmath3 of @xmath2 vertices of @xmath1 , there exists a rainbow tree @xmath0 in @xmath1 such that @xmath4 .
the minimum number of colors needed in a @xmath2-rainbow coloring of @xmath1 is the _
@xmath2-rainbow index of @xmath1 _ , denoted by @xmath5 . in this paper , we consider @xmath6-rainbow index @xmath7 of @xmath1 . we first show that for connected graph @xmath1 with minimum degree @xmath8 , the tight upper bound of @xmath7 is @xmath9)+4 $ ] , where @xmath10 is the connected @xmath11-dominating set of @xmath1 .
and then we determine a tight upper bound for @xmath12 and a better bound for @xmath13-free graphs .
finally , we obtain a sharp bound for @xmath6-rainbow index of general graphs .
* keywords * : @xmath6-rainbow index ; rainbow tree ; connected @xmath11-dominating set . |
the formation and evolution of massive elliptical ( passive ) galaxies in the universe is interesting for both studies of galaxy evolution and cosmology . in the former case ,
such galaxies present an observational challenge for hierarchical models of structure formation , as some form of feedback is required to suppress on - going star formation in such massive systems ( see @xcite and references therein ) .
for cosmology , such massive ellipticals can be used to directly constrain cosmological parameters ( e.g. , as standard candles ; @xcite ) and provide efficient tracers of the large scale structure ( lss ) in the universe @xcite . moreover ,
the ( relative ) ages of massive ellipticals , as a function of redshift , offers the possibility of directly constraining the hubble parameter , thus providing vital information on the expansion history of the universe and therefore , the equation of state of dark energy ( see @xcite for a discussion of the underlying concept , and @xcite ; @xcite and @xcite for observational constraints obtained from using this technique ) . in this paper , we revisit the techniques used to constrain cosmological parameters through the age redshift relationship of passive ( elliptical ) , massive galaxies . in detail
, we present a new analysis of the ages of luminous red galaxies ( lrgs ; @xcite ) , as selected from the sloan digital sky survey ( sdss ; @xcite ) .
the key difference present herein compared to other such work with sdss spectral data ( e.g. @xcite ; @xcite ; @xcite ) is that we calibrate , for the first time , the sdss spectra onto the well - known , and well - studied , lick / ids system ( @xcite ; @xcite ) , thus allowing us to exploit a host of previous work on this system including the simple stellar population ( ssp ) modeling in the literature ( @xcite ; @xcite ; ) and comparisons with lick measurements in the literature . in a companion paper , we will use the age
redshift data derived for sdss lrgs in this paper to obtain cosmological constraints .
in section 2 of this paper , we outline the sdss galaxy data we use to construct our age redshift relation , specifically luminous ( massive ) , quiescent galaxies .
we also discuss the stellar data available to calibrate these galaxies onto the lick / ids system ( see also appendix a ) . in section 3 ,
we provide details about the ssp models we have used to determine the ages , metallicities and @xmath4enhancements of our galaxies , using a markov chain monte carlo ( mcmc ) technique for parameter estimation .
we discuss our results in section 4 , including the effect of priors , and conclude in section 5 .
appendix a provides the extensive details on how we calibrate the sdss spectra onto the lick / ids system , including all the necessary corrections made to the data . in appendix b , we provide a review of tests we have performed to quantify the robustness of our spectral measurements , including comparisons with data from the literature .
we assume a concordance , flat @xmath7cdm cosmology , with @xmath8 and @xmath9 , where required .
the sloan digital sky survey ( sdss ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) is a photometric and spectroscopic survey , covering @xmath10 of the northern sky , using a 2.5 meter telescope at apache point observatory in new mexico , usa . the photometric survey consists of simultaneous observations of the sky using 5 optical filters ( _ u , g , r , i _ and _ z _ ) , providing a database of hundreds of millions of detected objects all with accurate photometric and astrometric calibrations @xcite .
the sdss spectroscopic galaxy survey consists of two samples of galaxies selected using different criteria ; namely the main sample @xcite and the lrg sample @xcite .
the sdss spectra of these galaxies span a wavelength range of @xmath11 with a median resolution of @xmath12 ( @xmath13 ) , which is approximately @xmath14 ( although this varies as a function of wavelength and is different for each fiber ) .
the sdss spectra are automatically reduced using dedicated software , which flux calibrates the spectra and references them to the heliocentric frame and converts to vacuum wavelengths .
the software also measures a redshift for each object and measures a series of spectral features consistent ( in their wavelength definitions ) as the standard lick indices ( see section 2.2 ) , but are not calibrated onto the lick systems as no attempt has been made to match to the resolution of the lick / ids system . for the analysis presented in this paper
, we do not use the lick indices measurements from the standard sdss pipeline , but instead determine our own line - strengths after matching the instrumental resolutions ( see section [ sec : lick_calib ] ) .
however , we do use redshifts ( ` z ` ) , velocity dispersions ( ` veldisp ` ) , magnitudes ( ` u , g , r , i , z ` ) and other derived quantities such as the _ r_-band de vaucouleurs radii ( ` devrad_r ` ) and the _ r_-band de vaucouleurs profile axis ratio ( ` devab_r ` ) from the standard sdss spectral pipeline , made available through the catalog archive server ( cas ) .
the spectral data used in this paper was obtained from the data archive server ( das ) . for our analysis ,
we only used cut i luminous red galaxies ( lrgs ) as outlined in @xcite .
this is achieved by selecting objects from cas using the ` target_galaxy_red ` flag , thus yielding a pseudo volume - limited sample of lrgs between @xmath15 . below @xmath16 contamination by low redshift star - forming galaxies
increases the space density of galaxies that satisfy the sdss lrg colour - colour selection , while above @xmath17 the space density of sdss lrgs decreases due to the @xmath18 break dropping out of the sdss @xmath19-band .
we do not use the cut ii lrg sample at @xmath20 as we require velocity dispersion measurements for each object and these are not supplied above this redshift .
therefore , we impose an upper redshift limit of @xmath21 in our sample and analysis .
we further restrict our sample to only include lrgs with ` specclass eq spec_galaxy ` ( spectrum classified as a galaxy ) , ` zstat eq xcorr_hic ` ( redshift obtained from cross - correlation with a template ) , ` zwarning eq 0 ` ( no warning flags on the redshift ) , ` eclass < 0 ` ( indicates an old stellar population ) , ` z < 0.4 ` ( redshift less than 0.4 ) and ` fracdev_r > 0.8 ` ( indicating a surface brightness profile best fit by a de vaucouleurs profile ) . based upon these selection criteria ,
we obtain a sample of approximately @xmath22 lrgs ( see figure [ fig : z_dist_quiescent ] for the redshift distribution of this sample ) . at low redshift
, our sample will be contaminated by bulges in late - type galaxies due to the fiber size ( 3 ) of the sdss spectrographs . to reduce this contamination , and increase the number of truly quiescent galaxies in our sample , we make further cuts based on the emission line properties of our lrg sample , e.g. , we use standard emission lines such as @xmath23 , @xmath24and @xmath25}}\lambda5007
$ ] @xcite .
if left unchecked , such nebular emission , or emission associated with low - ionization nuclear emission - line region ( liner ) activity , would confuse our interpretation of the ssp parameters derived from these objects . to combat this , we use the mpa - jhu spectral line data to select only those objects that are consistent with zero emission . in their original work , @xcite fitted the spectral energy distributions ( seds ) of @xcite to the sdss galaxy spectra in data release four ( dr4 ) , subtracted off the best fitting sed , and then measured the line - strengths of @xmath26$]@xmath27 , @xmath24 , @xmath28$]@xmath29 , @xmath30$]@xmath31 , @xmath23 and @xmath32$]@xmath33 .
this sed fitting approach has now been applied to the sdss data release seven ( dr7 ) dataset using the stellar population synthesis spectra of charlot & bruzual ( 2008 ) .
we use this latest dr7 dataset herein . in an approach similar to @xcite
we fit a two component function to the line - strength distributions of @xcite , that consists of a normal component to describe the quiescent lrgs , plus a log - normal component to describe the active objects .
the best fit is then used to determine the small zero - point offsets that exist in the dataset of @xcite from which we then select objects that are consistent with zero emission , at the @xmath34 level , in @xmath23 and @xmath26 $ ] , hence removing the need to make corrections to the lick index line - strengths for the possible presence of nebular emission . with these constraints on the emission - line characteristics of the lrg sample , we obtain approximately @xmath35 galaxies that we refer to as the quiescent lrg sample . while it is possible to use different , and more , emission lines in determining the quiescent sample , doing so results in a significant reduction of our sample size .
for instance , including @xmath24and @xmath28 $ ] in the joint constraint with @xmath23 and @xmath26 $ ] yields approximately @xmath36 lrgs over @xmath1 and further inclusion of @xmath30 $ ] and @xmath32 $ ] constraints yields approximately @xmath37 objects .
while the more restrictive sample selection ensures a more quiescent sample , the reduction in the total number of objects available hampers our ability to explore the evolution of the ssp parameters over the desired redshift range .
but , even selecting on @xmath23 and @xmath26 $ ] alone helps to ensure that we are not including objects with significant emission in @xmath24as show in figure [ fig : quiescent_ew_contours ] . in figure
[ fig : z_dist_quiescent ] we also show the redshift distribution of the quiescent lrgs sample . in order to select the same population of objects over the redshift range
@xmath38 we need to select on physical properties such as velocity dispersion and absolute luminosity .
aperture corrections are applied to the sdss velocity dispersion measurements following the approaches of @xcite and @xcite , see section [ sec : ap_corr ] . to select on luminosity
we have performed k - corrections using the code _ kcorrect _ code of @xcite and have k - corrected all galaxies to a redshift of 0.2 ( the median of our distribution ) in order to minimise errors in these k - corrections .
we have modeled the luminosity evolution in our lrgs using a simple linear fit to the volume limited part of the k - corrected , magnitude - redshift distribution and then extrapolated this to higher and lower redshifts , see figure [ fig : quies_lum_evolve_contour ] . from this , we obtain an evolutionary ( e ) correction which we apply to our quiescent lrg sample . a more detailed approach to determine the evolution correction using stellar population synthesis models is unlikely to significantly improve the accuracy of the e - correction given the difficulty that current models have in explaining the evolution of the sdss lrgs colours @xcite . after correcting velocity dispersions for aperture effects and performing k+e corrections to the magnitudes , we have produced four subsamples of quiescent lrgs ( see table 2 for numbers in each sample ) .
these subsamples span an intervals of @xmath41 and @xmath42 and are shown as the boxed regions in figure [ fig : quies_slices_contour ] .
objects satisfying these criteria are then co - added to produce high signal - to - noise spectra after binning into redshift intervals of @xmath43 .
this redshift interval is small enough that there is little evolution in the ages of the objects within the bin such that we are co - adding objects of sufficiently similar ages , e.g. in the case of a concordance cosmology the age evolution within a bin is approximately 0.2 gyr over the redshift range @xmath38 , and is big enough to yield a sufficient number of objects to reach s
/ n levels of approximately @xmath44 in generating the co - added spectrum we have adopted an approach similar to @xcite . briefly , after binning our sample in redshift , absolute magnitude and velocity dispersion we shift each spectrum to its restframe .
all spectra in a given bin are normalised to the median flux in the region @xmath45 and combined pixel - by - pixel using a weighted arithmetic mean where the weights are determined by the inverse variance of the flux in each pixel .
this approach means that not all spectra contribute equally to the co - added spectrum but it does mean that those pixels contaminated by residuals from imperfect sky - subtraction contribute less to the final stacked spectrum .
a long term programme undertaken by @xcite and @xcite , amongst others , has yielded a library of stellar spectra , obtained with the image dissector scanner ( ids , * ? ? ?
* ) on the shane 3 m telescope at the lick observatory .
these authors have also established a spectral index system used to investigate element abundances from low - resolution integrated spectra of extragalactic stellar populations .
twenty five absorption features in the wavelength range @xmath46 at @xmath47 resolution , have been identified that are sensitive to effective temperature ( @xmath48 ) , surface gravity ( @xmath19 ) and metallicity ( @xmath49 ) of a star .
the wavelength definitions of these features were chosen such that the broad absorption features found in elliptical galaxies could be studied and to minimise the effect of galaxy velocity dispersion on the measured line - strengths .
all features measured on the lick / ids system are defined by a central `` feature '' bandpass and two adjacent `` pseudocontinuum '' bandpasses , see appendix [ sec : lick_calib ] . from this continuum and feature
it is possible to measure a `` pseudo '' equivalent width ( ew ) .
it is not a true ew as the wavelength definitions for the bandpasses are fixed and , depending on the element abundances , instrumental resolution or velocity dispersion of the stellar population , the wings of an absorption feature may extend beyond the feature bandpass . as a consequence of adopting fixed wavelengths to determine the continuum and feature fluxes ,
the measured line - strength depends on the flux in the continua bandpasses as well as the feature bandpass .
the continua also contain absorption features which , when coupled to the effects of galaxy velocity dispersions and instrumental resolution , will help suppress its average flux
. other factors that effect the measured line - strength of absorption features are gas , dust and telluric features .
although dust has little effect due to the generally narrow nature of the features , gas can be a problem .
nebular emission will contaminate the measured line - strength of the balmer lines by filling in the `` natural '' absorption ( * ? ? ?
* and references therein ) and , when using the @xmath24absorption index , will lead to incorrectly determining older ages for stellar populations . while the @xmath24index line - strength is directly affected by balmer emission from ionized gas , some indices are indirectly affected by emission , e.g. the red sideband of @xmath50can be contaminated by @xmath51 $ ] .
telluric emission and absorption can also have a significant impact on the accuracy of the measured line - strengths , but stacking spectra in order to improve the s / n of the lick indices also allows us to minimise the impact from these telluric features .
by co - adding spectra at different redshifts , telluric features `` appear '' at different restframe wavelengths in the deredshifted spectra .
using the inverse variance of the flux when weighting each pixel allows us to minimise the impact from these features on the final stacked spectrum .
it should be noted that the sdss convention is to present all wavelengths as vacuum wavelengths .
therefore , the wavelength definitions of all lick indices were converted to vacuum wavelengths @xcite prior to performing any measurements of index line - strengths on spectra . we have used the publicly
available software indexf @xcite to measure our absorption feature line - strengths from our co - added stacked lrg spectra . to properly calibrate to the lick / ids
system it is necessary to have sdss observations of lick stars , or galaxies , to allow the zero - point offsets to be determined , see appendix [ sec : transform ] .
the library of lick stars that could be used in calibrating the sdss to the lick system have not been intentionally targeted by the sdss and therefore requires serendipitous observations of lick stars to allow the sdss to be placed on the lick system .
the diameter of the sdss fibres is @xmath52 with the median seeing at apache point of @xmath53 .
we have therefore chosen a cut - off of @xmath54 when matching sdss objects and lick star coordinates . in searching for lick stars in the sdss
we have included both _
special _ and _ survey _ plates in the search and have identified 13 stars from the lick library that match sdss objects to within @xmath55 ; 11 stars in m67 ( _ special plate _ ` # ` 0321 ) and 2 stars in ngc 7789 ( _ special plate _ ` # ` 2377 ) and are listed in table [ tab : lick_stars ] .
the cluster proper motion dispersion for m67 is @xmath56 and for ngc 7789 is @xmath57 , with only one star in table [ tab : lick_stars ] having a measured proper motion of @xmath58 for m67 f 164 . given the small size of the proper motions we have ignored this effect when matching the coordinates of sdss objects to the lick stars in these two clusters .
c*4@[email protected]@l redshift & & & & + interval & & & & + @xmath59 & 9&15 & @xmath60 & 9&71 & @xmath61 & 7&33 & @xmath62 & 6&79 & @xmath63 + @xmath64 & 9&05 & @xmath65 & 9&27 & @xmath66 & 9&14 & @xmath67 & 6&84 & @xmath68 + @xmath69 & 8&86 & @xmath70 & 7&74 & @xmath71 & 8&92 & @xmath72 & 6&76 & @xmath73 + @xmath74 & 8&28 & @xmath75 & 8&04 & @xmath76 & 7&99 & @xmath77 & 7&39 & @xmath78 + @xmath79 & 7&85 & @xmath80 & 8&08 & @xmath81 & 7&62 & @xmath82 & 6&39 & @xmath83 + @xmath84 & 7&80 & @xmath85 & 7&63 & @xmath85 & 7&29 & @xmath86 & 5&57 & @xmath87 + @xmath88 & 7&77 & @xmath89 & 7&54 & @xmath90 & 6&63 & @xmath89 & 5&94 & @xmath91 + @xmath92 & 7&94 & @xmath93 & 7&37 & @xmath94 & 6&84 & @xmath95 & 5&99 & @xmath96 + @xmath97 & 7&07 & @xmath98 & 6&60 & @xmath99 & 6&49 & @xmath100 & 5&36 & @xmath101 + @xmath102 & 6&94 & @xmath103 & 6&29 & @xmath104 & 5&92 & @xmath105 & 5&08 & @xmath106 + @xmath107 & 6&99 & @xmath66 & 6&66 & @xmath104 & 6&04 & @xmath108 & 5&17 & @xmath81 + @xmath109 & 6&38 & @xmath110 & 5&99 & @xmath111 & 5&52 & @xmath112 & 4&48 & @xmath113 + @xmath114 & 5&91 & @xmath115 & 5&60 & @xmath116 & 5&48 & @xmath94 & 4&92 & @xmath117 + @xmath118 & 5&80 & @xmath119 & 5&38 & @xmath120 & 4&82 & @xmath116 & 4&18 & @xmath121 + @xmath122 & 5&21 & @xmath123 & 4&83 & @xmath124 & 4&73 & @xmath125 & 4&23 & @xmath126 + @xmath127 & 4&80 & @xmath66 & 4&91 & @xmath128 & 4&01 & @xmath116 & 3&62 & @xmath129 + @xmath130 & 4&43 & @xmath131 & 4&23 & @xmath132 & 3&91 & @xmath133 & 3&48 & @xmath134 + @xmath135 & 3&59 & @xmath136 & 4&47 & @xmath137 & 3&56 & @xmath138 & 3&25 & @xmath139 + @xmath140 & 5&98 & @xmath141 & 3&92 & @xmath142 & 4&03 & @xmath143 & 2&91 & @xmath144 + | we present a detailed analysis of @xmath0 quiescent , luminous red galaxies ( lrgs ) selected from sloan digital sky survey ( sdss ) data release seven ( dr7 ) spanning a redshift range of @xmath1 .
these galaxies are co - added into four equal bins of velocity dispersion and luminosity to produce high signal to noise spectra ( @xmath2 ) , thus facilitating accurate measurements of the standard lick absorption
line indices . in particular , we have carefully corrected and calibrated these indices onto the commonly used lick / ids system , thus allowing us to compare these data with other measurements in the literature , and derive realistic ages , metallicities ( @xmath3 $ ] ) and @xmath4-element abundance ratios ( @xmath5 $ ] ) for these galaxies using simple stellar population ( ssp ) models .
we use these data to study the relationship of these galaxy parameters with redshift , and find little evidence for evolution in metallicity or @xmath4elements ( especially for our intermediate mass samples ) .
this demonstrates that our subsamples are consistent with pure passive evolving ( i.e. no chemical evolution ) and represent a homogeneous population over this redshift range .
we also present the age
redshift relation for these lrgs and clearly see a decrease in their age with redshift ( @xmath6gyrs over the redshift range studied here ) which is fully consistent with the cosmological lookback times in a concordance @xmath7cdm universe .
we also see that our most massive sample of lrgs is the youngest compared to the lower mass galaxies .
we provide these data now to help future cosmological and galaxy evolution studies of lrgs , and provide in the appendices of this paper the required methodology and information to calibrate sdss spectra onto the lick / ids system .
[ firstpage ] methods : observational galaxies : elliptical galaxies : abundances galaxies : evolution galaxies : fundamental parameters |
the data set of charged current quasi elastic ( ccqe ) events recently released by the miniboone collaboration @xcite provides an unprecedented opportunity to carry out a systematic study of the double differential cross section of the process , _ + ^12 - 5mu c ^- + x , averaged over the neutrino flux . the charged current elastic neutrino - nucleon process is described in terms of three form factors .
the proton @xmath0 and neutron @xmath1 vector form factors , @xmath2 and @xmath3 ( @xmath4 , @xmath5 being the four - momentum transfer ) , have been precisely measured up to large values of @xmath6 in electron - proton and electron - deuteron scattering experiments , respectively ( for a recent review , see , e.g. , ref.@xcite ) .
the @xmath6-dependence of the nucleon axial form factor @xmath7 , whose value at @xmath8 can be extracted from neutron @xmath9-decay measurements , is generally assumed to be of dipole form and parametrized in terms of the so called axial mass @xmath10 : f_a(q^2 ) = g_a ( 1 + q^2/m_a^2 ) ^-2 .
the world average of the measured values of the axial mass , mainly obtained from low statistics experiments carried out using deuterium targets , turns out to be @xmath11 gev @xcite , while the analyses performed by the k2k @xcite and miniboone @xcite collaborations using oxygen and carbon targets , respectively , yield @xmath12 .
it would be tempting to interpret the large value of @xmath10 reported by miniboonne and k2k as an _ effective _ axial mass , modified by nuclear effects not included in the oversimplified fermi gas model employed in data analysis .
however , most existing models of nuclear effects ( for recent reviews see ref.@xcite ) fail to support this explanation , suggested by the authors of ref.@xcite , a prominent exception being the model of ref.@xcite .
obviously , a fully quantitative description of the electron - scattering cross section , driven by the known vector form factors , is a prerequisite for the understanding of the axial vector contribution to the ccqe neutrino - nucleus cross section . over the past two decades
, the availability of a large body of experimental data has triggered the development of advanced theoretical descriptions of the nuclear electromagnetic response .
the underlying scheme , based on nuclear many - body theory and realistic nuclear hamiltonians , relies on the premises that i ) the lepton kinematics is fully determined and ii ) the elementary interaction vertex can be extracted from measured proton and deuteron cross sections .
the above _ paradigm _ has been successfully applied to explain the electron - nucleus cross section in a variety of kinematical regimes ( for a recent review of the quasi - elastic sector see ref.@xcite ) . however , in view of the uncertainties associated with the energy of the incoming beam , the identification of the reaction mechanisms and the determination of the interaction vertex , its extension to the case of neutrino scattering may not be straightforward .
the comparison between theoretical calculations and data , presented in section [ enua ] , suggests that the measured electron- and neutrino - nucleus cross sections can not be explained within the same scheme .
the difficulties associated with the description of the flux unfolded total cross section are discussed in section [ paradigm ] .
finally , in section [ conclusions ] i argue that the the _ paradigm _ succesfully employed to analyze electron - nucleus scattering is not suitable for application to processes in which the lepton kinematics is not fully determined , and must be significantly modified .
electron - nucleus scattering cross sections are usually analyzed at fixed beam energy , @xmath13 , and electron scattering angle , @xmath14 , as a function of the electron energy loss @xmath15 .
as an example , fig .
[ xsec_ee ] shows the double differential cross section of the process e + ^12 - 5mu c e^+ x , at @xmath16 mev and @xmath17 , measured at mit - bates @xcite .
the peak corresponding to quasi - elastic ( qe ) scattering , the bump at larger @xmath15 , associated with excitation of the @xmath18-resonance , and the region in - between , where the cross section is mainly arising from processes involving meson exchange currents ( mec ) , are clearly recognizable .
the three - momentum transfer @xmath19 turns out to be nearly constant , its variation over the range shown in the figure being @xmath20 5% . as a consequence
, the cross section of fig.[xsec_ee ] can be readily related to the linear response of the target nucleus to a probe delivering momentum @xmath21 and energy @xmath15 , defined as [ sqw ] s(*q * , ) = _ n | n | _ * k * a^_*k*+*q * a_*k * | 0 |^2 ( + e_0 - e_n ) . in the above equation ,
@xmath22 and @xmath23 represent the target initial and final states , with energies @xmath24 and @xmath25 , respectively , while @xmath26 and @xmath27 are the nucleon creation and annihilation operators .
the magnitude of the momentum transfer , @xmath28 450 mev , is large enough to make the impulse approximation ( ia ) scheme , in which the nuclear response of eq.([sqw ] ) reduces to @xcite s_ia(*q * , ) = d^3k de p_h(*k*,e ) p_p(*k*+*q*,-e ) , [ s : ia]safely applicable @xcite . in eq.([s : ia ] ) , the hole and particle spectral functions , @xmath29 and @xmath30 , describe the energy and momentum distributions of the struck nucleon in the initial ( hole ) and final ( particle ) states , respectively . the solid line of fig .
[ xsec_ee ] represents the results of a theoretical calculation of the qe contribution @xcite , carried out within the approach described in refs.@xcite using the hole spectral function of ref.@xcite and the recent parametrization of the vector form factors of ref.@xcite .
final state interactions ( fsi ) between the struck nucleon and the recoiling spectator system @xcite , whose main effect is a @xmath31 mev shift of the qe peak , have been also taken into account .
730 mev and electron scattering angle @xmath32 , plotted as a function of the energy loss @xmath15 @xcite .
the data points are taken from ref.@xcite . ]
it is apparent that height , position and width of the qe peak , mostly driven by the energy and momentum dependence of the hole spectral function , are well reproduced .
applying the same scheme employed to obtain the solid line of fig .
[ xsec_ee ] to neutrino scattering leads to the results shown in fig .
[ dsigma ] @xcite .
the data points represent the double differential ccqe cross section averaged over the miniboone neutrino flux , whose mean energy is @xmath33 mev , plotted as a function of the kinetic energy of the outgoing muon , @xmath34 , at different values of the muon scattering angle @xmath35 .
the solid lines show the results ( integrated over the @xmath36 bins ) obtained using the same spectral functions and vector form factors employed in the calculation of the electron scattering cross section of fig .
[ xsec_ee ] , and a dipole parametrization of the axial form factor with @xmath37 mev . comparison of figs .
[ xsec_ee ] and [ dsigma ] indicates that the electron and neutrino cross sections corresponding to the same target and _ seemingly _ comparable kinematical conditions ( the position of the qe peak in fig .
[ xsec_ee ] corresponds to kinetic energy of the scattered electron @xmath38 mev ) can not be explained using the same theoretical approach and the value of the axial mass resulting from deuterium measurements . in this instance
, the paradigm of electron scattering appears to conspicuously fail . , and a dipole parametrizaition of the axial form factor with @xmath37 mev @xcite . ]
the above conclusion , while being based on a calculation carried out within the scheme of refs.@xcite , is largely model independent .
theoretical approaches providing a quantitative description of the electron - nucleus cross section in the qe channel , are bound to predict ccqe neutrino - nucleus cross sections significantly below the miniboone data if the value of the axial mass is set to 1.03 gev . as a matter of fact ,
within the approach of refs.@xcite , the axial mass yelding the best @xmath39-fit to the flux integrated miniboone @xmath6-distribution , @xmath401.6 gev , turns out to be even larger than that reported in ref.@xcite .
the authors of ref.@xcite have shown that this value of @xmath10 also explains the muon energy spectrum and angular distribution obtained from integration of the double differential cross seciton of ref.@xcite .
-distribution obtained from the approach of refs.@xcite using @xmath401.6 gev , compared to the data of ref.@xcite . ] in fig .
[ q2:dist ] the theoretical results obtained using @xmath401.6 gev are compared to the distribution of the flux - averaged miniboone events , plotted as a function of the _ reconstructed _
@xmath6 , defined as q^2 = 2 e_e_(1- _ ) - m_^2 , where @xmath41 , @xmath42 and @xmath43 is the muon mass . note that the definition of @xmath6 , besides the measured kinematical variables @xmath34 and @xmath35 , involves the incoming neutrino energy @xmath44 , whose determination necessarily implies some assumptions .
figure [ sigmatot ] shows a comparison between the the miniboone _ flux unfolded _ total cross section and the results of the calculations of of ref.@xcite . it is apparent that in this case using @xmath45 gev leads to overestimating the data in the region of high energy ( @xmath46 mev ) , where the choice @xmath47 gev , resulting from he analysis of ref.@xcite , provides a better fit .
the different pattern emerging from fig .
[ sigmatot ] , compared to fig . [ q2:dist ] , clearly points to the uncertainty associated with the interpretation of flux averaged and flux unfolded data .
gev , respectively .
the data are taken from ref.@xcite . ]
a different scenario is suggested by the results of ref .
@xcite , whose authors obtain a quantitative account of the miniboone flux unfolded total cross section _ without increasing _ @xmath10 . within the model of ref .
@xcite , the main mechanism responsible for the enhancement that brings the theoretical cross section into agreement with the data is multi - nucleon knock out , leading to @xmath48 particle-@xmath48 hole ( @xmath48p-@xmath48h ) nuclear final states ( @xmath49 ) . within the approach of refs.@xcite , the occurrence of 2p-2h final states
is described by the continuum part of the spectral function , arising from nucleon - nucleon correlations @xcite .
it gives rise to the tail extending to large @xmath15 , clearly visible in fig .
[ xsec_ee ] . however , its contribution turns out to be quite small ( less than 10% of the integrated spectrum ) .
the analysis of the momentum distribution sum rule indicates that the contributions of @xmath48p-@xmath48h final states with @xmath50 are negligibly small @xcite .
according to the philosophy outlined in this paper , in order to firmly establish the role of multi - nucleon knock out in ccqe neutrino interactions the model of ref .
@xcite should be thoroughly tested against electron scattering data .
the theoretical and experimental results discussed in this paper suggest that the main difference involved in the analysis of neutrino - nucleus scattering , as compared to electron - nucleus scattering , lies in the flux average . unlike the electron cross section shown in fig .
[ xsec_ee ] , the flux averaged ccqe neutrino cross section at fixed energy and scattering angle of the outgoing lepton picks up contributions from different kinematical regions , where different reaction mechanisms dominate .
consider , as an example , the cross section at @xmath51 , the central value of the bin considered in the upper panel of fig .
[ dsigma ] , and @xmath52 mev , corresponding to the maximum of the measured cross section .
figure [ columbia ] displays the behavior of the bjorken scaling variable @xmath53 in these kinematical conditions and the miniboone flux as a function of the neutrino energy @xmath44 . in the kinematical conditions corresponding to maximum of the cross section displayed in the upper panel of fig .
[ dsigma ] , plotted as a function of the neutrino energy @xmath44 . for comparison ,
the miniboone neutrino flux is also shown . ]
it turns out that @xmath54 , corresponding to quasi - elastic kinematics , and @xmath55 , corresponding to the dip region , are associated with neutrino energies @xmath56 mev ( the mean energy of the miniboone flux ) and @xmath57 mev , respectively , and that 0.83 it follows that the flux averaged cross section picks up the contributions of quasi - elastic scattering and mec with about the same probability .
hence , it can not be described according to the paradigm successfully applied to electron scattering , based on the tenet that the lepton kinematics is fully determined . a _ new paradigm _ , suitable for studies of neutrino interactions ,
should be based on a more flexible model of nuclear effects , yielding a realistic description of the broad kinematical range associated with the relevant neutrino energies .
nuclear many - body theory provides a consistent framework for the development of such a model . besides single- and multi - nucleon
knock out , it should include the contributions of processes involving mec , which are long known to provide a significant enhancement of the electromagnetic nuclear response in the transverse channel @xcite .
it has to be emphasized that quasielastic scattering and processes involving mec lead to the same final state , and can not be distinguished in the miniboone analysis .
the occurrence of inelastic processes , leading to excitation of nucleon resonances and pion production should also be taken into account . as a final remark
, it has to be emphasized that a great deal of information could be obtained applying the new paradigm to the analysis of _ inclusive _ neutrino - nucleus cross sections , preferably , although not necessarily , through direct implementation of the resulting nuclear model in the monte carlo simulation codes .
this kind of analysis , which has been succesfully carried out for electron - nucleus scattering , may help to reconcile the different values of the axial mass obtained by different experiments , and would be largely unaffected by the problems associated with the possible misidentification of ccqe events , recently discussed in ref .
this work is based on results obtained in collaboration with p. coletti and d. meloni . a number of discussions with a.m. ankowski and c. mariani on issues related to the subject of this paper , are also gratefully acknowledged .
proceedings of the _ sixth international workshop on neutrino - nucleus interactions in the few - gev region ( nuint-09)_. eds .
f. sanchez , m. sorel and l. alvarez - ruso .
aip conference proceedings , vol .
1189 ( 2010 ) . | the availability of the double - differential charged - current neutrino cross section , measured by the miniboone collaboration using a carbon target , allows for a systematic comparison of nuclear effects in quasi - elastic electron and neutrino scattering .
the results of theoretical studies based on the impulse approximation scheme and state - of - the - art models of the nuclear spectral functions suggest that the electron cross section and the flux averaged neutrino cross sections corresponding to the same target and seemingly comparable kinematical conditions can not be described within the same theoretical scheme using the value of the nucleon axial mass obtained from deuterium measurements .
i analyze the assumptions underlying the treatment of electron scattering data , and argue that the development of a new _ paradigm _ , suitable for application to processes in which the lepton kinematics is not fully determined , will be required .
lepton - nucleus interaction , charged current neutrino interactions , nucleon axial form factor |
_ jacobi - type continued fractions _ , or _
j - fractions _ , correspond to power series defined by infinite continued fraction expansions of the form @xmath7 for some sequences @xmath8 and @xmath9 , and some ( typically formal ) series variable @xmath10 @xcite @xcite .
the formal series enumerated by special cases of the truncated and infinite j - fraction series of this form include _ ordinary _ ( as opposed to typically closed - form _ exponential _ ) _ generating functions _ for many one and two - index combinatorial sequences including the generalized factorial functions studied in the references and in the results from flajolet s articles .
the references provide infinite formal continued fraction series for the typically divergent ordinary generating functions of many special sequences , factorial functions , and combinatorial triangles arising in applications @xcite . [ [ j - fraction - expansions - of - ordinary - square - series - generating - functions ] ] j - fraction expansions of ordinary square series generating functions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in this article , we consider a @xmath1-series - related variant of the j - fractions in enumerating the square - power terms of @xmath0 , which we have discovered computationally ( see @xcite for integral representations ) .
in particular , for non - zero @xmath11 such that @xmath12 , we define the following jacobi - type j - fraction , @xmath13}(q , z)$ ] , for the ordinary square series generating function defined in @xcite as follows : @xmath14}(q , z ) & = \cfrac{1}{1-qz-\cfrac{q^2(q^2 - 1 ) z^2}{1-q(q^4+q^2 - 1)z- \cfrac{q^8 ( q^4 - 1 ) z^2}{1-q^3(q^6+q^4 - 1)z-\cfrac{q^{14}(q^6 - 1 ) z^2}{\cdots . } } } } \end{aligned}\ ] ] [ def_ciq_abiq_phqz_qhqz ] we define the component sequence functions , @xmath15}(q)$ ] and @xmath16}(q)$ ] , as follows for @xmath17 : @xmath18}(q ) & : = \begin{cases } q^{2h-3 } \cdot \left(q^{2h } + q^{2h-2 } - 1\right ) & \text { if $ h \geq 2 $ ; } \\ q & \text { if $ h = 1 $ ; } \\ 1 & \text { if $ h = 0 $ , } \end{cases } \\ \label{eqn_abhq_subseq_def } \operatorname{ab}_h^{[\sq]}(q ) & : = \begin{cases } q^{6h-10 } \cdot \left(q^{2h-2 } - 1\right ) & \text { if $ h \geq 2 $ ; } \\ 0 & \text { otherwise . } \end{cases } \end{aligned}\ ] ] we then define the @xmath3 convergent functions , @xmath19 , recursively through the component numerator and denominator functions given by , as @xmath20 .
similarly , @xmath21 in the remainder of the article . ]
@xmath22}(q ) \cdot z ) \operatorname{p}_{h-1}(q , z ) - \operatorname{ab}_{h}^{[\sq]}(q ) \cdot z^2 \operatorname{p}_{h-2}(q , z ) + \iverson{h = 1 } \\
\notag \operatorname{q}_h(q , z ) & = ( 1-c_{h}^{[\sq]}(q ) \cdot z ) \operatorname{q}_{h-1}(q , z ) -
\operatorname{ab}_{h}^{[\sq]}(q ) \cdot z^2 \operatorname{q}_{h-2}(q , z ) + ( 1-c_{1}^{[\sq]}(q ) \cdot z ) \iverson{h = 1 } + \iverson{h = 0}. \end{aligned}\ ] ] we claim , and subsequently prove in section [ section_proofs ] , that the truncated @xmath23-order power series expansions of the @xmath3 convergent functions , @xmath24 , defined in definition [ def_ciq_abiq_phqz_qhqz ] generate the so - termed `` _ _ square series _ _ '' powers defined in @xcite as @xmath25 for all @xmath26 .
more precisely , we state the next result as our main theorem which we carefully prove in section [ subsection_proofofmaintheorem ] of the article below . for @xmath30
, we show that the limiting convergent function , denoted by @xmath31}(q , z ) : = \sum_{n \geq 0 } q^{n^2 } z^n = \lim_{h \longrightarrow \infty}\ \conv_h(q , z ) , \end{aligned}\ ] ] which corresponds to the integral representations of the _ ordinary geometric square series _
cases defined in @xcite given by @xmath32}(q , z ) & = \int_0^{\infty } \frac{2 e^{-t^2/2}}{\sqrt{2\pi } } \left [ \frac{1 - z \cosh\left(t \sqrt{2 \log(q)}\right ) } { z^2 - 2z \cosh\left(t \sqrt{2 \log(q)}\right ) + 1 } \right ] dt , \end{aligned}\ ] ] exists , and that the limiting cases of this convergent function series converge uniformly as a function of @xmath33 with respect to the sequence of @xmath3 partial sums of defined below .
thus we are able to prove new infinite @xmath1-series identities generating special cases of @xmath13}(q , c \cdot q^{p})$ ] in section [ section_proofs ] , which arise in many applications to theta functions and special function series .
the primary new form of the infinite q - series expansions we prove in section [ section_proofs ] of the article has the form @xmath34}(q , z ) & = \sum_{i=1}^{\infty } \frac{(-1)^{i-1 } q^{(3i-4)(i-1 ) } { \left(q^2 ; q^2\right)_{i-1 } } z^{2i-2 } } { \sum\limits_{0 \leq j \leq
n < 2i } { \gkpsi{i}{j}_{q^2 } } { \gkpsi{i-1}{n - j}_{q^2 } } q^{2j } \cdot \left(-q^{2i-3 } \cdot z\right)^n}. \end{aligned}\ ] ] moreover , the @xmath3 partial sums of the infinite @xmath1-series expansion in corresponding to the @xmath3 convergent function , @xmath24 , is always at least @xmath23-order accurate in generating the coefficients , @xmath0 , of @xmath35 for @xmath28 .
in addition , we show that we can differentiate these new infinite square series expansions termwise with respect to @xmath33 thus allowing us to generate series for higher - order derivatives of the _ jacobi theta functions _ , which are generated as particular cases of the geometric square series , @xmath36 , in @xcite and which similarly arise in many well - known identities and classical expansions of special products and partition function generating functions .
more precisely , if we let @xmath38 denote an arbitrary sequence depending on the fixed parameter @xmath1 , we seek to explore the analogous @xmath1-series - related j - fraction expansions enumerating this sequence in the form of @xmath39 we can always solve for the sequences of implicit coefficients , @xmath40 and @xmath41 , in which ( at least formally ) generate the sequence of @xmath42 over @xmath33 .
typically we find that for more well - known @xmath1-series generating function forms it is easy to computationally `` _ _ guess _ _ '' exact formulas for these sequences , which are each implicit functions of the @xmath43 s , in many special case forms .
we conjecture , but do not offer conclusive proofs , that we have related @xmath1-series generating functions enumerating variants of the sequences , @xmath44 , @xmath45 , and @xmath46 over @xmath33 defined by the implicit subsequence listings given in table [ table_conclusions_specialj - fraction_conjsubseq_results ] .
examples of special series that result by generating the sequences , @xmath42 , in the table by the properties resulting from our conjectured j - fraction expansions include the following expansions @xcite @xcite : @xmath47 \\ \tag{$q$-exponential function } e_q(z ) & = \sum_{n=0}^{\infty } \frac{(1-q)^n z^n}{{\left(q ; q\right)_{n } } } \\ \tag{$q$-hypergeometric function } _ 1\phi_0(0 ; - ; q , z ) & = \sum_{n=0}^{\infty } \frac{z^n}{{\left(q ; q\right)_{n } } } = \frac{1}{{\left(z ; q\right)_{\infty } } } \\ \tag{partition function
generating function } _ 1\phi_0(0 ; - ; q , q ) & = \sum_{n=0}^{\infty } \frac{q^n}{{\left(q ; q\right)_{n } } } = \frac{1}{{\left(q ; q\right)_{\infty } } } \\ \tag{another $ q$-binomial theorem } _ 1\phi_0(0 ; - ; q , q ) & = { \left(z q^{-n } ; q\right)_{n}}. \end{aligned}\ ] ] for the first sequence case in the table , we can extend the methods of proof given in this article for the square series generating functions to find that @xmath48 other examples of new infinite series and q - series expansions for the generating functions of the sequences listed in table [ table_conclusions_specialj - fraction_conjsubseq_results ] , along with many other special cases not considered within the context of the applications cited in this article , are proved similarly . | we consider new series expansions for variants of the so - termed ordinary geometric square series generating functions originally defined in the recent article titled `` _ _ square series generating function transformations _ _ '' ( arxiv : 1609.02803 ) . whereas the original square series transformations article adapts known generating function transformations to construct integral representations for these square series functions enumerating the square powers of @xmath0 for some fixed non - zero @xmath1 with @xmath2 , we study the expansions of these special series through power series generated by jacobi - type continued fractions , or j - fractions
. we prove new exact expansions of the @xmath3 convergents to these continued fraction series and show that the limiting case of these convergent generating functions exists as @xmath4 .
we also prove new infinite @xmath1-series representations of special square series expansions involving square - power terms of the series parameter @xmath1 , the @xmath1-pochhammer symbol , and double sums over the @xmath1-binomial coefficients .
applications of the new results we prove within the article include new @xmath1-series representations for the ordinary generating functions of the special sequences , @xmath5 , and @xmath6 , as well as parallels to the examples of the new integral representations for theta functions , series expansions of infinite products and partition function generating functions , and related unilateral special function series cited in the first square series transformations article . : _ square series ; @xmath1-series ; j - fraction ; continued fraction ; sum of squares functions ; sum of divisors function ; theta function ; ordinary generating function . _ : _ 05a15 ; 11y65 ; 11b65 ; 40a15 . _ |
the concept of a possible thermal origin of charmonia was initially introduced @xcite to explain the measurement of the j/@xmath1/hadron ratio in nuclear collisions .
because of their large mass and small production cross section at thermal energies charm quarks are , however , not likely to be thermally produced .
on the other hand , significant production of charm quark pairs takes place in initial hard collisions .
this led to the idea of statistical hadronization of charm quarks @xcite , which has sparked an intense activity in this field @xcite . at about the same time
an independent effort @xcite based on a kinetic model has been developed to study charm production in heavy ion collisions at collider energies .
initial interest focussed on the available sps data on @xmath0 production in pb
pb collisions .
as we stress below , these data can indeed be described within the statistical approach , but only when assuming the charm production cross section to be enhanced beyond the perturbative qcd ( pqcd ) predictions . the largest differences between the results of the statistical coalescence scenario and earlier models are expected at collider energies . for example , for central au - au collision at rhic energy ,
matsui approach @xcite predicts a very strong ( up to a factor of 20 @xcite ) suppression of @xmath0 yields as compared to a direct production . in the present approach this suppression
is overcome by statistical recombination of @xmath0 mesons from the initially produced @xmath2 pairs , so that much larger yields are expected .
suppression at rhic energy is also predicted by the kinetic model @xcite .
] we therefore focus in this letter on the predictions of the model for open and hidden charm meson yields at rhic ( au au ) and lhc ( pb pb ) energies , with particular emphasis on the centrality dependence of rapidity densities at rhic .
we , furthermore , present predictions for the energy dependence of the production of hadrons with open and hidden charm .
the model assumes that all charm quarks are produced in primary hard collisions and equilibrate in the quark - gluon plasma ( qgp ) .
in particular , no @xmath0 mesons are preformed in the qgp , implying complete color screening @xcite .
there are indications from recent lattice qcd calculations that , already at sps energies , j/@xmath1 mesons can be dissociated in a deconfined medium via collisions with thermal gluons @xcite .
as noted earlier @xcite , the cross sections of charmed hadrons are too small to allow for their chemical equilibration in a hadronic gas .
we assume that open and hidden charm hadrons are formed at chemical freeze - out according to statistical laws .
consistent with the fact that , at sps energy and beyond , chemical freeze - out appears to be at the phase boundary @xcite , our model implies that a qgp phase was a stage in the evolution of the fireball .
the analysis of @xmath0 spectra at sps @xcite lends further support to the statistical hadronization picture where @xmath0 decouples at chemical freeze - out . a recent analysis of electron spectra at rhic
@xcite also strenghtens the case for an early thermalization of heavy quarks .
however , in that analysis it was pointed out that both the hydrodynamical approach and pythia reproduce the measured single - electron spectra .
as suggested meanwhile @xcite , the measurement of the elliptic flow ( @xmath3 ) of open charm hadrons and @xmath0 may help to disentangle the two scenarios .
recent work @xcite in extracting the quark @xmath3 from the one of hadrons may help towards this goal .
most important though , in our view , is a high - precision direct measurement of open charm . as expected , and as seen in ref .
@xcite , the @xmath4 spectra predicted by pythia and hydro are different in detail at low @xmath4 and differ manifestly at high @xmath4 ( @xmath5 mass of charm quark ) . on the other hand , a recent theoretical analysis @xcite indicates that charm quarks might not thermalize quickly because of their large mass . in statistical models
charm production needs generally to be treated within the framework of canonical thermodynamics @xcite .
thus , the charm balance equation required @xcite during hadronization is expressed as : @xmath6 here @xmath7 is the number of directly produced @xmath8 pairs and @xmath9 are modified bessel functions . in the fireball of volume @xmath10 the total number of open @xmath11 and hidden @xmath12
charm hadrons are computed from their grand - canonical densities @xmath13 and @xmath14 , respectively . the densities of different particle species in the grand canonical ensemble are calculated following the statistical model introduced in @xcite . all known charmed mesons and hyperons and their decays @xcite have to be included in the calculations .
the balance equation ( [ aa : eq1 ] ) defines a fugacity parameter @xmath15 that accounts for deviations of charm multiplicity from the value that is expected in complete chemical equilibrium .
the yield of open charm mesons and hyperons @xmath16 and of charmonia @xmath17 is obtained from : @xmath18 the above model for charm production and hadronization can be only used @xcite if the number of participating nucleons @xmath19 is sufficiently large .
taking into account the measured dependence of the relative yield of @xmath20 to @xmath0 on centrality @xcite in pb
pb collisions at sps energy the model appears appropriate @xcite if @xmath21100 . to calculate the yields of open and hidden charm hadrons for a given centrality and collision energy one needs to fix a set of parameters in eq.(1 ) and ( 2 ) :
\i ) a constant temperature of 170 mev and a baryonic chemical potential @xmath22 according to the parametrization @xmath23/(1+\sqrt{s_{nn}}[{\rm gev}]/4.3)$ ] @xcite are used for our calculations . for sps and rhic energies these thermal parameters
are consistent , within statistical errors , with those required to describe experimental data on different hadron yields @xcite . in table
[ aa : tab0 ] we present the values of these input parameters and the resulting chemical potentials for isospin , strangeness and charmness for the canonical ensemble calculations at sps , rhic and lhc .
.input values of @xmath24 and @xmath22 and output chemical potentials for isospin , strangeness and charmness for canonical ensemble calculations at sps , rhic and lhc .
[ cols="^,^,^,^",options="header " , ] the apparent weak dependence of @xmath0 yield on freeze - out temperature , seen in fig .
[ aa : fig3 ] , may be surprising .
this is particularly the case since in the thermal model considered in @xcite this dependence was shown to be very strong . in our model
this result is a consequence of the charm balance equation ( [ aa : eq2 ] ) .
the temperature variation leads , obviously , to a different number of thermally produced charmed hadrons as in @xcite , but this is compensated by the @xmath15 factor .
the approximate temperature dependence of @xmath15 and the @xmath0 yield are : @xmath25 as a result of the small mass difference in the exponent the @xmath0 yield exhibits only a weak sensitivity on @xmath24 .
this is in contrast to the purely thermal case where the yield scales with @xmath26 .
the results are summarized in table [ aa : tab3 ] , where we present the number of @xmath0 and @xmath27 mesons produced thermally ( @xmath28 and @xmath29 , respectively ) and within the frame of our statistical hadronization model .
also shown are the values of @xmath15 and the yield ratio @xmath30 for three different values of @xmath24 .
the compensations are evident in the last column of table [ aa : tab3 ] , where ratios for the two extreme temperatures are calculated .
the only exception is the ratio @xmath30 , which is obviously identical in the statistical hadronization scenario and in the thermal model and coincides , for t@xmath31170 mev , with the measured value at sps @xcite .
mesons at rhic for different rapidity window sizes .
the lines are calculations , the dots are experimental data from phenix collaboration @xcite ( the point for the central collisions is the upper limit for 90% c.l . ) . ]
most of our results presented above are obtained considering a one unit rapidity window at midrapidity , while results for the full volume were presented only for the sps .
unlike the kinetic model of thews et al .
@xcite , our model does not contain dynamical aspects of the coalescence process .
however , in our approach , the width of the rapidity window does influence the results in the canonical regime .
for the grand - canonical case , attained only at lhc energy , there is no dependence on the width of the rapidity window , due to a simple cancellation between the variation of the volume , proportional to the rapidity slice in case of a flat rapidity distribution , and the variation of @xmath7 , also proportional to the width of the rapidity slice . in fig .
[ aa : fig4 ] we present the centrality dependence of @xmath0 rapidity densities for rhic energy and for different rapidity windows @xmath32y from 0.5 to 3 .
the dependence on @xmath32y resembles that of the kinetic model @xcite , but is less pronounced .
the available data are not yet precise enough to rule out any of the scenarios considered . however , for the kinetic model @xcite , the cases of small @xmath32y seem to be ruled out by the present phenix data .
we stress in this context that the size of @xmath32y window has a potentially large impact on the results at sps energy .
it is conceivable that no charm enhancement is needed to explain the data if one considers a sufficiently narrow rapidity window for the statistical hadronization .
the results presented above were obtained under the assumption of statistical hadronization of quarks and gluons .
in addition , for charm quarks the yields were constrained by the charm balance equation that was formulated in the canonical ensemble .
we have further assumed that charm quarks are entirely produced via primary hard scattering and thermalized in the qgp .
no secondary production of charm in the initial and final state was included in our calculations .
final state effects like nuclear absorption of @xmath0 @xcite are also neglected .
we have demonstrated that the statistical coalescence approach yields a good description of the measured centrality dependence of @xmath0 production at sps energy , albeit with a charm cross section increased by a factor of 2.8 compared to current nlo perturbative qcd calculations .
rapidity densities for open and hidden charm mesons are predicted to increase strongly with energy , with striking changes in centrality dependence .
first rhic data on @xmath0 production support the current predictions , although the experimental errors are for the moment too large to allow firm conclusions .
we emphasize that the predictive power of our model relies heavily on the accuracy of the charm cross section , not yet measured in heavy - ion collisions .
the statistical coalescence implies travel of charm quarks over significant distances e.g. in a qgp .
if the model predictions will describe consistently precision data this would be a clear signal for the presence of a deconfined phase .
we acknowledge stimulating discussions with r.l . thews .
k.r . acknowledges the support of the alexander von humboldt foundation ( avh ) .
99 m. gadzicki , m.i .
gorenstein , [ hep - ph/9905515 ] .
p. braun - munzinger , j. stachel , [ nucl - th/0007059 ] .
p. braun - munzinger , j. stachel , [ nucl - th/0012064 ] .
gorenstein , a.p .
kostyuk , h. stcker , w. greiner , [ hep - ph/0010148 ] .
l. grandchamp , r. rapp , [ hep - ph/0103124 ] .
gorenstein , a.p .
kostyuk , h. stcker , w. greiner , [ hep - ph/0104071 ] .
kostyuk , m.i .
gorenstein , h. stcker , w. greiner , [ hep - ph/0110269 ] .
l. grandchamp , r. rapp , [ hep - ph/0205305 ] .
thews , m. schroedter , j. rafelski , [ hep - ph/0007323 ] .
thews , hep - ph/0206179 .
t. matsui , h. satz , .
r. vogt ( p. 250c ) in s.a .
bass et al . , .
f. karsch , aip conf .
602 ( 2001 ) 323 , hep - lat/0109017 .
wong , ; erratum ibid .
76 ( 1996 ) 4987 .
p. braun - munzinger , j. stachel , [ nucl - th/0112051 ] .
gorenstein , k.a .
bugaev , m. gadzicki , [ hep - ph/0112197 ] .
s. batsouli , s. kelly , m. gyulassy , j.l .
nagle , [ nucl - th/0212068 ] .
j. nagle , nucl - th/0306001 .
lin and d. molnar , nucl - th/0304045 .
dokshitzer , d.e .
kharzeev , [ hep - ph/0106202 ] .
j. cleymans , k. redlich , e. suhonen , .
review of particle physics , .
braun - munzinger , i. heppe and j. stachel , phys .
b465 ( 1999 ) 15 .
p. braun - munzinger , d. magestro , k. redlich , and j. stachel , phys .
b518 ( 2001 ) 41 , and references therein .
na38 collaboration , m.c .
abreu et al . , .
na50 collaboration , m.c .
abreu et al . , .
h. sorge , e. shuryak , i. zahed , .
na49 collaboration , s.v .
afanasiev et al . , [ nucl - ex/0205002 ] .
phobos collaboration , b.b .
back et al . , [ nucl - ex/0108009 ] ; [ nucl - ex/0211002 ] .
e877 collaboration , j. barrette et al . , [ nucl - ex/9412003 ] .
na50 collaboration , m.c .
abreu et al . , .
brahms collaboration , i.g .
bearden et al . , [ nucl - ex/0108016 ] ; [ nucl - ex/0112001 ] .
eskola , k. kajantie , p.v .
ruuskanen , k. tuominen , [ hep - ph/9909456 ] ; k.j .
eskola , k. kajantie , k. tuominen , [ hep - ph/0106330 ] .
r. vogt , hep - ph/0111271 ; hep - ph/0203151 .
r. vogt , heavy ion phys . 9
( 1999 ) 339 [ nucl - th/9903051 ] and refs . therein .
d. mikowiec , http://www.gsi.de/@xmath33misko / overlap/. p. braun - munzinger , j. cleymans , h. oeschler , k. redlich , [ hep - ph/0106066 ] .
na50 collaboration , m.c .
abreu et al . , ; .
j. gosset , a. baldisseri , h. borel , f. staley , y. terrien , .
braun - munzinger and k. redlich , .
na50 collaboration , m.c .
abreu et al . , .
wong , z.q .
r. rapp , e. shuryak , .
k. gallmeister , b. kmpfer , o.p .
pavlenko , [ hep - ph/9908269 ] .
blaizot , p.m. dinh , j .- y .
ollitrault , [ nucl - th/0007020 ] .
a. capella , a. b. kaidalov , d. sousa , [ nucl - th/0105021 ] .
thews , hep - ph/030205 .
adler et al . , phenix collaboration , nucl - ex/0305030 .
phenix collaboration , k. adcox et al .
, [ nucl - ex/0202002 ] .
r. averbeck , phenix collaboration , [ nucl - ex/0209016 ] .
leitch et al . , [ nucl - ex/9909007 ] . | we study the production of charmonia and charmed hadrons in nucleus - nucleus collisions at sps , rhic , and lhc energies within the framework of the statistical hadronization model .
results from this model are compared to the measured yields and centrality dependence of @xmath0 production at sps and rhic energies .
we explore and contrast the centrality dependence of the production of mesons with open and hidden charm at sps , rhic and lhc .
the sensitivity of the results to various input parameters is analyzed in detail for rhic energy .
, , , statistical hadronization , open and hidden charm hadrons , @xmath0 suppression or enhancement 25.75.-q , 25.75.nq , 25.75.dw |
interdisciplinary efforts in network science have considerably deepened our understanding about the structure and dynamics of the society @xcite .
this active development of the field is largely attributed to the huge amount of digital information that has become available due to rapid development of the information and communication technology ( ict ) .
many empirical studies have been conducted on social networks including those based on email @xcite , mobile phone call ( mpc ) @xcite , short - message communication , social networking services ( sns ) @xcite , and scientific collaborations @xcite . among these
, mpc data sets play a special role because mobile phones are frequently used in our daily life and the coverage of the service is almost 100% for adults in large part of the world . as a matter of fact , an analysis of the mpc data set @xcite has been successful in validating the `` strength of weak ties '' hypothesis proposed by granovetter @xcite , stating that if the tie between two persons is strong , then the overlap between their neighbors will be large .
this has far - reaching consequences for the overall structure of the social network : it suggests that the network consists of strongly wired communities connected by weak ties .
this structure was proven at a societal level by applying percolation analysis to a social network , in which the duration of a phone call between two individuals serves as a proxy for their social tie strength .
hereafter we will refer to these networks as having granovetterian structure .
this led to the attempts to constructing a model of social network formation reflecting the observed structures @xcite .
more recently , this model was generalized to reflect the multiplex character of social networks @xcite . in the original model by kumpula et al
@xcite two main mechanisms for the link - formation process were introduced , namely cyclic and focal closures , as they have been empirically observed to be the fundamental mechanisms for creating social ties @xcite . here
the cyclic closure describes the formation of a link with the neighbor of one s network neighbor or , in other words , with a friend of a friend .
the focal closure refers to making a link with the one sharing an attribute independently of the local network topology or geodesic distance between the individuals .
in addition to these two mechanisms , a link reinforcement was introduced to correspond to the general observation that social ties get stronger when they are used .
while in the previous studies the cyclic and focal closure mechanisms have been shown to play an important role in generating a granovetterian structure , much less is known about the role of the process in which a relationship becomes terminated . in the original model @xcite nodes ( and all their links ) were removed with a small probability or with a slow rate and new nodes without any links were added to the system to replace them . this way the asymptotic stationarity of the system with some average degree
could be maintained .
deletion of a node corresponds to the death of a person or giving up a service or moving far away so that all contacts break . in real human society , however , the termination of a relationship may occur in various ways and there is no empirical support that the deletion of a node would be the main mechanism for terminating links . even without a removal of a person , a relationship may end abruptly , for example , when a couple in intimate relationship suddenly break up .
according to an empirical study , an individual drops out one member of her relationships every @xmath0 months on average @xcite . apart from these rather sudden and drastic changes in human relationships , there is also a more gradual fading out of a friendship , which is typically seen when old but aging friends make less and less contact by time .
it is not clear whether such difference in tie terminations affects the emergent network properties . in this paper , we investigate the effects of various link termination mechanisms and show that the link aging may lead to a significantly different network from the one with abrupt link termination or link deletion .
the comparison between the models with different link termination mechanisms indicates that the link deletion reproduces the empirical findings from the mpc data set better than other mechanisms , implying that a link deletion plays a prominent role in the evolution of social networks among the link termination mechanisms .
this paper is organized as follows .
first we define the models with three different link termination mechanisms , including the original node deletion mechanism .
then we demonstrate that modular structures are more enhanced under the link aging than other mechanisms .
various local network features , such as degree distributions , link weight distributions , and link overlap , are investigated and compared with the mpc data set . finally , we present summary and discussion in the last section .
first we briefly review the original weighted social network model proposed in @xcite .
it considers an undirected weighted network of @xmath1 nodes , with links between them being updated by the following three mechanisms .
the first mechanism is _ local attachment _ ( la ) , in which node @xmath2 chooses one of its neighbors @xmath3 with probability proportional to @xmath4 that stands for the weight of the link between nodes @xmath2 and @xmath3 .
then node @xmath3 chooses one of its neighbors but @xmath2 , say @xmath5 , randomly with probability proportional to @xmath6 and if nodes @xmath2 and @xmath5 are not connected , they are connected with probability @xmath7 with a link of weight @xmath8 .
in addition all the involved links increase the weights by @xmath9 , whether a new link is created or not .
the second mechanism is _ global attachment _ ( ga ) , in which a node with no links or otherwise with probability @xmath10 makes a new link to a randomly chosen node with weight @xmath8 .
the third mechanism is _ node deletion _ ( nd ) , in which a node loses all its links with probability @xmath11 , equivalently to replacing the node with a new isolated node .
these three processes , i.e. , la , ga , and nd , are applied to all nodes at each time step .
the initial condition is set to be a network without any links and the network reaches a statistically stationary state after a sufficient number of updates .
this model turned out to show the granovetterian structure of the society with strongly wired communities connected by weak ties . in this paper
we extend the original model such that instead of nd , we consider two different link termination mechanisms : _ link deletion _ ( ld ) and _ aging_. in case of link deletion we assume an abrupt termination of links , such that each link is removed from the system with probability @xmath12 at each time step . here
we assume this probability to be independent of the link weight such that even a strong link can be suddenly terminated .
in contrast , for the link aging mechanism we assume a gradual degradation of the link weight , such that at each time step , weights of all the links are multiplied by an aging factor @xmath13 ( @xmath14 ) , where @xmath13 controls the speed of aging .
if the link weight becomes less than a threshold value , @xmath15 , the link is removed from the system . from now on , models with nd , ld , and aging mechanisms will be called the nd model , ld model , and aging model , respectively .
in figure [ fig : snapshot ] we show for comparison typical snapshots of the networks for three different link termination mechanisms with similar average degree .
all the networks show modular structures , where the links connecting communities are typically weaker than the intra - community links , thus indicating the existence of the granovetterian structure
. however , the strength of the modularity varies significantly .
the aging model shows clearly the strongest modularity among three networks while the ld model shows less modular structure than the others .
another interesting finding for the aging model is that the link weights within a community are more homogeneous .
furthermore , typical link weight in a community is inversely correlated with the community size .
this is contrasting to the other models , where each community has both weak and strong links . in the following ,
we investigate these observations quantitatively . ) .
color of links denotes the link weight .
strong and weak links are depicted in red and light blue , respectively .
note that different color scales are used for visibility .
for the aging mechanism , community structures are clearly visible and the link weights in each community are more homogeneous . in the networks using the nd and ld mechanisms , strong and weak links coexist in each community . ]
here we performed simulations by setting @xmath16 , @xmath17 , @xmath18 , @xmath19 , and @xmath20 .
for the nd model with @xmath21 a typical granovetterian structure is observed @xcite such that the average degree is @xmath22 . in order to compare the results with the ld and aging cases , we control the parameters for the ld and aging models so that average degrees are comparable with that of the nd model .
we adopted @xmath23 for ld model , and @xmath24 , @xmath25 , and @xmath26 for aging model .
note that a higher @xmath10 is used for the aging model , otherwise the fragmentation of the network occurs .
the results were obtained by running the simulations up to @xmath27 time steps and averaging over @xmath28 realizations .
the average degrees and other quantities are summarized in table [ table : degrees ] .
we note that qualitatively similar results are observed for reasonable ranges of parameters other than those specified above . .
* average degrees , clustering coefficients , and maximum modularity . *
the average degree @xmath29 , global clustering coefficient @xmath30 , maximum modularity for binary graph @xmath31 are summarized for the nd , ld and aging models together with the simulation parameters .
relative fluctuations of link weights in each community , @xmath32 , are also presented . [ cols="^,^,^,^",options="header " , ]
in this paper , we have investigated how the mechanism of link termination affects the emerging network structures by introducing and comparing three different models , namely node deletion ( nd ) , link deletion ( ld ) , and aging models .
first of all , we would like to emphasize that the granovetterian structure is robustly observed for all three models .
it supports the observation by the previous studies that cyclic and focal closures play key roles in generating the granovetterian structure .
on the other hand , there are significant differences between the nd , ld and aging models .
we find that aging promotes modular structure in networks having the granovetterian structure .
this is because inter - community weak links are often deleted while strong links existing within communities need a longer time to be removed , thus making the communities more persistent under the aging than under nd or ld .
it is also notable that the aging makes link weights in each community more homogeneous .
in contrast , with link deletions assumed in the nd or ld models communities are less modular and links have variable weights . with the ld model
it is even harder to maintain high modularity than with the nd model because the density of links in a community tends to decrease more rapidly with ld . local network properties for these three models , such as degree distribution , node strength distribution , and link overlaps ,
have also been investigated and compared with the mpc data set observations .
the results show that the nd and ld models reproduce the mpc data set while the aging model shows different profiles due to its highly modular structure with homogeneous links .
this implies that abrupt link deletion is the major mechanism for deleting ties in real society , while link aging , being presumably present , plays a minor role .
the difference between the nd and ld models are less clear , but we think that the ld model describes the reality better than the nd model . degree and node strength distributions for the ld model do not show a monotonically decaying behavior but have peaks at characteristic scales .
this matches better to our common sense that each person has some amount of resources devoted to communication although we do not have conclusive evidence from empirical data yet .
furthermore , an empirical analysis has shown that the relationships terminate with a significant rate even with individuals who have had a long term relationship @xcite .
thus , the ld model would be a good starting point for a simple model of social networks . in this paper , we have focused on non - overlapping communities but in reality , the communities may be overlapping as discussed in @xcite . in order to realize such overlapping communities ,
we have recently proposed a multilayer model , where the network in each layer is generated in a similar way to the original weighted social network model @xcite . in that case
the node deletion takes place in each layer independently , which corresponds to an event that a node leaves one of its communities .
this may also sound a reasonable assumption since we sometimes lose links not only to a person but collectively to his or her close friends .
although we expect most of the properties investigated in this paper to hold for the multilayer model , further studies on multilayer models would be interesting . in reality , the link termination process happens in more complicated and various ways and our model is at best a simplification .
although we assume that the link does not recover once it is removed , more than half of the real conflicts are reconciled @xcite . furthermore , the probability of link deletion should also depend on the age of the friendship @xcite . in twitter data set
, it was observed that `` unfollowing '' behavior depends on the local network topology such as reciprocity or existence of common friends @xcite . as these aspects are missing in our models the future investigations on these factors
are expected to make the model even more realistic .
two aspects should be emphasized finally .
first , when we compare our models with data , we have to restrict ourselves to the available sources , which means single channel communication records .
this is an obvious source of bias , as multiple channels are crucial in human communication , which may change some of the statistics qualitatively .
second , the three different mechanisms separately treated here in disjunct models are simultaneously present in a real human society .
the advantage of the present study is that we have gotten insight about their role in the formation of the social network .
y. m. appreciates hospitality at aalto university .
j. acknowledges financial support by aalto university postdoctoral program and by mid - career researcher program through the national research foundation of korea ( nrf ) grant funded by the ministry of science , ict and future planning ( 2014030018 ) , and basic science research program through the national research foundation of korea ( nrf ) grant funded by the ministry of science , ict and future planning ( 2014046922 ) .
j. t. and j. k. acknowledge support from eu grant no .
fp7 317532 ( multiplex ) .
the systematic simulations in this study were assisted by oacis @xcite .
in this appendix , a linear stability analysis of the link weight @xmath33 around the equilibrium value @xmath34 for the aging model is demonstrated . in order to evaluate the linear stability of @xmath33 around @xmath34
, we consider the case that one of the links , say the link between nodes @xmath2 and @xmath3 , is stronger than other links by @xmath35 while other links have the same weight as @xmath34 .
the expected amount of reinforcement on the link @xmath36 in one time step , @xmath37 , is given by @xmath38 the first , second , and third terms correspond to the case that the link @xmath36 is selected as the first , second , and third link in one la event , respectively .
la must start from the node @xmath2 or @xmath3 when the link @xmath36 is selected either as the first or the third link , which gives @xmath39 in the first and the third terms .
when the link @xmath36 is selected as the second link , the la event must start from the node except for @xmath2 and @xmath3 , which yields the factor @xmath40 in the second term .
if we assume @xmath41 and ignore the higher order terms of @xmath35 , the expected increase @xmath37 is calculated as @xmath42 on the other hand , the expected amount of the decrease due to aging , @xmath43 , is @xmath44 . when @xmath45 , the total change of the link weight is calculated using eq .
( [ eq : w_eq ] ) as @xmath46 which always has a negative sign of @xmath35 for @xmath47 , indicating the dynamics of @xmath35 is linearly stable around its equilibrium value @xmath34 .
this demonstrates that the link weight within communities gets closer to its equilibrium point with time .
since the equilibrium link weight is shared in a community , the link weights in a community become homogeneous . | in social networks of human individuals , social relationships do not necessarily last forever as they can either fade gradually with time , resulting in `` link aging '' , or terminate abruptly , causing `` link deletion '' , as even old friendships may cease . in this paper
, we study a social network formation model where we introduce several ways by which a link termination takes place .
if we adopt the link aging , we get a more modular structure with more homogeneously distributed link weights within communities than when link deletion is used . by investigating distributions and relations of various network characteristics , we find that the empirical findings are better reproduced with the link deletion model .
this indicates that link deletion plays a more prominent role in organizing social networks than link aging . |
a popular scenario to explain the complex physics of the high - temperature superconductors is based on the idea that the ground state of these materials exhibits some form of charge ordering tendency ( charge stripes , checkerboard order , etc . )
@xcite , which , in turn , can lead to incommensurate magnetism ( spin stripes ) . from the outset we state that we do not subscribe to this point of view . in order to illustrate the direction of our efforts
, we will take as an example the la cuprate family , of which the most studied representative is la@xmath0sr@xmath1cuo@xmath2 ( lsco ) ( @xmath4 is the hole doping ) .
this compound is commonly believed to show `` dynamic '' charge order near the special doping value @xmath5 , while static order ( although quite weak ) has been observed with additional nd co - doping , or upon the substitution la@xmath6ba . however
the rest of the lsco phase diagram at low doping @xmath7 does not exhibit any charge order @xcite making it rather hard to accept the universal concept that holes in antiferromagnets fundamentally tend to segregate into charge stripes or similar structures .
we start from the guiding principle that spin order is the primary phenomenon while charge order is an exception to the rule and could possibly occur under special circumstances only ( such as dopings corresponding to commensurate spin structures ) .
this leads to a natural candidate for the ground state - the spiral spin state , as pointed out quite a while ago @xcite .
we will seek to provide theoretical support for the validity of the spiral picture within the extended @xmath3 model , as well as explain two sets of phenomena observed in lsco : * the presence of incommensurate magnetic order ( with incommensurability = doping ) both in the metallic ( superconducting ) and insulating phases at low doping , and in particular the mysterious change of the incommensurability direction by 45@xmath8 at the metal - insulator transition point @xmath9 @xcite .
* the in - plane transport anisotropy , as high as 50% , observed in the insulating spin glass phase ( @xmath10 ) @xcite . the near equality of incommensurability and doping is usually considered as one of the successes of the charge stripe scenario , whereas the 45@xmath8 rotation could follow from considerations based on hole - pair checkerboard order @xcite .
the transport anisotropy of course also could be interpreted as a signature of ( fluctuating ) charge order , although no specific calculations have been performed @xcite . in the spirit of our alternative philosophy
we will show that a theory based on holes moving in an antiferromagnet and causing the formation of a spiral spin density wave can explain the above phenomena , giving in particular a _ quantitative _ value for the magnitude of the transport anisotropy .
the spiral theory is fermi liquid in nature @xcite , without any charge ordering tendencies , and stands on firm theoretical ground .
it also provides a unified and consistent picture of the relationship between incommensurate magnetism and transport anisotropy .
the spiral order is generated by the hole motion in the nel antiferromagnetic state , in an attempt of the magnetic background to partially relieve the frustration caused by the hopping @xcite .
this leads to the non - collinear configurations shown in fig . 1 .
parametrizing the magnetic order as : @xmath11 , @xmath12 , the angle of deviation from collinearity is @xmath13 , where @xmath14 is the spiral vector directed along the @xmath15 or @xmath16 lattice directions .
these are both co - planar configurations , and the unit vector @xmath17 is perpendicular to the spin plane . for low doping corresponding to small deviations from nel order , one finds that @xmath14 is proportional to doping : @xmath18 $ ] . here
@xmath19 is the hopping , @xmath20 is the spin stiffness , and @xmath21 is the quasiparticle residue at the points @xmath22 , corresponding to the minima of the hole dispersion at low doping .
it was realized shortly after the first several works on the spiral physics that the spiral state have a tendency towards some kind of instability in the charge sector ( phase separation ) @xcite .
indeed , within the @xmath3 model one finds a negative charge compressibility , @xmath23 , defined , as in fermi liquid theory , through a derivative of the ground state energy @xmath24 .
the presence of such an instability towards a hole segregated state would then mean that charge stripes in some form are likely to be present in the ground state .
under what conditions would the fermi liquid physics survive ?
a possibility that we have recently explored in detail is the presence of additional ( next nearest neighbor ) hoppings @xmath25 @xcite . for lsco
the values of these parameters are quite small : @xmath26 , where @xmath27 is the magnetic exchange , and @xmath28 . from now on .
] however we find that their presence is crucial for the stability of the system . in order to perform as accurate calculations as possible
we follow a two - step procedure : ( 1 . ) the one hole properties , such as @xmath21 and the low - energy dispersion @xmath29 near the nodal points are calculated in the self - consistent born approximation , and ( 2 . )
these are inserted into the many - body fermion - magnon low - energy vertices which are then treated in perturbation theory ( loop expansion ) in powers of @xmath30 .
we call this technique `` chiral perturbation theory '' , by analogy with qcd .
the perturbative parameter is not small , since @xmath31 , and thus @xmath32 .
however we find that the perturbation theory converges numerically extremely well @xcite , as if governed by the effective coupling constant @xmath33 .
are in units of @xmath27 .
inset : stability boundary at different levels of approximation ( loops ) , showing the good convergence . ] the results , presented in fig . 2 ,
indicate that fairly small values of @xmath25 stabilize the uniform @xmath16 spiral state .
the inset shows the transition line calculated in different loops ( powers of @xmath34 ) , confirming the good convergence and reliability of the results .
our results are in a way not surprising because dmrg work @xcite had found that charge stripes which are stable at @xmath35 become unstable ( and thus the system becomes uniform ) upon introduction of @xmath36 . in our theory
the stability depends physically on the shape of the fermi surface while at @xmath35 it is very elongated in the antinodal direction ( @xmath37 ) , upon introduction of @xmath38 it becomes more spherically symmetric ( for example @xmath39 for the lsco values ) , and this effective two - dimensionality leads to the stability of the uniform spiral state .
it should be noted that our calculations are valid for small doping @xmath40 , since we keep track of only the terms of order @xmath41 in the ground state energy , and thus the compressibility has the expansion @xmath42 , leading to a stability line which does not depend on doping . the ground state energy change ( relative to the undoped nel state ) due to the spiral formation with the two symmetries , satisfies the relation : @xmath43 .
this formula is correct to all perturbative orders that we have checked and can be traced to the different number of occupied hole pockets for the different spiral orientations @xcite . above the stability boundary
where @xmath44 , the @xmath16 state has lower energy , but one can show that the overall energy change relative to the _ doped _ collinear ( nel ) state is negative @xcite . ] .
the @xmath16 symmetry is the correct one for the metallic phase of lsco ( @xmath45 ) , where the elastic neutron scattering peaks are in a `` parallel '' pattern around @xmath46 @xcite . what happens if the system becomes an insulator ?
as is clear from the calculation of the energy @xcite , the @xmath16 state has lower energy only due to the fermi motion energy @xmath47 . on the other hand
if we let @xmath48 , which would be the case at the transition to the spin glass region @xmath49 where the holes are localized , then the @xmath15 state is selected ( consistent with neutron peaks in `` diagonal '' pattern @xcite ) .
of course this argument can be applied only up to the metal - insulator transition point , beyond which a detailed model for the spiral formation has to be constructed @xcite .
nevertheless it is clear that the spiral model can explain correctly the presence and the symmetries of the elastic neutron scattering peaks both in the metal and in the insulator .
the exact location of those peaks ( determined by @xmath14 ) is also in very good agreement with experiment @xcite .
we now turn to the insulating spin - glass region @xmath10 , where the presence of incommensurate magnetic peaks ( with @xmath15 symmetry ) is clearly related to the in - plane dc resistivity anisotropy @xmath50 @xcite ( @xmath51 and @xmath52 are the orthorhombic coordinates ) . from elastic neutron scattering @xcite the incommensurability is determined to be along the orthorhombic @xmath52 direction , meaning that in the spiral picture it is in the @xmath15 direction , as shown in fig .
experimentally @xmath53 at the lowest temperature @xmath54 and then decreases , disappearing completely around @xmath55 where the system becomes quasi - metallic . ) .
inset : maximum ( @xmath56 ) anisotropy as a function of @xmath57 . ]
since it is clear that the largest anisotropy is accumulated in the low - temperature region , we will concentrate on the strongly - localized , variable - range hopping ( vrh ) temperature range below approximately @xmath58 .
the resistivity of lsco is well fit by the mott 2d vrh formula @xmath59 , with characteristic values @xmath60 , depending on doping . in order to address the anisotropy problem
theoretically , one needs to develop a theory of spiral formation arising from the presence of ( randomly distributed ) localized holes .
recent works have developed such theories @xcite , and in particular we have shown @xcite that holes , localized around @xmath61 ions can `` carry '' spiral correlations with them that decay in dipolar fashion at long distances . at finite doping the dipoles can order , producing a spiral state as the one in fig .
1(a ) . within this framework
we have calculated the vrh conductivity anisotropy by calculating first the wave - function of a localized hole ( with localization length @xmath62 ) , in the presence of spiral correlations @xcite : @xmath63 .
this wave - function is anisotropic ( the angle @xmath64 is defined in fig .
1(a ) ) , which is physically natural since the hole `` feels '' a different environment depending on the lattice direction , due to the non - collinearity of the spiral state .
the strength of the anisotropy is controlled by the parameter @xmath57 , and @xmath65 for lsco ( where @xmath66 ) .
the overlap of the wave - functions discussed above leads to the resistivity anisotropy , shown in fig .
the magnitude of the anisotropy agrees extremely well with experiment @xcite .
we emphasize that there are no adjustable parameters in our theory , although it is certainly valid at low doping only ( and , similarly to the stability results of section 2 , the curve in fig .
3 is doping independent ) .
perhaps most importantly , our analysis shows that the transport anisotropy can be due to the underlying incommensurate magnetic correlations , rather than a tendency of the charges to self - organize .
the spiral spin density wave theory passes both the fundamental and phenomenological tests :
it is well theoretically supported in the context if the extended @xmath3 model , and is capable of describing magnetic and transport properties of lsco .
a novel finding is that the in - plane transport anisotropy fits well into the spiral picture on a quantitative level we consider this result particularly important because the anisotropy is a rather elusive quantity and usually hard to calculate consistently .
we should mention that the reported work essentially explored only the structure of the ground state ; how consistent the spiral picture would be with experiment at higher energy is not yet clear .
nevertheless all low - temperature lsco experiments we have looked at so far lead us to the conclusion that the physics at low doping is _
spin driven_. the recently observed `` magic '' doping concentrations , where the conductivity of lsco shows dips @xcite , are also due , in our view , to spin related phenomena ( such as special points where the spin structure becomes commensurate ) .
currently research is in progress to explore further the predictions of the spin density wave approach ( e.g. in the direction of including lattice effects and studies of magnetotransport ) .
the theory provides , we believe , a consistent picture of the complex interplay between spin and charge dynamics in the underdoped region of the cuprates . | we discuss the spiral spin density wave model and its application to explain properties of underdoped la@xmath0sr@xmath1cuo@xmath2 .
we argue that the spiral picture is theoretically well justified in the context of the extended @xmath3 model , and then show that it can explain a number of observed features , such as the location and symmetry of the incommensurate peaks in elastic neutron scattering , as well as the in - plane resistivity anisotropy .
a consistent description of the low doping region ( below 10% or so ) emerges from the spiral formulation , in which the holes show no tendency towards any type of charge order and the physics is purely spin driven .
address= institute of theoretical physics , swiss federal institute of technology ( epfl ) , + ch-1015 lausanne , switzerland address = school of physics , university of new south wales , sydney 2052 , australia |
a variation of a classical external parameter of a quantum system offers a way to manipulate quantum states . almost since the dawn of the quantum theory , it has been recognized that the slow variation of the parameter ensures the adiabatic time evolution , where the quantum state can be pinned to an eigenspace of system s hamiltonian @xcite .
later , the concept of quantum holonomy has been developed for adiabatic cycles on quantum systems . among them , the phase holonomy , where an adiabatic cycle induces a nontrivial change in the phase of a quantum state @xcite , is a textbook result nowadays @xcite .
recently , the quantum holonomy of exotic kind has been recognized in that an adiabatic cycle can induce changes both in the eigenvalue and the eigenspace of a stationary state @xcite .
such changes are also referred to as eigenenergy and eigenspace anholonomies @xcite .
spectral degeneracies are crucial for the quantum holonomy . the phase holonomy is associated with a spectral degeneracy point , where a structure mathematically identical to the magnetic monopole resides , in the parameter space @xcite . as for the exotic quantum holonomy
, it is shown , for a quantum kicked spin-@xmath0 , that degenerate points in the complexified parameter space play the central role @xcite .
such a non - hermitian degeneracy point is known as an exceptional point @xcite .
there have been considerable number of recent works on exceptional points in non - hermitian quantum physics @xcite .
it is natural to expect that the exceptional points govern the exotic quantum holonomy in general , once we accept the view that the anholonomies in eigenenergy and eigenspace are a manifestation of multiple - valuedness of the solution of the eigenvalue problem in the parameter space .
we remind the readers that an eigenvalue equation of a hamiltonian can be cast into an algebraic equation .
hence its multiple - valued solutions form a family , and the family coalesces at a branch point in the parameter space @xcite .
the encirclement of the branch point induces the permutation of the eigenvalues and eigenspaces in the family , which is to be identified as the complex - analytic origin of the exotic quantum holonomy .
this , however , is rather unexpected scenario , because the exceptional points emerge only when the hamiltonian is far from hermitian , in spite of the fact that the exotic quantum holonomy is induced by an adiabatic hermitian cycle .
hence the question is whether a family of hermitian hamiltonians that define an adiabatic cycle really `` encloses '' exceptional points . the aim of this manuscript is to offer another example of successful `` exceptional point picture '' for the exotic quantum holonomy , in quantum many - body systems .
we examine the lieb - liniger model , which describes bose particles confined in a one - dimensional space subject to the periodic boundary condition @xcite .
ushveridze showed that a non - hermitian extension of this model has an infinite number of exceptional points @xcite .
recently , it is shown that the lieb - liniger model exhibits the eigenenergy and eigenspace anholonomies @xcite .
hence , our purpose is to explain how the non - hermitian degeneracies of this model and the exotic quantum holonomy that occurs in hermitian hamiltonian is interrelated .
we here focus on the simplest case where the number of particles is two .
we believe that the present two - body study offers the foundation for the case of an arbitrary number of particles . the outline of this manuscript is the following .
we introduce the lieb - liniger model in section [ sec : hermite ] .
we also explain that an adiabatic hermitian cycle of this model induces the exotic quantum anholonomy @xcite .
we cover a non - hermitian extension of the lieb - liniger model @xcite in section [ sec : ndim ] .
we outline the analytic continuation of the quasi - momentum in section [ sec : wavenumber ] .
we show an association of the eigenenergy anholonomy with exceptional points in section [ sec : energy ] .
we explain the role of the exceptional points in the gauge theoretical formulation of eigenspace anholonomy in section [ sec : eigenspace ] .
we discuss the present result in section [ sec : discussion ] .
we summarize this manuscript in section [ sec : summary ] .
we review the two - body lieb - liniger model @xcite and its exotic quantum holonomy @xcite . throughout this manuscript
, we examine the system that consists of two identical bose particles confined within a one - dimensional space , which is @xmath1-periodic .
we assume that the two particles have a contact interaction whose strength is @xmath2 .
the system is described by the hamiltonian : @xmath3 where the units are chosen such that @xmath4 and the mass of a bose particle are @xmath5 .
we explain the standard method to solve the eigenvalue problem of @xmath6 @xcite .
we employ the bethe ansatz , where an eigenfunction is expressed by two plane waves that are associated with quasi - momentum ( rapidity ) @xmath7 ( @xmath8 ) .
the total momentum @xmath9 must be an integer , because the periodic boundary condition is imposed .
on the other hand , the difference in the quasi - momenta @xmath10 satisfies the condition that @xmath11 is an integer @xcite . throughout this paper
, @xmath12 denotes the principal value of the inverse tangent . for odd @xmath13
, @xmath14 satisfies @xmath15 whereas even @xmath16 implies @xmath17 we call and the bethe equations .
we look for the solution of the bethe equations for real @xmath2 .
let @xmath18 denote the solution that satisfies @xmath19 , and smoothly depends on @xmath2 in the real axis @xmath20 .
it suffices to examine the case where @xmath21 is a non - negative integer .
@xmath18 is either real or pure imaginary .
the latter case describes the `` clustering '' of two particles , which occurs only when @xmath22 and @xmath23 , or , @xmath24 and @xmath25 .
we here summarize the relevant facts on @xmath18 shown in ref .
@xcite : ( a ) @xmath26 , and accordingly @xmath27 for @xmath28 ; ( b ) for @xmath29 , @xmath30 and accordingly @xmath31 for @xmath23 ; ( c ) @xmath32 and @xmath33 satisfy and , respectively .
there remain freedoms to choose the signs of @xmath34 for @xmath22 and @xmath5 . here
we carry out the analytic continuation of @xmath35 through the lower half plane of @xmath2 from @xmath28 .
although this choice is arbitrary for our purpose , the present choice is consistent with the condition that the non - hermitian lieb - liniger model describes the ( forward-)time evolution correctly ( see , sec . [
sec : ndim ] ) . as a result
, we obtain @xmath36 for @xmath22 and @xmath5 .
the eigenstates of the two - body lieb - liniger model are specified by two quantum numbers @xmath37 and @xmath21 , which must be both even or odd .
the corresponding eigenenergy is @xmath38 we introduce a cycle @xmath39 in real @xmath2-space to investigate the exotic quantum holonomy .
the initial point of @xmath39 is @xmath40 .
we increase @xmath2 adiabatically during @xmath41 .
then , @xmath2 is suddenly flipped from @xmath42 to @xmath43 .
such a sudden flip has been investigated both in theory @xcite and experiments @xcite to approach super tonks - girardeau gas .
to finish @xmath39 , @xmath2 is adiabatically increased from @xmath43 to @xmath40 .
the exotic quantum holonomy is found in the two - body lieb - liniger model along the cycle @xmath39 @xcite .
we initialize the interaction strength as @xmath44 , and prepare the system to be in the eigenstate specified by quantum numbers @xmath45 .
as we increase @xmath2 adiabatically , the energy of the system follows @xmath46 , which increases monotonically .
when we arrive @xmath47 , we assume that we suddenly switch the value of @xmath2 from @xmath42 to @xmath43 , keeping the system remain unchanged . because of @xmath48 , the system is in the @xmath49 state after the switch .
as we increase @xmath2 from @xmath43 to @xmath40 adiabatically , the energy of the whole system arrives at @xmath50 , which does not agree with the initial energy .
this is the eigenenergy anholonomy of the two - body lieb - liniger model . because the hamiltonian is hermitian for real @xmath2 ,
the eigenenergy anholonomy implies the eigenspace anholonomy , i.e. , the initial and final state vectors correspond to different eigenenergies and are thus orthogonal .
we shall show , in the following sections , that the spectral degeneracies that are hidden in the complexified parameter space governs the exotic quantum holonomy of the two - body lieb - liniger model . to carry out this
, we introduce the complexification of the coupling strength @xmath2 .
we outline formal aspects of the consequence of the complexification in this section , and the details of the analytic continuations of relevant quantities will be explained in the following sections .
we refer ref .
@xcite for the theory of non - hermitian eigenvalue problem .
here we focus on the lower - half plane of @xmath2 .
this is just a matter of convention in our analysis since there is a symmetry about the real axis in the complex @xmath2-plane .
however , the present choice is suitable once we realize that the complexified lieb - liniger model describes the dissipative effect . in ref .
@xcite , it is shown that the presence of inelastic collisions implies that the imaginary part of the effective one - dimensional coupling constant must be zero or negative , i.e. , @xmath51 .
first of all , the complexification makes the lieb - liniger hamiltonian @xmath6 non - hermitian .
an immediate consequence is that the two eigenvalue problems for @xmath6 and @xmath52^{\dagger}$ ] become different because of the relation @xmath52^{\dagger } = h(g^*)$ ] .
an eigenvalue of @xmath52^{\dagger}$ ] is given by @xmath53^*$ ] , which may not be identical to @xmath54 . from the comparison of the spectrum sets of @xmath55 and @xmath52^{\dagger}$ ]
, there is a unique pair @xmath56 that satisfies @xmath57 where we assume that @xmath55 has no spectral degeneracy within the subspace specified by the total momentum @xmath37 .
this assumption holds in the vicinity of the real axis .
we note that the correspondence between @xmath58 and @xmath21 generally depends on the details of the analytic continuation and the value of @xmath2 .
we ignore the case @xmath59 because we focus on the neighborhood of the real axis of @xmath2 , and assume @xmath60 .
this implies @xmath61 where @xmath62 is either @xmath5 or @xmath63 , depending on how the analytic continuation of @xmath35 is carried out .
also , @xmath62 may depend on @xmath21 and @xmath2 .
eigenfunctions of @xmath6 can be obtained through a standard way with the help of the bethe ansatz .
we refer ref .
@xcite for details .
let @xmath64 be an eigenfunction that corresponds to the eigenenergy @xmath54 . because of the bose statistics , we will write down only the expressions of eigenfunctions in the region @xmath65 : @xmath66 where @xmath67 & \text { for even $ n$ , } \\ \frac{1}{\sqrt{2\pi } } { a^{(\mathrm{e})}}(k ) \sin\left[k ( x-\pi)/2\right ] & \text { for odd $ n$ , } \\ \end{cases}\end{aligned}\ ] ] and the following abbreviation is introduced : @xmath68 here we choose the normalization constants @xmath69 and @xmath70 being independent of @xmath21 in order to ensure that the eigenfunctions are continuous with respect to @xmath2 .
we turn to @xmath71 that is an eigenfunction of @xmath52^{\dagger}$ ] corresponding to the eigenvalue @xmath72 . because of @xmath52^{\dagger }
= { h}(g^*)$ ] , we find @xmath73 where @xmath74 & \text { for even $ n$ , } \\
\frac{1}{\sqrt{2\pi } } { a^{\mathrm{l}(\mathrm{e})}}(\tilde{k } ) \sin\left[\tilde{k}(x -\pi)/2\right ] & \text { for odd $ n$ , } \\ \end{cases}\ ] ] and we introduce another abbreviation @xmath75 we choose the following normalization condition @xmath76 this implies @xmath77 where @xmath62 was introduced in .
the exotic quantum holonomy in lieb - liniger model can be cast into the anholonomy of the quasi - momentum @xmath35 . as explained in sec .
[ sec : hermite ] , the cycle @xmath39 changes @xmath78 into @xmath79 .
we here examine the analytic continuation of @xmath35 to provide the basis of the analysis in the following sections .
the quasi - momenta @xmath35 s in the complex @xmath2-plane form riemann surfaces .
there are two kinds of riemann surfaces that correspond to @xmath35 s with even and odd @xmath21s , respectively , of the two - body lieb - liniger model .
we first look at the riemann surface that involves @xmath35 s with even @xmath21 .
we carry out the analytic continuation of @xmath35 .
when @xmath2 is positive and @xmath80 is sufficiently small , @xmath81 holds . expanding this equation with respect to the small parameter @xmath80
, we obtain @xmath82 where @xmath83 is applicable to negative @xmath2 and arbitrary @xmath21 .
note that makes sense only when the denominator of @xmath84 is non - zero .
as long as we can find a way to avoid the breakdown of this condition , provides a way to obtain @xmath35 with an arbitrary @xmath2 .
we examine the condition that the procedure above is inapplicable .
we assume that @xmath85 satisfies @xmath86 .
this implies that @xmath2 is a branch point of @xmath35 .
indeed , we obtain @xmath87 where @xmath88 as long as the denominator of @xmath89 is nonzero .
when @xmath90 vanishes , the branch point is of third or higher order .
for the two - body lieb - liniger model , the degree of all branch points is two , as is to be seen below .
we enumerate the branch points by solving the equation @xmath91 under the condition that @xmath13 is an integer .
this is equivalent to @xmath92 as long as @xmath93 holds . in the real axis
, the family @xmath35 s with even @xmath21 has only one branch point at @xmath94 , where two quasi - momenta @xmath32 and @xmath95 degenerate and the other quasi - momenta are not involved . accordingly the real branch point does not involve any spectral degeneracy .
complex branch points of quasi - momenta , which are obtained numerically , are depicted in figure [ fig : branchpoints ] .
all the complex branch points are confined in the region @xmath96 .
these complex branch points involve spectral degeneracies , because two quasi - momenta that are degenerate at a complex branch point provide different eigenenergies , except at the branch point .
this leads to a coalescence of the eigenspaces .
such complex branch points are called kato s exceptional points @xcite . in complex @xmath2-plane .
circles and squares correspond to even and odd @xmath21 , respectively .
when @xmath21 is even , we numerically obtain @xmath2 that simultaneously satisfies and .
there is only a single branch point @xmath94 on the real axis .
all other complex branch points are in the region @xmath96 .
we obtain the branch points for the odd @xmath21 case by solving and .
there is only a single branch point @xmath97 on the real axis .
all other complex branch points are in the region @xmath98 . , scaledwidth=46.0% ] our numerical result is consistent with the argument above in the sense that all complex branch points are of degree two .
we find that a complex branch point always involves @xmath32 , the quasi - momentum of the ground state .
as far as we see , there is no branch point that involves two `` excited states '' @xmath35 and @xmath99 with @xmath100 at the same time .
also , each @xmath35 with @xmath29 is involved with a complex branch point , which is denoted by @xmath101 .
namely , @xmath32 and @xmath35 degenerate at @xmath102 , where @xmath103 .
the numerical result can be explained qualitatively by a perturbation expansion around @xmath104 @xcite .
the quasi - momentum of the ground state diverges as @xmath105 at @xmath106 @xcite , where we ignore the small @xmath107 correction .
on the other hand , the quasi - momentum of the @xmath21-th excite state ( @xmath29 ) converges to a constant value , i.e. , @xmath108 at @xmath106 @xcite .
hence @xmath32 and @xmath35 coincides at @xmath109 , which is considered to be an approximation of @xmath101 in the lower half - plane .
although this argument is consistent only when @xmath21 is large enough , the present estimation seems to be applicable to smaller @xmath21 s ( see figure [ fig : branchpoints ] ) .
we also remark that the configuration of exceptional points in the upper half - plane is due to the other branch of @xmath32 , i.e. , @xmath110 at @xmath106 .
we construct a riemann sheet @xmath35 in the complex plane as follows . for real numbers @xmath111 and @xmath112 , @xmath113
is extended from @xmath114 along the line parallel to the imaginary axis , if there is no branch point in the interval between @xmath111 and @xmath115 .
branch cuts are chosen to be parallel to the imaginary axis .
we also require that the branch cuts do not traverse the real axis . when there is a branch point in the real axis , the corresponding branch cut is located in the upper half plane .
we depict some of the riemann sheets in fig [ fig : eps_in_even_sheets ] .
+ + + + we examine the quasi - momentum in the vicinity of @xmath116 ( @xmath29 ) .
a condition of the branch point @xmath117 implies @xmath118 and @xmath119 .
hence we have @xmath120 from , we conclude @xmath121 where the signs are determined from the numerical results , which are consistent with the fact @xmath122 and @xmath123 for @xmath124 . for the riemann surface that consists of @xmath35 s with odd @xmath21 , the situation is quite similar to the even @xmath21 case .
there is a real branch point , where @xmath33 and @xmath125 , which correspond to an equivalent eigenstate , degenerate at @xmath126 .
a complex branch point , which we denote @xmath101 ( @xmath29 ) , involves @xmath33 and @xmath35 .
there is no branch point that involves two excited states at a time . in the vicinity of @xmath101
, we have @xmath127 where the signs are determined from the numerical results , and are consistent with the fact @xmath128 and @xmath123 for @xmath124 .
it is straightforward to obtain the riemann surfaces of eigenenergies from the analysis of @xmath35 above . for a given @xmath37 , the quantum number of the center of mass
, we have a riemann surface that consists of @xmath46 for all possible @xmath21 s .
only even ( odd ) @xmath21 is possible for even ( odd ) @xmath37 @xcite .
let us take an example of the case of @xmath129 .
the eigenenergies @xmath130 with even @xmath21 form the corresponding riemann surface .
the branch point @xmath101 , which involves @xmath32 and @xmath35 ( @xmath29 ) , is introduced as the degeneracy point of @xmath131 and @xmath18 , as explained above .
also , @xmath101 is a branch point or exceptional point for the pair of eigenenergies of @xmath132 and @xmath130 . from
, we obtain a @xmath133-type behavior in the vicinity of @xmath101 : @xmath134 let us explain how we emulate the eigenenergy anholonomy by a closed contour in the complexified parameter space . more precisely , we compare the permutations induced by @xmath39 and closed cycles in the complex @xmath2-plane ( see , figure [ fig : c1c2 ] ) . the real cycle @xmath39 starts from @xmath44 and arrives at a point that is equivalent to @xmath40 , as explained in sec .
[ sec : ndim ] . note that @xmath2 passes @xmath135 during @xmath39 . on the other hand ,
the initial and final points of the closed cycles examined here are @xmath40 .
we show that this requires to include all relevant contribution from the exceptional points ( eps ) .
we start from `` @xmath136-ep approximation '' for integer @xmath136 .
firstly , we examine the contour that encloses only a single exceptional point @xmath137 ( `` @xmath5-ep approximation '' ) .
we explain the associated parametric evolution of each eigenenergy along the contour . as for @xmath132 and @xmath138
, they are exchanged each other after the completion of the closed cycle . on the other hand ,
the closed cycle does not change other eigenenergies .
this is because @xmath137 is the branch point that involves only @xmath132 and @xmath138 .
hence the resultant permutation is cyclic : @xmath139 which mimics the result of @xmath39 only for @xmath140 .
we note that such an approximation breaks down when we repeat the complex cycle .
a similar permutation between @xmath132 and @xmath130 occurs along a closed contour that enclose only @xmath101 ( @xmath29 ) .
we note that all @xmath5-ep case involves the ground energy @xmath132 .
secondly , we examine a closed contour that encloses only two different exceptional points , say , @xmath137 and @xmath141 ( `` @xmath142-ep approximation '' ) .
the result of the parametric evolution of each eigenenergy along the closed contour is the following cyclic permutation : @xmath143 hence the result of the single cycle , as for @xmath140 and @xmath144 , mimics the result of @xmath39 .
other cycles that involve two exceptional points induce a similar permutation .
we note again that all @xmath142-ep case involves the ground energy @xmath132 .
this is because all `` elementary '' branch points involve @xmath132 .
now it is straightforward to extend our analysis to @xmath136-ep cases ( `` @xmath136-ep approximation '' ) .
a closed contour that encloses @xmath136 branch points @xmath145 induces the following cyclic permutation of @xmath146 eigenenergies : @xmath147 which approximates the permutation induced by @xmath39 , as for lower - lying eigenenergies . in this sense , the limit @xmath148 provides us a closed cycle that emulates the eigenenergy anholonomy induced by @xmath39 .
our analysis suggests that the exceptional points offer `` elements '' of the eigenenergy anholonomy of the two - body lieb - liniger model .
this view is an extension of the previous result @xcite . by a thick line .
note that @xmath39 traverses the line @xmath149 .
the concatenation of @xmath39 and @xmath150 ( dashed curve ) encloses a single exceptional point @xmath151 , and offers the @xmath5-ep approximation . on the other hand , the concatenation of @xmath39 and @xmath152 ( dotted curve ) encloses two exceptional points @xmath151 and @xmath153 , and offers the @xmath142-ep approximation .
, scaledwidth=40.0% ]
in this section , we explain the role of exceptional points in the gauge theory that provides a unified formulation of the phase holonomy and the eigenspace anholonomy @xcite . in particular , we show an evaluation of the holonomy matrix @xmath154 ( see , below ) , which quantifies the eigenspace anholonomy , using the exceptional points .
this confirms our view that the exceptional points constitute a skeleton of the eigenspace anholonomy . to prepare this
, we make a brief review of the gauge theory of the eigenspace anholonomy in [ sec : gauge_theory ] .
we show that each exceptional point provides a `` local '' contribution to @xmath154 in [ sec : ep_local ] . by collecting these contributions
, we conclude this section ( [ sec : ep_global ] ) .
we outline the gauge theory of the eigenspace and phase anholonomies @xcite .
suppose that the system is initially in an eigenstate @xmath155 and the parameter is adiabatically deformed along a cycle @xmath156 .
let @xmath157 denote the final state induced by the adiabatic time evolution along @xmath156 .
we assume that the dynamical phase @xcite is removed from @xmath157 . a simple way to quantify the eigenspace anholonomy , which concerns about
the discrepancy between @xmath155 and @xmath157 , is to examine the overlapping integral or the holonomy matrix @xmath158 note that the adiabatic variation of the interaction strength @xmath2 does not vary @xmath37 , which is the quantum number of the center of mass .
hence it suffices to focus on the case @xmath159 : @xmath160 which is independent of @xmath37 , as is seen below .
also , @xmath161 is non - zero only when the oddness ( or evenness ) of @xmath37 , @xmath58 and @xmath21 is the same .
hence @xmath154 is consist of the even and odd blocks .
a gauge covariant expression of @xmath154 is @xmath162 where @xmath163 indicates the anti - path - ordered exponential , and @xmath164 and @xmath165 are gauge connections @xcite @xmath166 \\ { a^{\mathrm{d } } } _ { n',n}(g ) & \equiv \delta_{n',n}a_{\bar{k } n',\bar{k}n}(g ) , \end{split}\ ] ] which are independent of the total momentum @xmath37 , too .
two kinds of gauge invariants are involved in @xmath154 @xcite .
one is a permutation matrix and the other is the off - diagonal geometric phases @xcite . in the following ,
we impose the parallel transport condition @xcite for each eigenspace , i.e. , @xmath167 this makes the parametric evolution of the eigenvectors precisely describe the adiabatic time evolution except the dynamical phase .
hence it is also suitable to investigate analytic continuation of the adiabatic parameter for eigenfunction .
regardless of @xmath156 being closed or open , the parametric evolution of eigenvectors is described by the gauge connection @xmath168 as @xmath169_{n'n } .\ ] ] in particular , the second factor in vanishes @xmath170 we obtain the gauge connections from , and .
the diagonal elements are @xmath171 , & \text { for even $ n$ , } \\ \partial_g \left[i\ln\left\ { { a^{(\mathrm{e})}}(k)\left(1-\frac{\sin(\pi k)}{\pi k}\right)^{1/2}\right\}\right ] , & \text { for odd $ n$ } .
\end{cases}\end{aligned}\ ] ] the parallel transport condition implies @xmath172 where @xmath173 is a constant .
the off - diagonal elements of the gauge connections are @xmath174 where @xmath175 and @xmath176 are assumed , and @xmath177 we will obtain another expression of @xmath178 in appendix [ sec : dfunction ] : @xmath179 where @xmath180 is defined as @xmath181 } , \ ] ] and @xmath182 $ ] is the maximum integer less than @xmath183 .
we evaluate @xmath154 , deforming the integration contour @xmath156 .
the anti - path - ordered exponential in @xmath154 is decomposed into the contributions from the exceptional points . here
we focus on the contribution from the single exceptional point @xmath101 ( @xmath29 ) .
namely , we will evaluate the anti - path - ordered exponential of the gauge connection along the contour @xmath184 , which encircles the exceptional point @xmath185 along the clockwise direction with the radius @xmath80 ( see figure [ fig : epn ] ) .
in particular , we focus on the limit @xmath186 , i.e. , @xmath187 in the following . encircles an exceptional point @xmath101 with a radius @xmath80 ( see the main text ) . the initial point of the cycle @xmath188 is denoted by @xmath189 .
note that @xmath188 intersects a branch cut ( wavy line ) that emanates from @xmath101 .
the bold and dotted lines that connect @xmath40 and @xmath190 are @xmath191 and @xmath192 , respectively .
, scaledwidth=25.0% ] first , we show that the gauge connection @xmath168 is singular at the exceptional point .
assume that @xmath193 is small .
as explained in sec .
[ sec : wavenumber ] , the quasi - momenta @xmath194 and @xmath35 are degenerate at @xmath195 , i.e. , @xmath116 , where @xmath196 for even @xmath21 , and @xmath197 for odd @xmath21 . in appendix
[ sec : ddaroundep ] , we show that @xmath198 .\ ] ] on the other hand , implies @xmath199 ^ 2 - \left[k_{{n_{\mathrm b}}}(g)\right]^2 \right)^{-1 } = \frac{\pi^{1/2}}{2^{5/2 } k^{(n ) } \epsilon^{1/2 } } \left[1+\mathcal{o}(\epsilon^{1/2})\right ] .\ ] ] combining these factors , each of which diverges @xmath200 as @xmath201 , we find @xmath202 .\ ] ] accordingly , the gauge connection within the subspace spaned by @xmath203-th and @xmath21-th eigenstates is @xmath204 = i\frac{d_n}{\epsilon } r \left[1+\mathcal{o}(\epsilon^{1/2})\right ] , \ ] ] where @xmath205 .\ ] ] the leading term is proportional to @xmath206 and single - valued around @xmath195 .
the next leading term exhibits weaker divergence @xmath200 and multiple - valuedness .
other matrix elements of exhibit , at most , the weaker divergence @xmath200 .
we examine @xmath207 by expanding the anti - path - ordered exponential in : @xmath208 , \end{aligned}\ ] ] where @xmath209 indicates the anti - path ordering product of matrices . from the argument above
, the dominant part of the gauge connection has a block - diagonal structure @xmath210 + \mathcal{o}(\epsilon^{-\frac{1}{2 } } ) , \ ] ] where the matrix representation involves only the subspace that consists of @xmath211 with even @xmath21(@xmath212 ) .
since the circumference of @xmath213 is @xmath214 , only the leading term in contributes to @xmath207 : @xmath215 .\ ] ] it is straightforward to see @xmath216 \ ] ] from .
hence we obtain @xmath217 , \ ] ] which describes the parametric evolution of eigenvectors along @xmath213 .
the bound state @xmath218 evolves into @xmath219 .
the partner @xmath219 evolves into @xmath220 , where an extra phase factor @xmath221 is acquired .
other eigenvectors are remain unchanged .
it is straightforward to obtain @xmath222 for an arbitrary @xmath21(@xmath223 ) @xmath224 we depict a few of them : @xmath225 , \qquad m^{(6 ) } = \left [ \begin{array}{cccc|c } 0 & 0 & 0 & -1&0\\ 0 & 1 & 0 & 0&0\\ 0 & 0 & 1 & 0&0\\ 1 & 0 & 0 & 0&0\\ \hline 0 & 0 & 0 & 0 & 1 \end{array } \right ] .\ ] ] let us examine the analytic continuation of eigenvector @xmath226 , where @xmath40 is real and @xmath58 is even .
we extend @xmath226 along the cycle ( see figure [ fig : epn ] ) @xmath227 ( @xmath29 ) , where @xmath228 indicates the concatenation of paths @xmath229 and @xmath230 .
we note that all eigenvectors remain unchanged against the parametric evolution along @xmath231 , because @xmath232 encloses no branch point ( figure [ fig : epn ] ) .
hence , from , @xmath233 has the following expression @xmath234 next , we consider the effect of two exceptional points @xmath151 and @xmath153 using a contour @xmath235 ( see , figures [ fig : c1c2 ] and [ fig : epn ] ) . to carry out this ,
let us extend the eigenvectors in with @xmath236 along @xmath237 .
we find @xmath238 hence we obtain , using with @xmath239 , @xmath240 where @xmath241 furthermore , we examine the analytic continuation of @xmath242 along @xmath243 ( @xmath244 ) . in a similar way
above , we find @xmath245 where @xmath246 as shown in appendix [ sec : m2dots2 m ] , we obtain @xmath247 where the first and second term describe the shift of eigenstates @xmath248 , and the shift of eigenstate @xmath249 with a phase factor @xmath250 .
the other eigenstates remain unchanged by the closed contour @xmath251 .
hence this contour accurately describes the shift of eigenstates up to @xmath252-th excited states . in this sense
, @xmath251 emulates @xmath156 in the limit @xmath253 as for the exotic quantum holonomy induced by @xmath39 along the real axis of @xmath2 , i.e. , @xmath254 as @xmath253 .
first , we compare the present result with ref .
@xcite , where the correspondence between the exotic quantum holonomy and exceptional points is examined in families of quantum kicked spin-@xmath0 .
first of all , because a kicked spin is a periodically driven system , the exotic quantum holonomy of the eigenvalues and the eigenvectors of floquet operator , which is the time evolution operator during the period of a driving force , is investigated .
hence the physical context of the exotic quantum holonomy is slightly different from the one in autonomous systems .
on the other hand , these two models have the same relationship between the quantum holonomy and the exceptional points , as a whole .
for example , the multiple - valuedness of eigenvalues and eigenvectors is governed by the exceptional points .
the non - abelian gauge connection has a @xmath206-divergence around an exceptional point ( see ) , where @xmath80 is the distance from the exceptional point in the parameter space .
this divergence comprises the permutation of eigenvectors against a tiny loop around the exceptional point ( see ) .
however , we find a subtle difference on the analyticity of the gauge connection . as for the kicked spin ,
the gauge connection is single - valued in the parameter space .
we may say that the gauge connection has a degree-@xmath5 pole at an exceptional point of the kicked spin . on the other hand , as for the gauge connection of two - body lied - liniger model ,
an exceptional point is not only a divergent point , but also a branch point .
however the multiple - valuedness appears only in the higher - order correction terms about @xmath80 ( see ) .
second , it is certain that we should see whether the present observations apply to lieb - liniger model with an arbitrary number of particles , as we focus on the two - body case .
we may expect that a similar scenario on the interplay of the exotic quantum holonomy and the exceptional points can be applicable .
for example , according to the strong coupling expansion explained in section [ sec : wavenumber ] , the degree of exceptional points is @xmath142 and each exceptional point connects the ground state and an excited state , regardless of the number of particles @xcite .
on the other hand , however , there remain subtle points .
for example , as for two - body case , the repetitions of the adiabatic cycle @xmath39 and its inverse connect all eigenstates once we specify the total momentum .
we call the collection of such eigenstates a family @xcite . from the present analysis
, a family corresponds to a riemann surface of eigenenergy .
the analytic continuation of the interaction strength @xmath2 can connect any pair of eigenstates in a family .
however , as discovered in ref .
@xcite , there is an infinite number of families in three - body lieb - liniger model . for now , whether or not a family corresponds to a riemann surface of eigenenergy is unknown , because several families might be connected in a region far from the real axis of a riemann surface . in other words ,
the question is open as to whether there is any exceptional point that is `` inaccessible '' by the real cycles .
suppose that there is no such inaccessible exceptional point .
this implies one - to - one correspondence between a family and a riemann surface .
although this might suggest that the exceptional point picture obtained for the two body case is applicable to an arbitrary number of particles , another question is raised .
there is only a single riemann surface for a given total momentum when the number of particles is two .
the number of riemann surfaces , however , is infinite for the number of particle is three .
we do not know how such a proliferation of riemann surfaces against the increment of the number of particles is possible .
we have shown the direct link between the exotic quantum holonomy in eigenenergies and eigenspaces , and the exceptional points , which are degeneracy points in the complexified parameter space in two - body lieb - liniger model . with the help of bethe ansatz
, we examine the riemann surface of quasi - momentum .
all exceptional points in the lower half plane participate the eigenenergy anholonomy .
also the non - abelian gauge connection introduced for the eigenspace anholonomy exhibits divergent behavior around the exceptional point as well as tiny multiple - valuedness correction .
the exceptional points offer building blocks of the eigenspace anholonomy .
it remains to be seen how the current result is to be extended to systems with an arbitrary number of particles .
at wishes to thank satoshi ohya for discussion . this work has been partially supported by the grant - in - aid for scientific research of mext , japan ( grant numbers 22540396 and 21540402 ) .
we show how we obtain . in the following , we assume that @xmath21 is even .
it is straightforward to obtain a similar argument for odd @xmath21 . using the fact that @xmath175 satisfies , we find that the numerator and the denominator of satisfy @xmath256 and @xmath257 respectively . hence we obtain @xmath258 there remains the ambiguity of sign .
it is chosen so as to be consistent with the behavior of in the real axes : @xmath259 hence we obtain in the main text .
we examine the singular behavior of the gauge connection around the exceptional point @xmath101 ( @xmath29 ) .
we assume that @xmath193 is small .
two quasi - momenta @xmath35 and @xmath194 degenerates at @xmath260 , where @xmath261 for even @xmath21 , and @xmath262 for odd @xmath21 . we summarize and @xmath263 we choose that the branch cut emanating from @xmath101
is parallel to the imaginary axis , and is confined within the lower half plane .
namely , we suppose that @xmath264 in the riemann sheet where we are working .
we will examine @xmath265 in order to evaluate @xmath266 around @xmath101 .
we start from the real axis , where @xmath267 holds .
hence we expect @xmath268 holds around the region between the real axis and @xmath101 , as @xmath269 has no singular point there .
we examine @xmath270 in a similar way . as for real @xmath2
, we have @xmath271 for @xmath272 and @xmath273 for @xmath274 .
note that @xmath275 has a branch point at @xmath94 , and the corresponding branch cut locates at the imaginary axis within the upper half plane .
so we choose @xmath276 for @xmath277 and @xmath278 for @xmath274 .
hence we expect that @xmath279 is valid in the region between the real axis and @xmath101 .
we move to the vicinity of the exceptional point @xmath101 , where we obtain @xmath280 , \\ r_{{n_{\mathrm b}}}(g ) & = & -\frac{2^{3/2}}{\pi^{1/2 } } k^{(n ) } \epsilon^{1/2 } \left[1+\mathcal{o}(\epsilon^{1/2})\right ] . \ ] ] note that @xmath281 and @xmath270 are singular at @xmath101 , i.e. , @xmath195 . from @xmath101 , the real axis
is located in the direction @xmath282 , where and are expected to be valid .
we choose @xmath283 so as to satisfy @xmath284 in order to be consistent with . on the other hand
, there remains ambiguity of @xmath285 .
we resolve this using and .
we conclude @xmath286 which implies in the main text .
we prove by induction . note that ,
as for @xmath287 , implies eq . .
hence it suffices to prove for @xmath288 using the assumption that holds for @xmath289(@xmath223 ) . for simplicity , @xmath290 is denoted by @xmath291 . from the recursion relation ,
we have @xmath292 we find from @xmath293 hence , implies @xmath294 \delta_{n'',n ' } .
\nonumber\\\end{aligned}\ ] ] hence the proof of is completed .
10 url # 1#1urlprefix[2][]#2 born m and fock v 1928 _ z. phys . _ * 51 * 165 kato t 1950 _ j. phys .
. japan _ * 5 * 435 longuet - higgins h c 1975 _ proc .
london _ * a 344 * 147 mead c and truhlar d g 1979 _ j. chem .
_ * 70 * 2284 berry m v 1984 _ proc .
. london _ * a 392 * 45 wilczek f and zee a 1984 _ phys .
lett . _ * 52 * 2111 bohm a , mostafazadeh a , koizumi h , niu q and zwanziger z 2003 _ the geometric phase in quantum systems _ ( berlin : springer ) cheon t 1998 _ phys . lett . a _ * 248 * 285 tanaka a and miyamoto m 2007 _ phys
lett . _ * 98 * 160407 kim s w , cheon t and tanaka a 2010 _ phys .
a _ * 374 * 1958 kato t 1980 _ perturbation theory for linear operators _ ( berlin : springer - verlag ) chap ii corrected printing of the second ed heiss w d and steeb w h 1991 _ j. math .
phys . _ * 32 * 3003 heiss w 2004 _ czechoslovak journal of physics _ * 54 * 1091 see , e.g. , bender c , fring a , gnther u and jones h 2012 `` quantum physics with non - hermitian operator '' , _ j. phys . a : math .
* 45 * 440301 lieb e h and liniger w 1963 _ phys .
rev . _ * 130 * 1605 ushveridze a g 1988 _ j. phys .
a. _ * 21 * 955 yonezawa n , tanaka a and cheon t 2013 _ phys .
a _ * 87 * 062113 drr s , garca - ripoll j j , syassen n , bauer d m , lettner m , cirac j i and rempe g 2009 _ phys .
rev . a _ * 79 * 023614 cheon t and tanaka a 2009 _ europhys . lett .
_ * 85 * 20001 tanaka a , cheon t and kim s w 2012 _ j. phys . a : math .
* 45 * 335305 manini n and pistolesi f 2000 _ phys .
lett . _ * 85 * 3067 stone a j 1976 _ proc .
. london _ * a 351 * 141 arnold v i 1991 _ selecta mathematica , new series _ * 1 * 1 | an interplay of an exotic quantum holonomy and exceptional points is examined in one - dimensional bose systems .
the eigenenergy anholonomy , in which hermitian adiabatic cycle induces nontrivial change in eigenenergies , can be interpreted as a manifestation of eigenenergy s riemann surface structure , where the branch points are identified as the exceptional points which are degeneracy points in the complexified parameter space .
it is also shown that the exceptional points are the divergent points of the non - abelian gauge connection for the gauge theoretical formulation of the eigenspace anholonomy .
this helps us to evaluate anti - path - ordered exponentials of the gauge connection to obtain gauge covariant quantities . |
the melting curves of transition metals at pressures up to the megabar region are highly controversial , particularly for b.c.c
diamond anvil cell ( dac ) measurements find that the melting temperature @xmath3 increases by only a few hundred k over the range 1 - 100 gpa @xcite , while shock experiments indicate an increase of several thousand k over this range @xcite .
there have been several _ ab initio _
calculations on transition - metal @xmath3(p ) curves , and the predictions agree more closely with the shock data than with the dac data @xcite .
a challenging case is molybdenum , where there are very large differences between dac and shock data @xcite , and where the shock data reveal _ two _ transitions , the one at high pressure ( @xmath4 380 gpa ) being attributed to melting , and the one at low pressure ( @xmath4 210 gpa ) to a transition from b.c.c . to an unidentified structure @xcite .
we report here on new _ ab initio _ calculations of @xmath3(@xmath5 ) for mo , and on the @xmath1 and @xmath2 relations on the hugoniot .
we also report preliminary information that may help in searching for the unidentified high - p solid phase of mo .
we use density functional theory ( dft ) , which gives very accurate predictions for many properties of transition metals , including hugoniot curves @xcite .
dft molecular dynamics ( m.d . ) was first used to study solid - liquid equilibrium 12 years ago @xcite , and several different techniques are now available for using it to calculate melting curves . in such calculations
, no empirical model is used to describe the interactions between the atoms , but instead the full electronic structure , and hence the total energy and the forces on the atoms , is recalculated at each time step .
there have been earlier dft calculations on the melting of mo , but the techniques used were prone to superheating effects @xcite . in the present work , we use the so - called `` reference coexistence '' techniques @xcite , which does not suffer from this problem .
our work has several aims .
first , we want to improve on the reliability and accuracy of the predicted melting curve of mo obtained from dft ; second , we use dft to predict the @xmath1 and @xmath2 relations on the shock hugoniot ; third , we want to identify the unknown solid phase of mo observed in shock experiments .
our tests on the accuracy of dft for mo , and our extensive calculations of the mo melting curve will be reported in detail elsewhere @xcite , so we present only a summary here . however
, our very recent calculations on the shock hugoniot will be presented in more detail .
these are important , because temperature is very difficult to measure in shock experiments @xcite and dft gives a way of supplying what is missing in the shock data .
our search for the unidentified solid structure of mo is at the exploratory stage , but we present results for phonon frequencies as a function of pressure , which allow us to rule out some possibilities . in the following , we summarise our tests on the accuracy of the dft techniques ( sec .
[ sec : techniques_and_tests ] ) , and outline the reference coexistence technique . in sec .
[ sec : results ] we present our results for the dft melting curve and hugoniot of mo up to 400 gpa , and the study of the phonon frequencies . discussion and conclusions are in sec .
[ sec : discussion ] . .
solid and dashed curves show gga(pbe ) and lda(ca ) fp - lapw results , respectively ; short - dashed and dotted curves show gga(pbe ) and lda(ca ) paw calculations , respectively .
solid dots show experimental data @xcite . ]
our dft calculations are performed mainly with the projector augmented wave ( paw ) implementation @xcite , using the vasp code @xcite , since paw is known to be very accurate . the main uncontrollable approximation in dft is the form adopted for the exchange - correlation functional @xmath6 . to test the accuracy of paw , and the effect of @xmath6 ,
we have compared our predictions for the @xmath1 relation of the b.c.c .
mo crystal against low-@xmath7 experimental results ( fig .
[ fig : wien2kvasp ] ) .
the pressure predictions from paw using the perdew - burke - ernzerhof ( pbe ) @xcite and local - density approximation ( lda ) @xcite forms of @xmath6 deviate by @xmath8 % in opposite directions from the experimental data , but we adopt the pbe form , because the deviations in this case are rather constant .
the pbe results of fig .
[ fig : wien2kvasp ] were obtained with 4s states and below in the core and all other states in the valence set .
inclusion of 4s states in the valence set makes no appreciable difference to the pbe results .
we also tested the paw implementation itself by repeating the @xmath1 calculations with the even more accurate full - potential linearized augmented plane - wave ( fp - lapw ) technique @xcite , using the wien2k code @xcite .
as shown in fig .
[ fig : wien2kvasp ] , paw and fp - lapw results are almost identical . further confirmation for the accuracy of the paw implementation and the pbe functional comes from our comparisons of the calculated phonon dispersion relations for the ambient - pressure b.c.c .
crystal with experiment ( fig .
[ fig : vaspphonon ] ) . the `` reference coexistence '' technique for calculating _
ab initio _ melting curves has been described elsewhere @xcite , but we recall the main steps .
first , an empirical reference model is fitted to dft m.d .
simulations of the solid and liquid at thermodynamic states close to the expected melting curve .
then , the reference model is used to perform simulations on large systems in which solid and liquid coexist , to obtain points on the melting curve of the model . in crucial third stage ,
differences between the reference and dft total energy functions are used to correct the melting properties of the reference model , to obtain the _ ab initio _ melting curve . in the present work ,
the total energy function of the reference model is represented by the embedded - atom model ( eam ) @xcite , consisting of a repulsive inverse - power pair potential , and an embedding term describing the d - band bonding . the detailed procedure for fitting @xmath9 to dft m.d .
data will be reported elsewhere @xcite , but we note that we needed to re - fit the model in different pressure ranges . the simulations on solid and liquid mo in stable thermodynamic coexistence using the fitted reference model employed systems of 6750 atoms , and
we checked that the results do not change if even larger systems are used .
the protocols for preparing these simulated systems , and for achieving and monitoring stable coexistence were similar to those used in earlier work on the melting curve of cu @xcite .
the procedures for correcting the melting curve of the reference model , which depend on calculations of the the free energy differences between the reference and dft systems , have been described and validated in earlier work @xcite .
[ fig : melting ] shows the melting curve of our reference eam model , and the melting curve obtained from this by correcting for the differences between the dft and reference total - energy functions ; earlier _ ab initio_-based calculations of the mo melting curve due to moriarty and to belonoshko _
_ are also indicated @xcite .
we also show points on the melting curve from dac and shock measurements @xcite . the differences between reference and corrected melting curves are only a few hundred k , so that the corrected curve should be very close to the melting curve that would be obtained from the ( pbe ) exchange - correlation functional if no statistical - mechanical approximations were made . because we avoid approximations of earlier _
ab initio_-based work , our present results should be a more accurate representation of the dft melting curve . up to 100 gpa , the differences between our results and earlier dft work are rather small , and we confirm that dft gives a much higher melting slope than that given by dac experiments .
we obtain an accurate value of @xmath10 at @xmath11 by fitting our melting curve to the simon equation @xmath12 , with @xmath13 k , @xmath14 gpa , @xmath15 .
the resulting @xmath11 value of @xmath16 k is close to the accepted experimental value @xmath17 k. our @xmath10 value of @xmath18 k gpa@xmath19 at @xmath11 agrees with an older experimental value of @xmath20 k gpa@xmath19 @xcite .
our dft melting curve is consistent with the point obtained at @xmath21 gpa from shock measurements .
however , we stress that the temperature of the `` experimental '' point was not measured , but estimated by considering superheating corrections to the shock - wave data @xcite . because of this , it is important to try and corroborate the estimated experimental temperature , and this can be done by _ ab initio _ calculations , as we explain next .
_ ab initio _
relation on the hugoniot of mo up to pressure of 400 gpa ; dots reproduce experimental data of refs .
_ right _ : @xmath22 _ ab initio
_ relation on the hugoniot of mo up to pressure of 400 gpa ; dots reproduce experimental data of refs .
@xcite and triangles the same results but corrected for superheating effects as given by ref .
error bars show the uncertainties in pressure and temperature.,title="fig : " ] the pressure @xmath23 , volume @xmath24 and internal energy @xmath25 in the shocked state are related to the initial volume @xmath26 and internal energy @xmath27 by the well - known rankine - hugoniot formula @xcite : @xmath28 since the internal energy and pressure are given in terms of the helmholtz free energy @xmath29 by @xmath30 and @xmath31 , we can calculate the hugoniot from our dft simulations , provided we can calculate @xmath29 as a function of @xmath32 and @xmath7 .
so far , we have done this only for the b.c.c .
crystal in the harmonic approximation , in which @xmath33 . here
, @xmath34 is the free energy of the rigid perfect crystal , including thermal electronic excitations , and @xmath35 is calculated from the phonon frequencies @xmath36 ( @xmath37 the wavevector , @xmath38 the branch ) .
we calculate @xmath35 in the classical limit , in which @xmath39 per atom , with @xmath40 and @xmath41 is the geometric average of phonon frequencies over the brillouin zone .
the methods used for to calculate @xmath42 and the frequencies @xmath36 were similar to those used in our earlier work on fe ( ref .
@xcite ) . for a set of temperatures
, we calculated @xmath34 at a set of volumes , and fitted the volume dependence with a third - order birch - murnaghan equation @xcite .
the temperature dependence of the coefficients in this equation were then fitted with a third - order polynomial .
the phonon frequencies were calculated at 12 volumes in the range @xmath43 @xmath44/atom , as explained elsewhere @xcite .
the volume dependence of the average @xmath41 was then fitted with a third - order polynomial . to obtain @xmath45 and @xmath46 from our fitted free energy , for each value of @xmath24
we seek the @xmath7 at which the rankine - hugoniot equation is satisfied , and from this we obtain @xmath23 . for @xmath26 and @xmath27 , we used values from our gga calculations ; we checked that use of the experimental @xmath26 made no significant difference .
comparison of our calculated @xmath45 with measurements of hixson _ et al .
_ @xcite ( left panel of fig .
[ fig : hugoniot ] ) shows excellent agreement . in the right panel
, we compare our @xmath46 with both uncorrected results of hixson _ et al .
_ and also with results corrected for superheating , and our results confirm their temperature estimates . ) , 77 gpa ( @xmath47 ) , 136 gpa ( @xmath48 ) , 274 gpa ( @xmath49 ) and 400 gpa ( @xmath50 ) . ]
efforts have been made to identify the high-@xmath5/high-@xmath7 structure of mo indicated by shock experiments @xcite .
_ used their theoretical prediction of b.c.c . @xmath51
phase transition at @xmath52 gpa and @xmath53 k to suggest the h.c.p structure . however , later calculations locate this transition at higher pressures ( @xmath54 gpa ) , and temperature stabilisation of h.c.p . over b.c.c . below
melting seems improbable .
furthermore , there are recent claims that under pressure b.c.c mo transforms first to f.c.c . rather than h.c.p .
very recently , belonoshko _ et al . _
@xcite reported _ ab initio _
simulations on the f.c.c and a15 structures at temperatures and pressures near those of interest ( namely , @xmath55 gpa and @xmath56 k ) , and concluded that both structures are unlikely high-@xmath7 phases of mo .
we have performed _ ab initio _ calculations similar to those of belonoshko _ et al .
_ on the h.c.p . and @xmath57 structures ( @xmath58 gpa and @xmath59 k )
we find that temperature neither favours any of them respect to b.c.c .
, thus they must be excluded as well .
recently , a 2nd - order phase transition from cubic to rhombohedral has been observed in vanadium at @xmath60 gpa @xcite .
this appears to be related to an earlier _ ab initio _ prediction of a phonon softening in v @xcite .
this suggests that a similar structural transition might occur in mo .
we have used our calculated phonon dispersion relations of mo over the range @xmath0 gpa to test this .
[ fig : freqcomp ] depicts our results at 0 , 77 , 136 , 274 and 400 gpa , but we see that no phonon anomaly like that reported for v is observed at any wavevector .
this indicates that we should rule out structures based on small distortion of b.c.c .
an important outcome of the present work is improved dft calculations of the melting curve of mo over the pressure range @xmath0 gpa . in particular ,
our techniques allow us to avoid the superheating errors which appear to affect an independent recent dft study of mo melting @xcite .
the accuracy of our calculations is confirmed by the very close agreement with experiment for the melting temperature @xmath3 and the melting slope @xmath10 at ambient pressure .
the results fully confirm that the increase of @xmath3 by @xmath61 k over the range @xmath62 gpa predicted by dft is about 10 greater than that deduced from dac measurements .
a second important outcome is that our calculations of the temperature along the shock hugoniot support earlier temperature estimates based on experimental data but corrected for possible superheating effects @xcite .
this allows us to compare more confidently the point on the melting curve at @xmath63 gpa derived from shock data with our predicted melting curve , and we confirm that @xmath3 at this pressure is _ ca .
_ 8650 k. this is far above any reasonable extrapolation of the dac data . concerning the search for the unknown crystal structure of mo indicated by shock experiments to exist above _ ca . _
220 gpa , we have been able so far only to rule out some possibilities .
our calculations of the phonon dispersion realations in the b.c.c .
structure over the range @xmath0 gpa reveal no softening of any phonons , and no indication of any elastic instability .
this means that the new crystal structure does not arise from small distortion of b.c.c .
this is interesting in the light of the recent discovery of the elastic instability of b.c.c .
v above @xmath64 gpa , predicted initially by dft , and observed very recently in x - ray diffraction experiments @xcite .
it seems that the structural transition in mo is of a different kind .
the large conflict between the melting curve of mo derived from dac measurements on one side and from shock experiments and dft calculations on the other side must be due either to a misinterpretation of the dac data or to a combination of serious dft errors and misinterpretation of shock data . given the accuracy of dft that we have been able to demonstrate ( low - temperature @xmath1 curve , hugoniot @xmath1 curve , ambient-@xmath5 phonons ) , we think there is little evidence for significant errors in dft , which is also in good accord with shock data .
a possible explanation might be that the large temperature gradients and non - hydrostatic stress in dac experiments might give rise to flow of material giving the appearance of melting , even well below the thermodynamic melting temperature .
we also note recent evidence that temperature measurement in dac experiments may be subject to previously unsuspected errors @xcite , though probably not of the size needed to resolve the conflict by themselves .
the work was supported by epsrc grant ep / c534360 , which is 50% funded by dstl(mod ) , and by nerc grant ne / c51889x/1 . the work was conducted as part of a euryi scheme award to da as provided by epsrc ( see www.esf.org/euryi ) . | we report _ ab initio _ calculations of the melting curve and hugoniot of molybdenum for the pressure range @xmath0 gpa , using density functional theory ( dft ) in the projector augmented wave ( paw ) implementation .
we use the `` reference coexistence '' technique to overcome uncertainties inherent in earlier dft calculations of the melting curve of mo .
our calculated melting curve agrees well with experiment at ambient pressure and is consistent with shock data at high pressure , but does not agree with the high pressure melting curve from static compression experiments .
our calculated @xmath1 and @xmath2 hugoniot relations agree well with shock measurements .
we use calculations of phonon dispersion relations as a function of pressure to eliminate some possible interpretations of the solid - solid phase transition observed in shock experiments on mo . |
non - perturbative investigation of the qcd dynamics in the low energy region by means of the effective lagrangian approach has made considerable progress recently . in the very low energy region ( @xmath2 ) the chiral perturbation theory ( @xmath3pt ) @xcite based on the spontaneously broken chiral symmetry @xmath4 grew into a very successful model - independent tool for description of the green functions ( gf ) of quark currents and related low - energy phenomenology .
@xmath3pt is organized as a rigorously defined simultaneous perturbative expansion in small momenta and light quark masses .
recent calculations are performed at the next - to - next - to - leading order @xmath5 @xcite . in the intermediate energy region ( @xmath6 ) , however , the situation is less satisfactory .
the set of relevant degrees of freedom includes now the low lying resonances and because there is no mass gap existing in the spectrum , the effective theory in this region can not be constructed as a straightforward extension of the @xmath3pt low energy expansion . on the other hand , the considerations based on the large @xmath7 expansions together with high - energy constraints derived from perturbative qcd and operator product expansion ( ope ) allow to introduce another type of effective lagrangian description , corresponding to the leading order in @xmath8 and reflecting the basic features of qcd in the @xmath9 limit .
namely , the spectrum consisting of an infinite tower of free stable mesonic resonances exchanged in each channel requires infinite number of resonance fields in the @xmath10 symmetric lagrangian with interaction vertices suppressed by an appropriate power of @xmath11and ( since the @xmath12 expansion is correlated with semiclassical expansion ) only tree graphs have to be taken into account in the leading order . an approximation to this general picture consisting in limiting the number of resonance field to one in each channel and matching the resulting theory in the high energy region with ope is known as resonance chiral theory ( r@xmath3 t ) @xcite .
integrating out the resonance fields from the lagrangian of r@xmath3 t in the low energy region and subsequent matching with @xmath3pt has become a very successful tool for estimates of the resonance contribution to the values of the @xmath13 @xcite and @xmath5 @xcite low energy constants ( lec ) entering the @xmath14pt lagrangian .
though the usual chiral power counting fails within r@xmath3 t due to the presence of an additional heavy scale ( the mass of the resonances ) and the usual weinberg formula @xcite can not be generalized here ( because of the lack of a scale playing the role analogous to @xmath15 ) , it seems to be fully legitimate to go beyond the tree level r@xmath3 t and calculate the loops @xcite .
being suppressed by one power of @xmath8 , the loops allow to encompass such nlo effects in the @xmath8 expansion as resonance widths and final state interaction and to determine the nlo resonance contribution to lec ( and their running with the renormalization scale ) .
however , we have to be ready for both technical and conceptual complications connected with renormalization of the effective theory for which no natural organization of the expansion ( other than the @xmath8 counting ) exists .
especially , because there is no natural analog of the weinberg power counting in r@xmath3 t , we can expect mixing of the naive chiral orders in the process of the renormalization ( _ e.g _ the loops renormalize the @xmath16 lec and also counterterms of unusually high chiral orders are needed ) .
also , lack of appropriate protective symmetry can bring about appearance of new poles in the gf corresponding to new degrees of freedom which are frozen at the tree level .
the latter might be felt as a pathological artefact of the not carefully enough formulated theory , particularly because this extra poles might be negative norm ghost or tachyons @xcite . on the other hand , however , we could also try to take advantage of this feature and adjust the poles in such a way that they correspond to the well established resonance states . in the following
we will illustrate these problems in more detail . as an explicit example we use the one - loop renormalization of the propagator corresponding to the antisymmetric tensor field which originally describes the @xmath0 vector resonance ( @xmath17 meson ) at the tree level .
we will show that the loop corrections to the propagator could lead to the dynamical generation of various types of @xmath0 and @xmath18 states and that the appropriate adjustment of coupling constants allows us to generate in this way the one which could be identified with the @xmath19 meson .
the details of the calculations and further discussion will be provided in @xcite .
the @xmath0 resonance part of the r@xmath3 t lagrangian within the antisymmetric tensor formalism reads @xcite @xmath20 where @xmath21 with normalized @xmath22 generators @xmath23 , @xmath24 are antisymmetric tensor fields with appropriate quantum numbers and @xmath25 is the usual chiral covariant derivative .
@xmath26 is the interaction lagrangian which will be specified later .
the full antisymmetric tensor field propagator @xmath27 has in general the following tensor structure @xmath28 where @xmath29 are longitudinal and transverse projectors ( @xmath30 ) @xmath31 note that , for @xmath32 we can express @xmath33 as the polarization sums @xmath34 where @xmath35 and @xmath36 are the usual spin - one polarization vectors with mass @xmath37 .
the possible poles @xmath38 of @xmath39 and @xmath40 correspond therefore both to the spin - one states which couple to the fields @xmath41 and @xmath42 respectively and have therefore the same quantum numbers up to the parity . at lo in the @xmath8 expansion
the lagrangian ( [ lagrangian ] ) gives @xmath43 and @xmath44 so that the @xmath0 resonance multiplet appears as a pole in @xmath45 . because there is no pole in @xmath40 at this order , no additional @xmath46 state is propagated . beyond lo we get generally @xmath47 where the self - energies @xmath48 are of the order @xmath8 at least . in the next section we present the results of the calculation of the renormalized self - energies @xmath49 in the chiral limit at nlo for a concrete form of the interaction lagrangian @xmath26
a more systematic treatment will be given in @xcite .
in what follows we limit ourselves to the interaction lagrangian @xmath50 with at most two derivatives and up to two resonance fields .
writing explicitly only those terms that contribute to the one - loop self - energies we have @xcite , @xcite @xmath51\rangle + 2d_1\langle d_\beta u^\sigma \{\widetilde{r}_{\alpha \sigma } , r^{\alpha \beta } \}\rangle\ ] ] @xmath52 where @xmath53 . in the large @xmath7 limit
the couplings are @xmath54 and @xmath55 and apparently the intrinsic parity odd part is of higher order .
however , the trilinear vertices contributing to the one - loop self - energies are @xmath56 in both cases due to the appropriate power of @xmath57 accompanying @xmath58 .
therefore the operators with two resonance fields can not be eliminated using the large @xmath7 arguments .
also nonzero @xmath59 are required in order to satisfy the ope constraints for vvp gf at the lo @xcite . in order to cancel the infinite part of the one - loop self - energies we have to introduce a set of counterterms . because the interaction terms are @xmath60 we would expect ( by the analogy with @xmath3pt power counting ) these counterterms to have four derivatives at most . however , the nontrivial structure of the free resonance propagator ( namely the presence of the @xmath61 @xmath62 part ) results in the failure of this naive expectation .
in fact we need counterterms with up to six derivatives , namely @xmath63 the complete list of the counterterms and their infinite parts is postponed to @xcite .
let us only note that @xmath64 contains a new type of kinetic term @xmath65 .
provided such a term was included in the lo lagrangian from the very beginning , the propagator would have an additional pole in @xmath66 .
however , interpretation of such a pole as a @xmath1 state would be problematic .
according to the sign of @xmath67 this state would be either a tachyon or a negative norm ghost @xcite .
evaluating the one - loop feynman graphs and adding the polynomial contributions from the counterterms ( [ ct ] ) we get the @xmath3pt minimally subtracted self - energies @xmath68 .
the equation for the poles of @xmath45 has then an approximative perturbative solution @xmath69 corresponding to the original @xmath0 vector resonance with lo mass @xmath70 , which develops a mass correction and a finite width of the order @xmath71 due to the loops and which we identify with the @xmath72 meson .
this allows to re - parameterize perturbatively @xmath73 in terms of @xmath74 and @xmath75 and requiring further @xmath76 we get for @xmath77 @xmath78 @xcite @xmath79 \\ & & -\frac{40}9\left ( \frac{m_\rho } { 4\pi f_\pi } \right ) ^2d_3 ^ 2(x^2 - 1)^2% \widehat{j}(x ) \\ { \sigma } _
t^r(x ) & = & \frac 1\pi \frac{\gamma _ \rho } { m_\rho } % \sum_{i=0}^3b_ix^i+\frac{20}9\left ( \frac{m_\rho } { 4\pi f_\pi } \right ) ^2d_3 ^ 2\end{aligned}\ ] ] @xmath80 here we put for further numerical estimates @xmath81 , @xmath82 ( for @xmath83 we take the value from @xcite ) and we have introduced the re - scaled free parameters @xmath84 and @xmath85 with natural size @xmath86 in the large @xmath7 expansion . these can be expressed in terms of renormalization scale independent combinations of the renormalized counterterms couplings and @xmath87logs .
the loop functions @xmath88 and @xmath89 are given on the first sheet as @xmath90 , \label{loop_functions}\end{aligned}\ ] ] where we take the principal branch of the logarithm ( @xmath91 ) with a cut for @xmath92 .
on the second sheet we have then @xmath93 and similarly for @xmath94 @xmath95 .
the equation for the poles of @xmath40 has only non - perturbative solutions of the order @xmath96 .
the @xmath1 states corresponding to them therefore decouple in the @xmath9 limit , however , for physical values of @xmath74 , @xmath97 and @xmath98 ( and for reasonable @xmath86 values of the parameters @xmath85 and @xmath99 ) , the position of poles can lie well within the intermediate energy region we are interested in . the nature of the corresponding states , which is also controlled by the free parameters @xmath85 and @xmath99 , is rich and covers bound states @xmath100 , ( which also might be negative norm ghosts ) , tachyonic poles @xmath101 , resonance poles in the lower complex half - plane on the second sheet or even complex conjugated pair of lee - wick poles on the first sheet .
it is not straightforward to formulate general conditions for @xmath102 and @xmath99 under which there is _
no _ pole in @xmath103 at all , on the other hand we can rather easily arrange them to obtain the pole corresponding _
e.g. _ to the @xmath19 meson . this can be achieved in many ways _
e.g. _ for the choice @xmath104 , @xmath105 , @xmath106 and @xmath107 .
the plot of the denominator of @xmath108 for this particular choice for the section @xmath109 on the second sheet and the shape of @xmath110 on the first sheet for @xmath111 real are depicted in fig [ figo ] .
= 6.5 cm ( -158,18 )
we have illustrated the problems connected with loop calculations within r@xmath112 t using the one loop renormalization of the propagator of antisymmetric tensor field which describes @xmath0 resonance multiplet at the leading order of the large @xmath7 expansion as a concrete example .
we have found that new @xmath1 states of various nature ( including pathological ones like negative norm ghosts and tachyons ) can be dynamically generated .
for example , for a wide range of parameters @xmath85 , one such a state can be identified with @xmath19 meson . further discussion of possible applications of this feature is postponed to @xcite . | we discuss the renormalization of the @xmath0 vector meson propagator within resonance chiral theory at one loop . using the particular form of the interaction lagrangian
we show that additional poles of the renormalized propagator corresponding to @xmath1 degrees of freedom can be generated .
we give a concrete example of such an effect . |
dwarf novae are a subclass of cataclysmic variables undergoing outbursts lasting ( at least ) a few days during which their brightness increases by several magnitudes ( see e.g. * ? ? ?
* for a review ) .
these outbursts are widely believed to be due to a thermal / viscous disc instability ( see * ? ? ? * for a review ) .
the weakest point of this model aside from the assumption that angular momentum transport is due to viscosity ( i.e. is a local phenomenon accompanied by energy dissipation ) described by the @xmath0-prescription , is the approximate treatment of 2d effects at the disc edge .
there is in particular a debate about the outcome of the disc reaching the radius at which the 3:1 resonance occurs ( this may happen for low secondary to primary mass - ratios ) .
sph models ( see e.g. * ? ? ?
* ) treat accurately the dynamics of the disc and are in principle quite appropriate to deal with these effects ; they predict that when the 3:1 resonance radius is reached , the disc becomes tidally unstable , eccentric and precesses .
sph models are , however , limited in that it is difficult to include in them detailed microphysics ; the energy equation is often replaced by isothermal or adiabatic approximations ; this significantly affects the results as showed by @xcite .
nevertheless , the tidal instability , coupled with the standard thermal instability is also believed @xcite to be the reason for the long duration of superoutbursts during which superhumps are seen .
another point of debate is the amplitude of the torque @xmath1 due to the tidal forces , even far from the resonance . in a steady state ,
the radius of the accretion disc is determined by @xmath2 .
one often uses the prescription of @xcite , derived from the linear analysis of @xcite confirmed by numerical simulations by @xcite ( at least for radii not too close to the tidal truncation radius @xmath3 , at which the trajectories of test particle orbiting around the white dwarf intersect , leading to high dissipation * ? ? ?
this torque reads : @xmath4 where @xmath5 is the orbital period , @xmath6 the orbital separation and @xmath7 a numerical constant .
@xmath8 and @xmath9 are respectively the kinematic viscosity coefficient and the surface density .
the index @xmath10 was found to be close to 5 . in the particular case of u gem
, @xcite obtained a value for @xmath7 smaller by almost two orders of magnitude than that required for the disc to be truncated at the tidal radius in steady state .
they concluded that the tidal torque is very small everywhere except very close to the tidal radius @xmath3 where @xmath2 diverges .
if true , the tidal removal of the angular momentum would occur only at the disk s outer edge within a negligibly small radial extent @xcite .
we consider here various aspects of outer disc radius variations .
we first concentrate on disc sizes observed in eclipsing vy scl systems .
( see * ? ? ?
we then turn to su uma stars and soft x - ray transients ( sxts ) and examine for these systems the consequences of assuming that superhumps and superoutbursts are due to a tidal instability .
vy scl stars are a subgroup of cataclysmic variables which are usually in a bright state , and have occasionally low states during which their luminosity drops by more than one magnitude , bringing them into the dwarf - nova instability strip . yet , they do not have dwarf - nova outburst , even though the decline can be very gradual and prolonged , longer than the disc s viscous time .
it had been suggested that the apparent stability of these systems could be due to the irradiation of the inner disc edge @xcite , but @xcite showed that outbursts are unavoidable unless the disc disappears completely during quiescence , most probably truncated by the white - dwarf s magnetic field .
this translates into a requirement on the magnetic field strength that must be sufficient for the magnetospheric radius to be roughly equal to the circularization radius when the accretion rate is just equal to the critical rate below which instabilities would appear .
recently , @xcite observed the eclipsing vy scl star dw uma in a state intermediate between minimum and maximum .
eclipse mapping techniques allowing to reconstruct the disc luminosity profile , they found that the luminosity difference between the high and intermediate states is almost entirely due to a change in the accretion disc radius , from @xmath11 0.5 to @xmath12 times @xmath13 , the distance from the white dwarf to @xmath14 .
it is observed that in the intermediate state , the disc is entirely eclipsed by the secondary , while its outer parts are visible during the high state .
a possible explanation of this could be that the disc is not in equilibrium ( i.e. @xmath15 is not constant within the disc ) , so that radius variations are a consequence of mass redistribution inside the disc .
this , however , can not be the case , since , as noted by @xcite , the recovery from the low state takes @xmath11 4 months ; the disc has enough time to readjust its structure to changes of the mass - transfer rate .
the disc should then be quasi steady , and its radius close to the tidal radius if tidal torques were exerted only in very small annulus at @xmath16 .
observations show , however , a disc radius @xmath17 smaller than @xmath3 @xcite .
we used our disc - instability model code @xcite to simulate the accretion - disc properties of dw uma .
we used eq .
( [ torque ] ) with a large @xmath7 .
figure [ fig : evoldw ] shows the evolution of a system with parameters similar to those of dw uma , in which the white dwarf is magnetized strongly enough to prevent dwarf nova outbursts during intermediate and low states .
there are still some small oscillations left but these could be suppressed if the field were slightly stronger , and are probably not detectable .
we have assumed a slow decrease of the mass transfer rate @xmath18 until the disc almost disappears , followed by a rapid ( essentially for numerical reasons ) increase of @xmath18 .
the second panel shows that the disc is always close to equilibrium ( @xmath19 ) .
it can be seen that significant variations of the outer disc radius are obtained .
when the disc is still fainter by 1 magnitude than the maximum , the disc is 20% smaller than its maximal extension .
this is not quite the 50% variations that are claimed by @xcite , but is acceptable in view of the fact that we define the disc outer edge as the place where the surface density vanishes , whereas @xcite use a photometric definition .
these results show that the tidal coupling can not be negligible in the intermediate state , even at radii quite a bit smaller than the tidal truncation radius .
as mentioned earlier , the popular explanation for superoutbursts combines the thermal instability with a tidal instability that is supposed to arise when the disc reaches the 3:1 resonance radius ( see e.g. * ? ? ?
at this point , the disc would become eccentric and precess , allegedly causing the superhumps .
the tidal torque @xmath2 is supposed to increase by at least one order of magnitude , resulting in a corresponding enhancement of dissipation and angular momentum transport . according to this scenario the disc shrinks until the radius has decreased to an ( arbitrarily chosen ) critical value of the order of 0.35 times the orbital separation , and the superoutburst stops .
a sequence of several normal outbursts , during which the disc grows on average then follows until the next superoutburst .
this model is essentially based on sph simulations indicating that the disc does become eccentric and precesses when the 3:1 resonance radius is reached @xcite , and on the fact that su uma systems are found below the period gap , for systems in which the mass ratio @xmath20 is less than 1/3 the condition for the 3:1 resonance radius to be smaller than the tidal truncation radius . according to simulations by @xcite the critical value of @xmath20
above which no superoutburst occur is in fact rather 1/4 . however , two elements seem to have seriously put in doubt the viability of the thermal - tidal model of super- humps and outbursts .
first , @xcite found that the outcome of numerical simulations depends on the equation of state that has been used ; they found that eulerian models gave the same results as the sph codes in the isothermal approximation , but found very different results ( no superhumps ) when using the full thermal equation .
more recently @xcite discovered superhumps in the famous u gem 1985superoutburst .
the component masses of this prototypical binary are rather well constrained giving a rather large value of @xmath21 .
clearly the tidal instability can not apply to this system .
in addition , a permanent superhump was detected in tv col @xcite a binary with an estimated mass - ratio between @xmath22 and 0.93 @xcite , consistent with its 5.39 hour orbital period .
one could try to save the tidal model by arguing that u gem and tv col superhumps are phenomena different from those observed in su uma stars but such an argument would have to explain why , as remarked by @xcite , when plotted on superhump excess - period vs orbital period ( logarithmic ) plane , u gem and tv col fall on the linear extension of the relations defined by shorter period dwarf novae and permanent superhumpers ( see * ? ? ?
20 ) . clearly a new model for su uma stars is needed .
irradiation and enhanced mass transfer probably play a significant role @xcite , but one must find an explanation for the superhump and superoutburst phenomena keeping in mind that none of them seems to require the action of tidal forces .
in sxts , the mass ratio is usually very small , and one would expect to find superhumps in these systems , if indeed the explanation of this effect is related to the tidal instability . indeed there is observational evidence of modulations at periods slightly longer than the orbital period ( * ? ? ? * see also charles , these proceedings ) .
there are , however , several important differences between sxts and su uma s : first , because the mass ratios are extremely small , the disc should remain permanently eccentric , and second , the outer parts of the disc can remain unaffected by the thermal instability in systems where the disc is large .
finally , in sxts superhumps would arise from a modulation of the reprocessed flux by the changing disc area and not from an increase of viscous dissipation as in su uma stars @xcite .
as an example , for @xmath23 , the circularization radius is @xmath24 , where @xmath6 is the orbital separation , larger than the critical radius below which the tidal instability can not be maintained ( typically @xmath25 , see e.g. * ? ? ?
the disc can therefore never shrink enough for the tidal instability to stop , and it should remain in a permanently eccentric state .
this conclusion holds for values of @xmath20 below 0.10 , i.e. for most sxts .
therefore the tidal instability can not be the cause of sxt outbursts . for systems with periods
@xmath5 less than @xmath26 day , such as a 0620 - 00 , the whole disc is affected by the outburst .
one therefore expects that the superhump modulation , if related to the 3:1 tidal instability , should be visible in these systems .
systems for which a superorbital modulation has been claimed ( xte j1118 + 480 , nova musca 1991 , gro j0422 + 32 and gs 2000 + 25 ; * ? ? ? * ; * ? ? ?
* ) all have @xmath27 d. however , the disc is _ always _ larger than the 3:1 resonance radius as confirmed by the presence of the superhump in quiescence @xcite , hence outbursts in sxts can not be due to tidal interactions .
interestingly superhumps in quiescence have been also observed in su uma stars @xcite . in systems with long orbital periods
, the outer disc will not be affected by the outburst , except possibly by illumination effects .
if the tidal instability sets in , it will therefore always remain present , and superhumps are expected to be present both in quiescence or in outburst .
it turns out however that in quiescence , the outer disc is too cold to radiate efficiently even in the infrared : most of the luminosity originates from the central regions , despite the reduced emitting area .
figure [ fig : sxt ] shows the structure of a disc in a system with the orbital parameters of black - hole x - ray transient v404 cyg ( @xmath28 : 12 m@xmath29 ; @xmath30 : 0.7 m@xmath29 ; p : 78 hr ) ; the mass transfer rate was taken to be @xmath31 g s@xmath32 . in quiescence ,
the effective temperature is of order of 1000 k , and less than 0.1 % of the total luminosity is emitted by the outer parts of the disc ; any modulation of the light emitted by these cool regions would therefore be undetectable . in outburst
, the disc heats up as a result of irradiation from the primary ( we assumed here the same prescription for irradiation as in @xcite . a significant fraction of the disc luminosity in the infrared band should therefore be emitted by regions of the disc possibly affected by the tidal instability
however , the opacity is minimum for temperatures of the order of 3,000 k , typical in these regions that are affected by illumination , and it turns out that the optical depth is very small . as a consequence , the assumption of blackbody emission does no longer hold , and the colour temperature will be significantly higher than the effective temperature ; light will be shifted in the optical or blue band , and will again be diluted in the total light emitted by the inner disc .
we therefore do not expect to detect any superorbital modulation in these systems , in the hypothesis that it would originate from the outer parts of the accretion disc ( whether it is caused by a tidal instability or not ) .
the effects of the companion on the accretion disc are still unclear . even in the simple case where no strong resonance is present ,
observations of intermediate states of vy scl stars indicate that the tidal torques must be strong enough to truncate the disc at radii much smaller than the so called tidal truncation radius .
the presence of superhumps during the superoutburst of u gem casts doubt on the validity of the tidal - thermal model of these phenomena .
this conclusion is strengthened by theory and observations of superhumps in tv col .
if we make the assumption that superhumps are in some way related to the 3:1 resonance , we predict that superhumps should be observed in short period sxts in quiescence and in outburst , but not in long period systems .
the outburst mechanism should then be unrelated to this resonance , and the tidal - thermal instability model does not apply to these systems . | outer radius variations play an important role disc structure and evolution .
we consider theoretical and observational consequences of such variations in cataclysmic binaries and low - mass x - ray binaries .
we find that the action of tidal torques must be important well inside the tidal radius .
we also conclude that it is doubtful that the tidal - thermal instability is responsible for the superoutburst / superhump phenomena . , accretion , accretion discs , binaries : close , cataclysmic variables , x - ray binaries 97.10.gz , 97.30.qt , 97.80.gm , 97.80.jp |
linguistic typology aims to distinguish between logically possible languages and actually observed languages .
a fundamental building block for such an understanding is the _ universal implication _
these are short statements that restrict the space of languages in a concrete way ( for instance `` object - verb ordering implies adjective - noun ordering '' ) ; , and provide excellent introductions to linguistic typology .
we present a statistical model for automatically discovering such implications from a large typological database @xcite .
analyses of universal implications are typically performed by linguists , inspecting an array of @xmath0-@xmath1 languages and a few pairs of features .
looking at all pairs of features ( typically several hundred ) is virtually impossible by hand .
moreover , it is insufficient to simply look at counts .
for instance , results presented in the form `` verb precedes object implies prepositions in 16/19 languages '' are nonconclusive . while compelling , this is not enough evidence to decide if this is a statistically well - founded implication . for one ,
maybe @xmath2 of languages have prepositions : then the fact that we ve achieved a rate of @xmath3 actually seems really bad .
moreover , if the @xmath4 languages are highly related historically or areally ( geographically ) , and the other @xmath5 are not , then we may have only learned something about geography . in this work ,
we propose a statistical model that deals cleanly with these difficulties . by building a computational model ,
it is possible to apply it to a very large typological database and search over many thousands of pairs of features .
our model hinges on two novel components : a statistical noise model a hierarchical inference over language families . to our knowledge
, there is no prior work directly in this area . the closest work is represented by the books _ possible and probable languages _
@xcite and _ language classification by numbers _
@xcite , but the focus of these books is on automatically discovering phylogenetic trees for languages based on indo - european cognate sets @xcite .
[ sec : data ] [ fig : world ] [ cols="<,^,^,^,^,^,^ " , ]
we have presented a bayesian model for discovering universal linguistic implications from a typological database .
our model is able to account for noise in a linguistically plausible manner .
our hierarchical models deal with the sampling issue in a unique way , by using prior knowledge about language families to `` group '' related languages .
quantitatively , the hierarchical information turns out to be quite useful , regardless of whether it is phylogenetically- or areally - based .
qualitatively , our model can recover many well - known implications as well as many more potential implications that can be the object of future linguistic study .
we believe that our model is sufficiently general that it could be applied to many different typological databases we attempted not to `` overfit '' it to wals .
our hope is that the automatic discovery of such implications not only aid typologically - inclined linguists , but also other groups .
for instance , well - attested universal implications have the potential to reduce the amount of data field linguists need to collect .
they have also been used computationally to aid in the learning of unsupervised part of speech taggers @xcite .
many extensions are possible to this model ; for instance attempting to uncover typologically hierarchies and other higher - order structures .
we have made the full output of all models available at http://hal3.name/wals . | a standard form of analysis for linguistic typology is the universal implication .
these implications state facts about the range of extant languages , such as `` if objects come after verbs , then adjectives come after nouns . ''
such implications are typically discovered by painstaking hand analysis over a small sample of languages .
we propose a computational model for assisting at this process .
our model is able to discover both well - known implications as well as some novel implications that deserve further study .
moreover , through a careful application of hierarchical analysis , we are able to cope with the well - known sampling problem : languages are not independent . |
non - locality is in the heart of quantum world . from bell
s theorem@xcite , it is understood that no local hidden variable theory could replace quantum theory as a theory of physical world .
the existence of entangled states in composite quantum systems assures the non - local behavior of quantum theory .
generally , violation of bell s inequality is taken as the signature of non - locality .
entangled states play an important role to show the violation of bell s inequality .
however non - locality can be viewed from other perspectives .
for example , there exist sets of locally indistinguishable orthogonal pure product states@xcite , used to show non - locality without entanglement .
recently introduced measurement induced non - locality@xcite ( in short , min ) , is one of the ways to detect non - locality in quantum states by some locally invariant measurements .
locally invariant measurements can not affect global states in classical theory but this is possible in quantum theory .
so min is a type of non - local correlation which can only exists in quantum domain .
min is defined in such a way that it is non - negative for all states , invariant under local unitary and vanishes on product state . in this sense , min could be observed as a type of non - local correlation which is induced by certain measurement . although it is induced by some measurement ,
it is not a measure of quantum correlation in true sense .
but it can be treated as a measure of non - locality induced by some kind of measurement .
now a natural question arises about the shareability of quantum correlations in multipartite states . it may be monogamous or
may be polygamous .
it is known that entanglement is monogamous@xcite . in this work
we will investigate the monogamous behavior of min .
we will check the monogamy property for some classes of states in both three and four - qubit system . unlike discord@xcite
, here we will show that both the three qubit generalized ghz and w states satisfy monogamy relations , however there is violations of monogamy relation if we consider the generic whole class of pure three qubit states . in the case of pure four - qubit system
we consider the most important generic class of states .
it contains usual ghz , maximally entangled states in the sense of gour et.al .
most of the states satisfy monogamy relations .
there are two important subclasses of the generic class , say , @xmath0 and @xmath1 . in one subclass monogamy
holds but in another it does not . in particular , ghz state violates monogamy relation and w satisfies it .
therefore , monogamous relations of min are quite different to that of some important measures of correlation@xcite .
our paper is organized as follows - in the section ii we will review some of the basic properties of min . in section iii we will explain and discuss about some four - qubit classes of states which we will require in our work .
section iv is devoted to the notion of monogamy for min .
section v contains results on pure three qubit systems and section vi contains results on four qubit system .
consider @xmath2 be any bipartite state shared between two parties a and b. then min ( denoted by @xmath3 ) is defined by , @xmath4 where maximum is over all von neumann measurements @xmath5 which do not disturb @xmath6 the local density matrix of a , i.e. , @xmath7 and @xmath8 is taken as hilbert schmidt norm ( i.e. @xmath9^{\frac{1}{2}}$ ] ) .
it is in some sense , dual to that of geometric measure of discord .
physically , min quantifies the global effect caused by locally invariant measurements .
min has application in general dense coding , quantum state steering etc .
min vanishes for product state and remains positive for entangled states . for pure states min reduces to linear entropy like geometric discord@xcite .
explicit formula of min for @xmath10 system , @xmath11 ( if @xmath12 is non - degenerate ) system and an explicit upper bound for @xmath11 system were obtained by luo , fu in @xcite .
later mirafzali _ et.al .
_ @xcite formulate a way to reduce the problem of degeneracy in @xmath11 system and evaluate it for @xmath13 dimensional system .
min is invariant under local unitary , i.e. , in true sense , it is a non - local correlation measure .
the set of all zero min states is non - convex .
guo , hou @xcite derived the conditions for the nullity of min .
they have found that set of states with zero min is a proper subset of the set of all classical quantum states , i.e. , zero discord states .
min for classical quantum state vanishes if each eigen - subspace of @xmath12 is one dimensional .
it therefore reveals that non - commutativity is the cause of these kind of non - locality in quantum states .
recently , in @xcite , min have been quantified in terms of relative entropy to give it another physical interpretation .
however , in our work we have used the older definition of min .
suppose @xmath14 , @xmath15 be the hilbert spaces associated with the parties a and b respectively and @xmath16 , @xmath17 denote the hilbert space of linear operators acting on @xmath14 , @xmath15 with the inner product defined by @xmath18 .
we state two important results which we will use in our work .
+ * theorem 1*:(luo , fu @xcite ) _ let @xmath19 be any bipartite pure state with schmidt decomposition @xmath20 then @xmath21_. + + * theorem 2*:(luo , fu @xcite ) _ if @xmath22 be any state of @xmath23 dimensional system written in the form @xmath24 where @xmath25 i=0,1,2,3 } and @xmath26 are the orthonormal hermitian operator bases for @xmath16 and @xmath17 respectively with @xmath27 .
then @xmath28 where the matrix @xmath29 with @xmath30 being minimum eigenvalue of @xmath31 and @xmath32 with @xmath33_. + before going to discuss monogamy properties of measurement induced non locality , specifically for three and four qubit systems , we first mention some important classes of states in four qubit systems with some discussions on their entanglement behavior .
four qubit pure states can be classified in nine groups@xcite . among them the generic class is given by @xmath34 where @xmath35 , @xmath36 , @xmath37 and @xmath38 , @xmath39 .
consider two important subclasses @xmath0 and @xmath1 of the generic class @xmath40 which are defined as , @xmath41 and @xmath42 these two subclasses are important in the sense that @xmath0 is the maximally entangled class and @xmath1 have least amount of bipartite entanglement according to the definition of maximally entangled states , given by gour et .
al . in @xcite .
consider a pure bipartite state @xmath43 .
then the tangle is defined as @xmath44 where @xmath45 .
now for a pure state shared between four parties @xmath46 , three quantities @xmath47 are defined as : @xmath48 @xmath49 @xmath50 for the above two subclasses @xmath0 and @xmath1 we have @xmath51 , @xmath52 , @xmath53 and @xmath54 , @xmath55 , @xmath56 .
four - qubit ghz state belongs to the class @xmath1 .
monogamy is an important aspect in our physical world which restricts on the shareability of bipartite correlation .
entanglement is an example of quantum correlation which is monogamous w.r.t .
the tangle@xcite .
mathematically a correlation measure @xmath57 is said to be monogamous iff for any n - party state @xmath58 the relation @xmath59 holds for all @xmath60 .
now consider an n - party state @xmath61 .
let the locally invariant measurement is done on the party 1 .
then min is defined as @xmath62 , where @xmath63 is the optimal measurement done by the party 1 which does not change its local density matrix , i.e. , @xmath64 . on the other hand , since @xmath65 , the optimal measurement also does not change the local density matrices for all two - party reduced states of @xmath61 of the kind , @xmath66 .
then @xmath67 . so in case of polygamy @xmath68 and if @xmath69 then the state is monogamous w.r.t .
any three qubit pure state @xmath70 has a generic form @xmath71 , where @xmath72},\sum_{i}\lambda_{i}^{2}=1 $ ] @xcite .
this class includes ghz class .
w class can also be availed by putting @xmath73 .
for the general state @xmath70 , the reduced density matrix @xmath74 has the form @xmath75 then the correlation matrix @xmath76 , as obtained from the relation @xmath77 s being the pauli matrices , is @xmath78 other reduced density matrix @xmath79 and its corresponding correlation matrix could be written at once from @xmath22 and @xmath80 by only inter changing @xmath81 and @xmath30 .
coherent vector for both the reduced density matrices is @xmath82 .
+ clearly , @xmath83 iff @xmath84 . in case of
@xmath85 we have , @xmath86 @xmath87 @xmath88 where @xmath89 with @xmath90 .
now the minimum value of @xmath91 is 0.5 .
so in case of @xmath92 monogamy is violated for most of the states .
numerical simulation of @xmath93 states shows that around @xmath94 of three qubit generic states and around @xmath95 w class states satisfy monogamy .
+ when @xmath96 we have @xmath97 @xmath98 @xmath99 where @xmath100 , @xmath101 , @xmath102 , @xmath103 , @xmath104 , @xmath105 .
now the maximum value of @xmath106 is 0.5 .
hence three qubit pure states with @xmath96 satisfies the monogamy relation .
specifically , the three qubit generalized ghz class of pure states(@xmath107 ) is monogamous in the region @xmath108 @xmath109 ) and the monogamy relation holds good with equality if @xmath110 @xmath111 . in the three qubit generalized w - class states(@xmath112 )
we have @xmath113 . that is the monogamy relation holds with equality .
consider a four - qubit generic pure state @xmath46 which is shared between four parties @xmath114 ( i.e.,@xmath115 ) where @xmath116 we consider the two - qubit reduced density matrices @xmath117 and @xmath118 of @xmath119 .
each reduced density matrix is of the form @xmath120 where @xmath121 are some suitable functions of @xmath122 such that @xmath123 and @xmath124 .
we define 4 quantities @xmath125 . then for @xmath22 : @xmath126 ; for @xmath79 : @xmath127 and for @xmath118 : @xmath128 .
the elements of the correlation matrix @xmath80 can be obtained from @xmath129 .
eigenvalues of the matrix @xmath31 are of the form @xmath130 with @xmath131 .
the coherent vector @xmath33 as obtained from the relation @xmath132 is zero for all the three cases .
hence we have(by theorem 2 ) , @xmath133-\lambda_{3}\quad where\quad x = b , c , d\\ \end{split}\ ] ] and @xmath134 on the other hand , we can write the state @xmath135 in the form @xmath136 where @xmath137 are mutually orthonormal . since this is a pure state we have @xmath138 ( using theorem 1 )
numerical simulation for @xmath93 generic states shows that about 66% of them satisfies monogamy relation .
hence in general four - qubit generic class is not monogamous w.r.t .
so there exist quantum state whose locally shared non - locality exceeds the amount of globally shared non - locality .
+ [ fig:1 ] + _ @xmath139 simulation shows about 34% violation of monogamy for the generic class @xmath40 .
the red line marks the departure from monogamy_. + however for the sub classes @xmath0 and @xmath1 we can still check whether monogamy relation holds or not and if not then what is the amount of violation .
these things will be clear in the next few results . + + * theorem 3 * : _ consider a four - qubit system.then for any pure state @xmath2 belonging to the generic class @xmath40 we have @xmath140 where @xmath141 denotes any bipartite reduced density matrix of @xmath2 and @xmath142 s are the eigenvalues of @xmath31 ,
t being the correlation matrix of @xmath141_. + + _ proof _ : the proof of the theorem is very easy as it directly follows form theorem 2 keeping in mind that the eigenvalues of @xmath31 are all positive and the bloch vector @xmath143 for all bipartite reduced states of any generic state . + we will use this theorem in our proof of the the next two results .
@xmath144 + + * theorem 4 * : _ let @xmath145 be any four - qubit pure state belonging to the class @xmath0 . then @xmath146 _ + + _ proof _ : let us consider any four qubit pure state @xmath147 belonging to the class @xmath0 i.e. , @xmath148 , with @xmath149 . then , taking @xmath150 @xmath151 and utilizing the above restrictions , we get @xmath152 and @xmath153 .
the fruitful implementation of the result of previous theorem and simple algebraic manipulations using these results lead to the fact @xmath146.@xmath144 + + * theorem 5 * : _ let @xmath145 be any four - qubit pure state belonging to the class @xmath1.then @xmath154 _ .
+ + _ proof _ : let us consider any four - qubit pure state @xmath147 , belonging to the class @xmath1 i.e. , @xmath155 , with @xmath156 and @xmath157 .
now let us denote the sets of eigenvalues of @xmath158 and @xmath118 as @xmath159 , @xmath160 and @xmath161 respectively . without loss of generality we can consider , @xmath162,@xmath163 with @xmath164 .
this above consideration gives natural connections among the eigenvalues of @xmath79 and also @xmath118 .
these are @xmath165 and @xmath166 , @xmath163 with @xmath164 . observing the behavior of the eigenvalues and straight forward derivation
quite easily establish the desired result , i.e. , @xmath154.@xmath144 + + these two results give us information about the monogamous behavior of min in two important classes of four - qubit system .
since , @xmath167 in the whole class @xmath40 , therefore , min is monogamous in the subclass @xmath0 but it is polygamous in other class @xmath1 . further ,
all the states that are connected to these classes by lu , share the same fate .
four - qubit ghz state belongs to the class @xmath1 .
hence ghz and their lu equivalent states are not monogamous w.r.t min .
another two states(and obviously their lu equivalent states ) @xmath168 where @xmath169 and @xmath170 which maximize the tsallis @xmath171-entropy for different regions of @xmath171 @xcite , satisfy the monogamy relation of min . on the other hand the four - qubit cluster states
satisfies the monogamy relation as their two party reduced density matrices are completely mixed ( i.e. , min is zero ) . for four - qubit generalized w - states @xmath172 where @xmath173 , monogamy relation holds with equality , i.e.
, it can be easily shown ( as in the three - qubit case ) that @xmath174 .
hence for this type of states non - locality shows additive property with respect to each party .
this result of generalized w- class can be further extended to n - qubit system with the same conclusion .
thus we have explored the monogamy nature of min and found certain classes of states on which min shows monogamy relations . unlike geometric discord,@xcite min can be polygamous for pure state as revealed in some subclass of generic class . on the other hand , w - class seems to satisfy the monogamy relation with equality in n - qubit system .
so for the w - class min becomes additive in terms of sharing between the parties .
thus existence of monogamy of this type of correlation put a restriction on the amount of shared non - locality .
bell , physics 1 , 195,(1964 ) ; speakable and unspeakable in quantum mechanics , cambridge university press ( 1987 ) .
charles h. bennett , david p. divincenzo , christopher a. fuchs , tal mor , eric rains , peter w. shor , john a. smolin , and william k. wootters , phys .
a 59 , 1070 ( 1999 ) .
s. luo and s. s. fu , phys .
106 , 120401 ( 2011 ) .
valerie coffman , joydip kundu , william k. wootters , phys .
rev . a 61,052306 ( 2000 ) .
gian luca giorgi , phys .
a 84 , 054301 ( 2011 ) .
r. prabhu , arun kumar pati , aditi sen(de ) , and ujjwal sen , arxiv : 1108.5168v1 .
alexander streltsov , gerardo adesso , marco piani and dagmar bruss , arxiv : 1112.3967v1 . g. gour and n. r. wallach , j. math . phys .
51(11 ) , 112201 ( 2010 ) .
borivoje dakic , vlatko vedral and caslav brukner , phys .
105 , 190502 ( 2010 ) .
h. ollivier and w. h. zurek , phys .
88 , 017901 ( 2001 ) .
s. y. mirafzali , i. sargolzahi , a. ahanj , k. javidan , and m. sarbishaei , arxiv:1110.3499 .
yu guo , and jinchuan hou , arxiv:1107.0355v1 .
zhengjun xi , xiaoguang wang and yongming li , arxiv:1112.0736v1 .
f. verstraete , j. dehaene , b. de moor , h. verschelde , phys .
a 65 , 052112 ( 2002 ) .
a. acin , a. andrianov , e. jane , r. tarrach , j. phys .
34 ( 2001 ) , 6725 . | measurement induced non - locality was introduced by luo and fu as a measure of non - locality in a bipartite state . in this paper
we will discuss monogamy property of measurement induced non - locality for some three and four qubit classes of states .
unlike discord , we find quite surprising result in this situation .
both the generalized ghz and w states satisfy monogamy relations in the three qubit case , however , in general there are violations of monogamy relations in both the ghz - class and w - class states . in case of four qubit system , monogamy holds for most of the states in the generic class .
four qubit ghz does not satisfy monogamy relation , but w - state does .
we will also extend our results of generalized w - class to n - qubit . |
the possibility to describe in terms of hydrodynamic equations the evolution of a system where some physical quantities are not conserved is a challenging problem of non - equilibrium statistical mechanics .
several questions have to be faced as for example the validity of the underlying ( and in practice approximate ) kinetic theory , the choice of the hydrodynamical fields that are supposed to describe the relevant excitations in the problem , or the consistency of the method itself that is used to deduce the coarse - grained description from the kinetic theory .
much attention has been recently paid to such questions , mainly in the field of granular gas dynamics ( see e.g @xcite ) . in such systems ,
the kinetic energy is not conserved , while the linear momentum and number of particles are . however , even for low dissipation , the derivation of the hydrodynamic relations , based on a hard - sphere - like boltzmann equation is not a simple task and several approximations have to be invoked @xcite .
these difficulties lead to consider some simpler models by choosing ad - hoc collision term in the boltzmann equation .
the so - called maxwell and very hard particles ( vhp ) models @xcite are particularly interesting and reproduce some qualitative features of the granular gas of inelastic hard spheres @xcite .
another class of problems for which not only energy but also the density and momentum are not conserved is probabilistic ballistic annihilation ( pba ) .
in such a system , the particles move ballistically between collisions .
when two particles meet , they undergo an instantaneous collision and are removed from the system with probability @xmath0 or undergo an elastic scattering with probability @xmath1 . since collisions are assumed to be instantaneous , two body events only are taken into account .
the pba model was first introduced in one dimension in @xcite . in the limit @xmath2 , where density , momentum and kinetic energy are conserved ,
one recovers a system of hard spheres for which the hydrodynamic equations are well known @xcite .
the other limit @xmath3 ( pure annihilation ) has been the object of some work @xcite .
it was shown that in the long time limit the annihilation dynamics is exactly described by the boltzmann equation in dimensions higher than one @xcite .
this may qualitatively be understood by the fact that the density of the gas decays and , at late times , the packing fraction is very low .
this fact lead to conjecture that the boltzmann equation is an adequate description of pba at late times for @xmath4 @xcite .
given that @xmath0 may be considered as a perturbation parameter allowing to recover the elastic limit , the pba model is particularly interesting in view of testing the relevance and validity of the hydrodynamic description in general , which is a controversial issue and a long term goal of the present work .
the analytical treatment with usual hard sphere dynamics however appears to be quite involved @xcite , and we study here the simplified maxwell and vhp versions of pba .
the motivation is here is not only to test the ability of simplified kinetic models to mimic the hard sphere dynamics for a model far from equilibrium ( and with no conserved quantity , a more severe situation than that of granular gases ) but also to shed some light on some peculiar features obtained in the hydrodynamic study of ref .
in particular , this work exhibited divergent transport coefficients for a critical value of @xmath0 .
we will see that such singularities are absent in the simplified approaches , which may indicate that they are not associated with any physically relevant phenomenon .
it will also appear that maxwell and vhp approaches provide useful bounds for the hard sphere dynamics , so that similar inequalities as those found in @xcite concerning the scaling exponents can be obtained .
the paper is organized as follows . in section [ section2 ]
we introduce the boltzmann equation for both maxwell and vhp models of probabilistic ballistic annihilation , as well as the balance equations for the coarse - grained fields . in section [ section3 ]
we briefly describe the chapman - enskog scheme while section [ section4 ] is devoted to the maxwell model .
we first find the homogeneous state , and solve the corresponding homogeneous balance equations . to first order in the chapman - enskog expansion
we then study the effect of a small spatial inhomogeneity .
we follow the traditional route to compute the transport coefficients , which consists in truncating the first - order velocity distribution function to the first nonzero term in a sonine polynomial expansion @xcite .
we then show that this truncation does not constitute an approximation for the transport coefficients since they can be obtained by solving the maxwell model _ exactly _ to first order .
the vhp model is subsequently investigated in section [ section5 ] .
we first find the homogeneous cooling state , and then solve the corresponding homogeneous equations .
we implement monte carlo simulations in order to check the decay exponents found analytically .
next , we establish the transport coefficients to first order in the chapman - enskog expansion before presenting a comparison of the transport coefficients of the different models . in sec .
[ section6 ] we finally perform a linear stability analysis of the navier - stokes hydrodynamic equations around the spatially homogeneous state , and compare the results with pba of hard spheres .
we show as well that the second order decay rates may accurately be neglected .
our main findings and conclusions are summarized in section [ section7 ] .
since the underlying calculations of this paper are cumbersome , we present only the main steps in order to focus onto the more relevant results .
further technical details or explanations may be found in @xcite and for convenience , appendix [ appendixnotations ] contains a summary of the notations used .
the boltzmann equation for the one particle distribution @xmath5 of particles annihilating upon collision with probability @xmath0 reads @xmath6 + ( 1-p ) j_c[f , f ] , \label{boltzmann}\ ] ] where @xmath7 is the annihilation operator defined by @xmath8 = - \sigma^{d-1 } \phi(x ) v_t^{1-x } g(\mathbf{r},\mathbf{v}_1;t ) \int_{\mathbbm{r}^d } d \mathbf{v}_2 \ , v_{12}^x f(\mathbf{r},\mathbf{v}_2;t)\label{defja}\ ] ] and @xmath9 is the collision operator : @xmath10 = \sigma^{d-1 } \frac{\phi(x ) v_t^{1-x}}{s_d } \int_{\mathbbm{r}^d } d \mathbf{v}_2 \
, v_{12}^x \int d \widehat{\boldsymbol{\sigma}}\ , ( b^{-1 } -1 ) g(\mathbf{r},\mathbf{v}_1;t ) f(\mathbf{r},\mathbf{v}_2;t).\label{defjc}\ ] ] in these equations , @xmath11 denotes the spatial dimension , @xmath12 is the modulus of the relative velocity , @xmath13 is the solid angle surface , @xmath14 the euler gamma function , @xmath15 the time - dependent thermal velocity , @xmath16 , @xmath17 is the diameter of the particles , @xmath18 is a unit vector joining the centers of two particles and the corresponding integral is running over the solid angle .
finally , @xmath19 an operator acting on the velocities as follows : @xmath20 [ definitionoftheoperatorb ] the choice @xmath21 ( @xmath22 ) corresponds to the maxwell ( vhp ) model , respectively . for hard sphere dynamics
, that would correspond to @xmath23 , the relative velocity @xmath24 gives the rate of collision and its presence makes analytical progress difficult .
a convenient simplification @xcite to overcome this difficulty is to replace it by @xmath25 where @xmath26 is introduced for dimensional reasons .
the quantity @xmath27 which sets the relevant time scale in the problem can be freely chosen , and will be used in the following analysis to obtain the desired limiting behaviour in the limit @xmath28 ( see also @xcite for related considerations ) .
we also note that particles interacting with an inverse power - law potential are described by a kinetic equation of the same form as eq .
( [ defjc ] ) @xcite . in order to write hydrodynamic equations ,
we need to define local hydrodynamic fields : @xmath29 where @xmath30 , @xmath31 , and @xmath32 are the local number density , velocity , and temperature , respectively ( the latter definition being kinetic with no thermodynamic basis ) .
the definition of the temperature follows from the principle of equipartition of energy . in eq .
( [ deft ] ) , @xmath33 is the boltzmann constant and @xmath34 is the deviation from the mean flow velocity .
the balance equations follow from integrating the moments @xmath35 , @xmath36 , and @xmath37 with weight given by the boltzmann equation ( [ boltzmann ] ) . following the same route as in
@xcite we thus obtain @xmath38 , \\ & & \partial_t u_i +
\frac{1}{m n } \nabla_j p_{ij } + u_j \nabla_j u_i = - p \frac{1}{n } \omega[f , v_i f ] , \qquad i=1,\ldots , d , \\ & & \partial_t t + u_j \nabla_j t + \frac{2}{n k_b d } ( p_{ij } \nabla_i u_j + \nabla_j q_j ) = p \frac{t}{n } \omega[f , f ] - p \frac{m}{n k_b d } \omega[f , v^2 f],\end{aligned}\ ] ] [ balanceequations ] with implicit summation over repeated indices , @xmath39 , and @xmath40 = - \int_{\mathbbm{r}^d } d \mathbf{v}_1 \ , j_a[f , g].\ ] ] in the balance equations ( [ balanceequations ] ) , the pressure tensor @xmath41 and heat - flux @xmath42 are defined by @xmath43 @xmath44 where @xmath45 and @xmath46 @xmath47 as expected , when the annihilation probability @xmath2 , all three coarse - grained fields @xmath48 , @xmath49 , and @xmath50 are conserved .
the chapman - enskog method allows from eqs .
( [ balanceequations ] ) to build a closed set of equations for the hydrodynamic fields ( see e.g @xcite ) . for this purpose
, it is required to express the functional dependence of the pressure tensor @xmath41 and of the heat flux @xmath42 in terms of the hydrodynamic fields .
the chapman - enskog approach relies on two important assumptions .
the first one is the existence of a normal solution in which all temporal and spatial dependence of the distribution function @xmath5 may be expressed in terms of the hydrodynamic fields , @xmath51 $ ] .
the discussion of the relevance of this first assumption can be found elsewhere ( e.g in @xcite ) .
the second assumption is based on the separation of the microscopic time scale ( the average collision time between particles ) and macroscopic time scale ( the evolution of the hydrodynamic fields and their inhomogeneities ) .
this separation implies that the hydrodynamic fields are only weakly inhomogeneous , which allows for a series expansion in the gradients of the fields , @xmath52 , where each power of the formal small parameter @xmath53 is associated to a given order in spatial gradients .
the chapman - enskog method assumes the existence of an associated time derivative hierarchy : @xmath54 .
the insertion of these expansions in the boltzmann equation yields @xmath55 + ( 1-p ) j_c \left [ \sum_{l \geqslant 0 } \varepsilon^{l } f^{(l ) } , \sum_{l \geqslant 0 } \varepsilon^{l } f^{(l)}\right ] .
\label{expansion}\ ] ] the chapman - enskog solution is obtained upon solving the equations order by order in @xmath53 .
to zeroth order in the gradients , eq .
( [ expansion ] ) gives @xmath56 + ( 1-p ) j_c [ f^{(0)},f^{(0 ) } ] .
\label{eqorder0}\ ] ] this equation has a solution , describing the homogeneous state , and which obeys the scaling relation @xmath57 where @xmath58^{1/2}$ ] is the time dependent thermal velocity , and @xmath59 , @xmath60 . the existence of a scaling solution of the form ( [ eqscaling ] ) seems to be a general feature present in different but related contexts @xcite . this solution being isotropic , one has @xmath61 .
santos and brey @xcite showed that there exists a relationship between the homogeneous solutions of the maxwell model with @xmath62 and @xmath63 .
we shall here briefly reproduce their arguments .
it is possible to rewrite the boltzmann equation ( [ eqorder0 ] ) for @xmath21 under the form @xmath64 where @xmath65 , @xmath66 , @xmath67 , and @xmath68 is the removal collision frequency . integrating eq .
( [ bolhs1 ] ) over @xmath69 , the evolution of the number density is governed by @xmath70 , the solution being @xmath71 , where @xmath72 .
if we define @xmath73 and @xmath74 , then @xmath75 satisfies the boltzmann equation without annihilation ( i.e. , @xmath76 ) .
@xmath75 therefore evolves towards a maxwellian , and so does @xmath77 : we have @xmath78 .
since @xmath77 is isotropic , to zeroth order the pressure tensor ( [ pressure ] ) becomes @xmath79 , where @xmath80 is the hydrostatic pressure , and the heat flux ( [ heat ] ) becomes @xmath81 .
the balance equations to zeroth order read @xmath82 [ balanceequationstoorder0 ] where the decay rates are @xmath83,\\ \xi_{u_i}^{(0 ) } & = & \frac{1}{n v_t } \omega[f^{(0 ) } , v_i f^{(0 ) } ] , \qquad i=1,\ldots , d \label{decay0u } \\
\xi_t^{(0 ) } & = & \frac{m}{n k_b t d } \omega[f^{(0)},v^2 f^{(0 ) } ] - \frac{1}{n } \omega[f^{(0)},f^{(0)}].\end{aligned}\ ] ] [ maxwelldecayzero ] for antisymmetry reasons , one sees from eq .
( [ decay0u ] ) that @xmath84 .
the calculation of @xmath85 and @xmath86 are straightforward and give @xmath87 and @xmath88 .
we have written @xmath89 for @xmath90 .
the temperature of the maxwell model is therefore conserved in the scaling regime ( time independent thermal velocity @xmath26 ) .
in addition , one has @xmath91 where the subscript @xmath92 denotes a quantity evaluated in the homogeneous state , and @xmath93 is the decay rate for @xmath94 . note that eq .
( [ hcsboltzman ] ) was already established in sec .
[ sechcsmax ] .
[ secexactboltzce ] to first order in the gradients , the boltzmann equation ( [ expansion ] ) reads @xmath95 f^{(1 ) } = - [ \partial_t^{(1 ) } + \mathbf{v}_1 \cdot \boldsymbol{\nabla } ] f^{(0 ) } , \label{boltzmann1}\ ] ] the operator @xmath96 being defined by eqs .
( [ defj ] ) and ( [ defl ] ) .
the balance equations ( [ balanceequations ] ) to first order become @xmath97 [ balanceequations1 ] where the decay rates are given by @xmath98 , \\
\xi_{u_i}^{(1 ) } & = & \frac{1}{n v_t } \omega[f^{(0)},v_i f^{(1 ) } ] + \frac{1}{n v_t } \omega[f^{(1)},v_i f^{(0 ) } ] , \qquad i=1,\ldots , d , \\
\xi_t^{(1)}&= & - \frac{2}{n } \omega[f^{(0)},f^{(1 ) } ] + \frac{m}{n k_b t d } \omega[f^{(0 ) } , v^2 f^{(1 ) } ] + \frac{m}{n k_b t d } \omega[f^{(1)},v^2 f^{(0)}].\end{aligned}\ ] ] [ decay1 ] by definition @xmath99 is of first order in the gradients of the hydrodynamic fields ; for a low density gas @xcite @xmath100 the coefficients @xmath101 , @xmath102 , and @xmath103 depend on the fields @xmath48 , @xmath104 , and @xmath50 .
the hydrodynamic description of the flow requires the knowledge of transport coefficients , which may be determined from a sonine polynomial expansion of the first order distribution function .
in addition , the pressure tensor may be put in the form @xmath105 where @xmath106 is the ideal gas pressure , @xmath107 is the shear viscosity , and @xmath108 is the bulk viscosity which vanishes for a low density gas @xcite .
fourier s linear law for heat conduction is @xmath109 where @xmath110 is the thermal conductivity and @xmath111 a transport coefficient that has no analogue in the elastic case @xcite . the identification of eq .
( [ pressurephenomenological ] ) with eq .
( [ pressure ] ) using the result of the first order calculation yields @xmath112 similarly , the identification of eq .
( [ heatphenomenological ] ) with eq .
( [ heat ] ) using the first order calculation leads to @xmath113 the calculation follows the same route as in @xcite , and we obtain @xmath114 [ transportcoefficientsapprox ] where the thermal conductivity @xmath115 and shear viscosity @xmath116 coefficients for hard spheres ( used here to obtain dimensionless quantities ) are given by eqs .
( [ kappa0 ] ) and ( [ eta0 ] ) , respectively @xcite .
the dimensionless coefficients @xmath117 and @xmath118 are given by @xmath119 [ bigintegrals ] where @xmath120 , with @xmath80 .
note that the above relations are still exact within the chapman - enskog expansion .
the approximation consists in truncating the function @xmath99 to the first nonzero term in a sonine polynomial expansion : @xmath121 [ firstorderexpansion ] where @xmath122 , @xmath123 , and @xmath124 are the coefficients of the development , and @xmath125 is the maxwellian in the homogeneous state .
this allows one to compute the relations ( [ bigintegrals ] ) , and one finds ( see appendix [ appendix1 ] ) @xmath126 , \\ \nu_\kappa^ * & = & \phi^{\mathrm{m } } \frac{\sqrt{2 } \gamma(d/2)}{4 \pi^{(d-1)/2 } } \left [ \frac{d+2}{2 } + ( 1-p ) \frac{d-1}{d } \right].\end{aligned}\ ] ] the parameter @xmath89 governing the collision frequency may be freely chosen to allow for a relevant comparison with hard sphere dynamics ( see e.g. @xcite ) .
we choose @xmath127 such that the transport coefficients are normalized to one for @xmath28 , that is when all collisions are elastic .
it is remarkable that for the maxwell model a single parameter such as @xmath127 is sufficient to ensure normalization of all the transport coefficients ( this will not be the case in the vhp approach ) .
this leads to @xmath128 the above value turns out to be the same as the one obtained from the elastic limit of the maxwell model of granular gases @xcite . in the latter case ,
@xmath127 was chosen matching the temperature decay rate with that characterizing the homogeneous cooling state of inelastic hard spheres . with the choice ( [ choixphi ] )
the transport coefficients ( [ transportcoefficientsapprox ] ) become @xmath129 [ transportcoefficientsendapprox ] following the same route as in @xcite , the first - order distribution function ( [ f1 ] ) reads @xmath130 .
\label{f1end}\ ] ] by construction of the chapman - enskog method , the velocity moments of @xmath131 are given by those of the local equilibrium distribution @xmath77 .
it is then easy to show that the decay rates to first order ( [ decay1 ] ) are equal to zero [ therefore @xmath132 , where the operator @xmath133 is defined in appendix [ appendix2 ] ] . proceeding in a similar way as in @xcite , we obtain in appendix [ appendix2 ] the exact transport coefficients for the maxwell model , i.e. without any approximation on the form of @xmath99 .
this may be done by integrating the boltzmann equation ( [ boltzexacta ] ) over @xmath104 with weight @xmath134 and @xmath135 . with the choice for @xmath89 given by eq .
( [ choixphi ] ) , one finds the same transport coefficients as those given by eqs .
( [ transportcoefficientsendapprox ] ) .
this means that the truncation of @xmath99 to its first nonzero term in a sonine polynomial expansion is a harmless approximation when looking at the transport coefficients ( this is a peculiarity of the maxwell model ) .
in fact , it turns out that the transport coefficients depend only on the first term in the sonine polynomial expansion of @xmath99 @xcite .
for example , the heat current ( [ heat ] ) may be rewritten under the form @xcite @xmath136 where the first nonzero coefficients ( @xmath122 , @xmath123 ) ( that may depend on @xmath48 and @xmath50 ) in the sonine expansion are defined by eqs .
( [ firstorderexpansion ] ) . therefore the latter coefficients always give an exact result for the transport coefficients , but
the problem at hand is to calculate them exactly .
this turns out to be possible within the maxwell model . since the pressure tensor and the heat flux defined by eqs .
( [ pressurephenomenological ] ) and ( [ heatphenomenological ] ) , respectively , are of order @xmath35 in the gradients , their insertion in the balance equations ( [ balanceequations ] ) yields contributions of order 2 .
knowledge of the second order velocity distribution @xmath137 is therefore required in order to find the correct decay rates that contribute to navier - stokes order .
however , we show in sec .
[ section2order ] that for the maxwell gas those contributions are equal to zero for any annihilation probability @xmath0 .
the corresponding hydrodynamic navier - stokes equations are given by @xmath138 , \\ & & \partial_t u_i + \frac{1}{m n } \nabla_j p_{ij } + u_j \nabla_j u_i = - p v_t [ \xi_{u_i}^{(0 ) } + \xi_{u_i}^{(1 ) } ] , \qquad i=1,\ldots , d , \\ & & \partial_t t + u_i \nabla_i t + \frac{2}{n k_b d } ( p_{ij } \nabla_i u_j + \nabla_i q_i ) = - p t [ \xi_t^{(0 ) } + \xi_t^{(1)}].\end{aligned}\ ] ] [ balanceequationsend ] @xmath41 and @xmath139 are given by eqs .
( [ pressurephenomenological ] ) with @xmath140 , and ( [ heatphenomenological ] ) respectively . the rates @xmath141 , @xmath142 , and @xmath143
may be calculated using their definition ( [ balanceequations1 ] ) and the distribution ( [ f1end ] ) @xcite .
we find that all decay rates are equal to zero , except @xmath144 we thus have a closed set of equations for the hydrodynamic fields to the navier - stokes order .
integrating the boltzmann equation ( [ boltzmann ] ) over @xmath104 for @xmath22 , one obtains @xmath145 where @xmath146 and @xmath147 denotes the average of a function @xmath148 in the homogeneous cooling state ( hcs ) .
we have written @xmath149 for @xmath150 .
following the same route as in @xcite the boltzmann equation may be rewritten in the form @xmath151
\widetilde{f}(c_1 ) = \widetilde{f}(c_1 ) \int_{\mathbbm{r}^d } d \mathbf{c}_2 \ , c_{12}^2 \widetilde{f}(c_2 ) - \frac{1-p}{p } \frac{1}{s_d } \widetilde{i}[\widetilde{f},\widetilde{f}],\label{boltzmoments1}\ ] ] where @xmath152 \int_{\mathbbm{r}^{2d } } d \mathbf{c}_1 d \mathbf{c}_2 \int d \widehat{\boldsymbol{\sigma } } \ , c_{12}^2 \widetilde{f}(c_1 ) \widetilde{f}(c_2 ) } = \frac{\langle c_{12}^2 c_1 ^ 2\rangle}{\langle c_1 ^ 2 \rangle\langle c_{12}^2 \rangle},\label{alphae}\ ] ] and @xmath153 = \int_{\mathbbm{r}^{d } } d \mathbf{c}_1 \int d \widehat{\boldsymbol{\sigma } } \ , c_{12}^2 ( b^{-1 } -1 ) \widetilde{f}(c_1 ) \widetilde{f}(c_2).\ ] ] the limit @xmath154 of the boltzmann equation ( [ boltzmoments1 ] ) encodes a useful information for ballistically controlled dynamics @xcite : @xmath155 .
\label{boltzmoments0}\ ] ] next , we consider the first nonzero correction to the maxwellian in a sonine polynomial expansion of the hcs : @xmath156 , \label{firstsonine}\ ] ] where @xmath157 is the maxwellian and @xmath158 the second sonine polynomial @xcite .
( [ boltzmoments0 ] ) and ( [ alphae ] ) form a system of two equations for the two unknown @xmath159 and @xmath160 . making use of the relations ( [ app3eq5 ] ) ,
it is a straightforward task to compute the limit in the right hand side of eq .
( [ boltzmoments0 ] ) @xcite , which gives @xmath161 = - a_2 \frac{s_d}{\pi^{d/2 } } \frac{d^2(d+2)}{16}. \label{finaleone}\ ] ] using eq .
( [ firstsonine ] ) , one easily obtains from eq .
( [ alphae ] ) @xmath162 note that eqs .
( [ finaleone ] ) and ( [ finaletwo ] ) are exact relations for which all nonlinear contributions in @xmath160 were kept . however , those nonlinear terms cancel out in each case . making use of @xmath163 , the insertion of eqs .
( [ finaleone ] ) and ( [ firstsonine ] ) in ( [ boltzmoments0 ] ) gives @xmath164 = 1 + a_2 \frac{d(d+2)}{8 } \frac{1}{p}. \label{eqorder2}\ ] ] eq .
( [ eqorder2 ] ) admits two solutions , the first one being @xmath165 and the second one @xmath166/[d(d+2)p]$ ] .
the second solution is not physical since it diverges for @xmath62
. therefore @xmath165 and the hcs of the vhp model within the approximation ( [ firstsonine ] ) is described by the local maxwellian @xmath157 .
we also note that upon discussing the potential ambiguities resulting from such a linearization scheme in @xmath160 ( as done in @xcite ) , the same conclusion is reached . proceeding in a similar way
as already described , we obtain a set of equations formally identical to eqs .
( [ balanceequationstoorder0 ] ) and ( [ maxwelldecayzero ] ) .
the calculation of the decay rates gives @xmath84 , @xmath167 , and @xmath168 .
the hcs is therefore given by @xmath169 [ hcsvhp ] where the decay exponents are @xmath170 , @xmath171 , and the relaxation time @xmath172 .
in other words , we have @xmath173 these quantities do not depend either on @xmath127 nor on the annihilation probability @xmath0 .
the former result is an exact property of the dynamics under study ( the factor @xmath127 may be absorbed into a rescaling of time @xmath174 , leaving scaling exponents unaffected ) while the latter may _ a priori _ be an artifact of the approximations made ( it will however be shown below that the @xmath0 dependence if any is extremely weak ) . if we define the root - mean - square velocity by @xmath175 , then from the definition ( [ deft ] ) of the temperature @xmath176 , and from eq .
( [ secondeqac231 ] ) we have @xmath177 for long times , with @xmath178 .
the decay exponents @xmath179 and @xmath180 , as well as the decay exponents for the maxwell model , agree with the prediction of krapivsky and sire @xcite , and satisfy the scaling constraint @xmath181 , which essentially expresses the unicity of the relevant time scale in the problem . moreover , making use of the expression for the decay exponents of pba of hard spheres @xmath182 and @xmath183 obtained to linear order in @xmath160 and which are recalled in appendix [ appendixdecayexponentshs ] @xcite , it is easy to verify explicitly that the maxwell and vhp models provide bounds @xcite @xmath184 for all @xmath185 $ ] .
we emphasized however that the previous inequality have the status of `` empirical '' observations , and could not be anticipated from rigorous arguments .
we performed direct monte carlo simulations ( dsmc ) in order to verify the decay exponents of the vhp model .
the algorithm is similar to the one described in @xcite . for the sake of completeness
, we briefly outline the main steps of the algorithm .
we choose at random two different particles @xmath186 .
the time is then increased by @xmath187 where @xmath188 is the number of remaining particles . with probability @xmath0 the two particles are removed from the system , and with probability @xmath189 their velocities are modified according to eqs .
( [ definitionoftheoperatorb ] ) . as the fluctuations increase for small @xmath188
, it is necessary to average over several independent realizations in order to diminish the noise . a log - log plot of the density @xmath190 and the root - mean - squared velocity @xmath191 as a function of time gives the decay exponents ( see fig .
[ fig0a ] ) .
the dsmc results are in excellent agreement with the analytical predictions and the expected power - law behaviors are observed over several decades ( see .
[ fig0b ] ) . and @xmath180 ( inset ) in two dimensions for the vhp model ( @xmath22 ) . the analytical predictions @xmath192 and @xmath193 are shown by the continuous lines while the symbols stand for the dsmc results ( obtained from approximately @xmath194 independent runs and @xmath195 initial particles ) . from the above data
, it appears that the scaling relation @xmath196 is well obeyed ( the deviation from @xmath35 does not exceed @xmath197 ) and that the scaling exponents do not depend on @xmath0 .
note the small @xmath198 scale . ] and @xmath199 ( inset ) for @xmath200 and @xmath201 on a log - log scale .
the initial velocity distribution is gaussian .
@xmath202 ( resp .
@xmath188 ) is the initial ( resp .
remaining ) number of particles .
@xmath203 is the root - mean - square velocity at @xmath94 , whereas we write @xmath204 for @xmath205 .
the dashed straight line is a linear interpolation giving the decay exponent of the power - law , and the deviations to this law for large times is due to the low number of remaining particles . ]
the procedure is similar to the one followed within the maxwell model of sec .
[ secfirstordermax ] ( or @xcite ) , and we find @xmath206 [ systgde1 ] where @xmath207 + p \xi_t^{(0)*}\{-4\nu_\mu^ * + 3 p[\xi_n^{(0)*}+2\xi_t^{(0)*}]\ } $ ] , @xmath208 and @xmath209 , @xmath210 . truncating the function @xmath99 to the first term in a sonine polynomial expansion as it was the case for eqs .
( [ firstorderexpansion ] ) , the coefficients @xmath117 , @xmath118 , and @xmath211 may be calculated with the help of appendix [ appendix1 ] .
we find @xmath212,\\ \nu_\kappa^ * = \nu_\mu^ * & = & \phi^{\mathrm{vhp } } \frac{\sqrt{2 } \gamma(d/2)}{4 \pi^{(d-1)/2 } } \left [ p \frac{(d+2)(d+3)}{2 } + ( 1-p ) \frac{(d-1)(d+4)}{d } \right].\end{aligned}\ ] ] the free parameter @xmath149 setting the frequency collision has _ a priori _ no reason for being the same as for the maxwell model .
we choose this quantity such that @xmath213 , which means that the shear viscosity for the vhp gas is set for vanishing @xmath0 to coincide with the shear viscosity @xmath116 of hard spheres .
this allows for a better comparison of the transport coefficients for the maxwell , hard sphere , and vhp models .
other choices for @xmath149 are possible .
the condition @xmath214 leads to @xmath215 so that @xmath216 [ decay0vhp2 ] the first order distribution function reads @xmath217 .
\label{f1endvhp}\ ] ] where the transport coefficients are given by eqs ( [ systgde1 ] ) .
the decay rates to first order may be calculated using the definitions ( [ balanceequations1 ] ) and the distribution ( [ f1endvhp ] ) @xcite , which gives @xmath218 [ decayvhporder1end ] where @xmath219 the navier - stokes hydrodynamic equations are thus given by eqs .
( [ balanceequationsend ] ) with the decay rates ( [ decay0vhp2 ] ) and ( [ decayvhporder1end ] ) . for consistency ,
second order contributions in the gradients are again needed when evaluating the decay rates .
it will be shown in section [ section2order ] that they are small corrections to the previous relations .
we compare the transport coefficients for the maxwell , vhp , and hard sphere models ( the coefficients for the latter model being given in @xcite ) . figs .
[ fig1 ] , [ fig2 ] , and [ fig3 ] show @xmath220 , @xmath221 , and @xmath222 , as a function of the annihilation probability . as a function of the annihilation probability @xmath0 for the maxwell ( thin continuous line ) , vhp (
dashed line ) , and hard spheres models ( thick continuous line ) . ] as a function of the annihilation probability @xmath0 for the maxwell ( thin continuous line ) , vhp ( dashed line ) , and hard spheres models ( thick continuous line ) .
the vertical lines gives the value @xmath223 for which a divergence of the hard sphere transport coefficients @xmath221 and @xmath222 appears ( while the shear viscosity exhibits regular behavior , see fig . [ fig1 ] ) . ] as a function of the annihilation probability @xmath0 ( see fig .
[ fig2 ] for more details ) .
the maxwell model is not represented since in this case @xmath224 . ]
note that once we have chosen @xmath150 such that @xmath225 for @xmath28 there is no reason to expect @xmath226 in the same limit .
other choices would have been possible such as enforcing @xmath226 when @xmath28 . from figures
[ fig1 ] , [ fig2 ] , and [ fig3 ] it first appears that mawxell and vhp models capture the essential @xmath0 dependence of the `` hard sphere '' transport coefficients .
in addition , they provide in most cases lower and upper bounds for @xmath220 , @xmath221 and @xmath222 . however , as already pointed out in @xcite , for strong annihilation probability @xmath227 , the hard sphere thermal conductivity and `` fourier '' coefficient @xmath111 diverge ( see figs [ fig2 ] and [ fig3 ] ) which leads to a violation of the vhp upper bound for @xmath110 and @xmath111 in the vicinity of @xmath228 .
the fact that vhp and maxwell models lead to smooth and regular transport coefficients for all values of @xmath0 gives a hint that the hard sphere divergence obtained in previous work @xcite is a possible artifact of the underlying approximations and probably does not point towards a change of behavior nor a qualitative difference in the scaling or transport properties .
this point will be further discussed in the concluding section .
we finally note that an _ a priori _ similar deficiency was already reported for the maxwell model of inelastic hard spheres @xcite .
the hydrodynamic equations . ( [ balanceequationsend ] ) can not be solved analytically in general . however , their linear stability analysis allows one to answer the question of formation of spatial inhomogeneities .
the present study establishes under which conditions the homogeneous state is stable .
we consider here a small deviation from spatial homogeneity [ see eqs .
( [ hcsboltzman ] ) and ( [ hcsvhp ] ) ] and the linearization of eqs .
( [ balanceequationsend ] ) in the latter perturbation .
the procedure used here follows the same route as for granular gases @xcite or pba of hard spheres @xcite .
we define the deviations of the hydrodynamic fields from the homogeneous solution by @xmath229 , where @xmath230 . inserting this form in the navier - stokes - like equations yields differential equations with time - dependent coefficients . in order to obtain coefficients that do not depend on time
, it is necessary to introduce the new dimensionless space and time scales defined by @xmath231 } \mathbf{r}/2 $ ] , @xmath232 , as well as the dimensionless fourier fields @xmath233 , @xmath234 } \delta \mathbf{u}_\mathbf{k}(\tau)$ ] , and @xmath235 , where @xmath236 .
note that @xmath237 is defined ( up to a constant prefactor ) in units of the mean free path for a homogeneous gas of density @xmath238 .
the dimensionless time @xmath239 gives the accumulated number of collisions per particles up to time @xmath174 . since we will study both the maxwell and vhp systems , we recall here the general results valid for non - vanishing decay rates @xmath85 , @xmath86 , and @xmath240 .
making use of the dimensionless variables , the linearized hydrodynamic equations read @xmath241 \rho_{\mathbf{k}}(\tau ) + p \xi_n^{(0 ) * } \theta_\mathbf{k}(\tau ) + i k w_{\mathbf{k}_\parallel}(\tau ) = 0 , \label{stability8a } \\ \left [ \frac{\partial}{\partial \tau } - p \xi_t^{(0 ) * } + \frac{d-1}{d } \eta^ * k^2 \right ] \mathbf{w}_{\mathbf{k}_\parallel } + i \mathbf{k } \big [ \big ( 1-p\xi_u^ * \mu^ * \big ) \rho_\mathbf{k}(\tau ) + \big ( 1-p\xi_u^ * \kappa^ * \big ) \theta_\mathbf{k}(\tau ) \big ] = 0 , \label{stability8b } \\ \left [ \frac{\partial}{\partial \tau } - p \xi_t^{(0 ) * } + \frac{1}{2 } \eta^ * k^2 \right ] \mathbf{w}_{\mathbf{k}_\perp}(\tau ) = 0 , \label{stability8c}\\ \left [ \frac{\partial}{\partial \tau } + p \xi_t^{(0 ) * } + \frac{d+2}{2(d-1 ) } \kappa^ * k^2 \right ] \theta_\mathbf{k}(\tau ) + \left [ 2 p \xi_t^{(0 ) * } + \frac{d+2}{2(d-1 ) } \mu^ * k^2 \right ] \rho_\mathbf{k}(\tau ) + \frac{2}{d } i k w_{\mathbf{k}_\parallel}(\tau ) = 0 , \label{stability8d}\end{aligned}\ ] ] [ stability8 ] where the transverse mode @xmath242 appears to be decoupled from the other equations .
the longitudinal velocity field is given by @xmath243 , and @xmath244 is the unit vector along the direction given by @xmath245 .
the transversal velocity field @xmath246 consequently defines @xmath247 degenerated shear modes . upon direct integration
, we have @xmath248 , \label{solperp}\ ] ] where @xmath249 on the other hand , the longitudinal velocity field @xmath250 lies in the one dimensional vector space generated by @xmath245 .
hence there are three hydrodynamic fields to be determined , namely the density @xmath251 , temperature @xmath252 , and longitudinal velocity field @xmath253 .
the hydrodynamic matrix @xmath254 of the corresponding linear system reads off straightforwardly from eqs .
( [ stability8b])-([stability8d ] ) .
the corresponding eigenmodes are given by @xmath255 $ ] , @xmath256 , where @xmath257 are the eigenvalues of @xmath254 .
each of these three fields is a linear combination of the eigenmodes ; thus only the biggest real part of the eigenvalue @xmath257 has to be taken into account to discuss the limit of marginal stability of the different modes .
we define @xmath258 by the condition @xmath259=0 $ ] , i.e. , @xmath260^{1/2},\ ] ] and @xmath261 by @xmath262=0 $ ] , @xmath263 .
therefore if @xmath264 all _ rescaled _ modes are linearly stable . for @xmath265 $ ]
only the _ rescaled _ shear mode is linearly unstable ( the latter may however be non - linearly coupled to the other modes ) , and for @xmath266 all eigenvalues are positive which leads to instabilities .
however , it should be kept in mind that the previous discussion involves rescaled modes only , and should be connected to the original @xmath267 variable . indeed , for any real system ( for example a cubic box of volume @xmath268 ) the smallest wavenumber allowed for a perturbation is given by @xmath269 , which corresponds to the time - dependent dimensionless wavenumber @xmath270 where @xmath271 $ ] . since the density @xmath272 is a decreasing function of time , @xmath273 increases monotonously and there exists a time @xmath274 such that @xmath275 for @xmath276 .
the lower cut - off @xmath273 therefore eventually enters the region where the homogeneous solution is stable . for @xmath277 ,
the system is however not in a spatially homogeneous state , but it is nevertheless tempting to conclude that the perturbations will be damped for @xmath278 .
although this statement is not rigorously derived , we conclude here that an instability can only be a transient effect @xcite .
the time @xmath274 can be estimated from the condition @xmath279 . making use of the hypothesis of small spatial inhomogeneities
, we may replace the density @xmath272 appearing in the definition of @xmath280 by the homogeneous density @xmath238 given by eq .
( [ eq47a ] ) .
we obtain @xmath281^{1/\gamma_n } -1 \right\}. \label{bluea}\ ] ] is the transient instability alluded to easily observable in a simulation ?
a typical number of particles for molecular dynamics simulations is of the order of @xmath282 , and @xmath283 ( which corresponds to a rather low total initial packing function @xmath284 ) . for @xmath285 and @xmath200 eq .
( [ bluea ] ) gives @xmath286 .
making use of eq .
( [ eq47a ] ) to approximate the density , one obtains @xmath287 .
the density inhomogeneities therefore start to decrease after that the density decreased to only @xmath288 times its initial value , which for @xmath285 corresponds in average to only @xmath289 collisions per particle . for comparison purposes , inhomogeneities in granular gases
are observed after a few hundred collisions per particle @xcite . in order to observe the previous ( and presumably transient ) instabilities one would need molecular dynamics simulations with very large systems .
another condition is to have a large enough @xmath0 , which increases @xmath258 , see fig .
equivalently , increasing @xmath0 increases the divergence rate @xmath290 at fixed @xmath291 , see eq .
( [ modeperp ] ) . for sufficiently small @xmath0 ( or small system sizes ) eq .
( [ bluea ] ) does not have a positive solution because @xmath292 already for @xmath94 . to sum up
, the typical size of the inhomogeneities may grow as a function of time until @xmath293 but the subsequent evolution should drive the system back to a time dependent homogeneous regime . as already pointed out in sec .
[ sec : hydromax ] , it is necessary for consistency to include the second order decay rates in the first order navier - stokes equations .
their study is useful to establish the relevance of the second order contributions in the gradients [ that can be linear in @xmath294 or @xmath295 , or non - linear in @xmath296 , @xmath297 , etc . ] .
our analysis follows the method of ref .
@xcite , and we shall therefore not enter into much details . since we shall perform a linear stability analysis , the only terms in the second order decay rates that will contribute are linear in the gradients .
we therefore denote @xmath298 the part of the second order distribution @xmath137 that yields the linear contributions , and neglect any other term in @xmath137 .
the solution has the form @xmath299 , where @xmath300 and @xmath188 are to be determined . inserting the latter relation in the linear second order decay rates
[ that may be obtained from eqs .
( [ decay1 ] ) upon replacing @xmath99 by @xmath298 ] they take the form @xmath302 , where @xmath303 .
again , it is useful to resort to a first order sonine polynomial expansion for @xmath300 and @xmath188 , such that @xmath304 and @xmath305 , where @xmath306 and @xmath307 are the coefficients to be determined . while it is a rather straightforward task to see from eqs .
( [ decay1 ] ) that those second order contributions to the decay rates are equal to zero for the maxwell gas , the calculations for the vhp model are more involved . in the latter case
all decay rates are equal to zero except @xmath308 and @xmath309 , where @xmath310 .
the coefficients @xmath306 and @xmath307 can be put in a dimensionless form @xmath311 and @xmath312 . upon replacing the above expansions for @xmath298 and for the decay rates in the boltzmann equation , the coefficients @xmath313 and @xmath314 are solution of the equations @xmath315 c_t^{(2 ) * } & = & p \frac{1}{2 } \xi_n^{(0 ) * } c_n^{(2 ) * } + \frac{8}{d(d+2 ) } \kappa^ * y , \\ \left [ p \xi_n^{(0 ) * } + \nu_\xi^ * - p \frac{2(d+2)}{d+4 } \right ] c_n^{(2 ) * } & = & p \xi_t^{(0 ) * } c_t^{(2 ) * } + \frac{8}{d(d+2 ) } \mu^ * y,\end{aligned}\ ] ] where @xmath316 $ ] , and @xmath317}{\int_{\mathbbm{r}^d } d \mathbf{v } \ , v^4 s_2(c^2 ) \mathcal{m}(\mathbf{v } ) } = \phi \frac{\sqrt{2 } \gamma(d/2)}{4\pi^{(d-1)/2 } } \left [ p(d+2)(d+3 ) + ( 1-p ) 4 d \right].\ ] ]
the second order linear decay rates are given by @xmath318 $ ] and @xmath319 $ ] .
the dimensionless linearized hydrodynamic equations with nonzero second order decay rates are the same as eqs .
( [ stability8 ] ) except for eq .
( [ stability8c ] ) where one has to replace @xmath221 by @xmath320 and @xmath222 by @xmath321 .
consequently , the ratios @xmath322 and @xmath323 ( see fig . [ fig0c ] ) give a rough estimation of the relevance of the second order decay rates .
( @xmath222 ) to the linear second order dimensionless transport coefficients @xmath324 ( @xmath324 ) , for the vhp model .
these ratios do not depend on the the collision frequency @xmath127 . ] as seen from fig .
[ fig0c ] , the second order decay rates @xmath325 and @xmath326 are much smaller than the transport coefficients @xmath221 and @xmath222 , respectively .
they may therefore be neglected at least from a linear stability analysis point of view .
this approximation is increasingly more accurate as the annihilation probability is decreased .
note that our conclusions for the maxwell model of probabilistic ballistic annihilation are comparable to those for the granular gas @xcite . for the maxwell model , the temperature decay rate @xmath86 vanishes .
it follows from eq .
( [ modeperp ] ) that @xmath327 and the transverse mode is stable , which is confirmed by fig .
the maxwell model appears to be linearly stable for all values of the annihilation probability @xmath0 . and
the dispersion relation obtained from eq .
( [ modeperp ] ) is represented by a dashed line ( labeled @xmath290 ) whereas the three remaining relations are represented by continuous lines ( labeled @xmath329 ) .
the shear mode ( @xmath290 ) and sound modes ( which are on this figure such that @xmath330 when @xmath331 ) are degenerated twice . ] on the other hand , within the vhp approach , the decay rate @xmath332 .
the transverse mode may consequently be unstable for some wave - numbers @xmath291 of the perturbation ( see fig .
[ fig5 ] ) , which by nonlinear coupling to the other modes may lead to density inhomogeneities .
other modes than the shear may also be linearly unstable , when rescaled wave numbers are such that @xmath333 .
the thresholds @xmath258 and @xmath261 are shown in fig .
[ fig6 ] for the 3 models .
it appears again that the hard sphere quantity is bounded below by its maxwell counterpart and above by vhp .
note that the linear stability analysis does not suffer from arbitrariness related to the free parameter @xmath27 . and @xmath328 .
the dispersion relation obtained from eq .
( [ modeperp ] ) is represented by a dashed line ( labeled @xmath290 ) whereas the three remaining relations are represented by continuous lines ( labeled @xmath329 ) .
the first two biggest parallel modes are sound modes . ] and @xmath261 in dimensionless units as a function of the annihilation probability @xmath0 for @xmath328 .
s and vhp superscripts denote the hard spheres and very hard particles models , respectively . within the maxwell model , one has @xmath334 . ]
the imaginary part of the eigenvalues embodies the information on the propagation of the perturbations . in fig .
[ fig5 ] , we identify 3 different parallel modes for small enough @xmath291 ( @xmath335 ) . given that the shear mode is always @xmath247 times degenerated and that there are @xmath336 modes in total , none of the parallel modes are degenerated for low enough @xmath291 .
increasing @xmath291 up to the first bifurcation , the sound modes become degenerated and have a nonzero imaginary value .
the non - propagating sound modes thus have bifurcated into a pair of propagating modes .
since the eigenvalue for the transverse velocity field is always real , we shall study here only the imaginary part of the other eigenvalues .
we define @xmath337 such that for all @xmath338 all eigenvalues are real .
it means that only perturbations with small enough wave numbers @xmath339 such that @xmath340 are propagating .
[ fig7 ] shows @xmath337 as a function of the annihilation probability @xmath0 for the vhp , hard sphere , and maxwell models .
once more , the vhp and maxwell models appear as upper and lower bounds , respectively . from fig .
[ fig4 ] the maxwell sound modes are degenerated for all @xmath291 and therefore the sound modes of the maxwell model are always propagating , i.e. , @xmath341 . in the vhp case , fig .
[ fig5 ] shows a propagation gap for the sound modes , i.e. , a @xmath291- window with @xmath342 , where the sound modes are not degenerated .
this is confirmed by fig .
[ fig7 ] ( smaller inset ) .
a propagation gap in the sound mode dispersion relation has been predicted on the basis of the revised enskog theory for hard - sphere fluids @xcite .
such a gap has been observed in neutron scattering experiments of atomic liquids @xcite , of molten salts @xcite , or inelastic x - ray scattering of lipid bilayers @xcite for example . in dimensionless units as a function of the annihilation probability @xmath0 for @xmath328 .
the maxwell model is not represented since in this case @xmath341 for all @xmath0 .
the main inset shows the imaginary part of the eigenvalues in dimensionless units for the vhp model for @xmath328 and @xmath285 .
the smaller inset shows the propagation gap @xmath343 $ ] of the sound modes . ]
making use of the chapman - enskog scheme , we have derived in this paper the hydrodynamic equations governing the coarse - grained number density , linear momentum and kinetic energy density fields for an assembly of particles undergoing annihilating collisions with probability @xmath0 and an elastic scattering otherwise . in between collisions ,
the motion is ballistic . to this aim
, the relevant `` hard sphere''-like boltzmann equation has been simplified first into its maxwell , and second into its very hard particle ( vhp ) form . in both cases ,
the corresponding navier - stokes equations take the same form as in the initial hard sphere description and read @xmath344 with @xmath345 for the maxwell model , and @xmath346 in the vhp case [ the transport coefficients @xmath221 or @xmath222 are given by eqs .
( [ systgde1 ] ) ] .
our analysis showed that the maxwell and vhp simplifications , that are more amenable to analytic treatment , not only capture the essential features of hard sphere dynamics , but also provide lower and upper bounds for all comparable quantities .
some important differences should however be commented upon .
a first difference is that maxwell and vhp lead to regular transport coefficients for all values of the annihilation probability , whereas a divergence occurs for annihilating hard sphere thermal conductivity @xmath110 and fourier coefficient @xmath111 .
we concluded from this comparison that this divergence is presumably not physical and could result from the more stringent approximations put forward in the hard sphere computation .
it turns out that the hard sphere case is such that the velocity distribution is non - gaussian to zeroth order in spatial gradient , whereas it is gaussian in maxwell and vhp cases .
this fact could be at the root of the divergence observed in the transport coefficients .
we have also shown that the second order decay rates of the maxwell model are equal to zero , while for the vhp case they may accurately be neglected .
this analysis suggests that the second order decay rates of probabilistic ballistic annihilation of hard spheres may as well be accurately neglected , at least for small annihilation probabilities @xmath0 .
this approximation had been invoked in ref .
@xcite , without any control on its validity .
the second important difference between maxwell , hard sphere and vhp dynamics is that within the maxwell model , all fourier modes are found to be linearly stable .
this fact is intimately related to the non dissipative nature of the corresponding dynamics , an aspect which may be surprising at first : although particles are permanently removed from the system , the mean kinetic energy is conserved on average ( @xmath347 ) .
this may be considered as a deficiency of the maxwell ( over)simplification . on the other hand , vhp dynamics is such that the collision frequency increases with the velocity of a given population of particles , which in turn implies that the kinetic energy decreases faster than the number of particles , hence @xmath348 .
this dissipation is at the root of possible instabilities in the coarse - grained fields .
however , these instabilities manifest themselves for suitably rescaled fields , and we argued in section [ section6 ] that they should presumably only translate into transient instabilities for the `` real '' fields .
indeed , due to the decrease of density @xmath272 , an unstable fourier mode has a wavenumber increasing like @xmath349 , and eventually enters into a regime where damping should wash out the perturbation .
this feature presumably provides at least a linear saturation mechanism for instabilities , different from usual non - linear saturation effects , that may also play a _ transient _ role here if the initial conditions are sufficiently unstable [ in other words , if @xmath350 .
our stability analysis was nevertheless restricted to perturbations around the time dependent homogeneous state , so that strictly speaking , the effects of transient instabilities that may drive the system into a strongly modulated state are unclear at the moment .
this calls for a careful numerical ( molecular dynamics ) study of the coarse - grained fields , which is the purpose of future work .
this would also allow to question the validity of the hydrodynamic description , in a regime where the wave number is not much smaller than the inverse mean - free - path @xmath351 ( in the previous figures , @xmath291 is expressed in units of @xmath352 , up to a prefactor of order one ) .
we acknowledge useful discussions with a. barrat , a. puglisi , f. van wijland , p. visco , a. santos , v. garz , j.j .
brey , m.j .
ruiz - montero , j. piasecki , and i. bena .
this work was partially supported by the swiss national science foundation and the european community s human potential program under contract hprn - ct-2002 - 00307 ( dyglagemen ) .
we shall recall here some of the notations used through the paper . @xmath110 and @xmath111 are the transport coefficients appearing in fourier s linear heat conduction law ( [ heatphenomenological ] ) , and @xmath107 is the shear viscosity appearing in the pressure tensor ( [ pressurephenomenological ] ) . a quantity @xmath353 that is made dimensionless
is noted @xmath354 .
the corresponding dimensionless transport coefficients are written @xmath355 where @xmath356 @xmath357 are the thermal conductivity and shear viscosity coefficients for hard spheres , respectively .
the dimensionless coefficients @xmath117 , @xmath118 , and @xmath211 are given by @xmath358 [ bigintegralsappendix ] with @xmath359 and @xmath80 is the zeroth order pressure . in eqs .
( [ bigintegralsappendix ] ) , the operator @xmath96 is given by @xmath360 + ( 1-p ) l_c[f^{(0)},g],\label{defj}\ ] ] where @xmath361 & = & -j_a[f^{(0)},g ] - j_a[g , f^{(0)}],\\ l_c[f^{(0)},g ] & = & -j_c[f^{(0)},g ] - j_c[g , f^{(0)}],\\\end{aligned}\ ] ] [ defl ] @xmath362 being an arbitrary function .
the collision operator @xmath9 ( annihilation operator @xmath7 ) is defined by eq .
( [ defjc ] ) [ eq .
( [ defja ] ) ] .
the linear operator @xmath133 is defined by eq .
( [ defomega ] ) . the velocity distribution function
is denoted @xmath5 . in the scaling regime @xmath363 where @xmath59 .
the time dependent [ through @xmath364 thermal velocity is @xmath365 we note the maxwellian in the homogeneous cooling state by @xmath366 and the maxwellian by @xmath367 therefore , we obtain a similar relation to eq .
( [ scalingf ] ) : @xmath368 .
the decay rate for the field @xmath369 reads @xmath370 , where @xmath371 denotes the order in the chapman - enskog expansion .
the corresponding dimensionless decay rate is @xmath372
we shall recall results obtained in @xcite .
the density and temperature of the homogeneous cooling state for pba of hard spheres are given by @xmath373 where the decay exponents are @xmath374 , @xmath375 , and the relaxation time @xmath376 , where @xmath93 and @xmath377 are the decay rates at time @xmath94 .
the subscript @xmath92 denotes a quantity evaluated in the homogeneous state . making use of the explicit values of @xmath93 and @xmath93 found in @xcite and since @xmath378 one obtains @xmath379 \nu_0,\end{aligned}\ ] ] where the kurtosis @xmath160 of the velocity distribution is @xcite @xmath380
the expressions ( [ bigintegrals ] ) and ( [ lastone ] ) may be calculated with the help of the following relations .
let @xmath381 and @xmath382 be arbitrary functions , @xmath125 the maxwellian in the scaling regime , then @xmath383 = \sigma^{d-1 } \phi(x ) v_t^{1-x } \int_{\mathbb{r}^{2d } } d \mathbf{v}_1 d \mathbf{v}_2 \ ,
v_{12}^x f^{(0)}(\mathbf{v}_1 ) \mathcal{m}(\mathbf{v}_2 ) x(\mathbf{v}_2)\left [ y(\mathbf{v}_1 ) + y(\mathbf{v}_2 ) \right ] , \label{lemme1}\ ] ] and @xmath384 = - \sigma^{d-1 } \frac{\phi(x ) v_t^{1-x}}{s_d } \int_{\mathbb{r}^{2d } } d \mathbf{v}_1 d \mathbf{v}_2 \ , v_{12}^x f^{(0)}(\mathbf{v}_1 ) \mathcal{m}(\mathbf{v}_2 ) x(\mathbf{v}_2 ) \int d \widehat{\boldsymbol{\sigma } } ( b-1 ) \left [ y(\mathbf{v}_1 ) + y(\mathbf{v}_2 ) \right ] , \label{lemme2}\ ] ] where @xmath385 - j_a[g , f^{(0)}]$ ] and @xmath386 - j_c[g , f^{(0)}]$ ] for an arbitrary function @xmath362 .
let @xmath387 , then @xmath388}{\gamma(d/2 ) } , \label{app3eq5}\\ \int_{\mathbbm{r}^d } d \mathbf{x } \
, |\mathbf{x}|^n \ , \mathrm{e}^{-\alpha \mathbf{x}^2 } x_i x_j & = & \frac{\pi^{d/2}}{\alpha^{(d+n+2)/2 } } \frac{d+n}{2d } \frac{\gamma\left[(d+n)/2\right]}{\gamma\left[d/2\right ] } \delta_{ij}.\end{aligned}\ ] ] in the integrals below , the results when @xmath389 is absent are obtained upon multiplying the value of @xmath390 by two . finally for @xmath391 , @xmath392 , @xmath393 , @xmath394 , we have @xmath395}{\gamma\left [ ( n+d)/2 \right]}.\label{finally3}\end{aligned}\ ] ]
following the same route as in @xcite we may rewrite the right hand side of eq .
( [ boltzmann1 ] ) such that @xmath95 f^{(1 ) } = a_i \nabla_i \ln t + b_i \nabla_i \ln n + c_{ij } \nabla_i u_j + p \omega f^{(1)},\label{boltzexacta}\ ] ] where @xmath396 - \frac{k_b t}{m } \frac{\partial f^{(0)}}{\partial v_i } , \label{coeff1 } \\ b_i & = & - v_i f^{(0 ) } - \frac{k_b t}{m } \frac{\partial f^{(0)}}{\partial v_i } , \label{coeff2 } \\ c_{ij } & = & \frac{\partial}{\partial v_i}[v_j f^{(0 ) } ] - \frac{1}{d } \frac{\partial}{\partial v_k } [ v_k f^{(0 ) } ] \delta_{ij } , \label{coeff3}\end{aligned}\ ] ] and @xmath133 is a linear operator defined by @xmath397 - \frac{\partial f^{(0)}}{\partial v_i } v_t \xi_{u_i}^{(1)}[f^{0},g ] + \frac{\partial f^{(0)}}{\partial t } t \xi_t^{(1)}[f^{(0)},g ] .
\label{defomega}\ ] ] the quantity @xmath362 is either @xmath101 , @xmath102 , or @xmath103 , and the functionals @xmath141 , @xmath142 , and @xmath143 are obtained from eqs .
( [ decay1 ] ) upon replacing @xmath99 by @xmath362 . integrating the boltzmann equation ( [ boltzexacta ] ) over @xmath104 with weight @xmath134 and taking into account the symmetry properties of the coefficients ( [ coeff ] )
one obtains @xmath398 + ( 1-p ) \int_{\mathbbm{r}^d } d \mathbf{v } \ , m v_i v_j l_c[f^{(0)},f^{(1 ) } ] = \int_{\mathbbm{r}^d } d \mathbf{v } \ , m v_i v_j c_{kl}(\mathbf{v } ) \nabla_k u_l , \label{eqexact}\ ] ] where we have made use of the definition ( [ pressure1 ] ) for the pressure tensor .
the same definition further allows us to write @xmath399 = \xi_n^{(0 ) } p_{ij}^{(1)}(\mathbf{r},t ) , \label{eqexact1}\ ] ] and using additionally eq .
( [ lemme2 ] ) , and ( [ finally1 ] ) to ( [ finally3 ] ) : @xmath400 = \xi_n^{(0 ) } \frac{2}{d+2 } p_{ij}^{(1)}(\mathbf{r},t ) .
\label{eqexact2}\ ] ] finally @xmath401 where @xmath402 insertion of eqs .
( [ eqexact1 ] ) to ( [ eqexact3 ] ) in ( [ eqexact ] ) yields @xmath403 p_{ij}^{(1)}(\mathbf{r},t ) = - p^{(0 ) } \delta_{ijkl } \nabla_k u_l . \label{eqexact-2}\ ] ] the solution of eq .
( [ eqexact-2 ] ) is @xmath404 .
functional dependence analysis shows that @xmath405 , and since to zeroth order the temperature is conserved @xmath406 .
( [ eqexact-2 ] ) thus gives @xmath407 integrating the boltzmann equation ( [ boltzexacta ] ) over @xmath104 with weight @xmath408 and taking into account the symmetry properties of the coefficients ( [ coeff ] ) one obtains @xmath409 + ( 1-p ) \int_{\mathbbm{r}^d } d \mathbf{v } \ , \frac{1}{2 } m v^2 v_i l_c[f^{(0)},f^{(1 ) } ] \\
= \int_{\mathbbm{r}^d } d \mathbf{v } \
, \frac{1}{2 } m v^2 v_i a_{k}(\mathbf{v } ) \nabla_k \ln t + \int_{\mathbbm{r}^d } d \mathbf{v } \ , \frac{1}{2 } m v^2 v_i b_{k}(\mathbf{v } )
\nabla_k \ln t , \label{eqexact--2}\end{gathered}\ ] ] where we have made use of the definition ( [ heat ] ) for the heat flux to first order .
moreover , one finds @xmath410 = \xi_n^{(0 ) } q_{i}^{(1)}(\mathbf{r},t ) , \label{eqexact1 - -2}\ ] ] and using additionally eq .
( [ lemme2 ] ) , and ( [ finally1 ] ) to ( [ finally3 ] ) : @xmath411 = \frac{2(d-1)}{d(d+2 ) } \xi_n^{(0 ) } q_{i}^{(1)}(\mathbf{r},t ) .
\label{eqexact2 - -2}\ ] ] finally @xmath412 @xmath413 insertion of eqs .
( [ eqexact12 ] ) to ( [ eqexact3b2 ] ) in ( [ eqexact2 ] ) gives @xmath414 q_{i}^{(1)}(\mathbf{r},t ) = - \frac{d+2}{2 } \frac{p^{(0 ) } k_b}{m } \nabla_i t. \label{eqexact-2 - -2}\ ] ] the solution of eq .
( [ eqexact-22 ] ) is @xmath415 .
functional dependence analysis shows that @xmath416 and @xmath417 , therefore @xmath418 . in order to satisfy eq .
( [ eqexact-22 ] ) it is therefore required that @xmath419 and @xmath420 , edited by t. pschel and s. luding , lecture notes in physics vol 564 ( springer - verlag , berlin , 2001 ) .
dufty , recent .
* 2 * , 21 - 52 ( 2002 ) ( e - print : cond - mat/0108444 ) ; j.w .
dufty , adv .
syst . * 4 * , 397 - 406 ( 2001 ) ( e - print : cond - mat/0109215 ) .
n.v . brilliantov and t. pschel , _ kinetic theory of granular gases _
( oxford university press , 2004 ) .
dufty and j.j .
brey , e - print : cond - mat/0410133 .
m.h . ernst and e.m .
hendriks , phys .
* 70a * , 183 ( 1979 ) ; e.m .
hendriks and m.h .
ernst , physica * 120a * , 545 ( 1983 ) .
ernst , phys .
rep . * 78 * , 1 ( 1981 ) .
u. marini bettolo marconi and a. puglisi , phys .
e * 66 * , 011301 ( 2002 ) .
a. baskaran and j.w .
dufty , e - print : cond - mat/0410084 .
e. ben - naim and j. machta , cond - mat/0411743 .
a. santos , physica a * 321 * , 442 ( 2003 ) .
dufty , a. baskaran , and l. zogaib , phys .
e * 69 * , 051301 ( 2004 ) . v. garz and a. astillero , e - print : cond - mat/0404386 .
v. garz and j.m .
montanero , e - print : cond - mat/0411221 .
richardson , j. stat .
phys . * 89 * , 777 ( 1997 ) .
bogolubov , in _ studies in statistical mechanics _ , edited by j. de boer and g.e .
uhlenbeck , vol .
i ( north - holland publ .
co. , amsterdam , 1962 ) , p.5 .
cohen , in _ statistical mechanics of equilibrium and non - equilibrium _ , edited by j. meixner ( north - holland publ .
co. , amsterdam , 1962 ) , p.140 .
s. chapman and t.g .
cowling , _ the mathematical theory of non - uniform gases _ ( cambridge university press , cambridge , 1970 ) . y. elskens and h.l .
frisch , phys .
a * 31 * , 3812 ( 1985 ) .
e. ben - naim , s. redner , and f. leyvraz , phys .
lett . * 70 * , 1890 ( 1993 ) ; e. ben - naim , p.l .
krapivsky , f. leyvraz , and s. redner , j. phys .
chem . * 98 * , 7284 ( 1994 ) .
j. piasecki , phys .
e * 51 * , 5535 ( 1995 ) .
rey , m. droz , and j. piasecki , phys .
e * 57 * , 138 ( 1998 ) .
blythe , m.r .
evans , and y. kafri , phys .
lett . * 85 * , 3750 ( 2000 ) .
krapivsky and c. sire , phys .
lett . * 86 * , 2494 ( 2001 ) .
j. piasecki , e. trizac , and m. droz , phys .
rev . e * 66 * , 066111 ( 2002 ) .
e. trizac , phys .
lett . * 88 * , 160601 ( 2002 ) .
f. coppex , m. droz , j. piasecki , e. trizac , and p. wittwer , phys .
e * 67 * , 021103 ( 2003 ) .
f. coppex , m. droz , and e. trizac , phys .
e * 69 * , 011303 ( 2004 ) .
f. coppex , m. droz , and e. trizac , phys .
e * 70 * , 061102 ( 2004 ) .
e. trizac and p.l .
krapivsky , phys .
lett . * 91 * , 218302 ( 2003 ) .
a. santos and j.j .
brey , phys .
fluids * 29 * , 1750 ( 1986 ) .
r. soto , m. mareschal , and d. risso , phys .
lett . * 83 * , 5003 ( 1999 ) .
brey , j.w .
dufty , c.s .
kim , and a. santos , phys .
rev . e * 58 * , 4638 ( 1998 ) . j. ferziger and h. kaper , _ mathematical theory of transport process in gases _ , north - holland , amsterdam ( 1972 ) . v. garz and a. santos ,
_ kinetic theory of gases in shear flow _ ( kluver academic publisher , 2003 ) . f. coppex , m. droz , j. piasecki , and e. trizac , physica a * 329 * , 114 ( 2003 ) .
montanero and a. santos , granular matter * 2 * , 53 ( 2000 ) .
transient instabilities were also predicted for viscoelastic granular gases with velocity - dependent restitution coefficient , see n. brilliantov , c. saluea , t. schwager , and t. pschel , phys .
lett . * 93 * , 134301 ( 2004 ) .
i. goldhirsch and g. zanetti , phys .
lett . * 70 * , 1619 ( 1993 ) .
zuilhof , e.g.d .
cohen , and i.m .
de schepper , phys .
a * 103 * , 120 ( 1984 ) ; e.g.d .
cohen , i.m .
de schepper , and m.j .
zuilhof , physica b * 127 * , 282 ( 1984 ) ; e.g.d .
cohen , physica a * 194 * , 229 ( 1993 ) .
van well , p. verkerk , l.a .
de graaf , j .- b .
suck , and j.r.d .
copley , phys .
a * 31 * , 3391 ( 1985 ) ; a.a .
van well and l.a .
de graaf , phys .
rev . a * 32 * , 2396 ( 1985 ) .
mcgreevy and e.w.j .
mitchell , j. phys . c * 18 * , 1163 ( 1985 ) .
weiss , o .- j .
chen , e.e .
alp , s .- h .
chen , and h.w .
huang , biophys .
j. * 84 * , 3767 ( 2003 ) . | the hydrodynamic description of probabilistic ballistic annihilation , for which no conservation laws hold , is an intricate problem with hard sphere - like dynamics for which no exact solution exists .
we consequently focus on simplified approaches , the maxwell and very hard particles ( vhp ) models , which allows us to compute analytically upper and lower bounds for several quantities .
the purpose is to test the possibility of describing such a far from equilibrium dynamics with simplified kinetic models .
the motivation is also in turn to assess the relevance of some singular features appearing within the original model and the approximations invoked to study it .
the scaling exponents are first obtained from the ( simplified ) boltzmann equation , and are confronted against monte carlo simulation ( dsmc technique ) .
then , the chapman - enskog method is used to obtain constitutive relations and transport coefficients .
the corresponding navier - stokes equations for the hydrodynamic fields are derived for both maxwell and vhp models .
we finally perform a linear stability analysis around the homogeneous solution , which illustrates the importance of dissipation in the possible development of spatial inhomogeneities . |
high transverse momentum jets produced in heavy - ion collisions represent a valuable tool for studies of the properties of the hot parton plasma produced in the central rapidity region @xcite .
this is due to the fact that jets couple to the plasma causing the jet to broaden in momentum space and to lose energy .
the magnitude of momentum - space broadening and energy loss experienced by a parton depends on whether or not one assumes the matter to be hadronic or partonic in nature .
hence , it is one of the primary observables to ascertain experimentally whether or not the plasma has been produced . at very high energies
it is expected that hard bremsstrahlung processes dominate the light quark or gluon energy loss @xcite ; however , at intermediate energies the inclusion of both collisional and radiative processes is necessary in order to make phenomenological predictions .
here we present first results from real - time solution of the su(2 ) boltzmann - vlasov equation for locally isotropic and anisotropic plasmas which include both hard ( scattering ) and soft ( classical field ) processes .
we first demonstrate that in a locally isotropic plasma , one can obtain a cutoff - independent transport coefficient @xmath0 , which measures the square of the transverse momentum transfer per mean free path .
this measurement lays the groundwork for determining other jet transport properties like the energy - loss spectrum from real - time simulations .
we then perform a similar measurement in a locally anisotropic plasma and demonstrate that for a non - abelian plasma unstable modes can cause asymmetric broadening of jets . this may be relevant for recent measurements of di - hadron correlations which provide evidence for an _ asymmetric _ broadening of jet profiles in the plane of pseudorapidity ( @xmath7 ) and azimuthal angle ( @xmath8 ) , with @xmath9 , which has been called `` the ridge '' by experimentalists @xcite . here
we will show that this asymmetry could partly be caused by unstable plasma modes which are induced by the longitudinal expansion of the plasma . at the earliest times after an ultrarelativistic heavy - ion collision , before thermalization and hydrodynamic expansion , the plasma undergoes rapid longitudinal expansion .
this expansion can lead to an oblate anisotropic ( @xmath10 ) momentum distribution in the local rest frame @xcite .
it has been shown that instabilities develop @xcite in such anisotropic plasmas which lead to the formation of long - wavelength chromo - magnetic and chromo - electric fields .
these fields can then affect the propagation of hard jets and their induced hard radiation field .
in the abelian case soft transverse magnetic fields , @xmath11 , dominate all other field components and therefore in this case one expects the longitudinal pressure induced by unstables modes to be larger than the transverse pressure @xcite .
this pressure asymmetry causes asymmetric broadening of jets with larger broadening along the longitudinal or @xmath7 direction .
the situation is more complicated in non - abelian plasmas since then , in addition to generating large coherent @xmath11 and @xmath12 domains , one also generates large - amplitude longitudinal fields @xmath13 and @xmath14 .
it is therefore not obvious a priori that the pressure generated by other field components will result in an asymmetric broadening of the jet . through our numerical simulations we find that for oblate parton momentum distributions , that at different times either @xmath3 or @xmath15 with the net effect being a factor of 1.5 stronger longitudinal than transverse broadening
we solve the classical transport equation for hard gluons with su(2 ) color charge @xmath16 @xcite , with the color generators @xmath17 , including hard binary collisions @xmath18f={\cal c}\,,\end{aligned}\ ] ] where @xmath19 denotes the single - particle phase space distribution .
it is coupled self - consistently to the yang - mills equation for the soft gluon fields , @xmath20 with @xmath21 .
when the phase - space density is parametrically small , @xmath22 , the collision term is given by @xmath23 .
the matrix element @xmath24 includes all @xmath25 tree - level diagrams and color factors as appropriate for the su(2 ) gauge group .
we employ the test particle method to replace the continuous distribution @xmath26 by a large number of test particles , which leads to wong s equations @xcite @xmath27\,,\label{wong3}\end{aligned}\ ] ] for the @xmath28-th test particle , whose coordinates are @xmath29 , @xmath30 , and @xmath31 .
the time evolution of the yang - mills field is determined by the standard hamiltonian method @xcite in @xmath32 gauge .
our lattices have periodic boundary conditions and a lattice spacing @xmath33 , whose physical value is fixed by the length of the lattice @xmath34 .
dimensionless lattice variables scale such that when the number @xmath35 of lattice sites is changed , @xmath36 remains fixed ; see @xcite for details .
the theory without collisions as given by equations ( [ wong1]-[wong3 ] ) coupled to the lattice yang - mills equations was first solved in ref .
@xcite to study chern - simons number diffusion in non - abelian gauge theories at finite temperature .
it was applied later also to the problem of gauge - field instabilities in anisotropic su(2 ) plasmas @xcite .
our numerical implementation is based on the improved formulation detailed in ref .
@xcite where the non - abelian currents generated by the hard particle modes on the lattice sites are `` smeared '' in time .
this technique makes simulations in three dimensions on large lattices possible in practice . in this paper
we go beyond those earlier simulations by accounting also for hard ( short - distance ) collisions among particles .
the collision term is incorporated using the stochastic method @xcite .
scattering processes are determined by sampling possible transitions according to the collision rate in a lattice cell : @xmath37 with @xmath38 , where @xmath39 is the invariant mass of a gluon pair .
the total cross section is given by @xmath40 the momentum transfer is determined in the center of mass frame of the two colliding particles from the probability distribution @xmath41 in eq .
( [ totcross ] ) we have introduced an infrared cutoff @xmath42 for point - like binary collisions . to avoid double - counting , this cutoff should be on the order of the hardest field mode that can be represented on the given lattice , @xmath43 .
momentum transfers below @xmath42 are mediated by the classical yang - mills field ; a soft scattering corresponds to deflection of a particle in the field of the other(s ) .
note , that we use the color averaged expression for the collision term .
the color charge of a particle is hence not affected by a hard collision .
physically , the separation scale @xmath42 should be sufficiently small so that the soft field modes below @xmath42 are highly occupied @xcite .
on the other hand , @xmath42 should be sufficiently large to ensure that hard modes can be represented by particles and that collisions are described correctly by eq.([cterm1 ] ) .
furthermore , unstable modes arise in anisotropic plasmas ( see below ) , all of which should be located below @xmath42 . since @xmath44 , in practice , we choose @xmath45 to be on the order of the temperature for isotropic systems , and on the order of the hard transverse momentum scale for anisotropic plasmas . independently , one should have @xmath46 and @xmath47 : the first condition ensures that the relevant soft modes actually fit on the lattice while the latter corresponds to the continuum limit . here
, @xmath48 denotes the soft scale and is given by @xmath49 where @xmath50 is the number of colors and @xmath51 denotes the number density of hard gluons , summed over two helicities and @xmath52 colors .
also , @xmath53 is the typical momentum of a hard particle from the medium . as we have argued above
, we shall choose the inverse lattice spacing to be on the order of the temperature of the medium .
thus , with ( [ minf ] ) the continuum condition @xmath54 roughly translates into @xmath55 in order to satisfy this relation with @xmath44 , in our numerical simulations below we shall assume an extremely hot medium , @xmath56 . however , this should be viewed simply as a numerical procedure which ensures that the simulations are carried out near the continuum ( or weak - coupling ) limit .
we shall verify below that transverse momentum broadening of a high - energy jet passing through a thermal medium is independent of @xmath57 if the density and the ratio of jet momentum to temperature is fixed ; compare to eq.([<kt2>ll ] ) below .
one may therefore obtain a useful `` weak - coupling '' estimate of @xmath58 ( resp . for the related transport coefficient @xmath0 ) by extrapolating our measurements down to realistic temperatures .
we first consider a heat - bath of particles with a density of @xmath59 and an average particle momentum of @xmath60 gev .
the rather extreme `` temperature '' is chosen to satisfy the above conditions on @xmath61 lattices , assuming @xmath62 fm . for a given lattice ( resp .
@xmath42 ) we take the initial energy density of the thermalized fields to be @xmath63 where @xmath64 is a bose distribution normalized to the assumed particle density @xmath51 , and @xmath65 is the riemann zeta function .
this is equivalent to the energy density of bose - distributed particles with momenta below the separation momentum @xmath42 .
the initial spectrum is fixed to coulomb gauge and @xmath66 ( in continuum notation ) ; also , for simplicity we set @xmath67 at the initial time but electric fields build up quickly within just a few time steps .
we then measure the momentum broadening @xmath68 of high - energy test particles ( @xmath69 ) passing through this medium .
[ fig : ptnocoll ] shows that in the collisionless case , @xmath70 , the broadening is stronger on larger lattices , which accommodate harder field modes
. however , fig .
[ fig : ptcoll ] demonstrates that collisions with momentum exchange larger than @xmath71 compensate for this growth and lead to approximately lattice - spacing independent results even when @xmath42 varies by a factor of four .
gev , @xmath72 , @xmath50 , @xmath59 , @xmath73fm.,width=264 ] .
same parameters as in fig .
[ fig : ptnocoll].,width=264 ] figs .
[ fig : ptnocoll ] and [ fig : ptcoll ] show that the relative contributions to @xmath58 from soft and hard exchanges can depend significantly on @xmath42 , even for @xmath74 .
it is clear , therefore , that transport coefficients obtained in the leading logarithmic ( ll ) approximation from the pure boltzmann approach ( without soft fields ) will be rather sensitive to the infrared cutoff @xmath42 . fitting the difference of fig .
[ fig : ptcoll ] and fig .
[ fig : ptnocoll ] ( i.e. , the hard contribution ) to the ll formula @xmath75 gives @xmath76 , @xmath77 , @xmath78 for @xmath79 , @xmath80 , @xmath81 , respectively .
for the full calculation @xmath82 .
a related and frequently used transport coefficient is @xmath83 @xcite .
it is the typical momentum transfer ( squared ) per collision divided by the mean - free path , which is nothing but @xmath84 . from fig .
[ fig : ptcoll ] , @xmath85 gev@xmath86/fm for @xmath50 , @xmath87@xmath88 and @xmath89 .
our cut - off independent value for @xmath90 is in the range extracted from phenomenological analyses of jet - quenching data from rhic @xcite . in fig.[fig :
qhat_k ] we show that @xmath0 is indeed largely independent of the separation scale @xmath42 . in these simulations , test particles
were explicitly bunched into colorless jets , such that radiative energy loss does not contribute .
this explains why one does not need bremsstrahlung processes in the collision term to obtain a cutoff independent result .
note that the magnitude of @xmath91 gev@xmath92fm at @xmath93/@xmath88 and @xmath94 is smaller than the extrapolation obtained from eq .
( [ < kt2>ll ] ) with the constant under the logarithm fixed from the previous runs at @xmath95/@xmath88 and @xmath96 ; hence , @xmath97 effectively depends on the density and the jet momentum . as a function of @xmath42 at fixed @xmath93/@xmath88 and @xmath94.,width=264 ] in fig .
[ fig : qhat_t ] we verify that @xmath90 does not depend on the temperature @xmath57 so long as the particle density @xmath51 and the ratio of jet momentum to temperature , @xmath98 , is fixed .
thus , our measurement of @xmath99 gev@xmath86/fm may be considered as a weak - coupling extrapolation to realistic temperatures of @xmath100 mev and jet momenta of about 4.5 gev . as a function of @xmath57 at fixed @xmath95/@xmath88 and @xmath96.,width=283 ]
in heavy - ion collisions , locally anisotropic momentum distributions can emerge due to the longitudinal expansion .
such anisotropies generically give rise to instabilities @xcite ; see @xcite for simulations of unstable non - abelian plasmas within the present `` wong - yang - mills '' approach . here , we investigate their effect on the momentum broadening of jets , including the effect of collisions .
the initial momentum distribution for the hard plasma gluons is taken to be @xmath101 with @xmath102 .
this represents a quasi - thermal distribution in two dimensions with average momentum @xmath103 .
we initialize small - amplitude fields sampled from a gaussian distribution and set @xmath104 , for the reasons alluded to above .
the band of unstable modes is located below @xmath42 .
we find that binary collisions among hard particles reduce the growth rate of unstable field modes , in agreement with expectations @xcite .
however , for @xmath105 gev , @xmath106 fm , @xmath59 , @xmath72 , @xmath107/fm and @xmath108 , the reduction of the growth rate is only approximately @xmath109 , increasing to about @xmath110 when @xmath111 .
this is due to fewer available field modes and more randomizing collisions .
next , we add additional high momentum particles with @xmath112 and @xmath113 , respectively , to investigate the broadening in the @xmath114 and @xmath115 directions via the variances @xmath116 the quantity @xmath117 can be roughly associated with the ratio of jet correlation widths in azimuth and rapidity : @xmath118 . is directed along the beam axis , @xmath119 is transverse to the beam .
anisotropic plasma , @xmath120 lattice.,width=264 ] fig .
[ fig : pxyjet192 ] shows the time evolution of @xmath121 and of @xmath122 .
the strong growth of the soft fields sets in at about @xmath123 and saturates around @xmath124 due to the finite lattice spacing ( also see @xcite ) . outside the above time interval the ratio @xmath125 . during the period of instability , however , @xmath126 for both jet energies shown in fig .
[ fig : pxyjet192 ] .
we find approximately the same ratio for denser plasmas ( @xmath127 and @xmath128 ) .
reducing the number of lattice sites and scaling @xmath129 down to 8 gev gives @xmath130 .
these latter runs are rather far from the continuum limit and lattice artifacts are significant @xcite .
the explanation for the larger broadening along the beam axis is as follows . in the abelian case the instability generates predominantly transverse magnetic fields which deflect the particles in the @xmath115-direction @xcite . ) at various times ; the gauge potentials were transformed to coulomb gauge.,width=264 ] in the non - abelian case , however , on three - dimensional lattices transverse magnetic fields are much less dominant ( see , e.g. fig . 5 in @xcite ) although they do form larger coherent domains in the transverse plane at intermediate times than @xmath12 , fig .
[ fig : fts ] . longitudinal fields and locally non - zero chern - simons number @xmath131 emerge , also .
nevertheless , fig .
[ fig : ratios ] shows that @xmath3 , aside from @xmath2 .
hence , the field configurations are such that particles are deflected preferentially in the longitudinal @xmath115-direction ( to restore isotropy ) .
a third contribution to @xmath132 broadening in an expanding plasma , not considered explicitly here , is due to a longitudinal collective flow field which `` blows '' the jet fragments to the side @xcite .
this mechanism is also available for collision dominated plasmas with ( nearly ) isotropic momentum distribution .
however , rather strong flow gradients seem to be required to reproduce the observed broadening of midrapidity jets ( the flow velocity has to vary substantially _ within _ the narrow jet cone ) . in contrast
, color fields will naturally deflect particles with lower momentum by larger angles ( @xmath133 ) : the jet profile broadens even if the induced radiation is exactly collinear .
it is therefore important to determine , experimentally , whether the asymmetric broadening is related to the _ macroscopic _ collective flow or to an anisotropy of the plasma in the local rest frame .
more detailed simulations should account , also , for the fact that small-@xmath134 gluons are already correlated over large rapidity intervals at the initial time @xcite .
b.s . thanks o. fochler , p. romatschke and z. xu for many useful discussions .
the simulations were performed at the center for scientific computing of goethe university . b.s . also thanks the institute for nuclear theory at the university of washington for its hospitality . m.s . and b.s .
are supported by dfg grant gr 1536/6 - 1 .
r. baier , y. l. dokshitzer , a. h. mueller , s. peigne and d. schiff , nucl .
b * 484 * , 265 ( 1997 ) .
f. wang [ star collaboration ] , j. phys .
g * 30 * , s1299 ( 2004 ) ; j. adams _ et al .
_ [ star collaboration ] , phys .
lett . * 95 * , 152301 ( 2005 ) , phys .
c * 73 * , 064907 ( 2006 ) ; j. putschke , j. phys .
g * 34 * , s679 ( 2007 ) .
r. baier , a. h. mueller , d. schiff and d. t. son , phys .
b * 502 * , 51 ( 2001 ) .
p. romatschke and m. strickland , phys .
rev . * d68 * , 036004 ( 2003 ) ; p. arnold , j. lenaghan , and g. d. moore , jhep * 08 * , 002 ( 2003 ) ; s. mrowczynski , acta phys .
b * 37 * , 427 ( 2006 ) .
p. romatschke , phys .
rev . * c75 * , 014901 ( 2007 ) ; p. romatschke and m. strickland , phys .
d * 71 * , 125008 ( 2005 ) .
a. majumder , b. mller , and s. a. bass , phys .
lett . * 99 * , 042301 ( 2007 ) .
u. w. heinz , phys .
lett . * 51 * , 351 ( 1983 ) ; ann .
phys . * 161 * , 48 ( 1985 ) ; _ ibid . _
* 168 * , 148 ( 1986 ) ; h .- t . elze and u. w. heinz , phys .
rept . * 183 * , 81 ( 1989 ) ; j. p. blaizot and e. iancu , phys .
lett . * 70 * , 3376 ( 1993 ) ; nucl .
phys . * b417 * , 608 ( 1994 ) ; p.
f. kelly , q. liu , c. lucchesi , and c. manuel , phys .
lett . * 72 * , 3461 ( 1994 ) ; phys .
rev . * d50 * , 4209 ( 1994 ) ; d. f. litim and c. manuel , phys .
lett . * 82 * , 4981 ( 1999 ) ; p. arnold , g. d. moore and l. g. yaffe , jhep * 0301 * , 030 ( 2003 ) . s. k. wong , nuovo cim .
* a65 * , 689 ( 1970 ) .
j. ambjorn , t. askgaard , h. porter , and m. e. shaposhnikov , nucl .
. * b353 * , 346 ( 1991 ) ; a. krasnitz and r. venugopalan , nucl .
b * 557 * , 237 ( 1999 ) .
a. dumitru and y. nara , phys .
* b621 * , 89 ( 2005 ) . c. r. hu and b. mller , phys .
b * 409 * , 377 ( 1997 ) ; g. d. moore , c. hu and b. mller , phys . rev .
d * 58 * , 045001 ( 1998 ) .
p. danielewicz and g. f. bertsch , nucl .
* a533 * , 712 ( 1991 ) ; a. lang , h. babovsky , w. cassing , u. mosel , h. reusch , and k. weber , j. comp .
* 106 * , 391 ( 1993 ) ; z. xu and c. greiner , phys .
rev . * c71 * , 064901 ( 2005 ) . | we perform numerical simulations of the su(2 ) boltzmann - vlasov equation including both hard elastic particle collisions and soft interactions mediated by classical yang - mills fields . using this technique
we calculate the momentum - space broadening of high - energy jets in real - time for both locally isotropic and anisotropic plasmas . in both cases
we introduce a separation scale which separates hard and soft interactions and demonstrate that our results for jet broadening are independent of the precise separation scale chosen . for an isotropic plasma
this allows us to calculate the jet transport coefficient @xmath0 including hard and soft non - equilibrium dynamics . for an anisotropic plasma
the jet transport coefficient becomes a tensor with @xmath1 .
we find that for weakly - coupled anisotropic plasmas the fields develop unstable modes , forming configurations where @xmath2 and @xmath3 which lead to @xmath4 .
we study whether the effect is strong enough to explain the experimental observation that high - energy jets traversing the plasma perpendicular to the beam axis experience much stronger broadening in rapidity , @xmath5 , than in azimuth , @xmath6 . |
cellular automata ( ca ) has been one of the most studied fields in the research of complex systems .
various patterns has been generated by choosing the rule tables .
wolfram@xcite has classified these patterns into four rough categories : class i ( homogeneous ) , class ii ( periodic ) , class iii ( chaos ) and class iv ( edge of chaos ) .
the class iv patterns have been the most interesting target for the study of ca , because it provides us with an example of the self - organization in a simple system and it is argued that the possibility of computation is realized by the complexity at the edge of chaos@xcite . + a more detailed classification of ca , has been carried out mainly for the elementary cellular automata ( 3-neighbor and 2-state ca)@xcite , in which the pattern is studied quite accurately for each rule table .
however the number of rule table in n - neighbor and k - state cellular automata grows like @xmath3 .
therefore except for a few smallest combinations of the @xmath4 and @xmath5 , the numbers of the rule tables become so large that studies of the ca dynamics for all rule tables are impossible even with the fastest supercomputers .
+ on the other hand the rule table of edge of chaos is rare in the whole ca rule table space , therefore it is important to find a set of parameters by which the pattern classes could be classified , and to determine a phase diagram of ca in these parameter space , even if it is a qualitative one .
+ langton has introduced @xmath1 parameter and argued that as @xmath1 increases the pattern class changes from class i to class ii and then to class iii . and in many cases , class iv behavior is observed between class ii and class iii patterns@xcite .
the @xmath1 parameter represents rough behavior of ca in the rule table space , but finally does not sufficiently classify the quantitative behavior of ca .
it is well known that different pattern classes coexist at the same @xmath1 .
which of these pattern classes is chosen , depends on the random number .
the reason or mechanism for this is not yet known ; we have no way to control the pattern classes at fixed @xmath1 . and
the transitions between a periodic to chaotic pattern classes are observed in a rather wide range of @xmath1 . in ref.@xcite ,
a schematic phase - diagram was sketched .
however a vertical axis was not specified .
therefore , it is has been thought that more parameters are necessary to arrive at a more quantitative understanding of the rule table space of the ca .
+ in this article , we will report a mechanism which discriminates the pattern classes at a fixed @xmath1 .
the mechanism is closely related to the structure of the rule tables and is expressed by the numbers of rules which breaks strings of quiescent state . for n - neighbor and k - state ca
, it is found that in the region @xmath6 , the class ii , class iii and class iv pattern classes coexist .
+ this property is studied quantitatively by introducing a new parameter @xmath2 , which is taken to be orthogonal to @xmath1 . in the region
@xmath6 , the maximum of @xmath2 correspond to class iii rule tables while minimum of @xmath2 , to class ii or class i rule tables .
therefore the transition of the pattern classes takes place somewhere between these two limits without fail .
if we determine the region of @xmath2 , where the transitions of the pattern classes take place , we could obtain the phase diagram in @xmath7 plane .
+ the determination of the phase diagram is carried out for 5-neighbor and 4-state ca . in this case ,
phase boundary is not sharp but has some range in @xmath7 plane .
the region has a gentle slope as a function of @xmath1 , and extends over the range @xmath8 .
this means that for this ca the edge of chaos could be found at least in this range in @xmath1 .
+ in section 2 , we will briefly summarize our notations and present a key discovery , which leads us to the understanding of the structure of the rule table and pattern classes .
it strongly suggested that the rules which break strings of the quiescent states play an important role for the pattern classes .
+ in section 3 , we classify rule tables according to the destruction and construction of strings of the quiescent states , and carry out the replacements of the rules to change the chaotic pattern class into periodic one and vice versa while keeping @xmath1 fixed .
the reason why the patten classes changes by the replacements is discussed , and we will show that in the region @xmath9 , the change of the pattern classes takes place without fail by the replacements .
+ in section 4 , the result obtained in section 3 is studied quantitatively by introducing a new parameter @xmath2 . using @xmath2 and @xmath1
, we determine the phase diagram in the @xmath7 plane for 5-neighbor and 4-state ca .
+ section 5 is devoted to conclusions and discussions .
in order to make our arguments concrete , we focus mainly on the one - dimensional 5-neighbor and 4-state ca in the following , however , the qualitative conclusions hold true for other cas .
this point will be discussed in subsections 3.2 .
+ we will briefly summarize our notation of ca@xcite . in our study , the site consists of 150 cells having the periodic boundary condition .
the states are denoted as @xmath10 .
the @xmath11 represents the time step which takes an integer value , and the @xmath12 is the position of cells which range from @xmath13 to @xmath14 .
the @xmath10 takes values @xmath15 and @xmath16 , and the state @xmath13 is taken to be the quiescent state .
the set of the states @xmath10 at the same @xmath11 is called the configuration .
+ the configuration at time @xmath17 is determined by that of time @xmath11 by using following local relation , @xmath18 the set of the mappings @xmath19 is called the rule table .
the rule table consists of @xmath20 mappings , which are selected from a total of @xmath21 elements .
+ the @xmath1 parameter is defined as@xcite @xmath22 where @xmath23 is the number in which @xmath24 in eq.[r_table ] is not equal to @xmath13 . in other words
the @xmath1 is the probability that the rules do not select the quiescent state in next time step . in the following we set the rule tables randomly with the probability @xmath1 .
we choose @xmath25 rules randomly , and set @xmath26 in the right hand side of eq .
[ lambda ] . for the rest of the @xmath23 rules
, the @xmath24 picks up the number @xmath27 randomly .
the initial configurations are also set randomly .
+ the time sequence of the configurations is called a pattern . the patterns are classified roughly into four classes established by wolfram@xcite . it has been known that as the @xmath1 increases the most frequently generated patterns change from homogeneous ( class i ) to periodic ( class ii ) and then to chaotic ( class iii ) , and at the region between class ii and class iii , the edge of chaos ( class iv ) is located . in order to find the reason why the different pattern classes are generated with the same @xmath1 , we have started to collect rule tables of different pattern classes , and tried to find the differences between them .
we have fixed at @xmath28 @xmath29 , because at this point the chaotic , edge of chaos , and periodic patterns are generated with a similar ratio . by changing the random number , we have gathered a few tens of the rule tables and classified them into chaotic , edge of chaos , and periodic ones .
+ in this article , a pattern is considered the edge of chaos when its transient length@xcite is longer than @xmath30 time steps .
+ first , we study whether or not the pattern classes are sensitive to the initial configurations .
we fix the rule table and change the initial configurations .
the details of the patterns depend on the initial configurations , but the pattern classes are not changed@xcite .
thus the difference of the pattern classes is due to the differences in the rule tables , and the target of our inquiry has to do with the differences between them .
+ for a little while , we do not impose a quiescent condition ( qc)@xcite , + @xmath31 , because without this condition , the structure of the rule table becomes more transparent .
this point will be discussed at the footnote 4 in section 4 .
+ after some trial and error , we have found a strong correlation between the pattern classes and the qc . for class ii patterns , the probability of the rule table , which satisfies the qc is much larger than that of the class iii patterns .
this correlation has suggested that the rule @xmath32 , which breaks the string of the quiescent states with length 5 , pushes the pattern toward chaos .
we anticipate that the similar situation will hold for the strings of quiescent states with length 4 . + we go back to the usual definitions of ca . in the following we discuss ca under qc , @xmath31 .
we study the correlation between the number of the rules of eq.[d4 ] and the pattern classes : @xmath33 these rules break length 4 strings of the quiescent states , and will also push the pattern toward chaos .
we denote the total number of rules of eq.[d4 ] in a rule table as @xmath34 .
we have collected 30 rule tables and grouped them by the number @xmath34 .
we have 4 rule tables with @xmath35 , 13 rule tables with @xmath36 , 9 rule tables with @xmath37 and 4 rule tables with @xmath38 .
when @xmath39 , all rule tables generate chaotic patterns , while when @xmath38 , only periodic ones are generated . at @xmath36 and @xmath37 ,
chaotic , edge of chaos , and periodic patterns coexist .
examples are shown in the fig.[pattern_at_lambda0.44 ] .
the coexistence of three pattern classes at @xmath36 is seen in fig.[pattern_at_lambda0.44](b ) , fig.[pattern_at_lambda0.44](c ) and fig.[pattern_at_lambda0.44](d ) and that of @xmath37 is exhibited in fig.[pattern_at_lambda0.44](e ) , fig.[pattern_at_lambda0.44](f ) and fig.[pattern_at_lambda0.44](g ) .
+ the strong correlation between @xmath34 and the pattern classes has been observed in this case too , as anticipated .
these discoveries have provided us with a key hint leading us to the hypothesis that the rules , which break strings of the quiescent states , will play a major role in the pattern classes .
in order to test the hypothesis of the previous section , we classify the rules into four groups according to the operation on strings of the quiescent states . in the following , greek characters in the rules represent groups @xmath40 while roman , represent groups @xmath27 .
+ group 1 : @xmath41 .
+ the rules in this group break strings of the quiescent states . + group 2 : @xmath42 .
+ the rules of this group conserve them .
+ group 3 : @xmath43 .
+ the rules of this group develop them .
+ group 4 : @xmath44 .
+ the rules in this group do not affect string of quiescent states in next time step .
+ the sum of the numbers of the group 1 and group 2 rules is 256 , while that of group 3 and group 4 rules is 768 . the number of each group of rules included in the rule table is determined mainly by the probability @xmath1 , therefore it suffers from fluctuation due to randomness . + the group 1 rules are further classified into five types according to the length of string of quiescent states , which they break .
these are shown in table [ destruc ] .
[ destruc ] .
the classification of the rules in group 1 . [ cols="^,^,^,^,^,^,^ " , ] using these results for @xmath45 and @xmath46 , we calculate @xmath2s for each critical combination ( @xmath47 ) . for @xmath48 , they are shown in fig.[phase_diagram_0.6]a for 101 data points .
the number distributions of them are displayed in fig.[phase_diagram_0.6]b . from these 101 data , we calculate average , @xmath49 and standard deviation @xmath50 of the distribution of @xmath47 .
similar calculations are carried out for other @xmath1 points .
the results for @xmath49 , @xmath51 and maximum of @xmath2 ( @xmath52 ) determined by the coefficients @xmath45 and @xmath46 are shown in the fig.[phase_diagram ] .
this is a phase diagram in @xmath7 plane .
+ the phase diagram is natural in the sense that as @xmath1 increases , the proportion of the chaotic rule table region increases in the total @xmath2 range .
the two lines @xmath51 in the figure , indicate that about @xmath53 of the critical combinations are located in this region , if the normal distribution is assumed for @xmath47 .
we call this region as transition region .
we think that the class ii and class iii patterns would not be separated by a line in @xmath7 plane . because the transition region in fig .
[ phase_diagram ] is a mapping of the hyper - surface , which separates the two pattern classes at a fixed @xmath1 in 1024 dimensional space , into the normal line of the regression plane in the 3-dimensional space .
therefore the two parameters will not be enough to draw the boundary by a line .
+ however we should like to notice that in spite of the simple approximations and assumptions , the @xmath47s distribute within a rather narrow portion of total @xmath2 range as shown in fig .
[ phase_diagram ] . at @xmath48 , the total range of @xmath2 is @xmath54 .
all the @xmath47 distribute within the region @xmath55 , which is roughly @xmath56 of the total region .
the region @xmath51 occupies only about @xmath57 of whole range of @xmath2 .
the similar situation is observed for the other @xmath1 points .
+ the fig.[phase_diagram ] is still qualitative but rule table space of ca is classified much better by using @xmath1 and f. and it provides us with a deeper understanding of the structure of the ca rule table space .
the existence of the transition region in @xmath58 , means that the edge of the chaos could be found at least in this region , which is rather wide range in @xmath1 .
+ let us discuss the distribution of the rule tables around the transition region .
all the class iv rule tables are located around transition region as fig .
[ phase_diagram_0.6 ] shows , but converse is not true . in this region ,
the three pattern classes coexist and the number densities of the class ii plus class iii rule table are considerably larger than that of class iv rule tables . outside of the transition region , the probability of finding the class iv rule table decreases rapidly and the pattern classes began to be classified only by the position in the @xmath7 plane ; in the @xmath59 region , class iii rule tables dominates while in the @xmath60 region , class ii or class i ca dominates .
the similar distributions of the rule tables are observed at @xmath61 , 0.75 , 0.7 , 0.5 , 0.4 , 0.3 and 0.2 .
+ we proceed to the investigation of the end points of the transition region . in the subsection 3.2 , we have find the region of @xmath1 , where both chaotic limit and periodic limit coexist . in the 5-neighbor and k - state ca , it is @xmath62 . in @xmath63 region ,
not all the group 2 rules could be replaced by the group 1 rules . therefore the maximum of @xmath64 could not become 256 , and it deceases to zero as @xmath1 approaches to zero .
then the maximum of @xmath2 , ( @xmath52 ) also decreases to zero toward @xmath65 .
+ conversely in @xmath66 region , not all the group 1 rules could be replaced by the group 2 rules .
the minimum of n(g1 ) and therefore the minimum of @xmath2 , ( @xmath67 ) could not becomes 0 .
the line @xmath67 increases until @xmath68 . in fig.[phase_diagram ] , we have schematically shown the @xmath52 and @xmath67 lines with dotted lines . + the points where the transition region crosses the @xmath67 and @xmath52 lines determine the end points of the coexistence of the two limits , and also the end points of existence of the class iv rule tables . we have already found that the transition region is found in the region @xmath8 . therefore the crossing points are outside of this region in the case of 5-neighbor 4-state ca . +
at a fixed @xmath1 , the patten classes of the ca could be controlled by the numbers of the group 1 rules , which has been denoted by @xmath64 .
the maximum of @xmath69 corresponds to chaotic limit , and @xmath70 , to the periodic limit . in the n - neighbor and k - state ca ,
these two limit exist in the region @xmath71 . therefore in this @xmath1 region
, we could control the patten classes by changing @xmath64 without fail .
the method for it , is the replacements of eq .
[ del_num_rp ] .
this provides us with a new method to obtain the rule table of edge of chaos .
however , we should like to comment that the fluctuation due to the random number will make the statements less rigorous . +
this property is studied quantitatively by introducing a new parameter @xmath2 .
the maximum of the @xmath2 corresponds to the chaotic limit and minimum of it , the periodic limit . by changing the @xmath2 using the replacements
, we could find the region , where the transition of the class ii and class iii patterns takes place .
thus we could obtain the phase diagram of the ca in @xmath7 plane .
+ in this article , we have applied the analysis to 5-neighbor and 4-state ca . in this case
the group 1 rules are further classified into 5 types as shown in table [ destruc ] , and the phase diagram is obtained in fig.[phase_diagram ] .
the transition region has a rather gentle slope as a function of @xmath1 , and it extend at least from @xmath72 until @xmath61 , which is wider than @xmath73 .
this explains why the edge of chaos is found in the wide range in @xmath1 for this ca
. it may be interesting whether or not the slope of the transition region depends on the models .
+ in the replacement experiments , we have found the edge of chaos ( very long transient lengths ) in many cases . the examples are shown in fig.[replace1 ] . sometimes they are observed in a rather wide range in @xmath74 or @xmath75 .
this indicates that in many cases , the transitions are second - order like . but the widths in the ranges of @xmath74 or @xmath75 are different from each other , and there are cases where the widths are less than one unit in the replacement of rp2 ( first - order like ) .
it is very interesting to investigate under what condition the transition becomes first - order like or second - order like .
the mechanism of the difference in the nature of the transition is an open problem and may be studied by taking into account effects of group 2 , 3 , and 4 rules . in these studies
another new parameters may be found and a more quantitative phase diagram may be obtained . + it is also interesting to compare the detailed dynamics of class iv ca at different points of rule table space in @xmath7 plane by applying wuencshe s method@xcite or by using computational mechanics@xcite .
+ these issues together with finding the points where the transition region crosses @xmath52 and @xmath67 lines in fig.[phase_diagram ] , and the nature of the phase transition at these points will be addressed in the forthcoming publications .
9 s. wolfram , physica d 10(1984 ) 1 - 35 .
s. wolfram , physica scripta t9(1985 ) 170 - 185 .
c.g.langton , physica d 42(1990 ) 12 - 37 .
m. mitchell , j.p.crutchfield and p.t .
hraber , santa fe institute studies in the science of complexity , proceedings volume 19 .
reading , ma : , addison - wesley .
online paper , http://www.santafe.edu/ mm/ paper-abstracts.html#dyn-comp-edge .
j.e.hanson and j.p.crutchfield , physica d 103(1997),169 - 189 .
a. wuensche , complexity vol.4/no.3(1999 ) 47 - 66 .
c.g.langton , physica d 22(1986 ) 120 - 149 .
w.li , n.h.packard and c.g.langton , physica d45(1990 ) 77 - 94 .
t. suzudo , crystallisation of two - dimensional cellular automata , complexity international , vol .
on line journal , + http://www.csu.edu.au/ci/vol06/suzudo/suzudo.html .
see appendix . | it is shown that for the n - neighbor and k - state cellular automata , the class ii , class iii and class iv patterns coexist at least in the range @xmath0 . the mechanism which determines the difference between the pattern classes at a fixed @xmath1
is found , and it is studied quantitatively by introducing a new parameter @xmath2 . using the parameter f and @xmath1 , the phase diagram of cellular automata is obtained for 5-neighbor and 4-state cellular automata .
+ pacs : 89.75.-k complex systems |
squeezing of the radiation field is one of the distinct features of the quantum theory of light @xcite . squeezed states of light , being nonclassical in origin @xcite , have a reduced variance in one of the quadrature components of the electric field .
it is well known that the resonance fluorescence from a driven atomic system can serve as a source of squeezed radiation .
theoretical investigations on two - level and three - level atoms demonstrated squeezing either in the total variance of phase quadratures or in the frequency ( spectral ) components of the fluorescence radiation @xcite . for a driven two - level atom , walls and zller first predicted that squeezing can occur in the in - phase / out - of - phase quadrature component of the fluorescent light [ 4a ] .
the noise spectrum was shown to exhibit single- and two - mode squeezing in the weak- and strong - excitation regimes @xcite .
some experimental evidences of squeezing have also been reported in the phase - dependent spectra of two - level atoms @xcite .
unlike in two - level systems , two - photon coherences play a significant role in the dynamics of three - level atoms driven by coherent fields @xcite .
_ examined the role of atomic coherences and studied the maximum squeezing that can be obtained in the fluorescence from three - level systems @xcite .
a detailed study by gao _
@xcite has shown that ultranarrow squeezing peaks may appear in the spectrum of driven three - level atoms in @xmath0- and v - type configurations .
however , the role of two - photon coherence is seen to destroy the spectral - component squeezing in the fluorescent field @xcite .
one of the interesting developments in the study of resonance fluorescence is the possibility of modifying spectral properties of the atoms via quantum interferences in spontaneous decay channels .
the interference in spontaneous emission occurs when the atomic transitions are coupled by same vacuum modes .
the early work of agarwal on this subject demonstrated population trapping and generation of quantum coherence between the excited states in a v - type atom @xcite . since the fluorescence properties of a driven atomic system result from its spontaneous emission , studying the influence of interference in such processes has become an important topic of research @xcite .
much attention has been paid to study the fluorescence spectrum of driven atoms @xcite .
all these theoretical studies assume non - orthogonal dipole moments of the atomic transitions for the interference to exist in decay processes @xcite . however , in real atomic systems , it is difficult to meet this condition .
different schemes were later proposed to bypass the condition of non - orthogonal dipole moments @xcite .
experimentally , coherence between the ground states arising from spontaneous emissions has been reported using electron spin polarization states in quantum dots @xcite and zeeman sub - levels in atomic systems @xcite .
the squeezing characteristics of the emitted fluorescent light was also discussed extensively @xcite . in a three - level v - type atom interacting with a coherent field ,
the interference is shown to enhance the spectral component squeezing in the presence of standard / squeezed vacuum @xcite .
_ studied the squeezing spectrum of a four - level atom in the @xmath1-type scheme and showed that unusual squeezing properties appear due to interference if the fluorescent field is detected on the slow decaying transition @xcite .
further , gonzalo _ et al . _
have examined a driven three - level atom of the @xmath1 configuration with particular attention to the squeezing in spectral components @xcite .
recently , the effects of spontaneously generated interferences have been investigated in the context of enhancing self - kerr nonlinearity @xcite , soliton formation @xcite , and preserving bi - partite entanglement @xcite . in this paper
, we consider a four - level atom in the y - type configuration interacting with two coherent fields ( as shown in fig .
the excited atomic states are assumed to be near degenerate and decay spontaneously via the same vacuum modes to the intermediate state .
the atom in the intermediate state can make spontaneous transitions to the ground state .
since the cascade decays from the excited atomic states lead to an emission of the same pair of photons , quantum interference exists in decay processes .
the role of the interference was investigated in the fluorescence spectrum of this system in ref .
@xcite . in the present work ,
we study the squeezing spectrum and examine the interference effects on squeezing properties of the fluorescence fields .
the paper is arranged as follows . in sec .
ii , we present the atomic density matrix equations , describing the interaction of a y - type atom with two coherent fields , when the presence of quantum interference in decay channels is included . the formula for the squeezing spectrum
is then derived using atomic correlation operators in sec iii . in sec .
iv , we analyze the numerical results of the squeezing spectrum and identify the origin of interference effects using the dressed - state picture .
finally , the main results are summarized in sec .
we consider a four - level atom in the y - type configuration as shown in fig .
1 . in this scheme , the atom has two closely lying excited states @xmath2 and @xmath3 with energy separation @xmath4 .
it is assumed that the excited atomic states are coupled by common vacuum modes to decay spontaneously to the intermediate state @xmath5 with rates @xmath6 and @xmath7 .
the atom in the intermediate state @xmath5 is further allowed to undergo spontaneous emissions to the ground state @xmath8 with decay rate @xmath9 .
the direct transitions between the excited states @xmath10 and that between the excited and ground states @xmath11 of the atom are forbidden in the dipole approximation .
we assume that the transition frequencies @xmath12 of the upper transitions differ widely from that of the lower transition @xmath13 .
this leads to a situation in which the vacuum modes coupling the upper and lower atomic transitions are totally different .
in addition to spontaneous decays , two coherent fields are applied on the atom as shown schematically in fig .
1 . the upper transitions @xmath14 are driven by a coherent field of frequency @xmath15 ( amplitude @xmath16 ) and another field of frequency @xmath17 ( amplitude @xmath18 ) couples the lower transition @xmath19 .
the rabi frequencies of the atom - field interaction are denoted as @xmath20 , @xmath21 , and @xmath22 with @xmath23 being the dipole moment of the atomic transition from @xmath24 to @xmath25 .
the system is studied in the interaction picture using time independent hamiltonian @xmath26 here , the operators @xmath27 represent the atomic population operators for @xmath28 and transition operators for @xmath29 . to include decay processes in the dynamics
, we use the master equation framework . with the inclusion of the decay terms
, the time evolution of the density matrix elements in the interaction picture obeys @xcite @xmath30 @xmath31 @xmath32 @xmath33 @xmath34 @xmath35 \rho_{23 } + i \omega_2 ( \rho_{33 } - \rho_{22 } ) \nonumber \\ & & -
i \omega_1 \rho_{21 } - i \omega_3 \rho_{24 } - p \sqrt{\gamma_1 \gamma_2}~ \rho_{13},\end{aligned}\ ] ] @xmath36 @xmath37 \rho_{14 } + i \omega_1 \rho_{34 } - i \omega_3 \rho_{13 } \nonumber \\ & & - p \sqrt{\gamma_1 \gamma_2}~ \rho_{24},\end{aligned}\ ] ] @xmath38 \rho_{24 } + i \omega_2 \rho_{34 } - i \omega_3 \rho_{23 } \nonumber \\ & & - p \sqrt{\gamma_1 \gamma_2}~ \rho_{14}. \label{rho9}\end{aligned}\ ] ] here , @xmath39 corresponds to the detuning between the atomic frequency @xmath40 of the @xmath41 transition and the frequency of the applied field @xmath16 .
similarly , @xmath42 denotes the detuning of the field acting on the lower transition . in writing eqs .
( [ rho1])-([rho9 ] ) , we have assumed that the trace condition @xmath43 is satisfied .
the cross - coupling term @xmath44)-([rho9 ] ) .
it reflects the fact that population can be transferred between the excited states by the vacuum field .
when @xmath45 , the decays from the excited states @xmath2 and @xmath3 are coupled and the interference effects are maximal . if the atomic dipole moments are orthogonal @xmath46 , there is no interference effect in spontaneous emission . for convenience in the calculation of the squeezing spectrum , we rewrite the density matrix equations ( [ rho1])-([rho9 ] ) in a more compact matrix form by the definition @xmath47 substituting eq .
( [ psidef ] ) into eqs .
( [ rho1])-([rho9 ] ) , we get the matrix equation for the variables @xmath48 @xmath49 where @xmath50 is the @xmath51-th component of the column vector @xmath52 and the inhomogeneous term @xmath53 is also a column vector with non - zero components @xmath54 in eq .
( [ matrix ] ) , @xmath55 is a 15@xmath5615 matrix whose elements are time independent and can be found explicitly from eqs .
( [ rho1])-([rho9 ] ) .
the steady - state solutions of the density matrix elements can be obtained by setting the time derivative equal to zero in eq .
( [ matrix ] ) : @xmath57
we now proceed to the study of the squeezing spectra of the driven atom .
since the atom is driven by two coherent fields , each field induces its own atomic dipole moment which then generates a scattered field .
however , the fields scattered by the upper and lower transitions in the atom will have no correlations because the applied fields @xmath58 are of quite different carrier frequencies @xmath59 .
assuming that the point of observation lies perpendicular to the atomic dipole moments , the positive - frequency part of the fluorescent fields in the radiation zone can be written as @xmath60\exp(-i\omega_a\hat{t } ) , \nonumber \\
\vec{e}^{(+)}_b(t ) & = & g(r ) \vec{\mu}_{34 } a_{43}(\hat{t } ) \exp(-i\omega_b\hat{t } ) , \label{electric}\end{aligned}\ ] ] where @xmath61 , @xmath62 , @xmath63 , and @xmath64 is the distance of the detector from the atom .
the index @xmath65 @xmath66 in eq .
( [ electric ] ) refers to the fluorescent light of central frequency @xmath15 @xmath67 . in squeezing measurements
, the two - time expectation value of a particular quadrature component of the electric field is the quantity of interest .
we consider squeezing in the fluorescent light exclusively emitted by the upper and lower transitions in the atom .
the slowly varying quadrature components with phase @xmath68 are defined as @xmath69 the spectrum of squeezing is defined by the fourier transformation of the normal and time - ordered correlation of the quadrature component @xmath70 : @xmath71 where @xmath72 , @xmath73 represents the time ordering operator , and the steady - state limit @xmath74 is considered . the calculations of the correlation functions in eq .
( [ spect ] ) can be carried out easily with the help of the quantum regression theorem and the density matrix equations ( [ matrix ] ) .
for this purpose , we introduce column vectors of two - time averages @xmath75^{t } , m , n = 1,2,3,4 .
\label{corre}\end{aligned}\ ] ] here , @xmath76 are the deviations of the atomic operators from its steady state values eq .
( [ steady ] ) .
according to the quantum regression theorem @xcite , the column vectors ( [ corre ] ) satisfy @xmath77 now , following the procedure explained in refs .
@xcite for the time ordering of operators in eq .
( [ spect ] ) , the squeezing spectrum can be obtained as @xmath78 e^{i2(\theta + \omega_a
r / c ) } \label{specta } \\ & & + \sum_{k = 1}^{15 } \lim_{t \rightarrow \infty } \left [ \hat{m}_{11,k } \left ( { |\vec{\mu}_{13}|}^2 \hat{u}^{31}_k(t,0 ) + \vec{\mu}_{13}.\vec{\mu}_{23 } \right .
. \nonumber \\ & & \left .
\times \hat{u}^{32}_k(t,0 ) \right ) + \hat{m}_{12,k}\left ( { |\vec{\mu}_{23}|}^2 \hat{u}^{32}_k(t,0 ) \right .
\nonumber \\ & & + \left . \left .
\vec{\mu}_{13}.\vec{\mu}_{23}\hat{u}^{31}_k(t,0)\right ) \right ] \bigg\ } , \nonumber\end{aligned}\ ] ] where @xmath79 denotes the @xmath80 element of the matrix @xmath81 $ ] .
similarly , as a function of @xmath82 for the parameters @xmath83 , @xmath84 , @xmath85 , @xmath86 , @xmath87 , and @xmath88 ( a ) and @xmath89 ( b ) .
the solid ( dotted ) curves are for @xmath90 @xmath91 . for clarity , the dotted curve in
( a ) has been displaced by 2 units along the @xmath82-axis.,title="fig:",height=226 ] as a function of @xmath82 for the parameters @xmath83 , @xmath84 , @xmath85 , @xmath86 , @xmath87 , and @xmath88 ( a ) and @xmath89 ( b ) .
the solid ( dotted ) curves are for @xmath90 @xmath91 . for clarity , the dotted curve in
( a ) has been displaced by 2 units along the @xmath82-axis.,title="fig:",height=226 ] as a function of @xmath82 for the parameters @xmath83 , @xmath84 , @xmath92 , @xmath86 , @xmath87 , and @xmath88 .
the solid ( dotted ) curves are for @xmath90 @xmath91 . for clarity , the dotted curve has been displaced by 2 units along the @xmath82-axis.,height=226 ] @xmath93\bigg\ } , \label{spectb}\end{aligned}\ ] ] with the elements of the matrix @xmath94 as defined in eq .
( [ specta ] ) .
the squeezing spectra of the fluorescence fields can be obtained numerically using eqs .
( [ specta ] ) and ( [ spectb ] ) . in the following ,
we assume equal dipole moments @xmath95 and decay rates @xmath96 for the upper atomic transitions . all the frequency parameters such as decay rates , detuning , and
rabi frequencies are scaled in units of @xmath97 .
the numerical results are presented by considering both the presence @xmath98 and absence @xmath91 of quantum interference . * a. squeezing spectrum : @xmath99 * we first consider spectral squeezing in the fluorescence field generated by the upper atomic transitions . a selected quadrature component @xmath68
is said to exhibit spectral squeezing if the squeezing spectrum is negative , @xmath100 , at a certain frequency @xmath82 .
we analyze the squeezing spectrum @xmath99 calculated using eq .
( [ specta ] ) for the case of strong driving fields @xmath101 @xcite . in the calculations , we assume @xmath102 and scale the spectrum in units of @xmath103 .
the numerical results of the spectrum @xmath99 are displayed in fig . 2 for the in - phase @xmath104 quadrature component @xcite .
it is seen that the presence of quantum interference @xmath98 enhances ( three times ) the spectral squeezing @xmath105 in outer sidebands [ see fig .
2(a ) ] . for a suitable choice of parameters as in fig .
3 , the squeezing can be enhanced in inner sidebands as well .
note that the amount of maximal squeezing due to quantum interference is appreciable only when the excited atomic levels decay much slower than the middle level @xmath106 a physical understanding of the numerical results can be obtained if the squeezing spectrum ( [ specta ] ) is rewritten as a sum of two contributions , i.e. , @xmath107 . here
, the terms @xmath108 and @xmath109 referred to as noninterfering and interfering terms , are expressed as @xmath110,\end{aligned}\ ] ] as a function of @xmath82 from different contributions : @xmath111 [ dotted curve ( a ) ] , @xmath112 [ solid curve ( a ) ] , @xmath113 [ dotted curve ( b ) ] , and @xmath114 [ solid curve ( b ) ] . the parameters are those used to produce fig .
2(a ) with @xmath90 . for clarity ,
the dotted curve has been displaced by 2 units along the @xmath82-axis.,title="fig:",height=226 ] as a function of @xmath82 from different contributions : @xmath111 [ dotted curve ( a ) ] , @xmath112 [ solid curve ( a ) ] , @xmath113 [ dotted curve ( b ) ] , and @xmath114 [ solid curve ( b ) ] .
the parameters are those used to produce fig .
2(a ) with @xmath90 . for clarity ,
the dotted curve has been displaced by 2 units along the @xmath82-axis.,title="fig:",height=226 ] to explore further the origin of interference induced effects , we go to the dressed state description of the atom - field interaction .
the dressed states are defined as eigenstates of the time independent hamiltonian @xmath120 in eq .
( [ ham ] ) . in the general parametric conditions , it is difficult to find analytical solutions to the eigenvalue problem @xmath121 .
however , the eigenvalues @xmath122can be obtained numerically by solving a quartic equation .
the eigenstates @xmath123 can be expanded in terms of the bare atomic states as @xmath124 where the expansion coefficients are explicitly given by @xmath125 with the normalization constant @xmath126^{\frac{1}{2 } } .
\nonumber\end{aligned}\ ] ] in order to interpret the numerical results , we consider transitions between the dressed states with the inclusion of decay processes .
the allowed transitions between the dressed states @xmath127 give the peaks in the fluorescence spectrum at the frequencies @xmath128 .
specifically , for the parameters of fig .
2 , the eigenvalues ( in units of @xmath97 ) obtained numerically are @xmath129 , @xmath130 , @xmath131 , and @xmath132 . the squeezing found at the outer sideband in fig .
2 can be seen as arising from the transitions @xmath133 . in the high field limit @xmath134 , the squeezing spectrum @xmath119 can be worked out to be @xmath135 } { \gamma_{\alpha\beta}^2 + { ( \omega \mp \omega_{\alpha \beta})}^2 } , \label{dressa}\ ] ] where the subindex + ( - ) stands for the positive @xmath136 [ negative @xmath137 part of the spectrum and @xmath138 represents the population of the dressed state @xmath139 .
the term @xmath140 denotes the decay rate of coherence between the dressed states and is given explicitly in appendix . by using the numerical values of expansion coefficients ( [ ecoeff ] ) , the formula eq . ( [ dressa ] ) accounts very well for the spectral squeezing displayed in fig .
2 . as a function of @xmath82 for the parameters @xmath83 , @xmath84 , @xmath85 , @xmath141 , @xmath142 , and
@xmath143 the solid ( dotted ) curves are for @xmath90 @xmath91.,height=226 ] * b. squeezing spectrum : @xmath144 * we next consider the spectrum of fluorescence field generated by the lower atomic transitions . the squeezing spectrum calculated using eq .
( [ spectb ] ) is displayed in fig .
5 for the in - phase @xmath104 quadrature @xcite .
the spectrum is scaled in units of @xmath145 .
we also assume @xmath146 and consider the case of fast decaying upper levels @xmath147 of the atom .
the effect of quantum interference @xmath98 is now seen to reduce the squeezing in spectral components as evident from fig .
5 . for slow decay of the upper levels
@xmath148 , the reduction in squeezing is not appreciable ( graph not shown ) . as a function of the interference parameter @xmath149 .
the other parameters for the calculation are the same as in fig .
5.,height=226 ] in order to understand this result , the spectrum is further analyzed in the dressed state picture .
the contribution to the squeezing spectrum ( [ spectb ] ) originating from the dressed - state transitions @xmath150 can be given as @xmath151 , \label{dressb}\end{aligned}\ ] ] where the different terms have their meanings as in eq .
( [ dressa ] ) . for the parameters of fig . 5
, the numerical values of eigenvalues ( in units of @xmath97 ) obtained are @xmath152 , @xmath153 , @xmath154 , and @xmath155 .
the spectrum eq .
( [ dressb])is a pair of lorentzians centered at the frequencies @xmath156 . since the numerators of the lorentzians are negative , the graph ( fig .
5 ) has negative peaks of height proportional to the inverse of decay rate @xmath140.the reduction in squeezing observed in fig .
5 may be traced to the increase in the decay rate @xmath140 of the dressed atomic transition . to this end
we study the variation of the decay rate @xmath140 as a function of the interference parameter @xmath149 . as seen in fig .
6 , the decay rate attains a maximum in the presence of full quantum interference @xmath157 , thereby reducing the spectral squeezing observed in fig .
finally , we note that the spectrum @xmath144 exhibits pure two - level squeezing for weak applied fields @xmath158 independent of the quantum interferences . the spectrum in this case
is centered around the laser frequency @xmath159 and the maximum squeezing is obtained for the out - of - phase quadrature @xmath160 similar to two - level atoms @xcite .
in this paper , we investigated the squeezing spectrum of the resonance fluorescence from a driven y - type atom when the presence of interference in spontaneous decay channels is important .
in particular , we considered the atom to be driven by two coherent fields and examined the squeezing spectrum of fluorescence radiation from both the upper- and lower transitions in the atoms .
it was shown that the decay - induced interference enhances squeezing in the spectrum of upper transitions for off - resonance and strong driving fields . a detailed analysis using dressed - states
was presented to bring out the role of interferences .
further , the interference was also shown to degrade the spectral squeezing if the fluorescence is detected on the lower atomic transitions .
this has been explained as due to a fast decay of the dressed state in the atom . in the secular approximation ,
the time evolution of the coherence @xmath161 between the dressed states @xmath162 and @xmath163 obeys @xmath164 with @xmath165 obtained by diagonalizing the hamiltonian ( [ ham ] ) . the decay rate @xmath166 is given by @xmath167 where @xmath168 @xmath169 @xmath170 @xmath171 99 j. mod . opt .
34 ( 1987 ) ( 6/7 ) , special issue on squeezed light , edited by r. loudon and p.l .
knight ; j. opt .
b * 4 * ( 1987 ) ( 10 ) , special issue on squeezed light , edited by h.j .
kimble and d.f . walls .
hong , and l. mandel , j. opt .
b 4 ( 1987 ) 1574 .
squeezing , photon antibunching , and sub - poissonian photon statistics are essential phenomena demonstrating the non - classical states of radiation . for a discussion of the photon statistics of non - classical fields ,
lai , v. buek , and p.l .
knight , phys .
a 43 ( 1991 ) 6323 ; fam le kien , anil k. patnaik , and k. hakuta , phys . rev . a 68 ( 2003 ) 063803 .
( a ) d.f .
walls and p. zller , phys .
( 1981 ) 709 ; ( b ) l. mandel , phys .
49 ( 1982 ) 136 ; p. zhou and s. swain , phys .
a 59 ( 1999 ) 3745 ; z. ficek and s. swain , j. opt .
b 14 ( 1997 ) 258 . m.j .
collet , d.f .
walls , and p. zller , opt .
52 ( 1984 ) 145 .
p. zhou and s. swain , phys .
a 59 ( 1999 ) 841 ; p. zhou and s. swain , phys .
a 59 ( 1999 ) 1603 .
lu , s. bali , and j.e .
thomas , phys .
81 ( 1998 ) 3635 ; h.z .
zhao , z.h .
lu , a.m. bacon , l.j .
wang , and j.e .
thomas , phys .
a 57 ( 1998 ) 1427 . f.a.m .
de oliveira , b.j .
dalton , and p.l .
knight , j. opt .
b 4 ( 1987 ) 1558 ; w. vogel and r. blatt , phys .
a 45 ( 1992 ) 3319 ; z. ficek , b.j .
dalton , and p.l .
knight phys .
a 50 ( 1994 ) 2594 ; p.a .
lakshmi and g.s .
agarwal , phys .
a 32 ( 1985 ) 1643 ; z. ficek , b.j .
dalton , and p.l .
knight , phys . rev . a 51 ( 1995 ) 4062 .
dalton , z. ficek , and p.l .
knight , phys .
a 50 ( 1994 ) 2646 .
gao , f.l .
li , and s.y .
zhu , phys .
lett . a 335 ( 2005 ) 110 .
s.y.gao , f.l.li , and d.l.cai , j.phys .
b 40 ( 2007 ) 3893 .
g. s. agarwal , _ quantum statistical theories of spontaneous emission and their relation to other approaches _ , springer tracts in modern physics : quantum optics ( springer - verlag , berlin , 1974 ) , p. 94 .
cardimona , m.g .
raymer , and c.r .
stroud , j.phys .
b 15 ( 1982 ) 65 ; s.y .
zhu and m.o .
scully , phys .
76 ( 1996 ) 388 ; p. zhou and s. swain , phys .
77 ( 1996 ) 3995 ; m. macovei , j. evers , and c.h .
keitel , phys .
91 ( 2003 ) 233601 .
li and s.y .
zhu , phys .
a 59 ( 1999 ) 2330 ; z. ficek and s. swain , phys .
a 69 ( 2004 ) 023401 ; f.l .
zhu , and a.q .
ma , j.mod . opt .
48 ( 2001 ) 439 .
r. arun , phys .
a 77 ( 2008 ) 033820 .
gao , f.l .
li , and s.y .
zhu , phys .
a 66 ( 2002 ) 043806 ; m.a .
antn , f. carreo , and o.g .
caldern , opt .
commun . 234 ( 2004 ) 281 .
gao , and s.y .
zhu , phys .
a 67 ( 2003 ) 063818 .
isabel gonzalo , m. a. antn , f. carreo , and oscar g. caldern , phys .
a 72 ( 2005 ) 033809 . a. k. patnaik and g. s. agarwal , phys .
rev . a 59 ( 1999 ) 3015 ; a. k. patnaik and g. s. agarwal , j. mod .
opt . 45 ( 1998 ) 2131 ; z. ficek and s. swain , phys .
a 69 ( 2004 ) 023401 .
gurudev dutt , j. cheng , b. li , x. xu , x. li , p. r. berman , d. g. steel , a. s. bracker , d. gammon , s. e. economou , r. b. liu , and l. j. sham , phys .
94 ( 2005 ) 227403 .
d. g. norris , l. a. orozco , p. barberis - blostein , and h. j. carmichael , phys .
105 ( 2010 ) 123602 .
xiang - an yan , li - qiang wang , bao - yin yin , and jian - ping song , j. opt .
b 26 ( 2009 ) 1862 .
liu - gang si , xin - you l , xiangying hao , jia - hua li , j.phys .
b 43 ( 2010 ) 065403 .
sumanta das and g. s. agarwal , phys .
a 81 ( 2010 ) 052341 .
m. lax , phys . rev .
172 ( 1968 ) 350 . for weak driving fields @xmath158 , the spectrum @xmath99 generated by the upper transitions does not show squeezing in any quadrature in the absence and in the presence of quantum interference .
the spectral squeezing appears in out - of - phase @xmath160 quadrature and more generally other phase quadratures @xmath172 as well .
however , the in - phase quadrature exhibits more squeezing than the out - of - phase quadrature . | the squeezing spectrum of the resonance fluorescence is studied for a coherently driven four - level atom in the y - type configuration .
it is found that the squeezing properties of the fluorescence radiation are modified significantly when quantum interference of the spontaneous decays channels is included .
we show a considerable enhancement of steady - state squeezing in spectral components for strong and off - resonant driving fields . the squeezing may be increased in both the inner and outer sidebands of the spectrum depending upon the choice of parameters .
we also show that the interference can degrade the spectral squeezing by increasing the decay rates of atomic transitions . an analytical description using dressed states
is provided to explain the numerical results
. + + keywords : resonance fluorescence ; squeezing ; quantum interference |
one important constraint on the moduli space of vacua of supersymmetric gauge theories @xcite is that the massless fermions in the low energy theory should have the same flavor anomalies as the fundamental fields , i.e.the t hooft consistency conditions should be satisfied @xcite . these conditions are considered a particularly stringent test on the spectrum of massless fermions , which is usually obtained from symmetry arguments and renormalization group flows .
it is found that for some theories the classical moduli space @xmath1 or a suitable quantum modified version @xmath2 of it satisfies t hooft consistency conditions at every point .
other theories fail t hooft s test at some vacua , and it is believed that in the infrared , these theories correspond to the weak coupling sector of a dual theory with a different gauge group and matter content @xcite . in this paper , sufficient conditions on the fundamental chiral field content @xmath3 for a theory to satisfy the t hooft conditions is established .
this is usually done by first finding a basic set of gauge invariants @xmath4 and the constraints among them , then checking point by point the matching of flavor anomalies between the fundamental fields and the composites at each point on the moduli space .
this procedure involves tedious calculations , and does not offer a systematic approach to the problem of determining whether @xmath1 ( or a suitable quantum modified version of it ) gives the right description of the infrared sector or not .
our results establish simple sufficient conditions on the fundamental fields that guarantee t hooft consistency conditions will be satisfied .
the novelty of our approach is that no explicit calculation of anomalies is required at any time ; it is not even required to know what the basic gauge invariants are .
this paper fills in the details omitted in ref .
the outline of the paper is as follows .
we devote sec .
[ sec : vacua ] to a review of the connection between @xmath1 and the algebraic quotient of @xmath5 under the action of the complexification @xmath6 of the gauge group .
the results in @xcite are rederived and additional information related to the structure of gauge orbits in @xmath5 and the geometry of @xmath1 is provided .
methods to determine the dimension of @xmath1 before finding the invariants and their constraints are presented . a number of examples illustrating how naive expectations fail to be true in special situations is also given in this section . in sec .
[ sec : uv ] , we analyze anomaly matching at points on the classical moduli space .
the proof of the anomaly matching theorem makes use of the connection between @xmath1 and the vector space @xmath5 spanned by the fundamental matter fields @xmath3 provided by the map @xmath7 , @xmath8 .
knowledge of the algebraic geometric construction of @xmath1 and details related to this map is essential . in sec .
[ sec : ir ] we give a rigorous proof of the fact that the flavor anomalies of the massless modes at the vacuum @xmath9 equal those of the full vector space @xmath10 of the gauge invariant composites ( i.e. ignoring all constraints ) when @xmath11 is the set of critical points of an invariant superpotential .
we apply this result to : ( i ) extend the matching theorem to cases where the superpotential @xmath12 is not trivial , ( ii ) globalize the point - by - point result of the matching theorem , and ( iii ) prove that anomaly matching conditions are compatible with integrating out fields .
our results guarantee the matching of flavor anomalies between the uv and @xmath1 for the large family of @xmath13-confining theories introduced in @xcite and those theories obtained from them by integrating out matter fields , which have a quantum modified moduli space . as an application , in sec . [
sec : examples ] we analyze in detail the well known case of supersymmetric qcd .
we first repeat the analysis of ref .
@xcite , and find all points at which anomalies match by performing explicit calculations , then show how the matching follows readily from our results with virtually no calculations .
extensions of the results to dual theories will be described elsewhere .
in this section , we will review some of the properties of supersymmetric vacua in supersymmetric gauge theories , and their connection with algebraic geometry .
many of the results are well known in the physics or mathematics literature . including the results here
will allow us to define our notation , and also to make the paper more self - contained .
the physical objects that we will consider in this paper are supersymmetric gauge theories with gauge group @xmath14 , where @xmath14 is the direct product of a compact connected semisimple lie group and ( possible ) @xmath15 factors .
we will assume that all the @xmath15 s are compact , i.e. that the fields have rational charges .
the action of the gauge theory is @xmath16 = \int d^4x\ , d^4 \theta\ \phi^\dagger e^v \phi + \int d^4x\ , d^2\theta\ \left[{1\over 4 g^2 } \tr w^\alpha w_{\alpha } + w(\phi ) + c.c . \right]\ ] ] where @xmath17 is a set of chiral superfields transforming as a ( reducible ) representation of @xmath14 , @xmath18 is the gauge chiral superfield , @xmath19 is the gauge coupling constant , @xmath10 is the gauge vector superfield , @xmath20 @xmath21 are the generators of the lie algebra of @xmath22 , and @xmath12 is the superpotential . @xmath10 and @xmath23 are related by @xmath24 . for simplicity of notation , we have assumed that @xmath14 is a simple lie group with coupling constant @xmath19 ; if not , there is a different coupling constant for each simple factor in @xmath22 . the action eq .
( [ 2.1 ] ) will in general also have a global flavor symmetry group @xmath25 , and a @xmath26 symmetry .
the set of inequivalent vacua of a supersymmetric gauge theory is referred to as the moduli space .
the classical moduli space @xmath1 is determined by studying gauge inequivalent constant field configurations that are critical points of the superpotential @xmath12 , and satisfy the @xmath27-flatness condition @xmath28 where @xmath29 is the lie algebra of @xmath14 , and has dimension @xmath30 .
the action eq .
( [ 2.1 ] ) has a larger invariance than the gauge symmetry @xmath14 .
it is also invariant under local transformations of the form @xmath31 with @xmath32 chiral superfields @xcite .
this implies invariance under @xmath6 , the complexification of @xmath14 .
this result was used in ref .
@xcite to show that the moduli space of supersymmetric vacua is an algebraic variety .
we will review the analysis given in ref .
@xcite here because we need additional details about the construction of the moduli space and the structure of @xmath6-orbits not discussed previously .
any classical supersymmetric vacuum configuration can have only a constant expectation value for the scalar component of the superfield .
thus classical supersymmetric vacua are a subset of @xmath5 , the vector space of all constant field configurations @xmath17 .
@xmath5 has dimension @xmath33 , the number of chiral superfields .
if @xmath17 is a point in @xmath5 , the @xmath6-orbit of @xmath17 will be denoted by @xmath34 . if @xmath17 is a critical point of the superpotential @xmath12 , then so are all points of @xmath34 since @xmath12 is @xmath6-invariant .
the set of critical points of @xmath12 in @xmath5 will be denoted by @xmath35 .
the mathematical objects that we will consider in this paper are affine algebraic sets ( a special case of varieties ) over the field of complex numbers .
let @xmath36 $ ] be the ring of polynomials in @xmath37 complex variables @xmath38 .
let @xmath39 , @xmath40 be a finite set of polynomials in the @xmath37 variables . then the algebraic set @xmath41 is the set @xmath42 .
it can be thought of as a curve in @xmath43 given implicitly by the polynomial equations @xmath44 .
the ideal @xmath45 is the ideal generated by the polynomials @xmath46 , i.e. the set of polynomials of the form @xmath47 , where @xmath48 are arbitrary polynomials in @xmath36 $ ] .
clearly , any polynomial in @xmath45 vanishes at all points on @xmath49 .
the set of polynomials that vanish on an algebraic set @xmath50 ( such as @xmath49 ) will be denoted by @xmath51 , and forms a finitely generated ideal of @xmath36 $ ] .
in general , @xmath52 , but equality need not hold .
it is possible to define the algebraic set @xmath50 as the zero set of polynomials @xmath53 , @xmath54 that generate @xmath51 . in this case , @xmath55 , and @xmath56 .
if this is the case we say that the equations @xmath57 , @xmath58 _ correctly define _ @xmath50 .
an irreducible algebraic set is one that can not be written as a proper union of two algebraic sets .
any algebraic set can be written as a finite union of irreducible algebraic sets in a unique way .
the tangent space @xmath59 at a point @xmath60 of the algebraic set @xmath61 correctly defined by polynomials @xmath62 is the vector subspace @xmath63 of @xmath64 .
the dimension @xmath65 of an irreducible algebraic set @xmath50 is @xmath66 .
there are alternative equivalent definitions of @xmath59 and different ways of calculating dimensions of irreducible algebraic sets ( see @xcite ) .
the natural complex valued functions on an algebraic set @xmath50 are the restrictions to @xmath50 of polynomials in its ambient vector space . these are called regular functions , and the set of regular functions is the coordinate ring @xmath67 $ ] .
note that two polynomials @xmath68 and @xmath69 in @xmath70 $ ] define the same regular function on @xmath50 if and only if @xmath71 .
this defines an equivalence relation in @xmath70 $ ] and , clearly , @xmath67
= \c[x_1, ... ,x_n]/i(x).$ ] regular functions from the algebraic set @xmath61 to the algebraic set @xmath72 are naturally defined to be those that can be written as the restriction to @xmath50 of @xmath73 polynomials in @xmath64 .
all the geometric properties of @xmath50 are encoded in @xmath67 $ ] , and @xmath50 can be constructed from @xmath67 $ ] .
this fact plays a key role in the discussion of the next subsection .
it is natural to use the zariski topology in studying algebraic geometry .
the zariski - closed sets of @xmath43 are algebraic sets , by definition .
the fact that this correctly defines a topology in @xmath64 is non trivial , and is a consequence of the hilbert basis theorem , which implies that an infinite intersection of algebraic sets is an algebraic set . in the rest of this paper
, we will mainly use the zariski topology , so open and closed sets will always mean with respect to the zariski topology .
in a few places , we will also need to use the more familiar metric - topology on @xmath43 , and we will state this explicitly . the zariski topology might seem a little strange to readers used to thinking about the more familiar metric topology . in @xmath74 , for example , the zariski - closed sets are @xmath74 , a finite set of points , or the null set .
thus the zariski - closure of the set of integers in @xmath74 is the entire space @xmath74 . in contrast , the metric - closure of the set of integers is itself .
regular functions @xmath75 from the algebraic set @xmath76 into the algebraic set @xmath77 are continuous in the induced zariski topology , whereas most other functions which are continuous in the metric topology fail to be zariski continuous .
thus , the zariski topology allows us to get the strongest results when dealing with regular function . as an example ,
a regular function @xmath78 that vanishes on the integers @xmath79 , must vanish on the closure @xmath80 , which is a fancy way of saying that the only complex polynomial with infinite roots is the trivial one .
the natural groups to study in algebraic geometry are linear algebraic groups .
let @xmath81 be the space of all @xmath82 complex matrices .
@xmath83 is the set @xmath84 in @xmath81 .
@xmath85 is a linear algebraic group if it is the intersection of @xmath83 with an algebraic set in @xmath81 .
the group @xmath86 is not an algebraic group , because the constraint that the matrix be unitary is not a polynomial relation ; it involves complex conjugation .
the complexification of @xmath86 is the group @xmath87 , which is an algebraic group since the condition that the determinant be unity is a polynomial relation in the elements of the matrix .
the complexification of all compact connected semisimple lie groups are algebraic groups .
the vacuum structure of supersymmetric gauge theories will depend on @xmath6 , the complexification of the gauge group @xmath14 .
@xmath6 is an algebraic group , but @xmath14 need not be .
it is instructive at this point to go over the well - known example of susy qed , which is a @xmath15 supersymmetric gauge theory with two chiral superfields @xmath88 and @xmath89 with charges @xmath90 , respectively .
the set of all constant field configurations is @xmath91 for sqed . in sqed ,
the group @xmath92 is the set of transformations @xmath93 , where @xmath94 , @xmath95 , as can be seen from eq .
( [ gi ] ) .
the @xmath6 orbits are illustrated in fig .
[ fig : sqed ] .
they are the curves @xmath96 , the curve @xmath97 , the curve @xmath98 , and the origin @xmath99 .
the decomposition of @xmath5 into @xmath6 orbits is quite complicated , even for the simple example of sqed . @xmath6
orbits need not be closed .
the orbits @xmath100 and @xmath101 are closed , and the orbits @xmath102 and @xmath103 are not .
the closure of the @xmath6 orbit of @xmath17 is denoted by @xmath104 .
a very useful result is that the closure of a @xmath6 orbit in the zariski topology is equal to its closure in the metric topology .
the closure of the two orbits @xmath102 and @xmath103 are the curves @xmath105 , and @xmath106 , where @xmath88 and @xmath89 are no longer restricted to be non - zero . the closure @xmath107 of orbits @xmath34 of an algebraic group are irreducible algebraic sets .
the closures of @xmath6-orbits are given by the polynomial equations @xmath108 for @xmath109 , @xmath110 for @xmath111 , @xmath112 for @xmath113 , and @xmath114 for @xmath115 , each of which is an irreducible algebraic set .
the power of algebraic geometry lies in the interplay between geometric properties of the algebraic set @xmath50 and algebraic properties of its coordinate ring @xmath67 $ ] .
the key properties of the ring @xmath67 $ ] are : ( i ) @xmath67 $ ] has no nilpotents ( that is , @xmath116 $ ] , @xmath117 implies @xmath118 ) , and ( ii ) @xmath67 $ ] is finitely generated .
it turns out ( @xcite , chap .
i , sect 2 , thm 1 ) that these are also sufficient conditions for a given ring @xmath119 over @xmath74 to be isomorphic to the coordinate ring of some algebraic set @xmath120 , @xmath121 $ ] .
the construction of @xmath122 is in fact very simple : take a set @xmath123 of generators of @xmath119 . in general , @xmath119 is not a free algebra ; there are polynomial equations among the generators @xmath124 .
one shows that @xmath125 , the equality being valid only in the case of a free algebra .
now assume @xmath126 , the @xmath6 invariant algebraic set of critical points of a polynomial @xmath6 invariant superpotential @xmath12 , and let @xmath33 be the dimension of @xmath5 .
there is a natural representation of @xmath6 on @xmath127 $ ] , namely @xmath127 \ni f \to g \circ f , g \circ f(x ) \equiv f(g^{-1}x ) , g \in g$ ] . under this representation
the homogeneous polynomials of degree @xmath128 form invariant subspaces . for reductive linear algebraic groups (
@xmath6 is reductive if any regular representation is completely reducible ) , the subring @xmath127^g$ ] of @xmath6 invariant polynomials is finitely generated , as follows from thm .
4.1.1 in @xcite .
as @xmath127^g$ ] does not have nilpotent elements , it can be thought of as the coordinate ring of an algebraic set @xmath129^g}$ ] . if @xmath130 is a minimal set of generators of @xmath127^g$ ] , then @xmath131 is the algebraic set defined by the constraints between the @xmath132 s .
there is a natural regular map @xmath133 by @xmath134 obtained by evaluation of the generating polynomials .
the pair @xmath135 is called algebraic ( or categorical ) quotient of @xmath35 under the action of @xmath6 in the mathematics literature .
the basic property of the quotient is that , by construction , @xmath136 \to \c[u^w]^g$ ] is an isomorphism ( as algebras over @xmath137 ) between the coordinate ring of @xmath138 and the ring of invariant polynomials in @xmath35 .
therefore , given any algebraic variety @xmath139 and @xmath6-invariant regular map @xmath140 , there exists a unique regular map @xmath141 such that @xmath142 , a property that uniquely determines the quotient .
further details of this construction , including properties one through three quoted without proof below , can be found in problems 11.2.6 - 1,2 in @xcite , and also in @xcite . using these
, we prove other results required to understand the ultraviolet - infrared global anomaly matching .
we assume @xmath6 is reductive , as is always the case in physically interesting theories .
property p5 establishes the anticipated contact with supersymmetric moduli spaces : points in @xmath143 are in one to one correspondence with closed orbits in @xmath35 .
as the latter are in one to one correspondence with physically inequivalent classical supersymmetric vacua , @xmath138 is the classical moduli space @xmath1 of the theory ( [ 2.1 ] ) .
this picture breaks if @xmath6 fails to be reductive ( subsection [ subsecf ] ) .
the fiber of @xmath144 , @xmath145 is in general a reducible algebraic set .
it can be written as the union @xmath146 of @xmath6-invariant irreducible algebraic sets .
( note that @xmath147 are @xmath6-invariant , but need not be a single @xmath6-orbit . )
then @xmath148 is a closed @xmath6 invariant subset of @xmath145 , and contains the unique closed @xmath6 orbit .
one can now connect the algebraic construction of @xmath138 discussed above with the classical moduli space @xmath1 of supersymmetric vacua @xcite . in ref .
@xcite it is proven that every @xmath6 orbit contains a ( unique ) @xmath27-flat point if and only if the orbit is closed .
every fiber contains exactly one closed @xmath6 orbit , and thus exactly one @xmath27-flat point which is also a critical point of @xmath12 .
thus the gauge - inequivalent field configurations that satisfy the @xmath27-flatness condition and @xmath149 are in one - to - one correspondence with fibers @xmath150 .
this means that @xmath151 is isomorphic to @xmath1 the classical moduli space of the supersymmetric gauge theory , the result shown in ref .
@xcite . in studying anomaly
matching conditions , one needs to compute anomalies in the full theory ( uv theory ) and for the massless modes ( ir theory ) .
the uv anomalies are computed using the fields @xmath17 , i.e. using @xmath5 , and the ir anomalies are computed using the classical moduli space @xmath1 or its quantum modification @xmath2 .
the classical moduli space @xmath152 , so the anomaly matching conditions require knowledge of the structure of the @xmath6-orbits in @xmath5 .
we will collect here a number of useful results from refs .
@xcite that will be needed later in this article .
some of them have already been discussed earlier in this section .
the basic setting is the algebraic quotient @xmath153 of the variety @xmath35 given by the critical points of @xmath12 under the action of a reductive algebraic group @xmath6 .
( p1 ) : : @xmath154 is surjective ( p2 ) : : if @xmath155 is closed and @xmath6-invariant , then @xmath156 is closed in @xmath138 and @xmath157 is the algebraic quotient of @xmath158 under the action of @xmath6 .
( p3 ) : : if @xmath159 is any collection of closed @xmath6-invariant subsets of @xmath35 , then @xmath160 .
( p4 ) : : if @xmath35 is irreducible , @xmath138 is irreducible .
the proof follows from ( p1 ) and lemma a.1.16 of ref .
( p5 ) : : there is a unique closed @xmath6 orbit in every fiber of the map @xmath154 . ( p6 ) : : two points @xmath161 lie in the same fiber if and only if @xmath162 .
consists of a single point . ] from now on , we shall assume that @xmath35 is irreducible .
then @xmath163 is irreducible , and we can use the results of chap . 1 sec .
6.3 of @xcite .
the dimensions of various objects are defined as follows : @xmath30 : : : the real dimension of the lie algebra @xmath164 of the gauge group @xmath14 , which is equal to the complex dimension of @xmath165 . @xmath33 : : : the dimension of @xmath5 .
@xmath166 : : : the dimension of @xmath35 .
@xmath167 : : : the number of generators of the ring of @xmath6-invariant polynomials @xmath127^g$ ] .
@xmath168 : : : the dimension of the classical moduli space @xmath169 .
@xmath128 : : : the maximum dimension of a @xmath6-orbit in @xmath35 .
( note that @xmath6-orbits can have different dimensions , so @xmath128 is the maximum possible dimension . also , @xmath170 . )
@xmath171 : : : the minimum dimension of a fiber in @xmath35 .
let @xmath172 be a basis for @xmath29 , the lie algebra of the gauge group @xmath14 .
the dimension of a @xmath6 orbit through @xmath17 equals the rank of the @xmath173 matrix @xmath174 with columns @xmath175 .
note that the entries of @xmath174 are polynomials of degree one in @xmath17 .
let @xmath176 , @xmath177 be the set of all @xmath178 minors of @xmath174 , and @xmath179 the set of polynomials in @xmath17 obtained by taking the determinants of the elements of @xmath180 .
the rank of @xmath174 equals the maximum value of @xmath181 for which there is a polynomial in @xmath179 which does not vanish at @xmath17 .
the set @xmath182 of points in @xmath5 whose orbits have dimension less than @xmath181 is the closed set @xmath183 , obtained by requiring that all @xmath178 minors have zero determinant .
the complement is then an open set .
hence `` maximum rank '' of a matrix valued function on an algebraic set with regular entries is then an example of an _ open condition , _
i.e. it defines an open set .
the same type of argument shows that the set of smooth points of an irreducible algebraic set @xmath184 is open in @xmath184 . these results together with a straightforward application of @xcite chap . 1
, sec 6.3 imply ( p7 ) : : the following is a list of non - empty open subsets of @xmath35 and @xmath1 .
quantities with a @xmath185 belong to @xmath1 , and without a @xmath186 belong to @xmath35 .
the inclusion sign means the left - hand side is a non - empty open set contained in the right - hand side , but equality need not hold .
+ ; ; @xmath187 ; ; @xmath188 . ; ; @xmath189 . ; ; @xmath190 . ; ; @xmath191 . using these non - empty open sets , and
the result that non - empty open subsets of an algebraic set are dense , and any two non - empty open sets have a non - empty intersection ( lemma a.1.12 in @xcite ) allows one to prove ( p8 ) : : the dimension of @xmath1 equals that of @xmath35 minus the minimum dimension of a fiber . proof : take a point @xmath192
. then @xmath193 . (
p9 ) : : the minimum dimension of fibers is greater than or equal to the maximum dimension of @xmath6-orbits .
proof : take @xmath194 .
as @xmath154 is @xmath6-invariant , @xmath195 , so that @xmath196 .
( here @xmath197 denotes the tangent space of @xmath50 at the point @xmath60 . )
we finally quote an important theorem due to knop @xcite , and prove a result that is used in sec .
[ sec : examples ] * theorem 1 [ knop ] * ( ref .
@xcite ) if @xmath6 is semisimple , @xmath198 has maximum dimension , and @xmath199 is smooth , then @xmath200 is surjective .
* theorem 2 * if @xmath201 ( maximal ) and @xmath202 is closed , then * @xmath203 * @xmath204 .
if also @xmath205 then * @xmath206 is surjective . finally , if @xmath207 is also a smooth point * @xmath208 and @xmath209 is smooth .
proof : assume ( a ) is false , and pick any point @xmath210 . because of @xmath211 , @xmath212 .
therefore there is a sequence @xmath213 and a point @xmath214 such that @xmath215 . given any point @xmath216 , @xmath217 , so @xmath218 .
thus @xmath202 is a proper subset of @xmath219 , which implies that @xmath220 ( lemma a.1.18 of @xcite ) , i.e. @xmath221 which is a contradiction , since @xmath6-orbits have maximum possible dimension @xmath128 .
\(b ) @xmath222 and ( a ) implies @xmath223 , and ( b ) follows using @xmath224 .
\(c ) by luna s theorem @xcite @xmath202 closed implies there is an tale slice at @xmath207 . in the case
dim @xmath225 the existence of the slice implies that any curve @xmath226 through @xmath227 can be lifted to a curve @xmath228 in @xmath35 satisfying @xmath229 and @xmath230 .
this is equivalent to the surjectivity of @xmath206 .
\(d ) if @xmath207 is smooth then dim @xmath231 . from ( b ) and ( c ) we have : @xmath232 thus , these are all equalities and ( d ) follows .
conditions ( c ) and ( d ) ensure that the tangent to the moduli space is isomorphic to the massless modes of the theory determined from the ultraviolet lagrangian at the classical level , i.e. by looking at the massless modes of the quadratic part of the lagrangian in unitary gauge . we have already discussed this case .
the space @xmath50 is @xmath233 , since there are two complex fields @xmath88 and @xmath89 . the ring @xmath67 $ ] is the ring of polynomials in two indeterminates , @xmath234 $ ] .
the ring @xmath67^{g}$ ] is the set of gauge invariant polynomials @xmath235 $ ] , freely generated by the single polynomial @xmath236 .
the algebraic set @xmath184 defined by @xmath237=\c[q\tilde q ] = \c[x]^g$ ] is thus @xmath74 .
the map @xmath238 takes @xmath239 to the point @xmath240 in @xmath184 .
the fiber of a point @xmath241 is the closed orbit @xmath100 discussed in sec .
[ sec : sqed ] .
the fiber of @xmath242 is @xmath243 .
note that @xmath244 is a smooth algebraic set , with no singular points .
nevertheless the fiber of @xmath242 is different in structure from that of the points @xmath245 .
the fibers @xmath246 are closed irreducible and smooth one - dimensional algebraic sets for @xmath245 .
@xmath247 is the reducible and singular algebraic set @xmath248 , and contains three @xmath6 orbits , @xmath102 , @xmath103 and @xmath101 .
it can be written as the union of two irreducible components , @xmath249 given by @xmath112 and @xmath250 given by @xmath110 .
their intersection , @xmath251 is the origin @xmath252 , and is the unique closed @xmath6 orbit in @xmath247 .
the smooth points @xmath139 of @xmath247 form the open set @xmath253 .
the complement @xmath254 is the origin @xmath101 and is a smooth closed algebraic set that contains the unique closed @xmath6 orbit .
@xmath255 is surjective at all points of @xmath256 and at all points of the fiber @xmath247 except @xmath257 .
thus it is possible in this case to choose a point in @xmath247 where @xmath255 is surjective ; however this point is not in the unique closed orbit @xmath257 in the fiber .
now consider adding a superpotential @xmath258 to sqed .
the critical points of @xmath12 are @xmath259 , so that @xmath35 is a single point , and is @xmath6-invariant .
note that in this example , the critical points of @xmath12 _ do not _ form a complete fiber .
the ideal @xmath260 is the set of all polynomials of the form @xmath261 , and @xmath262 is the set of all polynomials of the form @xmath263 , where @xmath264 , @xmath19 and @xmath265 are arbitrary polynomials in their arguments .
the classical moduli space @xmath1 without a superpotential is the complex plane @xmath74 given by @xmath266 .
including @xmath12 restricts one to @xmath267 given by the algebraic set @xmath242 in @xmath1 . since @xmath12 is gauge invariant , one can rewrite @xmath12 in terms of the gauge invariant polynomials used to describe the moduli space . in the present example , @xmath268 .
however , when regarded as a function of the gauge invariants rather than the fundamental fields , the critical points of the resultant @xmath12 do not correctly describe the moduli space in the presence of a superpotential . in our example , @xmath269 has no critical points , whereas @xmath270 contains the single point @xmath271 .
one can obtain @xmath270 by minimizing @xmath272 , instead of @xmath269 .
however , in general , one can not obtain the equations defining the moduli space @xmath270 by minimizing a superpotential in the space of gauge invariants .
an interesting example is susy qed with fields @xmath273 and @xmath274 both with charge @xmath275 .
this is an anomalous theory , but still provides a useful example that illustrates the structure of fibers .
similar results can be found in more complicated anomaly free theories ( such as susy qcd with @xmath276 ) .
the space @xmath5 is @xmath233 , and @xmath277=\c[q_1,q_2]$ ] .
there are no gauge invariant polynomials other than constants , so @xmath277^g = \c$ ] , and the moduli space @xmath278 is the zero dimensional space consisting of a single point @xmath279 .
the orbits of all points other than @xmath257 are radial lines ( see fig .
[ fig : sqedp ] ) , and are not closed . the orbit of @xmath257 is a single point , and is the unique closed orbit .
the fiber @xmath280 is the entire plane @xmath233 , and is irreducible . in this case @xmath281 , @xmath282 , @xmath283 , @xmath284 , @xmath285 and @xmath286 .
note that @xmath287 , so that the minimum dimension of a fiber can be strictly greater than the maximum dimension of an orbit .
this example shows that theorems 1 and 2 are the most we can say about surjectivity of @xmath255 above smooth points of @xmath1 . consider a @xmath15 theory with three fields @xmath274 , @xmath288 and @xmath289 with charges 2 , 1 and @xmath290 ( an anomalous theory ) .
the space @xmath5 is @xmath291 .
the gauge invariant polynomials are generated by @xmath292 and @xmath293 .
the moduli space @xmath1 and @xmath10 are both @xmath233 , since there are no relations among @xmath102 and @xmath103 .
@xmath1 is smooth everywhere .
consider the orbit of @xmath294 .
it is the set of all points @xmath295 with @xmath95 , and has dimension one , i.e. equal to the dimension of the gauge group .
@xmath296 so @xmath255 is not surjective even though the orbit has maximal dimension and lies above a smooth point .
note that theorem 1 does not apply here because the gauge group is not semisimple , and theorem 2 does not apply here , because the orbit is not closed .
this is the example studied in detail in refs .
the gauge group is @xmath297 , with @xmath298 matter fields @xmath299 , @xmath300 , @xmath301 in the fundamental @xmath302 representation of @xmath303 , and @xmath298 matter fields @xmath304 , @xmath305 , @xmath306 in the @xmath307 representation of @xmath303 .
the ultraviolet space @xmath5 is a vector space of dimension @xmath308 .
the complexification of @xmath303 is @xmath309 , under which @xmath299 and @xmath310 are in the fundamental and its dual representation . in a vector space @xmath10 by @xmath311 we define the dual and conjugate representations on @xmath312 by @xmath313 and @xmath314 . they agree only when the representation of @xmath315 on @xmath10 is unitary .
note that the dual representation is defined so as to make @xmath316 invariant , which is not the case for the conjugate of a non - unitary representation . ]
a set of generators of @xmath277^g$ ] , the coordinate ring of gauge invariant polynomials are the mesons and baryons , @xmath317 these polynomials span the vector space @xmath10 .
the structure of the classical moduli space @xmath318 depends crucially on the value of @xmath298 . *
( i ) * @xmath319 : : : if @xmath320 , @xmath103 and @xmath321 are identically zero .
@xmath277^g$ ] is freely generated by @xmath322 , @xmath323 and @xmath324 .
this example illustrates a non - trivial case of dimension counting .
consider the point @xmath17 given by @xmath325 the @xmath326 orbit of @xmath17 is closed and of maximum dimension , @xmath327 . from theorem 2
we obtain @xmath328 . *
( ii)*@xmath329 : : : in this case @xmath330 , and @xmath277^g$ ] is not a free algebra , as its generators are constrained by the single relation @xmath331 this gives a hypersurface in @xmath332 , and the dimension of the moduli space is @xmath333 .
this number can also be obtained by applying theorem 2 to the closed @xmath6 orbit of the point @xmath334 * ( iii)*@xmath335 * + 1 with a superpotential * : : : when @xmath336 the fields ( [ qcd ] ) are subject to the following algebraic constraints @xmath337 where cof@xmath338 is the matrix of cofactors .
assume we add a superpotential @xmath339 , which is a mass term for the @xmath340 flavor .
the set @xmath35 of critical points of @xmath12 is just the space @xmath5 for @xmath341 , naturally embedded in @xmath5 for @xmath336 by setting the components of @xmath88 and @xmath89 for the @xmath340 flavor to zero .
@xmath342 is the intersection of the moduli space eq .
( [ cmsn+1 ] ) with the subspace @xmath343 this reproduces eq .
( [ nf = n ] ) , as anticipated by p2 above .
this example is from @xcite .
the abelian group @xmath344 of complex numbers under addition ( which is the complexification of the group @xmath14 of real numbers under addition ) is a simple example of a non - reductive group .
we will consider the representation on @xmath345 given by @xmath346 .
note that @xmath347 is an invariant subspace with no invariant complement .
@xmath348 is the linear algebraic non - reductive subgroup of @xmath349 of upper triangular , determinant one matrices of the form @xmath350 acting on @xmath233 as the restriction of the fundamental representation of @xmath349 .
the orbits @xmath351 are closed one dimensional lines @xmath352 when @xmath353 . for @xmath354
the orbits are points @xmath355 for each value of @xmath356 ( see fig . [
fig : nonred ] ) .
every @xmath6 orbit is closed .
the fibers are the horizontal lines @xmath357 .
the @xmath358-axis @xmath354 is a fiber which contains an infinite number of closed orbits . if this example were an acceptable supersymmetric gauge theory , the algebraic quotient @xmath359 would not equal the classical moduli space @xmath1 , since the fiber @xmath354 contains infinitely many closed orbits , i.e. infinitely many inequivalent supersymmetric vacua
* theorem 3 : * let @xmath1 be the classical moduli space of a supersymmetric gauge theory with gauge group @xmath22 and flavor symmetry @xmath25 and superpotential @xmath12 .
it is assumed that the gauge theory has no gauge or gravitational anomalies , and the flavor symmetries have no gauge anomalies .
let @xmath360 be a point in the classical moduli space .
assume there is a point @xmath361 in the fiber @xmath362 of @xmath9 such that * @xmath6 ( the complexification of @xmath22 ) is completely broken at @xmath207 , so that @xmath363 .
* @xmath364 and @xmath365 is surjective
. if a subgroup @xmath366 is unbroken at @xmath9 , then the t hooft consistency conditions for the @xmath367 flavor anomalies and the @xmath315 gravitational anomalies are satisfied . for the purposes of the proof , it is convenient to write the original flavor symmetry as @xmath368 , where @xmath369 is the @xmath369-symmetry , and @xmath370 now contains only non-@xmath369 symmetries .
we first prove anomaly matching when @xmath371 and @xmath372 , and then prove consistency for anomalies that include the @xmath369 symmetry .
( note that the unbroken @xmath369 symmetry might be a linear combination of the original @xmath369 symmetry and some generator in @xmath370 . )
the extension to @xmath373 follows simply from the results of sec .
[ sec : ir ] .
since @xmath315 is unbroken at @xmath9 , @xmath9 is @xmath315-invariant @xmath374 the map @xmath375 commutes with the flavor symmetries , so @xmath376 thus , by ( a ) and ( b ) @xmath377 this implies that given any @xmath378 , there is a unique @xmath379 such that @xmath380 where the minus sign is chosen for convenience .
it is straightforward to check that the map @xmath381 given by @xmath382 is a lie - algebra homomorphism , @xmath383 \right ) = \left [ g \left ( h_1 \right ) , g \left ( h_2 \right ) \right].\ ] ] this allows us to define a new `` star '' representation of @xmath384 in @xmath5 @xmath385 since @xmath386 , the new @xmath384 representation on @xmath387 defined by @xmath388 agrees with the original one .
thus the @xmath389-anomalies computed at @xmath390 are the same as the @xmath391-anomalies at the same point .
@xmath392 is an invariant subspace under @xmath393 , and the restriction of @xmath393 to @xmath392 is the adjoint action by @xmath394 .
this can be seen by direct computation .
take any element @xmath395
. then @xmath396 } \phi_0 = ad_{\frak g \left ( h \right ) } \ , { \frak g } \phi_0,\ ] ] since the flavor and gauge symmetries commute , and using eq .
( [ 3.1 ] ) .
the space @xmath5 can be broken up into the tangent space to the @xmath6-orbit @xmath397 and its invariant complement , @xmath398 , since @xmath6 is reductive . by ( b ) , the map @xmath365 is a bijective linear map from @xmath398 to the tangent space @xmath399 of the moduli space @xmath1 at @xmath9 , and commutes with @xmath389 .
thus the action of @xmath389 on @xmath398 is equivalent to the action of @xmath391 on @xmath387 , by the similarity transformation @xmath400 given by @xmath365 restricted to @xmath398 .
one can write @xmath401 where the second form shows the structure of @xmath389 on @xmath402 .
the action of @xmath391 on @xmath5 has been labeled by the subscript uv , and the action on the moduli space has been labeled by ir .
one can now compare anomalies in the uv and ir theories using the two different forms for @xmath389 . since the adjoint representation is real , the @xmath403 flavor anomaly and @xmath389 gravitational anomaly are equal to the anomalies in the infrared theory .
all that remains is the proof that the @xmath404 and @xmath389 anomalies of @xmath5 equal the @xmath405 and @xmath391 anomalies of @xmath5 .
let @xmath406 be any three elements of @xmath384 .
then @xmath407 the last three lines vanish because the original theory had no gauge and gravitational anomalies , and the flavor symmetries have no gauge anomalies .
thus the @xmath408 and @xmath404 anomalies are the same .
similarly the @xmath389 and @xmath391 anomalies agree since @xmath409 is traceless because there is no gravitational anomaly . thus t hooft s consistency condition for the flavor anomalies is satisfied .
we now prove the matching theorem for anomalies involving the @xmath369-charge using an argument similar to the one presented above .
the @xmath369-charge acting on @xmath5 is given by the matrix @xmath410 .
the @xmath369-charge is defined acting on chiral superfields , and so is the charge of the scalar component .
anomalies are computed using the fermionic components , so it is convenient to define a new charge @xmath411 which we will call fermionic @xmath369-charge , defined by @xmath412 the anomaly can be computed by taking traces over the chiral superfields of @xmath411 .
the reason for making the distinction between @xmath410 and @xmath411 is that the map @xmath154 from @xmath5 to @xmath1 commutes with @xmath413 , but does not commute with @xmath414 .
assume that @xmath369 is unbroken at @xmath415 .
then by an argument similar to that above , it is possible to define a `` star '' @xmath369-charge , @xmath416 , @xmath417 which has the form @xmath418 under the decomposition of @xmath5 into @xmath419 . as in eq .
( [ 3.a ] ) , we have used the subscripts uv and ir to denote the @xmath369 charges in the ultraviolet and infrared theories . note that @xmath400 is the same matrix in eqs .
( [ 3.a],[3.b ] ) , given by @xmath365 restricted to @xmath398 .
the fermionic @xmath369-charge is then given by @xmath420 where in the last equality we have used the fact that fermion @xmath369 charge @xmath421 in the infrared theory . compute the trace of @xmath422 in @xmath5 , @xmath423 the @xmath369-charge has no gauge anomaly , so @xmath424 , for any @xmath425 .
here the first term is the matter contribution to the anomaly , and the second term is the gaugino contribution .
the absence of gauge anomalies implies that odd powers of @xmath426 vanish when traced over the matter fields , since there is no gaugino contribution to these anomalies .
thus we find @xmath427 the block diagonal form of @xmath428 eq .
( [ 14 ] ) , gives @xmath429 the @xmath430 anomaly @xmath431 in the uv theory is given by adding the matter and gaugino contributions @xmath432 the @xmath430 anomaly @xmath433 in the ir theory is given by @xmath434 since there are no gauginos in the low energy theory .
combining eq .
( [ 3.5][3.12 ] ) , one sees immediately that the uv and ir anomalies are equal , @xmath435 .
it is straightforward to check that the gravitational @xmath369 anomaly , and the @xmath436 and @xmath437 anomalies match .
one finds from eqs .
( [ 3.a],[3.b ] ) that @xmath438 which is the matching condition for the gravitational @xmath369 anomaly , when rewritten as @xmath439 . for the @xmath440 anomaly : @xmath441 where the last line follows using the fact that there are no @xmath442 anomalies and that the @xmath443 anomaly cancels when the matter and gaugino contributions are added . using the block diagonal forms of @xmath444 and @xmath445
, one has @xmath446 which gives the anomaly matching condition for the @xmath440 anomaly when combined with the previous result .
similarly , for the @xmath447 anomaly : @xmath448 the block diagonal forms of @xmath416 and @xmath389 give @xmath449 comparing with the previous equation shows that the @xmath447 anomalies are the same in the uv and ir theories .
the results of the previous section allow one to study the matching of anomalies between the ultraviolet and infrared theories at certain points in the classical moduli space . in this section , we derive some results that allow us to relate the anomaly matching conditions at different points on the moduli space .
the moduli space is no longer restricted to be the classical moduli space @xmath1 .
we consider the case where the moduli space @xmath450 is an algebraic curve in an ambient vector space @xmath10 given as the critical points of a superpotential @xmath12 with @xmath369-charge two , @xmath451 where @xmath452 denotes a point in @xmath10 , and we will use the notation @xmath453 , @xmath454 , etc . the tangent space to @xmath450 at @xmath9 , @xmath455 , is defined by @xmath456 in all the cases we are interested in , @xmath12 is a polynomial in @xmath452 and eq .
( [ ir:2 ] ) correctly describes the algebraic set , so that eq .
( [ ir:1 ] ) agrees with the algebraic geometry notion of the tangent space .
assume that a subgroup @xmath315 of the flavor symmetry group @xmath25 is unbroken at a point @xmath457 .
the invariance of the superpotential @xmath12 under @xmath25 implies that @xmath458 where @xmath459 is the matrix for the @xmath315 transformation in the representation @xmath369 of the fields @xmath452 .
differentiating this equation twice with respect to @xmath452 and evaluating at the @xmath315-invariant point @xmath9 gives @xmath460 which shows that @xmath461 is a @xmath315 invariant tensor that transforms as @xmath462 under @xmath315 , where @xmath369 is the @xmath315 representation of the low energy fields @xmath463 .
the tangent space to @xmath450 at @xmath9 is the null - space of @xmath464 .
one can write @xmath465 as the direct sum of the tangent space and its orthogonal complement in @xmath10 .
then @xmath464 provides a non - singular invertible map from @xmath466 into its dual , so that @xmath466 transforms as a real representation of @xmath315 .
this immediately implies that the @xmath315 anomalies computed using the flat directions @xmath467 agree with those computed using the entire vector space @xmath10 .
a similar result holds for the anomalies involving the @xmath369 charge .
let @xmath468 be the @xmath369-charge of @xmath469 , so that @xmath470 since @xmath12 has @xmath369 charge two .
differentiating twice with respect to @xmath452 shows that @xmath471 which can be written in the suggestive form @xmath472 + \left [ r_j - 1 \right ] \right ) } w_{ij } ( \hat \phi_0 ) = w_{ij } ( \hat \phi_0 ) .\ ] ] @xmath473 is the @xmath369 charge of the fermionic component of the chiral superfield .
thus eq .
( [ ir:3 ] ) shows that @xmath466 transforms like a real representation under @xmath474 , the fermionic @xmath369 charge .
thus the @xmath369 anomalies , and @xmath475 anomalies can be computed at @xmath9 using @xmath10 instead of @xmath455 .
the result can be summarized by * theorem 4 : * let @xmath476 be a moduli space described by the critical points of a flavor invariant superpotential @xmath12 of @xmath369-charge two .
then the anomalies of an unbroken subgroup @xmath477 at a point @xmath457 can be computed using the entire space @xmath10 , instead of the tangent space of @xmath450 . if the anomaly matching conditions between the uv and ir theories for @xmath315 are satisfied at @xmath9 , they are also satisfied at all points of any moduli space @xmath478 given by the critical points of any @xmath479 ( including @xmath480 and @xmath481 ) , and at which @xmath315 is unbroken
. this result tells us that for moduli spaces described by a superpotential , the precise form of the moduli space is irrelevant .
the only role of possible quantum deformations is to remove points of higher symmetry from the moduli space .
one interesting application of this result is to prove that anomaly matching conditions are compatible with integrating out heavy fields .
assume that one has a theory with a moduli space @xmath2 described by a superpotential @xmath482 .
now perturb the uv theory by adding a tree level mass term @xmath483 to the superpotential .
@xmath484 is gauge invariant , and can be written as a polynomial @xmath485 of the gauge invariant composites @xmath452 of the ir theory . if the uv theory contains no singlets , then @xmath486 is linear in the basic gauge invariant composite fields @xmath452 . from this
, it immediately follows that the effective superpotential of the massive theory is given by @xmath487 where @xmath488 is the superpotential in the absence of a mass term , since a linear term in the fields is equivalent to a redefinition of the source .
the anomalies in the ir theory for any unbroken subgroup are unaffected by the change in the moduli space due to the addition of the mass term .
they are still obtained by tracing over the whole space @xmath10 . in the uv theory
, one should trace not over the whole space @xmath5 , but only over the modes that remain massless when @xmath489 is turned on .
but it is easy to see that the massive modes in the uv theory form a real representation of the unbroken symmetry .
the argument is the same as that used in the ir theory , except that @xmath490 is replaced by the ( constant ) matrix @xmath491 .
thus the mass term does not introduce any modifications to the anomaly in the uv or ir theory .
thus one finds that if the t hooft conditions are verified for a theory , they are also valid for any theory obtained by integrating out fields by adding a mass term . the same argument
is used to extend theorem 3 to the case where there is a tree level invariant superpotential @xmath12 .
assume @xmath366 is unbroken at @xmath492 , and @xmath493 satisfies the conditions of theorem 3 .
then theorem 4 applied to the @xmath494 invariant subset @xmath495 gives @xmath496 because @xmath207 is a fixed point of the @xmath497 action and @xmath35 is the set of critical points of an @xmath497 invariant superpotential .
( here @xmath498 denotes the @xmath497 anomaly computed using the vector space @xmath139 . )
thus @xmath499 where the first and third equalities are shown in the proof of theorem 3 .
we finally show how to use theorem 4 to globalize the point - by - point result of the matching theorem for theories with a moduli space given by a superpotential . for a generic supersymmetric gauge theory , the representation @xmath500 of the gauge group on @xmath5
is made up of @xmath501 tensor copies ( flavors ) of representations @xmath502 , @xmath503 and the non - anomalous flavor symmetry group is @xmath504 the most non - trivial anomaly check is to show that the anomalies for the entire flavor group @xmath25 match at the origin . from this , anomaly matching at all points of the moduli space follows by another application of theorem 4 .
the non - trivial anomalies are @xmath505 , @xmath506 , and @xmath507 , where the @xmath508 include the @xmath369-symmetry .
an @xmath505 anomaly is proportional to the @xmath128-symbol for the @xmath509 group .
since the @xmath128-symbol is non - zero for @xmath510 , the @xmath509 anomaly can be computed by looking at an @xmath510 subgroup .
this is a standard trick for computing the anomalies of representations of a general lie group . to show that the @xmath505 anomaly matches , it is sufficient to find a point on the moduli space that leaves an @xmath510 subgroup of @xmath509 unbroken .
the @xmath505 anomaly must then also match at the origin .
similarly , the @xmath506 anomaly matching can be proven by considering a point on the moduli space which leaves a @xmath15 subgroup of @xmath509 and @xmath511 unbroken .
there is a subtlety here : the unbroken @xmath511 generator can be a linear combination of the original @xmath508 generator and some of the ( broken ) @xmath512 generators .
however , having already proven that the @xmath505 anomaly matches , @xmath513 anomaly matching implies @xmath506 anomaly matching .
the @xmath514 anomaly matching is proven by finding a point where three @xmath515 s are unbroken , where again @xmath515 is a linear combination of the original @xmath15 s and @xmath509 generators .
this procedure might seem complicated , but it is not that involved in practice . in the case of supersymmetric qcd , we will show explicitly how one can prove anomaly matching for @xmath516 by considering just one point on the moduli space ( and its charge conjugate partner ) .
the construction of the moduli space of supersymmetric gauge theories can be highly non - trivial when there are many gauge invariant combinations of the fundamental fields .
the results of the preceding sections help simplify the analysis of the structure of the moduli space , and of anomaly matching between the fundamental and massless composite fields .
we treat the following problems : * determining the flavor isotropy group @xmath517 of the vacuum @xmath518 , i.e , the maximal unbroken subgroup of @xmath25 at @xmath209 . * setting sufficient conditions for the @xmath517 anomalies in @xmath5 to match the corresponding anomalies in @xmath519 * setting sufficient conditions for anomalies to match at _ every _ vacuum of the ( possibly quantum deformed ) moduli space .
three different situations can be considered : * the basic gauge invariants of the given theory are not known . *
the basic gauge invariants are known , but not the constraints among them . * both the invariant and the constraints are known . in situation
( i ) theorems 2 and 3 tell us that if @xmath361 is smooth , totally breaks the complexification @xmath6 of the gauge group @xmath22 , and has a closed orbit @xmath520 ( equivalently , there is a d - flat point in @xmath521 ) , then the anomalies do match between @xmath5 and @xmath522 , @xmath523 this solves problem b. to determine the flavor isotropy group @xmath517 we note from theorem 2 in section [ sec : vacua ] that for @xmath207 as above @xmath524 , and @xmath517 is isomorphic to the @xmath525 isotropy group at @xmath207 .. ] in situation ( ii ) we can solve a and b in the same way as in situation ( i ) , but we have the alternative of directly determining @xmath517 , for which only the gauge invariants and not the constraints among them are needed .
in situation ( iii ) we have an additional way of dealing with problem b , using the fact that the gauge group @xmath22 is semisimple .
as @xmath1 is known we can calculate its dimension .
suppose dim @xmath526 dim @xmath527 and a point @xmath207 is found such that : @xmath528 is smooth , @xmath529 breaks completely @xmath6 .
then the @xmath530 anomalies match between @xmath5 and @xmath531 .
this follows from theorems 1 , 2 and 3 .
note that it is crucial that @xmath207 break completely the _ complexification _ of the gauge group . consider the example of supersymmetric qcd with @xmath532 .
the classical moduli space is the span of the unconstrained fields @xmath322 , so @xmath533 and dim @xmath526 dim @xmath527 .
the point @xmath207 of coordinates @xmath534 for @xmath535 , totally breaks @xmath536 , but not @xmath537 .
@xmath1 is smooth everywhere , in particular @xmath538 is a smooth point .
however , anomalies do not match between the uv and @xmath539 .
now consider the point @xmath540 with coordinates @xmath541 for @xmath542 .
this point breaks completely the _ complexification _ of the gauge group , and it is mapped onto a smooth point .
theorems 1 , 2 and 3 predict anomaly matching , which is straightforward from the fact that @xmath543 is the diagonal @xmath544 .
+ theorem 4 in section [ sec : ir ] gives an answer to problem c. whenever @xmath1 is the set of critical points of an invariant superpotential , the matching of the @xmath545 anomalies at the vacua @xmath546 implies the matching of @xmath545 anomalies at any vacuum @xmath452 where @xmath547 . in particular , we can take @xmath548 , where the flavor group is not broken at all .
the points @xmath549 can be chosen such that the matching of the @xmath545 anomalies at the origin ( where the tangent space is the full ambient vector space ) implies the matching of the full flavor group anomalies there . applying again theorem 4 we prove that every point of @xmath1 will pass t hooft s test .
this extends to @xmath2 if @xmath2 also comes from a flavor invariant superpotential .
this idea can be used to prove anomaly matching for the large family of s - confining theories of ref @xcite , and for those obtained from them by integrating out any number of flavors .
the latter may have a quantum modified moduli space , which can be embedded in the same ambient vector space of the s - confining theory , and is described with an invariant superpotential ( section [ sec : ir ] ) .
+ below , we extend the explicit anomaly matching calculations in supersymmetric qcd done in @xcite to a generic point in moduli space , and show that anomalies match precisely at those points predicted by our theorems .
the amount of calculations using both approaches is contrasted .
the quarks and antiquarks transform as the @xmath551 of @xmath552 , and so can be treated together as @xmath553 flavors of doublets .
the flavor symmetry is @xmath554 .
the basic gauge invariant field is the meson @xmath555 the transformation properties of the fundamental fields and composites under the flavor symmetry is listed in table [ table : n=2 ] .
.flavor representation of the fundamental and composite fields for @xmath552 gauge theory with @xmath553 fundamentals . [ cols="^,^,^ " , ] [ table : anom ] for any value of @xmath556 and @xmath557 the classical moduli space of supersymmetric qcd is the set of critical points of the flavor invariant superpotential @xmath558 . the effective superpotential is therefore @xcite @xmath559 and the quantum moduli space @xmath2 agrees with @xmath1 .
the matching of the anomalies of the @xmath560 flavor subgroup unbroken at the `` baryon '' point of the previous subsection , together with that of the `` antibaryon '' point ( with @xmath561 ) with unbroken flavor subgroup @xmath562 imply the matching for the full flavor group at the origin ( theorem 4 ) .
another application of theorem 4 proves t hooft s conditions are satisfied at any vacuum in the moduli space .
according to the discussion of section [ sec : ir ] , adding a tree level mass term @xmath563 to @xmath12 gives the effective superpotential @xmath564 the critical points of eq .
( [ nf = na ] ) give the quantum moduli space @xmath565 of the @xmath566 theory , where @xmath567 and the determinant extends to the light flavors @xmath568 . applying theorem 4 to @xmath488 and @xmath489 proves that anomaly matching for the @xmath557 theory _ implies _ anomaly matching _ at every point of the quantum deformed moduli space of the @xmath566 theory_. we can therefore avoid the explicit checks done at isolated points of @xmath2 in @xcite and be sure the result holds at every point .
note that , although @xmath2 of the @xmath566 theory can be described as the set of critical points of a superpotential when embedded in the vector space of gauge invariants of the @xmath569 theory , @xmath1 can not
. in particular , anomalies do not match at the point @xmath570 , which belongs to @xmath1 but not to @xmath2 .
our analysis for qcd extends to other s - confining theories , an exhaustive list of which can be found in @xcite . for all s - confining theories @xmath571 can be described by a flavor invariant superpotential , so a finite number of points @xmath572 satisfying the hypothesis of the matching theorem is enough to prove anomaly matching at _ every _ point of their moduli space and at the quantum deformed moduli space of the theories obtained from them by integrating out a heavy fields .
explicit verification of this matching both for the s - confining theory and the quantum deformed one is a formidable task .
even the determination of a complete set of basic gauge invariants is sometimes nontrivial .
we give one simple example of an s - confining theory .
@xmath573 is the group that leaves invariant the 2-form @xmath576 this is the complexification of the group @xmath577 .
just one point @xmath207 is required to prove anomaly matching in the s - confining @xmath578 theory with @xmath574 fundamentals @xmath575 , @xmath579 for @xmath580 .
this point totally breaks @xmath6 , and its orbit is the closed set @xmath573 , naturally embedded in the vector space @xmath581 of @xmath582 matrices .
for supersymmetric gauge theories with a reductive gauge group @xmath6 , the classical moduli space @xmath1 is the algebraic quotient of the algebraic set of critical points of the superpotential under the action of @xmath6 .
there are known bounds for its dimension , which can be determined without knowing the basic invariant polynomials when closed orbits of maximum dimension are found .
the anomaly matching theorem can be used to show that the t hooft consistency conditions are satisfied at points that totally break the complexification of the gauge group and either have a closed orbit or are mapped onto a smooth point of @xmath1 .
anomalies will match at _ every _ point of @xmath2 if a few suitable points satisfying the above hypothesis are found and both @xmath1 and @xmath2 can be described as the set of critical points of a flavor invariant superpotential .
anomaly matching for a theory _ implies _ anomaly matching for those theories obtained from it by integrating out matter , even when these have a quantum modified moduli space .
anomalies match for the large family of s - confining theories and those obtained from them by integrating out a flavor , which have a quantum modified moduli space .
the explicit anomaly computations found in the literature can often be avoided , if one uses the results discussed here
. these results also allow one to anticipate if @xmath1 or a quantum deformed @xmath2 describes correctly the ir massless modes , or if an alternative description ( such us a dual theory ) is required .
extensions of our results to cases where the complexification of the gauge group is not totally broken and applications to dual theories is currently under study .
we are indebted to m.l .
barberis , m. hunziker and n. wallach , for extensive discussions on algebraic geometry , and to n. wallach for giving us a copy of his book @xcite prior to publication .
we are also grateful to g. schwarz for telling us about knop s theorem , and for discussion about when @xmath255 can be surjective .
this work was supported in part by a department of energy grant doe - fg03 - 97er40546 .
k. intriligator , r.g .
leigh , n. seiberg , phys .
d50 ( 1994 ) 1092 , k. intriligator and n. seiberg , nucl .
b444 ( 1995 ) 125 . for a more detailed list of references ,
see k. intriligator and n. seiberg hep - th/9509066 and m.e .
peskin hep - th/9702014 . | the structure of the moduli space of @xmath0 supersymmetric gauge theories is analyzed from an algebraic geometric viewpoint .
the connection between the fundamental fields of the ultraviolet theory , and the gauge invariant composite fields of the infrared theory is explained in detail .
the results are then used to prove an anomaly matching theorem .
the theorem is used to study anomaly matching for supersymmetric qcd , and can explain all the known anomaly matching results for this case .
c # 1lie ( # 1 ) # 1 ^-1 ( ( # 1 ) ) # 1#2- # 2pt |
the existence of coherent superpositions is a fundamental postulate of quantum mechanics but , apparently , implies very counterintuitive consequences when extended to macroscopic systems .
this problem , already pointed out since the beginning of quantum theory through the famous schrdinger cat paradox @xcite , has been the subject of a large scientific debate which is still open and very active .
nowadays there is no doubt about the existence of quantum superpositions .
indeed this effect has been demonstrated in a number of experiments involving microscopic systems ( photons @xcite , electrons @xcite , neutrons @xcite , atoms @xcite , molecules @xcite , _ etc . _ ) .
however , at least in principle , the standard theory of quantum mechanics is valid at any scale and does not put any limit on the size of the system : if you can delocalize a molecule then nothing should forbid you to delocalize a cat , apart from technical difficulties .
such difficulties are usually associated with the impossibility of isolating the system from its environment , because it is well known that any weak interaction changing the state of the environment is sufficient to destroy the initial coherence of the system . in this work
we are interested in the ideal situation in which we have a macroscopic mass or a macroscopic charge perfectly isolated from the environment and prepared in a quantum superposition of two spatially separated states . without using any speculative theory of quantum gravity or sophisticated tools of quantum field theory , we propose a simple thought experiment based on particles interacting via semiclassical forces . surprisingly a simple consistency argument with relativistic causality is enough to obtain a fundamental result which , being related to gravitational and electric fields , indirectly tells us something about quantum gravity and quantum field theory . the result is the following : assuming that a macroscopic mass @xmath0 is prepared in a superposition of two states separated by a distance @xmath1
, then any experiment discriminating the coherent superposition from a classical incoherent mixture requires a minimum time @xmath2 , proportional to the mass and the separation distance .
analogously for a quantum superposition of a macroscopic charge @xmath3 , such minimum time is proportional to the associated electric dipole @xmath4 . in a nutshell ,
experiments testing macroscopic superpositions are possible in principle , but they need to be slow . for common experiments involving systems below the planck mass and the planck charge
this limitation is irrelevant , however such time can become very important at macroscopic scales . as an extreme example , if the center of mass of the earth were in a quantum superposition with a separation distance of one micrometer , according to our result one would need a time equal to the age of the universe in order to distinguish this state from a classical statistical mixture .
clearly this limitation suggests that at sufficiently macroscopic scales quantum mechanics can be safely replaced by classical statistical mechanics without noticing the difference .
the fact that large gravitational or electromagnetic fields can be a limitation for the observation of quantum superpositions is not a new idea . in the past decades , several models of spontaneous localization @xcite
have been proposed which , going beyond the standard theory of quantum mechanics , postulate the existence a gravity induced collapse at macroscopic scales .
remaining within the domain of standard quantum mechanics , the loss of coherence in interference experiments due to the emission of electromagnetic radiation has been already studied in the literature @xcite .
similarly , the interaction of a massive particle with gravitational waves @xcite and the dephasing effect of time dilation on internal degrees of freedom @xcite have been considered as possible origins of quantum decoherence . for what concerns our thought experiment ,
a similar setup can be found in the literature where the interference pattern of an electron passing through a double slit is destroyed by a distant measurement of its electric field .
this thought experiment can be traced back to bohr as quoted in @xcite , was discussed by hardy interviewed in @xcite and appears as an exercise in the book by aharanov and rohrlich @xcite .
moreover , recently different experiments involving interacting test particles have been proposed in order to discriminate the quantum nature of the gravitational field from a potentially classical description @xcite .
the original contribution of our work is that , imposing the consistency with relativistic causality , our thought experiment allows the derivation of a fundamental minimum time which is valid for _ any _ possible experiment involving macroscopic superpositions . in this sense
our bounds represent universal limitations having a role analogous to the heisenberg uncertainty principle in quantum mechanics . for this reason , while our results could be observable in advanced and specific experimental setups @xcite , their main contribution is probably a better understanding of the theory of quantum mechanics at macroscopic scales . for charged particles
we propose two different measurements for testing the coherence .
the first requires to accelerate the charge , and our bound on the discrimination time is due to the entanglement with the emitted photons . in the second ,
the bound can be instead ascribed to the presence of the vacuum fluctuations of the electromagnetic field . on the other hand
we also find an equivalent bound associated to quantum superposition of large masses .
what is the origin of this limitation ?
the analogy suggests that the validity of our bound could be interpreted as an indirect evidence for the existence of quantum fluctuations of the gravitational field , and of quantum gravitational radiation .
this work is structured in the following way : in section [ experiment ] we propose our thought experiment which suggests a minimum discrimination time for any macroscopic quantum superposition . in section [ time ] we derive a quantitative bound .
finally in sections [ ent ] and [ qed ] we check the consistency of our results with an explicit analysis of a charged particle interacting with the electromagnetic field . here
we propose two different measurements , and show that in both cases they are able to check the coherence only if their duration satisfies our fundamental limit .
consider the thought experiment represented in fig . [ gedanken ] , and described by the following protocol .
the protocol can be equivalently applied to quantum superpositions of large masses or large charges .
* protocol of the thought experiment * 1 .
alice has at disposal , in her laboratory , a massive / charged particle in a macroscopic superposition of a `` left '' and a `` right '' state : @xmath5 the wave functions of the two states are @xmath6 and @xmath7 , where @xmath8 is the relative separation of the superposition .
bob is in a laboratory at a distance @xmath9 from alice and containing a charged / massive test particle prepared in the ground state of a very narrow harmonic trap .
bob freely chooses between two options : doing nothing ( detector = _ off _ ) , or removing the trap ( detector = _ on _ ) . in the first case
the state of test particle remains unchanged while , in the second case , the dynamics is sensitive to the local newton / coulomb field generated by alice s particle and the global state will eventually become entangled .
if the detector is _ off _ , the initial quantum superposition is preserved , while if the detector is _ on _ the generation of entanglement eventually destroys the coherence of the reduced state of alice .
3 . alice performs an arbitrary measurement in her laboratory with the task of discriminating the coherent superposition from a statistical incoherent mixture of the two states @xmath10 and @xmath11 .
for example she could make an interference experiment , a measurement of the velocity , a projection on the symmetric / antisymmetric basis @xmath12 , _ etc._. the specific details of the experiment are irrelevant . depending on the result of the measurement
, alice deduces the choice of bob ( _ i.e. _ if the detector was _ on _ or _ off _ ) .
clearly , the previous thought experiment constitutes a communication protocol in which bob can send one bit of information to alice .
moreover , for a large enough mass @xmath0 or for a large enough charge @xmath3 , the test particle of bob can become entangled with alice s particle in an arbitrarily short time .
but then , apparently , bob can send a message to alice faster than light violating the fundamental principle of relativistic causality .
how can we solve this paradox ?
let us make a list of possible solutions : 1 .
it is impossible to prepare a macroscopic superposition state or to preserve its coherence because of some unknown intrinsic effect lying outside the theory of quantum mechanics .
once the superposition is created the particle is entangled with the gravitational / electromagnetic field and the reduced state of the particle is mixed . independently on the measurement of bob , alice always finds the state of the particle incoherent unless the entanglement process can be in some way reversed .
3 . independently on the presence or absence of entanglement , alice can not locally discriminate instantaneously whether the superposition is coherent or not .
more quantitatively we have that , if bob is able to generate entanglement in a time @xmath13 and if @xmath14 is the time necessary to alice for performing her discrimination measurement , then relativistic causality requires @xmath15 therefore , whenever entanglement can be generated in a time @xmath16 , we get a non - trivial lower bound on @xmath14 .
anomalous decoherence effects @xcite ( as _ e.g. _ the penrose spontaneous localization model ) are important open problems in the foundations of quantum mechanics and can not be excluded a priori . up to now
however their existence was never experimentally demonstrated and therefore , instead of closing our discussion by directly invoking point a ) , we try to remain within the framework of quantum mechanics and check if points b ) or c ) are plausible solutions .
the reader who is familiar with the field of open quantum systems may find the option b ) very natural . in standard non - relativistic quantum mechanics , the formation of entanglement between a system and its environment
is widely accepted as the origin of any observed form of decoherence .
indeed this approach has also been used to explain the decoherence of moving charged particles , mainly focusing to the double - slit interference experiment @xcite .
it has been recognized by previous works that in a double - slit experiment there is a limit to the charge of the particle above which photons are emitted due to the acceleration associated to the interference paths @xcite . for large charges then , the particle entangles with the emitted photons and this effect can destroy the interference pattern .
the reader can then notice that also in our case the particle needs to be accelerated when it is put in the superposition , and if it is charged it will radiate and can become entangled with the emitted photons .
similarly , an accelerated mass generates gravitational radiation and can become entangled with the emitted gravitons .
however , in appendix [ rad ] we prove that , if the accelerations are slow enough , the resulting quantum state of the electromagnetic field has almost overlap one with the vacuum , and therefore the particle does not get entangled with the emitted photons because no photons at all are emitted . the same argument
can be repeated for the gravitational radiation in the linear approximation .
the reader may now think that the particle in the superposition is entangled at least with its static coulomb electric field .
however , as we show in details in appendix [ dof ] , the static coulomb electric field is not a propagating degree of freedom since it vanishes in absence of electric charges .
thus , the hilbert space associated to the coulomb field is the same hilbert space of the particle , whose reduced state remains pure . what is then the solution of the paradox ?
a reasonable answer appears to be the final option c ) .
basically , even if the state of the particle is pure and coherent , alice can not instantaneously test this fact with a local experiment in her laboratory .
notice that the hypothesis c ) is weaker than the others and can logically coexist with a ) and b ) . clearly in every case in which a ) or b ) are valid alice can not make any useful experiment because decoherence has already happened .
therefore we conclude that the weaker and most general solution to the paradox is the fundamental limitation exposed in point c ) . in sections [ ent ] and [ qed ] ,
we propose two different measurements and show that they are both consistent with this limitation .
in the previous section we argued that relativistic causality requires a fundamental limitation : alice s discrimination experiment must be slow . but
how slow it has to be ? by construction any thought experiment of the class described before gives a lower bound on the discrimination time @xmath14 whenever @xmath16 . in what follows we are going to optimize over this class of experiments .
we anticipate that this approach leads to the following two bounds which constitute the main results of this work . * ( i ) minimum discrimination time for quantum superpositions of large masses * + given a particle of mass @xmath0 prepared in a macroscopic quantum superposition of two states separated by a distance @xmath1 , it is impossible to locally discriminate the coherent superposition from an incoherent mixture in a time ( up to a multiplicative numerical constant ) less than @xmath17 where @xmath18 is the planck mass @xmath19 * ( ii ) minimum discrimination time for quantum superpositions of large charges * + given a particle of charge @xmath3 prepared in a macroscopic quantum superposition of two states separated by a distance @xmath1 , it is impossible to locally discriminate the coherent superposition from an incoherent mixture in a time ( up to a multiplicative numerical constant ) less than @xmath20 where @xmath21 is the planck charge @xmath22 before giving a derivation of the previous results , we stress that both the bounds and are relevant only for @xmath23 and @xmath24 .
indeed for systems below the planck mass / charge , even if the bounds are formally correct , their meaning is trivial since any measurement of the state must at least interact with both parts of the superposition and this process requires at least a time @xmath25 .
let us first focus on the superpositions of massive particles and give a proof of the bound ( the proof of is analogous and will be given later ) .
it is easy to check that , for a sufficiently narrow trap ( detector = _ off _ ) the test mass of bob is insensitive to the gravitational force of alice s particle and remains stable in its ground state ( see appendix [ strength ] for details ) . on the contrary ,
if the trap is removed , the test mass will experience a different force depending on the position of alice s particle .
the two corresponding hamiltonians are : @xmath26 where @xmath27 is the mass of bob s particle , and @xmath28 and @xmath29 are the different gravitational forces associated to the left " and right " positions of alice s particle .
their difference @xmath30 where @xmath31 is the mass of alice s particle , determines the dipole force sensitivity that bob should be able to detect in order to induce the decoherence of the reduced state possessed by alice . given the initial state of the test mass @xmath32 , it is easy to check that entanglement can be generated in a time @xmath33 whenever the different time evolutions associated to @xmath34 and @xmath35 drive the test mass into almost orthogonal states , _
i.e. _ @xmath36 such time depends on the initial state @xmath37 and on the loschmidt echo operator @xmath38 which after two iterations of the baker - campbell - hausdorff formula can be written as @xmath39\;.\ ] ] neglecting the complex phase factor @xmath40 , @xmath41 is essentially a phase space displacement operator of the form @xmath42 , where @xmath43 are the shifts in position and momentum , respectively .
since the initial state @xmath44 of the test mass is the ground state of a very narrow harmonic trap , it will correspond to a localized gaussian wave packet which is very noisy in momentum and therefore we may focus only on the position shift and compare it with the position uncertainty @xmath45 of the initial state ( see appendix [ strength ] for a detailed proof ) .
we can argue that entanglement is generated only after a time @xmath46 such that @xmath47 apparently bob can generate entanglement arbitrarily quickly by reducing the position uncertainty @xmath45 .
however there is a fundamental limit to the localization precision which is set by the planck length .
it is widely accepted that no reasonable experiment can overcome this limit @xcite : @xmath48 from eq . ,
substituting eq . and using the minimum @xmath45 allowed by the constraint , we get @xmath49 as we have explained in the previous section , relativistic causality implies the inequality involving alice s measurement time @xmath14 and the entanglement time @xmath13 .
such inequality provides a lower bound on @xmath14 only if @xmath50 while it gives no relevant information for @xmath51 .
therefore we parametrize @xmath9 in terms of @xmath13 and a dimensionless parameter @xmath52 : @xmath53 using this parametrization , from eq . , we get @xmath54 from the causality inequality we have @xmath55 optimizing over @xmath52 we get @xmath56 this is , up to a multiplicative numerical constant , the bound given in eq . .
the calculation in the case in which we have a test charge instead of a test mass is almost identical .
the only difference is that eq .
is replaced by the coulomb counterpart @xmath57 where @xmath58 and @xmath59 are the charges of alice s and bob s particles , respectively , while the localization limit is replaced by bob s particle charge radius @xcite @xmath60 more details on the minimum localization of a macroscopic charge are given in appendix [ loc ] . from eq.s and , repeating exactly the previous argument one finds @xmath61 which is , up to a multiplicative numerical constant , the bound given in eq . .
in section [ time ] we have proved that relativistic causality requires that any measurement alice can perform to test the coherence of her superposition must require a minimum time , depending on the mass or charge of her particle . here
we focus on the electromagnetic case , and propose two different measurements to check the coherence of the superposition .
the first is a simplified version of the experiment proposed in ref .
let alice s particle have spin @xmath62 , and suppose that her superposition is entangled with the spin , i.e. is replaced by @xmath63 let now alice apply a spin - dependent force , that vanishes if the spin is up , while brings the particle from @xmath64 to @xmath65 if the spin is down .
if bob does not perform the measurement , the final state of alice s particle is @xmath66 where @xmath67 on the contrary ,
if bob induces a collapse of the wave - function , the final state is @xmath68 and alice can test the coherence with a measurement on the spin .
however , this protocol requires the particle to be accelerated if it has spin down and needs to be moved from @xmath64 to @xmath65
. then it will radiate , and it can entangle with the emitted photons . a semiclassical computation of the emitted radiation can be found in appendix [ rad ] .
. shows that such radiation is indistinguishable from the vacuum state of the field iff the motion lasts for at least the time required by our previous bound .
in section [ ent ] we have provided an example of experiment able to test the coherence .
the protocol requires to accelerate the charge , and if its duration is too short , the charge radiates and entangles with the emitted photons . the reader could now think that the bound on the time could be beaten with an experiment that does not involve accelerations .
an example of such experiment could seem to be a measurement of the canonical momentum of alice s particle . in this section ,
we first show that this measurement is indeed able to test the coherence of the superposition and then we estimate the minimum time necessary to perform it .
the canonical momentum of a charged particle coupled to the electromagnetic field is not gauge invariant , and therefore can not be directly measured .
alice can instead measure directly the velocity of her particle , that is gauge invariant .
however , its relation with the canonical momentum now contains the vector potential .
even if there is no external electromagnetic field , the latter is a quantum - mechanical entity , and is subject to quantum vacuum fluctuations .
then , the fluctuations of the vector potential enter in the relation between velocity and momentum .
if alice is not able to measure the field outside her laboratory , she can measure only the velocity of her particle ( see appendix [ dof ] for a detailed discussion ) , and can reconstruct its canonical momentum only if the fluctuations are small .
we show that in an instantaneous measurement these fluctuations are actually infinite .
however , if alice measures the average of the velocity over a time @xmath69 , they decrease as @xmath70 , and can be neglected if @xmath69 is large enough . this minimum time
is found consistent with the bound derived in section [ time ] . in order to simplify our formulas , in this section and in the related appendices we put as in @xcite @xmath71 these constant
will be put back into the final result .
let us first show that alice can test the coherence with a measurement of the canonical momentum of her particle .
let the particle be in the coherent superposition of two identical wave - packets centered in different points , with wave - function @xmath72 where @xmath73 is an arbitrary phase .
the probability distribution of the canonical momentum @xmath74 is the modulus square of the fourier transform of the wave - function : @xmath75 and she can test the coherence of the superposition from the interference pattern in momentum space generated by the cosine . indeed , an incoherent statistical mixture would be associated to the probability distribution @xmath76 where the cosine squared is replaced by @xmath77 , its average over the phase @xmath73 .
notice from that , in order to be actually able to test the coherence , alice must measure the canonical momentum with a precision of at least @xmath78 where @xmath79 .
this precision increases with the separation of the wave - packets , e.g. for @xmath80 , it is @xmath81 .
let now alice s particle carry an electric charge @xmath3 .
we want to take into account the quantum vacuum fluctuations of the electromagnetic field , so quantum electrodynamics is required .
the global hilbert space is then the tensor product of the hilbert space of the particle @xmath82 with the hilbert space of the field @xmath83 .
the reader can find in appendix [ emf ] the details of the quantization .
the position and canonical momentum operators of alice s particle @xmath84 and @xmath85 still act in the usual way on the particle hilbert space alone , so that the argument of subsection [ momentum ] remains unchanged .
the full interacting hamiltonian of the particle and the electromagnetic field is @xmath86 where @xmath87 is the vector - potential operator @xmath88 of with the coordinate @xmath89 replaced with the position operator @xmath84 , and @xmath90 is the free hamiltonian of the electromagnetic field defined in . due to the minimal - coupling substitution
, the operator associated to the velocity of the particle is @xmath91=\frac{1}{m}\left(\hat{\mathbf{p}}-q\;\hat{\mathbf{a}}\left(\hat{\mathbf{x}}\right)\right)\;,\ ] ] that contains the operator vector - potential , and acts also on the hilbert space of the field .
the canonical momentum can be reconstructed from the velocity with @xmath92 if the second term in the rhs can be neglected . with the help of the commutation relations , a direct computation of the variance of @xmath93 on the vacuum state of the field
gives @xmath94 that has a quadratic divergence for @xmath95 due to the quantum vacuum fluctuations .
this divergence can be cured averaging the vector potential over time with a smooth function @xmath96 .
we must then move to the heisenberg picture , where operators explicitly depend on time , and we define it to coincide with the schrdinger picture at @xmath97 , the time at which alice measures the velocity . since the divergence in does not depend neither on the mass nor on the charge of alice s particle and is proportional to the identity operator on the particle hilbert space @xmath98 , it has nothing to do with the interaction of the particle with the field . then the leading contribution to the result
can be computed evolving the field with the free hamiltonian @xmath90 only , i.e. with @xmath99 defining the time - averaged vector potential as @xmath100 its variance over the vacuum state of the field is now @xmath101 where @xmath102 is the fourier transform of @xmath96 .
taking as @xmath96 a normalized gaussian function of width @xmath69 centered at @xmath97 : @xmath103 we get as promised a finite result proportional to @xmath70 : @xmath104 then , if alice estimates one component of the canonical momentum ( say the one along the @xmath105 axis ) with the time average of the velocity taken with the function @xmath96 , she commits an error of the order @xmath106 comparing with the required precision to test the coherence , the minimum time required is @xmath107 in agreement with the bound imposed by relativistic causality alone .
in this work we have studied the limitations that the gravitational and electric fields produced by a macroscopic particle impose on quantum superposition experiments .
we found that , in order to avoid a contradiction between quantum mechanics and relativistic causality , a minimum time is necessary in order to discriminate a coherent superposition from an incoherent statistical mixture .
this discrimination time is proportional to the separation distance of the superposition and to the mass ( or charge ) of the particle . in the same way
as the heisenberg uncertainty principle inspired the development of a complete theory of quantum mechanics , our fundamental and quantitative bounds on the discrimination time can be useful for the development of current and future theories of quantum gravity .
moreover , despite an experimental observation of our results clashes with the difficulty of preparing superpositions of masses above the planck scale , the current technological progress on highly massive quantum optomechanical and electromechanical systems provides a promising context @xcite for testing our predictions .
we thank seth lloyd , leonardo mazza , and andrea tomadin for fruitful discussions .
gdp thanks m. c. sormani for useful comments . 99 j. a. wheeler , w. h. zurek , quantum theory and measurement ( princeton , 1983 ) .
g. taylor , proc .
soc . math .
sci . * 15 * , 114 ( 1909 ) .
b. dopfer , ph.d .
thesis , ( university of innsbruck , 1998 ) .
o. donati , g. p. missiroli and g. pozzi , am .
* 41 * , 639 ( 1973 ) .
a. tonomura , j. endo , t. matsuda , t. kawasaki , h. ezawa , am .
j. phys * 57 * . 117
a. zeilinger , r. ghler , c. g. shull , w. treimer , and w. mampe , rev .
* 60 * , 1067 ( 1988 ) .
b. p. anderson , m. a. kasevich , science * 282 * , 5394 ( 1998 ) . c. monroe , d. m. meekhof , b. e. king , d. j. wineland , science * 272 * , 1131 ( 1996 ) .
m. arndt , o. nairz , j. vos - andreae , c. keller , g. van der zouw , a. zeilinger , nature * 401 * , 680 ( 1999 ) .
s. eibenberger , s. gerlich , m. arndt , m. mayor , j. txen , phys .
phys . * 15 * , 14696 ( 2013 ) .
r. penrose , general relativity and gravitation * 28 * , 581 - 600 ( 1996 ) g. c. ghirardi , a. rimini , and t. weber , phys .
d * 34 * , 470 ( 1986 ) .
f. karolyhazy , il nuovo cimento a * 42 * , 390 ( 1996 ) . l. disi , phys . rev . a * 40 * , 1165 ( 1989 ) .
a. bassi , k. lochan , s. satin , t. p. singh and h. ulbricht , rev .
phys . * 85 * , 471 ( 2013 ) .
g. baym and t. ozawa , pnas * 106 * , 3035 ( 2009 )
breuer and f. petruccione , phys .
a * 63 * , 032102 ( 2001 ) m. p.
blencowe , phys .
lett . * 111 * , 021302 ( 2013 ) f. suzuki and f. queisser , j. phys .
: conf . ser . *
626 * , 012039 ( 2015 )
jaekel , b. lamine , a. lambrecht , s. reynaud and p. m. neto , _ beyond the quantum _
( world scientific , 2007 ) , ch .
10 , 125 - 134 i. pikovski , m. zych , f. costa , c. brukner , nature physics , 1745 ( 2015 ) .
m. schlosshauer , _ elegance and enigma _ , springer berlin ( 2011 ) .
y. aharanov , d. rohrlich , _ quantum paradoxes _ , wiley - vch , weinheim ( 2005 ) .
d. kafri , j. m. taylor , g. j. milburn , new j. phys .
* 16 * , 065020 ( 2014 ) .
m. bahrami , a. bassi , s. mcmillen , m. paternostro , h. ulbricht , arxiv:1507.05733 ( 2015 ) . o. romero - isart , a. c. pflanzer , f. blaser , r. kaltenbaek , n. kiesel , m. aspelmeyer and j. i. cirac , phys .
107 * , 020405 ( 2011 ) .
w. marshall , c. simon , r. penrose , and d. bouwmeester , phys .
letts . , * 91 * , 130401 ( 2003 ) .
i. pikovski , m. r. vanner , m. aspelmeyer , m. s. kim and c. brukner , nature physics * 8 * , 393 , ( 2012 ) .
m. bawaj , c. biancofiore , m. bonaldi , f. bonfigli , a. borrielli , g. di giuseppe , l. marconi , f. marino , r. natali , a. pontin , g. a. prodi , e. serra , d. vitali and f. marin , nature comm . * 6 * , 7503 ( 2015 ) .
r. schnabel , phys .
a * 92 * , 012126 ( 2015 ) m. scala , m. s. kim , g. w. morley , p. f. barker , and s. bose , phys .
lett . * 111 * , 180403 ( 2013 ) . c. wan , m. scala , s. bose , a. c. frangeskou , atm a. rahman , g. w. morley , p. f. barker , m. s. kim , arxiv:1509.00724 ( 2015 ) .
h. salecker and e. p. wigner , phys . rev . * 109 * , 571 ( 1958 ) . c. a. mead , phys .
rev . * 135 * , b849 ( 1964 ) .
s. hossenfelder , living rev .
relativity 16 , 2(2013 ) .
s. weinberg , _ the quantum theory of fields , volume i : foundations _ , cambridge university press ( 1995 ) j. d. jackson , _ classical electrodynamics _
, 3rd edition , wiley ( 1999 ) c. cohen - tannoudji , j. dupont - roc and g. grynberg , _ photons and atoms : introduction to quantum electrodynamics _ , wiley ( 1997 ) s. m. barnett and p. m. radmore , methods in theoretical quantum optics , oup ( 1997 )
in this appendix we show that for bob a measurement of the position is always better than a measurement of the momentum , i.e. it permits to distinguish the force difference @xmath108 in a shorter time .
let @xmath109 be the frequency of the harmonic trap .
the spatial width of its ground state is given by @xmath110 this ground state is insensible to the force difference @xmath108 iff the displacement that it generates is less than @xmath45 , i.e. @xmath111 eliminating @xmath109 with , the inequality becomes hence @xmath112 which is the condition we have to enforce to ensure that bob s detector is ineffective when switched off .
suppose then that , after switching on the detector , bob tries to distinguish the two states of alice by a measurement of @xmath113 : accordingly the momentum spread @xmath114 of his initial state must be lower than the displacement in momentum @xmath115 given by .
recalling that heisenberg s uncertainty principle @xmath116 is saturated by a gaussian pure state , the minimum time after which bob can distinguish is @xmath117 on the other hand , from the minimum discrimination time with a measurement of @xmath105 is @xmath118 the reader can check that implies @xmath119 , i.e. if the trap is strong enough to be insensible to the force difference , for bob it is always better to measure the position of his particle rather than its momentum .
in this appendix we prove that the minimum width over which a charge @xmath3 greater than the planck charge @xmath120 can be localized is its charge radius .
let us suppose to use a harmonic trap of frequency @xmath109 to localize the charge .
one could think that in principle , with a strong enough trap , the charge can be arbitrarily localized .
however , from the larmor formula @xcite we know that a classical particle with charge @xmath3 following a harmonic motion of frequency @xmath109 and width @xmath45 loses into electromagnetic radiation a power @xmath121 in the quantum case , the charge radiates until it gets to the ground state of the trap , where it can not radiate anymore since there are no other states with a lower energy to go .
however , if the trap is very strong , its ground state is very localized , and therefore has a great uncertainty in velocity .
since any moving charge generates a magnetic field , this velocity uncertainty generates a large uncertainty in the magnetic field , resulting in a large entanglement between the state of the particle and the state of the field .
qualitatively , this happens when the energy classically radiated in a period becomes greater than @xmath122 , the energy of the first excited state . combining with
, this happens exactly when the localization @xmath45 becomes smaller that the charge radius : @xmath123 then , if we want the reduced state of the particle to remain pure , we can localize it only up to the limit in .
we recall here the basics of the quantization of the electromagnetic field .
more details can be found in ref .
@xcite .
we denote with @xmath124 an operator in the schrdinger picture , and with @xmath125 its heisenberg - picture counterpart .
the two pictures are defined to coincide for @xmath97 , i.e. @xmath126 .
we recall that in the heisenberg picture the operators are evolved with the full interacting hamiltonian .
an hamiltonian formulation of electrodynamics requires the introduction of the scalar and vector potentials @xmath127 and @xmath128 .
it is convenient to fourier - transform with respect to @xmath89 .
the potentials are related to the electric and magnetic fields by @xmath129 we choose the coulomb gauge , in which the divergence of the vector potential is set to zero at the operator level : @xmath130 it is now convenient to define the ladder operators @xmath131 satisfying the constraint @xmath132 as a consequence of . the definition in can be inverted : @xmath133 the ladder operators satisfy the equal - time canonical commutation relations @xmath134&=&\pi^{ij}(\mathbf{k})\;(2\pi)^3\delta^3(\mathbf{k}-\mathbf{q})\\ \left[\hat{a}^i(\mathbf{k},t),\;\hat{a}^{j}(\mathbf{q},t)\right]&=&\left[\hat{a}^{i\dag}(\mathbf{k},t),\;\hat{a}^{j\dag}(\mathbf{q},t)\right]=0\ ; , \quad\end{aligned}\ ] ] where @xmath135 is the projector onto the subspace orthogonal to @xmath136 : @xmath137 the vacuum state of the field @xmath138 is defined as the state annihilated by all the schrdinger - picture annihilation operators : @xmath139 besides , the @xmath140 together with their hermitian conjugates @xmath141 generate the whole observable algebra of @xmath83 .
maxwell s equations determine the time evolution of the ladder operators : @xmath142 where @xmath143 is the operator associated to the current density of the quantum system interacting with the electromagnetic field .
is easily integrated : @xmath144 where we have imposed the heisenberg and schrdinger pictures to coincide at @xmath97 , i.e. @xmath145 . in the free case , i.e. when the current vanishes at the operator level ( @xmath146 ) , the relation between the two pictures is given by the free hamiltonian @xmath147 i.e. @xmath148
we introduce now the set of coherent states of the electromagnetic field .
our formalism is analogue to the one of refs .
@xcite , where the reader is referred for further details . for any function @xmath149
subject to the constraint @xmath150 define the unitary displacement operator @xmath151\equiv\exp\left(\int\left(f_i(\mathbf{k})\,\hat{a}^{i\dag}(\mathbf{k})-f_i^*(\mathbf{k})\,\hat{a}^i(\mathbf{k})\right)\frac{d^3k}{(2\pi)^3}\right)\;,\\ & \hat{d}^\dag[f]=\hat{d}[-f]\;,\end{aligned}\ ] ] acting on the ladder operators as @xmath152\;\hat{a}^i(\mathbf{k})\;\hat{d}[f]=\hat{a}^i(\mathbf{k})+f^i(\mathbf{k})\;.\ ] ] their composition rule is @xmath153\;\hat{d}[g]=\hat{d}[f+g ] \times \ ; \nonumber \\ & & \times\exp{\left(\frac{1}{2}\int\left(f^i(\mathbf{k})\,g_i^*(\mathbf{k})-f_i^*(\mathbf{k})\,g^i(\mathbf{k})\right)\frac{d^3k}{(2\pi)^3}\right)}\;. \qquad\end{aligned}\ ] ] we can now define the coherent states with a displacement operator acting on the vacuum state of the field : @xmath154|0\rangle\in\mathcal{h}_f\;,\ ] ] that are eigenstates of the annihilation operators : @xmath155 their overlap is @xmath156
in this appendix we consider a classical charged particle following an accelerated trajectory and coupled to the quantum electromagnetic field .
we show that the radiation classically emitted by the particle induces on the quantum field a displacement operator , and if the initial state is the vacuum , it is brought to a coherent state ( see appendix [ coh ] for the definition ) .
it turns out that , if the accelerations of the particle are smooth enough , this coherent state has overlap almost one with the vacuum , i.e. it is not distinguishable from it .
then , for a slow motion the particle does not radiate photons at all , and therefore it does not get entangled with the field .
let @xmath157 be the classical current density associated to the trajectory of the particle .
looking at the time evolution equation for the ladder operators , and recalling and , it is easy to show that such evolution is provided by a displacement operator , i.e. @xmath158\;e^{i\hat{h}_ft}\;\hat{a}^i(\mathbf{k})\;e^{-i\hat{h}_ft}\;\hat{d}[f]\;,\ ] ] where @xmath159 since the ladder operators generate the whole observable algebra of @xmath83 , if the fields starts in the vacuum , its time - evolved state is the coherent state @xmath160 .
its overlap with the vacuum can be computed with : @xmath161 we now consider a point particle carrying charge @xmath3 that starts in @xmath162 at @xmath97 , and in a time @xmath163 is brought to the position @xmath164 with a trajectory described by @xmath165 .
the current density is then @xmath166 where @xmath167 is the particle velocity . for wavelengths large with respect to the extension of the motion ,
i.e. for @xmath168 we want to look at the state of the field after the particle has reached the new position @xmath164 , i.e. for @xmath169 .
since the velocity @xmath170 vanishes for @xmath171 and @xmath172 , the displacement of becomes @xmath173 where @xmath174 is the fourier transform of the velocity .
putting into , the overlap becomes @xmath175 for simplicity we consider a one - dimensional motion , and we put the @xmath105 axis in the direction of @xmath176 . as an example , we take @xmath177 satisfying the conditions @xmath178 the fourier transform of the velocity is @xmath179 and the overlap @xmath180 where @xmath181 is the sine integral function @xmath182 looking at , the dominant contribution to the integral in comes from the region @xmath183 .
the approximation in is then valid iff @xmath184 , i.e. if the motion is not relativistic .
the final result tells us that for a fixed distance @xmath1 , no photons are radiated if the motion lasts for at least @xmath185 then , alice can always create the coherent superposition without entangling her particle with the emitted photons provided she has enough time to do it .
besides , if alice wants to perform the measurement of section [ ent ] , she needs at least a time to bring the state @xmath64 back to @xmath65 if she does not want to entangle with the emitted photons .
in this appendix we explain in detail why when alice s charged particle is in the superposition , despite it generates a static electric field that depends on its position , it is still not entangled with the field , and the global state is a product with the field part in the vacuum .
the first maxwell s equation reads @xmath186 and completely determines the electric potential operator @xmath187 in terms of the charge density operator @xmath188 : @xmath189 putting together and , the electric field is given by @xmath190 then , the longitudinal ( i.e. proportional to @xmath136 ) component of the electric field operator is determined by the charge - density operator , and acts on the hilbert space of the particle alone .
therefore , even if the field is in its vacuum state , the expectation value of the electric field is the static coulomb electric field generated by the expectation value of the charge density , and hence depends on the particle wave - function .
this means that the state of the field alone does not contain all the information on the electric field , since its longitudinal component is encoded into the state of the particle . seen from a different perspective , the longitudinal component of the electric field is not a dynamical propagating degree of freedom , since it vanishes in absence of external charges and is completely determined by them , so there is no hilbert space associated to it .
the hilbert space of the field contains only the degrees of freedom associated to the electromagnetic radiation , i.e. the magnetic field and the transverse ( orthogonal to @xmath136 ) component of the electric field .
then in a product state with the field part in the vacuum only these components are in the vacuum mode , while there can be a static electric field depending on the state of the particle .
in this appendix we explain in detail why alice can measure only the velocity of her particle , and not its canonical momentum , if she is constrained to remain in her laboratory , which has the size of the support of the wave - function of the particle .
the wave - function @xmath191 of a particle carrying electric charge @xmath3 coupled to an electromagnetic field is invariant under the joint gauge transformation @xcite @xmath192 the canonical momentum @xmath193 is not gauge invariant , but transforms in the heisenberg picture as @xmath194 and therefore alice can not measure it directly . the reader can easily check using and that the velocity given by is gauge invariant , as it has to be .
alice can then measure directly the velocity , and reconstruct from it the canonical momentum .
however , the relation between them ( eq . ) contains the vector potential , that from is not gauge invariant , and can not be directly measured . in the coulomb gauge , it is possible to invert and express the vector potential in terms of the magnetic field , that is gauge invariant and can be actually measured by alice : @xmath195 putting together and , we get @xmath196 however , reconstructing the canonical momentum from the velocity with requires alice to measure the magnetic field in the whole space .
even if she can allow for some error in the reconstruction , the region in which she has to measure the field increases with the charge @xmath3 , and can extend well outside the support of the wave - function . | we consider a thought experiment where the preparation of a macroscopically massive or charged particle in a quantum superposition and the associated dynamics of a distant test particle apparently allow for superluminal communication .
we give a solution to the paradox which is based on the following fundamental principle : any local experiment , discriminating a coherent superposition from an incoherent statistical mixture , necessarily requires a minimum time proportional to the mass ( or charge ) of the system . for a charged particle , we consider two examples of such experiments , and show that they are both consistent with the previous limitation . in the first
, the measurement requires to accelerate the charge , that can entangle with the emitted photons . in the second
, the limitation can be ascribed to the quantum vacuum fluctuations of the electromagnetic field . on the other hand , when applied to massive particles our result provides an indirect evidence for the existence of gravitational vacuum fluctuations and for the possibility of entangling a particle with quantum gravitational radiation . |
synchronization in the sense of coordinated behavior in time is essential for any efficient organization of systems , natural as well as artificial ones . in artificial systems like factories the sequence of production processes
should be synchronized in a way that time - and space consuming storage of input or intermediate products is avoided .
the same applies to natural systems like networks of cells which obviously perform very efficiently in fulfilling a variety of functions and tasks . in a more special sense
, the synchronized behavior refers to oscillators with almost identical individual units , in particular to phase oscillators with continuous interactions as described by the kuramoto model @xcite .
these sets of limit - cycle oscillators describe synchronization phenomena in a wide range of applications @xcite .
one of these applications is pattern formation in chemical oscillatory systems @xcite , described by reaction - diffusion systems .
a well - known example for such a chemical system is the belousov - zabotinsky reaction , a mixture of bromate , bromomalonic acid and ferroin , which periodically changes its color corresponding to oscillating concentrations of the red , reduced state and the blue , oxidized state . in these systems , expanding target - like waves and rotating spiral waves are typical patterns @xcite . as kuramoto showed ( as e.g. in @xcite ) , these dynamical systems can be well approximated by phase oscillators if the interaction is weak . for reaction - diffusion equations with a nonlinear interaction term he predicted circular waves with strong similarities to the experimental observation of target patterns . in these systems , defects or impurities
seem to play the role of pacemakers , driving the system into a synchronized state .
therefore he treated pacemakers as local `` defect '' terms leading to heterogeneities in the reaction - diffusion equations .
in this paper we consider two types of pacemakers . in the first part they are introduced as `` defects '' in the sense that they have a different natural frequency from the rest of the system , whose oscillators have either exactly the same frequency or small random fluctuations about some common average by assumption , where in the second case the pacemaker s frequency is clearly different from the values of a typical fluctuation
actually , we will not consider the latter case , since small fluctuations do not lead to any qualitative change .
the ad - hoc distinction of the pacemaker may reflect natural and artificial systems with built - in impurities like those in the belousov - zhabotinsky system .
we have recently studied these systems for nearest neighbor interactions @xcite and regular networks in @xmath0 dimensions for periodic and open boundary conditions . for such systems
we have shown that it is the mean distance of all other nodes from the pacemaker ( the so - called depth of the network ) that determines its synchronizability .
periodic boundary conditions facilitate synchronization as compared to open boundary conditions . here
we extend these results to larger interaction range .
the effect of an extended interaction range on oscillatory systems without pacemaker were considered in @xcite .
we analytically derive the entrainment frequency for arbitrary interaction range between next - neighbor and all - to all interactions .
we derive the entrainment window for all- to - all couplings analytically , for intermediate couplings numerically .
the entrainment window depends non - monotonically on the interaction range , so that the synchronization transition is reentrant . in our system
the reentrance can be explained in terms of the weight of the interaction term and the normalization of the coupling strength . in the second part
we consider a system of oscillators without a `` defect '' , all oscillators on the same footing up to the difference in their natural frequencies .
an obvious choice for the natural frequencies would be a gaussian distribution or another random distribution to describe fluctuations in natural frequencies in otherwise homogeneous systems , without impurities .
such systems have been studied by blasius and tnjes @xcite for a gaussian distribution .
the authors showed that in this case an asymmetric interaction term is needed to have synchronization , this time driven by a dynamically established pacemaker , acting as a source of concentric waves .
this result shows that conditions exist under which a system establishes its own pacemaker in a `` self - organized '' way .
in contrast to the random distribution of natural frequencies we are interested in a ( deterministic ) gradient distribution , in which the natural frequencies linearly decrease over a certain region in space .
such a deterministic gradient alone is probably not realistic for natural systems , but considered as a subsystem of a larger set of oscillators with natural frequency fluctuations it may be realized over a certain region in space .
here we suppress the fluctuations and focus on the effect of the gradient alone . as it turns out ,
the asymmetry in the natural frequency distribution is sufficient for creating a pacemaker as center of circular waves in oscillator - phase space , without the need for an asymmetric term in the interaction .
the oscillator with the highest natural frequency becomes the source of circular waves in the synchronized system .
we call it dynamically induced as its only inherent difference is its local maximum at the boundary in an otherwise `` smooth '' frequency distribution
. the type of pattern created by the pacemaker has the familiar form of circular waves . beyond a critical slope in the natural frequencies ,
full synchronization is lost .
it is first replaced by partial synchronization patterns with bifurcation in the frequencies of synchronized clusters , before it gets completely lost for too steep slopes .
the outline of the paper is as follows . in the first part we treat the pacemaker as defect and
derive the common entrainment frequency for arbitrary interaction range and topology ( section [ sec.i ] ) .
next we determine the entrainment window , analytically for all - to all coupling and numerically for intermediate interaction range ( section [ sec.i1 ] ) . here
we see the reentrance of the transition as a function of the interaction range ( section [ sec.i2 ] ) . in the second part ( section [ sec.ii ] ) we consider dynamically induced pacemakers , first without asymmetric interaction term , for which we analytically derive the synchronization transition as a function of the gradient in the frequencies ( sections [ sec.ii1 ] ) . in section [ sec.ii2 ] we add an asymmetric term , so that the pattern formation is no longer surprising due to the results of @xcite , but the intermediate patterns of partial synchronization are different . in section [ sec.iv ]
we summarize our results and conclusions .
the system is defined on a network , regular or small - world like . to each node @xmath1 , @xmath2 , we assign a limit - cycle oscillator , characterized merely by its phase @xmath3 , which follows the dynamics @xmath4 with the following notations .
the frequency @xmath5 denotes the natural frequencies of the system . in this system
we treat the pacemaker as a defect .
it is labelled by @xmath6 and has a natural frequency that differs by @xmath7 from the frequency of the other oscillators having all the same frequency @xmath5 .
@xmath8 denotes the kronecker delta .
the constant @xmath9 parameterizes the coupling strength .
we consider regular networks and choose @xmath10 @xmath11 the distance between nodes @xmath12 and @xmath1 , that is @xmath13\right ) \label{def : distance}\ ] ] on a one - dimensional lattice with periodic boundary conditions . in two or higher dimensions it is the shortest distance in lattice links .
the parameter @xmath14 tunes the interaction range .
alternatively we consider @xmath15 as the adjacency matrix on a small - world topology : @xmath16 if the nodes @xmath12 and @xmath1 are connected and @xmath17 otherwise ) .
moreover , @xmath18 is the degree of the @xmath1-th node , it gives the total number of connections of this node in the network .
this system was considered before in @xcite for nearest - neighbor interactions ( @xmath19 ) on a @xmath0-dimensional hypercubic lattice and on a cayley tree . here
we extend the results to long - range interactions via @xmath20 . in the appendix of @xcite we derived the common entrainment frequency @xmath21 in the phase - locked state , for which @xmath22 for all @xmath1 , to be given as @xmath23 in the rotated frame , in which the natural frequency @xmath5 is zero .
in particular such a result is obtained directly from system ( [ def : model ] ) , applying the transformation to the rotating frame @xmath24 @xmath25 , multiplying the @xmath1-th equation by @xmath26 , summing over all equations and then using the fact that @xmath27 , because the symmetry of the adjacency matrix @xmath28 and the antisymmetry of the @xmath29-function .
whenever @xmath30 is independent of @xmath1 as it happens for periodic boundary conditions and homogeneous degree distributions , in particular for @xmath31 ( all - to - all coupling ) , the common frequency ( in the non - rotated frame ) is given by @xmath32 independent on @xmath33 . in the limit @xmath34 the system synchronizes at @xmath5 .
+ analytical results for the entrainment window were derived for open and periodic boundary conditions and @xmath35 in @xcite . for @xmath31 ,
the result is the same as for @xmath36 , periodic boundary conditions and @xmath37 @xmath38 this is seen as follows .
since for all - to - all couplings the mean - field approximation becomes exact , it is natural to rewrite the dynamical equations ( [ def : model ] ) in terms of the order parameter @xmath39 defined according to @xmath40 this quantity differs from the usual order parameter just by the fact that the sum runs over @xmath41 .
eq.s([def : model ] ) then take the form @xmath42 using @xmath43 for all @xmath44 as one of the possible solutions and the fact that the @xmath29-function is bounded , these equations imply the entrainment window as given by ( [ eq : critic ] ) with @xmath45 the total number of nodes minus @xmath46 . as we have shown in @xcite ,
the same dependence in terms of @xmath45 holds for a @xmath0-dimensional hypercubic lattice with @xmath47 nodes per side and nearest neighbor couplings , for which @xmath48 , so that the entrainment window for nearest neighbor interactions and on a regular lattice decays exponentially with the depth of the network in the limit of infinite dimensions @xmath0 ; for random and small - world networks the same type of decay was derived in @xcite .
the intermediate interaction range @xmath49 can not be treated analytically . here
the numerical integration of the equations ( [ def : model ] ) ( via a fourth - order runge - kutta algorithm with time step @xmath50 and homogeneous initial conditions ) shows that the entrainment window is even smaller than in the limiting cases @xmath31 or @xmath51 for otherwise unchanged parameters
( that is fixed size @xmath45 , dimension @xmath0 , coupling @xmath52 ) . for these cases formula ( [ eq : critic ] ) gives an upper bound on the entrainment window , as one can easily see from eq.[def : model ] for @xmath53 because @xmath54 , @xmath55 . actually the entrainment window becomes smaller for intermediate interaction range , a feature leading to the reentrance of the synchronization transition as function of @xmath33 , as we shall see in the figures [ fig : reentalpha ] and [ fig : reentkandp ] below .
reentrance phenomena are observed in a variety of phase transitions , ranging from superconductor - insulator transitions as function of temperature @xcite and noise - induced transitions @xcite to chaotic coupled - map lattices as function of the coupling @xcite . in @xcite
a reentrance phenomenon is discussed as an artifact of the approximation .
reentrance of phase transitions is challenging as long as it appears counterintuitive .
for example , it is counterintuitive when synchronization is first facilitated for increasing coupling and later gets blocked when the coupling exceeds a certain threshold , so that synchronization depends non - monotonically on the coupling . as we see in fig .
[ fig : reentalpha ] for a one - dimensional system with periodic boundary conditions and for various system sizes @xmath45 , the entrainment window @xmath56 depends non - monotonically on the interaction range , parameterized by @xmath33 .
this dependence is easily explained by looking at the total weight coming from the sum of interactions .
this contribution depends non - monotonically on @xmath33 , as @xmath57 decreases for increasing @xmath33 , while the coupling strength increases under the same change of @xmath33 , so that there is a competition between the two factors leading to a minimum of the weight for a certain value of @xmath33 . in contrast , if we scale the coupling strength by a constant factor @xmath58 , independently of the degree of the node @xmath1 , the overall weight of the sum decreases for increasing @xmath33 and along with this , the drive to the synchronized state . as it is shown in the inset of fig .
[ fig : reentalpha ] , the threshold monotonically decreases with @xmath33 if the coupling strength is normalized by @xmath59 independently of the actual degree , as expected from our arguments above .
the second inset shows the @xmath45-dependence of the value of @xmath33 for which synchronization is most difficult to achieve , it converges to @xmath60 for large @xmath45 .
similar non - monotonic behavior is observed if the interaction range is tuned via the number @xmath61 of nearest neighbors on a ring rather than via @xmath33 , see fig . [
fig : reentkandp]a and for a small - world topology @xcite , see fig .
[ fig : reentkandp]b . here
the small - world is constructed by adding random shortcuts with probability @xmath62 from a regular ring with nearest neighbor interactions ( @xmath63 ) to an all - to - all topology ( @xmath64 ) @xcite .
, for a one - dimensional lattice with @xmath65 sites , periodic boundary conditions and for interaction strength between the oscillators @xmath1 and @xmath12 suppressed by the power @xmath33 of their distance @xmath11 .
the main plot indicates the reentrance of the synchronization transition .
the upper inset shows the size dependence of @xmath66 , for which the entrainment window becomes minimal , it saturates at @xmath67 .
the lower inset shows the monotonic dependence on @xmath33 for constant coupling strength @xmath58 .
the figure was obtained for homogeneous initial conditions .
the qualitative features remain the same for random initial conditions.,scaledwidth=47.0% ] ( a ) and of the probability @xmath62 for adding shortcuts in a small - world topology ( b ) for a one - dimensional lattice with periodic boundary conditions and @xmath65 sites .
average values in ( b ) are taken over @xmath68 realizations , error bars correspond to the standard deviation . again
the non - monotonic behavior indicates reentrance of the synchronization transition.,title="fig:",scaledwidth=47.0% ] ( a ) and of the probability @xmath62 for adding shortcuts in a small - world topology ( b ) for a one - dimensional lattice with periodic boundary conditions and @xmath65 sites .
average values in ( b ) are taken over @xmath68 realizations , error bars correspond to the standard deviation .
again the non - monotonic behavior indicates reentrance of the synchronization transition.,title="fig:",scaledwidth=47.0% ]
in view of dynamically induced pacemakers we study the dynamical system @xmath69 without a pacemaker .
the interaction term @xmath70 is given by @xmath71\ ; , \label{def : gamma}\ ] ] where @xmath72 corresponds to antisymmetric interaction , @xmath73 to asymmetric interaction . in the following
we consider the cases @xmath74 and @xmath75 .
the value @xmath75 is chosen for convenience in order of comparing our results with those of blasius and tnjes @xcite , but different values of @xmath73 lead to qualitatively the same features .
+ we focus on lattices with periodic boundary conditions in one and two dimensions .
moreover we consider only nearest neighbor interactions , so that in @xmath76 we have @xmath77 @xmath25 and @xmath78 @xmath25 in @xmath79 .
the natural frequencies @xmath80 are chosen from a gradient distribution@xmath81 where @xmath82 is the network distance between the oscillators @xmath1 and @xmath12 , given by ( [ def : distance ] ) in one dimension and by the shortest path along the edges of the square lattice in two dimensions .
the oscillator @xmath6 is the oscillator with the highest natural frequency . for a one - dimensional lattice with periodic boundary conditions ( ring ) and @xmath65 oscillators , choosing @xmath83 for simplicity , @xmath84 , with @xmath85 for @xmath45 even and @xmath86 for @xmath45 odd . in the following we choose @xmath45 as even ; the case of @xmath45 odd is very similar .
next we derive the common frequency in case of phase - locking @xmath87 , @xmath25 . summing the equations of ( [ eq1 ] ) and using the antisymmetry of @xmath70 for @xmath72
, we obtain @xmath88 = \delta\omega \frac{n}{2}\ ; , \end{array}\ ] ] leading to @xmath89 so that @xmath90 for @xmath34 .
it is convenient to rewrite the differential equations ( [ eq1 ] ) in terms of phase lags @xmath91 , so that the equation for the oscillator at position @xmath92 reads @xmath93 for the other oscillators we obtain @xmath94 \\ \rightarrow \ ; \sin{(\theta_2 ) } = 2\frac{\omega-\omega_1}{k } + \sin{(\theta_1 ) } \end{array } \;\;\ ; , \ ] ] @xmath95 \\ \rightarrow \ ; \sin{(\theta_3 ) } = 2\frac{\omega-\omega_2}{k } + \sin{(\theta_2 ) } \end{array } \;\;\ ; , \ ] ] so that in general @xmath96 \\ \rightarrow \ ; \sin{(\theta_{i } ) } = 2 \frac{\omega-\omega_{i-1}}{k } + \sin{(\theta_{i-1 } ) } \end{array}\ ] ] or @xmath97 the former sum can be performed as @xmath98= \\ = \frac{2}{k } \left ( \omega - \delta\omega\right ) ( i-1 ) + \delta\omega \frac{2}{k } \frac{i(i-1)}{n}\ ; , \end{array}\ ] ] from which @xmath99 or , using eq.([eq : common ] ) , @xmath100 + \frac{\delta\omega}{k } \frac{2i(i-1)}{n } = \\ = \frac{\delta\omega}{k } \ ; \left [ ( 1 - 2i ) \frac{n+2}{2(n+1 ) } + \frac{2i(i-1)}{n}\right ] \;.
\end{array } \label{eq : critic1}\ ] ] for given @xmath45 , we first determine the position @xmath101 at which the right - hand - side of eq.([eq : critic1 ] ) takes its minimum ( maximum ) value as a function of @xmath1 .
this value is constrained by the left - hand - side of ( [ eq : critic1 ] ) to be out of @xmath102 $ ] .
for @xmath1 treated as a continuous variable the derivative of eq.([eq : critic1 ] ) with respect to @xmath1 yields @xmath103 on the other hand @xmath104 and @xmath105 so that for @xmath106 , @xmath107 is certainly satisfied . upon inserting @xmath101 from ( [ eq:4a ] )
we have @xmath108 the critical threshold is then given by @xmath109 this relation is plotted in fig .
[ fig.theonum ] , where the full line represents the theoretical prediction , while the full dots correspond to the threshold values , numerically determined by integrating the system of differential eq.s ( [ eq1 ] ) , starting with @xmath110 .
( the case of @xmath110 violates the constraint ( [ eq.11a ] ) in agreement with the fact that for @xmath110 the system degenerates to the former case of having a pacemaker as a defect and two oscillators with natural frequencies @xmath111 . )
furthermore it should be noticed that eq.([eq : critic_fin ] ) scales as @xmath59 for large values of @xmath45 , the large-@xmath45 behavior is , however , not visible for the range of @xmath45 , ( @xmath112 ) , plotted in fig .
[ fig.theonum ] . , full line corresponds to the analytic prediction , full dots to numerical results for @xmath113 , the symmetry breaking parameter is chosen as @xmath74.,scaledwidth=47.0% ] for a ring with maximal distance @xmath114 and given @xmath45 and @xmath52 , the slope @xmath115 in the natural frequencies determines whether synchronization is possible or not .
the oscillators synchronize if the values of @xmath45 lead to a ratio @xmath116 below the bound @xmath117 of eq.([eq : critic_fin ] ) , named as in section [ sec.i ] , eq.([eq : critic ] ) , but with @xmath7 now standing for the maximal difference between the highest and the lowest natural frequencies .
if for given @xmath45 the slope , parameterized by @xmath7 , exceeds this threshold , no entrainment is possible . for @xmath34
the allowed slope goes to zero , the system does no longer synchronize . to further characterize the synchronization patterns
, we measure the stationary frequency @xmath118 of the individual oscillators @xmath1 , defined according to @xmath119 as well as its average @xmath120 and variance @xmath121^{1/2}$ ] .
the average frequency and variance are introduced to characterize the intermediate patterns of partial synchronization and to distinguish it from the case @xmath75 in section [ sec.ii2 ] . in all numerical calculations we set from now on @xmath122 .
[ fig.former7 ] shows the natural frequencies @xmath123 ( represented as full line ) and the stationary @xmath124 frequencies ( shown as dots ) for four slopes
@xmath125 , respectively , as function of the position on a linear chain of @xmath126 oscillators , of which we show only one half due to the symmetric arrangement .
the case of @xmath127 corresponds to full synchronization , while the other three figures represent partial synchronization of two @xmath128 and more clusters .
it should be noticed that the two clusters for @xmath128 would violate the condition ( [ eq : critic_fin ] ) if they were isolated clusters of @xmath68 oscillators each with a slope determined by @xmath129 , but due to the nonlinear coupling to the oscillators of the second cluster , the partial synchronization is a stationary pattern .
[ fig.bif ] displays the frequencies @xmath118 as function of the slope , again parameterized by @xmath7 .
here we have chosen @xmath130 .
we clearly see the bifurcation in frequency space starting at the critical value of @xmath131 and ending with complete desynchronization , in which the stationary frequencies equal the natural frequencies . for frequency synchronization
the phases are locked , but their differences to the phase of say the oscillator @xmath83 ( @xmath132 ) increases nonlinearly with their distance from @xmath6 , as seen in fig .
[ fig.former8 ] .
therefore the distance between points of the same @xmath133 decreases with @xmath1 . in two dimensions we simulated @xmath134 oscillators on a square lattice with periodic boundary conditions and the oscillator @xmath6 with the highest natural frequency placed at the center of the square lattice .
the natural frequencies of the oscillators are still given by eq.([eq2 ] ) .
their spatial distribution looks like a square pyramid centered at the middle of the square lattice .
[ fig.former9 ] displays @xmath135 on this grid .
it exhibits stationary patterns after @xmath136 steps of integration , for six choices of @xmath7 , all above the synchronization threshold .
we see some remnants of synchronization , most pronounced in fig .
[ fig.former9]a , in which a circular wave is created at the center at @xmath6 , and coexists with waves absorbed by sinks at the four corners of the lattice .
the projection of fig .
[ fig.former9]a on one dimension corresponds to fig .
[ fig.former7]b with a bifurcation into two cluster - frequencies .
note that it is again the oscillator with the highest natural frequency that becomes the center of the outgoing wave , while the corners with the lowest natural frequencies @xmath137 become sinks .
although fig .
[ fig.former9]f shows almost no remnant of synchronization , it is interesting to follow the time evolution towards this `` disordered '' pattern via a number of snapshots , as displayed in fig .
[ fig.former10 ] after @xmath138 [ @xmath139 ( a ) , @xmath140 [ @xmath141 ( b ) , @xmath142 [ @xmath143 ( c ) , @xmath144 [ @xmath145 ( d ) , @xmath146 [ @xmath147 ( e ) , @xmath148 [ @xmath149 ( f ) integration steps @xmath150 [ @xmath151 time steps in units of natural frequency(@xmath152 of eq.[eq2 ] ] .
while the pattern of ( a ) would be stationary in case of full synchronization , here it evolves after iterated reflections to that of ( f ) .
the pattern of ( f ) is stationary in the sense that it stays `` disordered '' , on a `` microscopic '' scale it shows fluctuations in the phases .
the evolution takes a number of time steps , since the interaction is mediated only via nearest neighbors and not of the mean - field type ( all - to - all coupling ) .
( full lines ) and measured , stationary frequencies @xmath118 ( full dots ) as function of the position @xmath1 of the oscillators on a one - dimensional lattice , for @xmath153 , periodic boundary conditions , and four values of the maximal natural frequency @xmath7 .
the oscillator with maximal natural frequency @xmath7 is at position @xmath83 .
only half of the oscillators up to index @xmath68 are shown because of the symmetric arrangement .
the symmetry breaking parameter is chosen as @xmath72.,scaledwidth=47.0% ] , the maximal natural frequency .
numerically measured frequencies @xmath118 ( normalized such that @xmath154 @xmath155 @xmath156 $ ] ) as function of @xmath7 for a one - dimensional lattice with @xmath130 and periodic boundary conditions .
the symmetry breaking parameter is chosen as @xmath157.,scaledwidth=47.0% ] ( from the oscillator at @xmath158 ) as function of the oscillator index @xmath1 for a one - dimensional lattice , with @xmath153 and periodic boundary conditions , for different values of the maximal natural frequency @xmath7 and @xmath72 .
the difference depends nonlinearly on the distance from the oscillator at @xmath158 .
the `` steps '' in these plots are due to the projection of the variables @xmath159 onto the interval @xmath160.,scaledwidth=47.0% ] on a two - dimensional lattice with @xmath153 oscillators per side .
the oscillator @xmath6 with the maximal natural frequency is placed at the center the lattice .
we plot @xmath161 for every oscillator @xmath1 .
the colors vary between white , corresponding to @xmath162 , and black , corresponding to @xmath46 .
the plots are realized after @xmath136 integration steps and for different values of @xmath7 [ therefore after different time in natural units of @xmath7 ] : a ) @xmath164 [ @xmath165 , b ) @xmath166 [ @xmath167 , c ) @xmath168 [ @xmath169 , d ) @xmath170 [ @xmath171 , e ) @xmath172 [ @xmath173 , f ) @xmath68 [ @xmath174 . a cross section through the pattern of a )
would reveal a frequency distribution as in fig.[fig.former7]b.,scaledwidth=47.0% ] f for symmetry breaking parameter @xmath74 .
all plots are realized for @xmath175 , after a different number of integration steps [ time in natural units of @xmath7 ] : a ) @xmath138 [ @xmath139 b ) @xmath140 [ @xmath141 , c ) @xmath142 [ @xmath143 , d ) @xmath144 [ @xmath145 , e ) @xmath146 [ @xmath147 , f ) @xmath148 [ @xmath149 . only the pattern of f )
is stationary on a global scale , i.e. , up to local fluctuations in the phases.,scaledwidth=47.0% ] next let us consider system ( [ eq1 ] ) for @xmath75 , for which we expect pattern formation from the results of @xcite due to the broken antisymmetry of @xmath29 via a @xmath176-term in the interaction .
the main difference shows up in the partial synchronization patterns above the synchronization threshold .
as it is evident from fig.s [ fig.former3 ] and [ biforc2 ] , the size of the one synchronized cluster , filling the whole lattice below the threshold , shrinks with increasing @xmath7 , since the oscillators with the highest natural frequencies decouple from the cluster , but the remaining oscillators do not organize in other synchronized clusters , the cluster - frequency does no longer bifurcate as before . along with this , the two - dimensional stationary patterns change to those of fig .
[ fig.former5 ] , whereas the evolution towards fig .
[ fig.former5]f is now displayed in fig.s [ fig.former6 ] with snapshots taken after @xmath138 [ @xmath139 ( a ) , @xmath177 [ @xmath178 ( b ) , @xmath179 [ @xmath180 ( c ) , @xmath181 [ @xmath182 ( d ) , @xmath183 [ @xmath184 ( e ) , and @xmath185 [ @xmath186 ( f ) integration steps [ @xmath151 as above ] .
a further manifestation of the difference in the partial synchronization patterns is seen in the average frequency @xmath187 and the variance @xmath188 as function of @xmath7 ( fig .
[ fig.former11 ] ) . above the synchronization transition
the average frequency remains constant for @xmath74 , but increases for @xmath75 in agreement with fig.s [ fig.former3 ] and [ biforc2 ] , the variance increases in both cases ( @xmath74 and @xmath75 ) above the transition , so that it may serve as order parameter .
therefore we can tune the synchronization features via the slope of the gradient .
the main qualitative feature , however , is in common to both systems with and without antisymmetric coupling : the oscillator with the highest natural frequency becomes the center of outgoing circular waves , it is dynamically established as pacemaker .
the patterns , seen here in case of full and partial synchronization , are quite similar to those predicted by kuramoto @xcite for reaction - diffusion systems with continuous diffusion terms , and to those experimentally observed in chemical oscillatory systems @xcite . .
the frequency of the synchronized cluster decreases and the cluster size shrinks , until all oscillators keep their natural frequency in a completely desynchronized state.,scaledwidth=47.0% ] . above the synchronization threshold
the size of the original synchronized cluster shrinks with @xmath7 .
oscillators outside this cluster stay isolated in contrast to fig .
[ fig.bif].,scaledwidth=47.0% ] .
we use the same parameter choice as in fig .
[ fig.former9 ] .
now fig.s a ) , b ) , c ) correspond to full synchronization , while one - dimensional cross sections through d ) , e ) , and f ) would reveal distributions of @xmath118 as in fig.[fig.former3 ] c)-d).,scaledwidth=47.0% ] , otherwise same as in fig .
[ fig.former10 ] .
this time the plots are realized after @xmath138 [ @xmath139 ( a ) , @xmath177 [ @xmath178 ( b ) , @xmath179 [ @xmath180 ( c ) , @xmath181 [ @xmath182 ( d ) , @xmath183 [ @xmath184 ( e ) , and @xmath185 [ @xmath186 ( f ) time steps [ time in units of @xmath7 ] . only the disordered pattern of ( f ) is stationary up to local fluctuations in the phases.,scaledwidth=47.0% ] ) as function of the maximal natural frequency @xmath189 , in case of a one - dimensional lattice with @xmath153 oscillators and for values of the symmetry breaking parameter @xmath72 and @xmath190 .
the raise in the variance from zero indicates the transition to the desynchronized phase.,scaledwidth=47.0% ]
for pacemakers implemented as defect we have extended former results on the entrainment frequency and the entrainment window to arbitrary interaction range , analytically for @xmath31 , numerically for intermediate @xmath49 . for large dimensions
the entrainment window decays exponentially with the average distance of nodes from the pacemaker so that only shallow networks allow entrainment .
the synchronization transition is reentrant as function of @xmath33 ( or @xmath61 , the number of neighbors on a ring , or @xmath62 , the probability to add a random shortcut to the ring topology ) .
the entrainment gets most difficult for @xmath67 and large @xmath45 , while its most easily achieved for next - neighbor and all - to - all couplings .
this reentrance is easily explained in terms of the normalization of the coupling strength .
for the same system without a pacemaker , but with a gradient in the initial natural frequency distribution the oscillator with the highest natural frequency becomes the center of circular waves , its role as a pacemaker is dynamically induced without the need for an asymmetric term in the interaction . here
we analytically determined the synchronization transition as function of the gradient slope . above some threshold , full synchronization on one or two - dimensional lattices
is lost and replaced by partial synchronization patterns .
these patterns depend on the asymmetry parameter @xmath191 .
for @xmath74 we observe a bifurcation in frequency space , for @xmath75 the one synchronized cluster shrinks in its size , before for even steeper slopes synchronization is completely lost . for artificial networks these results may be used to optimize the placement and the number of pacemakers if full synchronization is needed , or to control synchronization by tuning the slope of natural frequency gradients .
winfree , _ the geometry of biological time _ , ( springer - verlag new york 1980 ) ; j. buck , nature * 211 * , 562 ( 1966 ) ; t.j .
walker , science * 166 * , 891 ( 1969 ) ; i. kanter , w. kinzel , and e. kanter , europhys . lett . * 57 * , 141 ( 2002 ) ; b. blasius , a. huppert , and l. stone , nature ( london ) * 399 * , 354 ( 1999 ) ; j.a .
acebrn , l.l .
bonilla , c.j .
prez - vicente , f. ritort , and r. spigler , rev .
* 77 * , 137 ( 2005 ) . c. anteneodo , s.e .
de s. pinto , a.m. batista , and r.l .
viana , phys .
e * 68 * , 045202 ( 2003 ) ; c. anteneodo , s.e .
de s. pinto , a.m. batista , and r.l .
viana , phys .
e * 69 * , 029904(r ) ( 2004 ) . | we study phase entrainment of kuramoto oscillators under different conditions on the interaction range and the natural frequencies . in the first part
the oscillators are entrained by a pacemaker acting like an impurity or a defect .
we analytically derive the entrainment frequency for arbitrary interaction range and the entrainment threshold for all - to - all couplings . for intermediate couplings
our numerical results show a reentrance of the synchronization transition as a function of the coupling range . the origin of this reentrance can be traced back to the normalization of the coupling strength . in the second part
we consider a system of oscillators with an initial gradient in their natural frequencies , extended over a one - dimensional chain or a two - dimensional lattice . here
it is the oscillator with the highest natural frequency that becomes the pacemaker of the ensemble , sending out circular waves in oscillator - phase space .
no asymmetric coupling between the oscillators is needed for this dynamical induction of the pacemaker property nor need it be distinguished by a gap in the natural frequency . |
massive ob stars play an important role in astrophysics ; their winds and ejecta shape the morphology and ionization structure of their local environment , and provide a source of heavy elements which serve as the building blocks of planetary systems .
b[e ] stars comprise a heterogeneous group of massive stars which all exhibit similar observational properties , yet range in evolutionary age from the pre- to post - main sequence @xcite .
b[e ] supergiants ( b[e]sgs ) represent one sub - class of the b[e ] phenomenon and are notable in that at least some share similar photometric , spectroscopic , and rotational velocity properties as luminous blue variables ( lbvs ) @xcite , which suggests b[e]sgs might be precursors to lbvs .
b - type stars which exhibit strong h i balmer emission , forbidden and low excitation emission lines , commonly fe ii , [ fe ii ] , and [ o i ] , and evidence of warm ( t@xmath0 1000 k ) dust are assigned the classification of b[e ] @xcite .
b[e]sgs are known to exhibit a hybrid spectrum characterized by broad uv resonance absorption lines and a wealth of narrow emission lines @xcite .
@xcite suggested that a non - spherical wind model consisting of a cool , equatorial component , and a hot , polar component best explained these observations .
subsequent polarimetric observations seem to confirm the presence of an axisymmetric circumstellar environment in b[e]sgs @xcite , the base of which may sometimes be clumpy @xcite .
while polarimetric observations indicate the presence of an asymmetric , clumpy wind in some lbvs and lbv - like objects @xcite , it is not yet clear whether they exhibit a clumpy , axi - symmetric geometry similar to that observed for b[e ] stars ( see e.g. @xcite ) , which might be expected if b[e]sgs are truly precursors of lbvs . to date , four b[e ] stars have been identified in the small magellanic cloud ( smc ) , while eleven have been identified in the large magellanic cloud ( lmc ) . owing to the known distance of the smc and lmc and the observed luminosities of the fifteen smc and lmc b[e ] sources , it is well accepted that all are b[e]sgs @xcite . in this paper
, we report the serendipitous discovery of the fifth known b[e ] star in the smc .
the star was initially classified by @xcite as a compact h ii source ( [ ma93 ] 1116 ) likely characterized by broad h@xmath1 emission .
a later optical photometric survey by @xcite classified the same target ( ngc346:kwbbe 200 ) as a classical be star residing in the smc cluster ngc 346 , which is known to be a rich source of current and recent star formation @xcite .
our observations of the source were obtained in the context of assessing the formation and evolution of classical be stars in ngc 346 and other clusters in the smc and lmc @xcite .
our spectroscopic observations were obtained on 15 november 2004 using the cerro tololo inter - american observatory ( ctio ) 4 m telescope hydra 138 fiber multi - object spectrograph .
we used the 527 lines mm@xmath2 kpgl3 grating and a 200@xmath3 m slit plate , yielding a dispersion of 1.75 at 5500 ( r @xmath03000 ) .
we obtained two 3600 second exposures with ngc 346:kwbbe 200 on a fiber ; for each exposure , we also assigned fourteen fibers to locations void of stellar sources to serve as probes of background sky emission .
our ngc 346 exposures were bracketed by observations of a penray lamp exposure to wavelength calibrate our data .
standard bias frames and dome flats were obtained at the beginning of each night of our run ; sky flats were also obtained to calibrate the fiber - to - fiber throughput of hydra .
following bias subtraction using standard iraf techniques , we used pieter van dokkum s _ l.a.cosmic _ routine to identify and remove cosmic ray artifacts from our images . remaining standard reduction steps , including flat fielding , aperture extraction , and wavelength calibration were done within the _ hydra _ iraf package . to subtract the strong nebular lines present in our data
, we evaluated the use of various combinations of our fourteen sky fibers in an iterative manner .
following our best subtraction , we only see evidence of residual nebular contamination from [ o iii ] lines at 4959 and 5007 , at an intensity of 2% of the total sky emission in these transitions .
we do not observe other narrow ( gfwhm @xmath02.9 ) emission profiles at other wavelengths , which would be indicative of a nebular origin ; hence , we estimate that any residual nebular components still present in our data should contribute @xmath42% to the equivalent widths quoted in this paper
. we also present archival _
spitzer space telescope _ infrared array camera ( irac ; @xcite ) 3.6 , 4.5 , 5.8 , and 8 @xmath3 m observations of ngc 346:kwbbe 200 , obtained as part of program 63 ( houck pi ) .
the data were reduced via the spitzer pipeline s14.0 software .
photometry was extracted from post - pipeline mosaic images using standard iraf _
aperture photometry techniques ; uncertainties in the absolute flux calibration are less than 5% @xcite .
the continuum normalized optical spectrum of ngc 346:kwbbe 200 from 3935 - 5365 and 6000 - 6540 is presented in figures [ blue ] and [ red ] respectively .
continuum signal to noise levels range from @xmath020 at 4200 to @xmath070 at 5200 and 6200 .
ngc 346:kwbbe 200 is clearly characterized by a wealth of emission lines ; we used line identifications reported by @xcite and @xcite to identify the strongest lines detected at the resolution of our data , which are labeled in the figures and tabulated in table [ lines ] .
the dominant species seen in emission is fe ii , although lines of [ fe ii ] , [ o i ] , mg i , and ti ii also are present .
ngc 346:kwbbe 200 is also characterized by strong h i emission , from h@xmath5 ( figure [ blue ] ) through h@xmath1 ( figure [ ha ] ) , with all lines clearly exhibiting p - cygni profiles indicative of a strong outflowing wind .
the h@xmath1 line is particularly strong and broad ; we measured our second exposure to have a net equivalent width of @xmath0 -267 and wings extending to -2100 km s@xmath2 and + 2190 km s@xmath2 .
similarly , we measured h@xmath1 in our first exposure to have a net equivalent width of -266 and wings extending to -2100 km s@xmath2 and + 2100 km s@xmath2 .
the dominant source of uncertainty in these measurements is not photon statistics , but rather given the presence of extended electron scattering wings and the curvature of the spectral response of the grating+detector near h@xmath1 , our choice of continuum placement .
spectra of emission- and non - emission line stars located in other hydra fibers were used to calibrate the typical spectral response behavior near h@xmath1 , providing a constraint on the appropriate order of fitting function to use to fit the continuum region near h@xmath1 for ngc 346:kwbbe 200 . to further explore the nature of ngc 346:kwbbe 200
, we plot the star on a near - ir 2-color diagram ( black star ; figure [ 2cd ] ) , along with likely magellanic cloud classical be stars from @xcite ( red triangles ) and @xcite ( green triangles ) , a potential lmc herbig ae / be star ( blue cross ) from @xcite , and known smc and lmc b[e]sgs ( blue circles ) from @xcite , @xcite , and the 2mass catalog .
ngc 346:kwbbe 200 s near - ir color is clearly consistent with that observed for other smc / lmc b - type stars characterized by dusty circumstellar envelopes , and is inconsistent with the typical color of classical be stars , which are characterized by gaseous circumstellar disks @xcite .
we next construct ngc 346:kwbbe 200 s spectral energy distribution ( sed ) , using all available photometry .
its observed ( b - v ) and ( v - i ) colors , 0.351 and 0.366 @xcite , are substantially redder than the expected intrinsic colors of main sequence ( -0.30 @xmath4 ( b - v)@xmath6 @xmath4 -0.07 , @xcite ; -0.44 @xmath4 ( v - i)@xmath6 @xmath4 -0.11 , @xcite ) or supergiant - type ( -0.23 @xmath4 ( b - v)@xmath6 @xmath4 0.0 , @xcite ; -0.37 @xmath4 ( v - i)@xmath6 @xmath4 0.0 , @xcite ) b stars . as such , we have assumed that ngc 346:kwbbe 200 is characterized by a _ minimum _
e(b - v ) reddening of 0.35 , and de - reddened its u- through k - band photometry using r@xmath7 = 2.74 @xcite and the standard extinction curves of @xcite .
the resultant sed incorporating u - band through spitzer irac 8 @xmath3 m photometry is shown in figure [ sed ] .
an excess of ir emission is evident both at near - ir ( j - band ) wavelengths , likely originating from free - free and bound - free emission from hydrogen in a circumstellar envelope ( or disk ) , and at irac - band ir wavelengths , indicating the presence of warm dust in a circumstellar envelope ( or disk ) ( see e.g. @xcite ) . to crudely characterize this dust component
, we have overlayed a planck function corresponding to a dust temperature of t = 800k in figure [ sed ] , which we believe best represents the observed sed trend .
we have shown that ngc 346:kwbbe 200 s optical spectrum is dominated by emission lines from fe ii , [ o i ] , and [ fe ii ] .
it exhibits broad p - cygni profile h i emission lines , indicative of a strong outflowing wind ; the strength of its h@xmath1 equivalent width , @xmath0 -267 is rivaled only by other known b[e]sgs @xcite and some lbvs @xcite .
the star s circumstellar environment is clearly characterized by the presence of both gas and warm dust , as diagnosed from its near - ir colors ( figure [ 2cd ] ) and near- and ir - sed excess ( figure [ sed ] ) .
these observational properties are characteristic of b[e ] stars @xcite , and inconsistent with the expected behavior of classical be stars ; thus , we suggest that ngc 346:kwbbe 200 should be re - classified as the fifth known b[e ] star in the smc . to constrain the evolutionary status of ngc 346:kwbbe 200 , we compared its de - reddened photometry to standard kurucz model atmospheres @xcite using log ( z / z@xmath8 ) = -1.0 and log ( g ) = 3.5 .
as shown in figure [ sed ] and summarized in table [ param ] , the sed data are best represented by a model having t@xmath9 = 19,000 k , and r@xmath10 = 14 r@xmath8 ( figure [ sed ] ) .
given the significant uncertainty over the exact e(b - v ) reddening associated with ngc 346:kwbbe 200 , we caution that these stellar parameters should only be considered initial estimates ; the presence of a higher e(b - v ) reddening than assumed here ( 0.35 ) will inflate t@xmath9 and modify r@xmath10 . nonetheless , assuming a standard relationship that l@xmath10 = ( r@xmath10/r@xmath8)@xmath11 ( t@xmath9/t@xmath8)@xmath12 ) , we can estimate the luminosity of ngc 346:kwbbe 200 to be log ( l / l@xmath8 ) @xmath04.4 ( table [ param ] ) .
these stellar parameters correspond to a crude spectral classification of b3[e ] ii . using these crude stellar parameters , we plot ngc 346:kwbbe 200 on a hertzsprung - russell ( hr ) diagram ( black circle ) , along with other known smc , lmc , and galactic b[e]sg ( green , red , and blue triangles ) and pre - main sequence herbig b[e ] ( hereafter haeb[e ] ) stars ( yellow squares ) in figure [ hrd ] .
the canonical ( non - accreting ) evolutionary tracks @xcite and zero - age main sequence ( zams ) for 9 and 15 m@xmath8 stars at a metallicity appropriate for the smc / lmc , z = 0.001 ( black lines ) , and for 3 and 5 m@xmath8 stars at a metallicity appropriate for the galactic haebe stars , z = 0.020 ( red lines ) , are also plotted in figure [ hrd ] , along with the birthline for the z = 0.020 models ( solid light blue line ) and isochrones for the z = 0.001 models ( dashed green , blue , and yellow lines ) .
it is clear that ngc 346:kwbbe 200 is well above the birthline of late - type b stars ( 5 m@xmath8 ) and is most similar to the lower luminosity magellanic cloud b[e]sgs stars reported by @xcite .
if it were a pre - ms star , inspection of figure [ hrd ] indicates that ngc 346:kwbbe 200 would be positioned at a youthful position along the the z = 0.001 15 m@xmath8 evolutionary track , and intersect the log(t ) = 4.25 isochrone . such a young , high - mass object
would likely still be deeply embedded in its natal star formation envelope , in contradiction to the star s observed optical v - band magnitude and our easy detection of its optical spectrum .
@xcite provided a nice discussion of the classification of b[e ] stars and established numerous criteria for determining whether a b[e ] star is a b[e]sg , haeb[e ] , compact planetary nebula b[e ] , symbiotic b[e ] , or an `` unclassified '' b[e ] star , which we briefly review here . to be considered a b[e]sg , @xcite suggest stars should obey two primary criteria : a1 ) exhibit the b[e ] phenomenon ; and a2 ) be supergiants with log ( l / l@xmath8 ) @xmath13 4.0 ; and four secondary criteria , two of which are : b1 ) optical spectroscopic evidence of mass loss ; and b2 ) typically exhibit small photometric variations of order 0.1@xmath14 . to be considered a haeb[e ]
, @xcite suggest stars should obey three primary criteria : a1 ) exhibit the b[e ] phenomenon ; a2 ) be associated with a star forming region ; and a3 ) exhibit evidence of accretion via inverse p cygni spectroscopic line profiles ; and three secondary criteria : b1 ) exhibit a luminosity of log ( l / l@xmath8 ) @xmath15 4.5 ; b2 ) exhibit large , irregular photometric variations ; and b3 ) exhibit evidence of warm and cool dust in their seds .
note that some of the criteria for b[e]sg and haeb[e ] designations overlap ; moreover , a growing body of evidence suggests that some b[e]sgs may also exhibit significant , large amplitude photometric variability @xcite , indicating that this behavior is not exclusive to haeb[e ] stars .
ngc 346:kwbbe 200 s strong p - cygni profile h i emission is indicative of mass - loss ; in our spectrum , we observe no evidence of inverse p - cygni profiles which would be indicative of infall .
as shown in figure [ hrd ] , ngc 346:kwbbe 200 s crudely derived luminosity , log ( l / l@xmath8 ) @xmath04.4 , and location on a hr diagram is more consistent with the star being a low luminosity b[e]sg than a high luminosity haeb[e ] star .
the star does reside in a young cluster , portions of which are actively forming low - mass stars @xcite , which seemingly agrees with the aforementioned `` a2 '' criteria of haeb[e ] stars ; however , other well known highly evolved massive stars reside in the region ( e.g. the wolf - rayet / lbv hd 5980 ; aka ngc 346:kwbbe 13 @xcite ) hence this particular criteria is probably not a useful discriminant in this situation . as such ,
given the available data presently available for ngc 346:kwbbe 200 , we suggest the star is a b[e]sg , similar to the other 4 known smc and 11 lmc b[e ] identified to date @xcite .
ngc 346:kwbbe 200 is a b - type star whose optical spectrum is dominated by fe ii , [ o i ] , [ fe ii ] , and strong p - cygni h i emission lines , and exhibits clear evidence of having a circumstellar envelope characterized by the presence of gas and warm ( t@xmath16 @xmath0800 k ) dust . based on these observational properties , we suggest that ngc 346:kwbbe 200 is a b[e]sg star , representing the fifth such object identified to date in the smc .
our crude estimate of the star s luminosity , log ( l / l@xmath8 ) @xmath04.4 , its location on a hr diagram , and its observed line profile morphologies suggests it is most likely to be a b[e ] supergiant .
we recommend several observational approaches be pursued to further constrain the evolutionary status of ngc 346:kwbbe 200 .
mid- to far - ir photometric observations would allow one to search for the presence of cool dust , which is one of the defining characteristics of haeb[e ] stars @xcite and would not be an expected characteristic of b[e]sgs .
moreover , high resolution optical and uv spectroscopic observations would facilitate the derivation of a more reliable estimate of ngc 346:kwbbe 200 s spectral type and luminosity , hence providing more conclusive evidence of the post - main sequence evolutionary status suggested in this paper .
optical photometric monitoring , to search for and characterize variability , and measurements of the star s rotational velocity would aid efforts to explore evolutionary links between the b[e]sg and lbv phases of massive star evolution .
we thank ted gull and aki roberge for helpful discussions about these results .
we also thank our referee , jorick vink , for providing useful feedback which improved the content and presentation of this paper .
support for this project was provided by nasa npp and gsrp fellowships to jpw ( nnh06cc03b , ngt5 - 50469 ) , a nasa ltsa grant nag5 - 8054 and a research corporation cottrell scholar award to ksb , and a nsf grant ( ast-0307686 ) to jeb .
this work is based in part on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
we have also made use of the simbad database operated at cds , strasbourg , france , and the nasa ads system .
-267 ) , line - profile morphology ( p - cygni ) , and broad electron scattering wings ( @xmath0 -2100 km s@xmath2 and @xmath0 + 2190 km s@xmath2 ) of the h@xmath1 line in ngc 346:kwbbe 200 is similar to that observed in other b[e]sgs @xcite .
[ ha ] ] sgs ( blue circles ) from @xcite , @xcite , and the 2mass catalog are shown on a near - ir 2-color diagram .
ngc 346:kwbbe 200 s near - ir color clearly coincides with that of other dusty magellanic cloud ( b[e ] and herbig ae / be ) stars and is inconsistent with the observed near - ir colors of magellanic cloud classical be stars . [ 2cd ] ] @xmath0 800k , dust .
the optical photometry are consistent with a kurucz model atmosphere having t@xmath9 = 19,000k and r@xmath10 = 14 r@xmath8 .
the extrapolated stellar luminosity , log ( l / l@xmath8 ) @xmath04.4 , and observed line profile morphologies suggest that ngc 346:kwbbe 200 is a b[e]sg .
[ sed ] ] sgs ( red , green , and blue triangles respectively ) and galactic haebe stars ( yellow squares ) .
temperature and luminosity values were taken from @xcite and references therein for all of these ancillary sources , except for the luminosity of mwc 1080 @xcite and v594 cas @xcite , whose tabulated values within @xcite did not match their quoted literature origin and/or could not be otherwise verified . also overplotted are the canonical evolutionary tracks and zams for 9 and 15 m@xmath8 ( z = 0.001 ) stars ( black lines ) and 3 and 5 m@xmath8 ( z = 0.020 ) stars ( red lines ) , the birthline for the z = 0.020 models ( light blue line ) , and isochrones for the z = 0.001 models ( dashed green , blue , and yellow lines ) .
ngc 346:kwbbe 200 is well above the birthline of late - type b stars ( 5 m@xmath8 ) , and appears to be most similar to the lower luminosity magellanic cloud b[e]sgs stars reported by @xcite .
furthermore , its observed photometric and spectroscopic properties do not indicate that the star is highly embedded in a natal star formation envelope , as would be expected if it were truly located at the log(t ) = 4.25 isochrone of the 15 m@xmath8 pre - ms evolutionary track of @xcite .
[ hrd ] ] lccccc wavelength & line & ew ( ) & comment + + 3970 & h @xmath5 & & p - cygni + 4101 & h @xmath17 & 0.4 & p - cygni + 4233 & fe ii & -1.3 & + 4340 & h @xmath18 & -5.2 & p - cygni + 4417 & fe ii & & + 4491 & fe ii & & blend@xmath19 + 4508 & fe ii & & + 4515 & fe ii & & + 4522 & fe ii & & blend@xmath11 + 4534 & ti ii + fe ii & & blend@xmath20 + 4549 & fe ii + ti ii & & blend@xmath12 + 4555 & fe ii & & + 4584 & fe ii & & + 4629 & fe ii & & + 4861 & h @xmath21 & -29.3 & p - cygni@xmath22 + 4921 & he i & & abs + 4924 & fe ii & -2.3 & + 5015 & he i & & abs + 5018 & fe ii & -3.5 & + 5158 & [ fe ii ] & & + 5169 & fe ii & & p - cygni ?
+ 5184 & [ fe ii ] & & + 5198 & fe ii & -1.0 & + 5235 & fe ii & -1.2 & + 5276 & fe ii & & + 5317 & fe ii & -2.4 & + 5363 & fe ii & & + 5535 & fe ii & & + 6148 & fe ii & -0.4 & + 6238 & fe ii & & + 6248 & fe ii & -0.7 & + 6300 & [ o i ] & -1.2 & + 6318 & mg i & & + 6363 & [ o i ] + fe ii & & blend@xmath23 + 6385 & fe ii & -0.8 & + 6417 & fe ii & & + 6433 & fe ii & & + 6456 & fe ii & & + 6516 & fe ii & & + 6563 & h @xmath1 & -267 & p - cygni@xmath24 + lc parameter & value + + u & 14.834 @xmath250.048 + b & 15.957 @xmath250.048 + v & 15.606 @xmath250.038 + i & 15.240 @xmath250.148 + j & 14.551 @xmath250.046 + h & 13.855 @xmath250.043 + k & 12.772
@xmath250.032 + 3.6 @xmath3 m & 10.53 @xmath250.05@xmath19 + 4.5 @xmath3 m & 9.76 @xmath250.04@xmath19 + 5.8 @xmath3 m & 8.85 @xmath250.04@xmath19 + 8.0 @xmath3 m & 7.81 @xmath250.03@xmath19 + t@xmath9 & @xmath019000 k + r@xmath10 & @xmath014 r@xmath26 + log ( l@xmath10/l@xmath26 ) & @xmath04.4 + t@xmath16 & @xmath0800 k + | we present new optical spectroscopic and archival spitzer irac photometric observations of a b - type star in the smc cluster ngc 346 , ngc 346:kwbbe 200 .
we detect numerous fe ii , [ o i ] , and [ fe ii ] lines , as well as strong p - cygni profile h i emission lines in its optical spectrum .
the star s near - ir color and optical to ir sed clearly indicate the presence of an infrared excess , consistent with the presence of gas and warm , t @xmath0800 k , circumstellar dust . based on a crude estimate of the star s luminosity and the observed spectroscopic line profile morphologies , we find that the star is likely to be a b - type supergiant .
we suggest that ngc 346:kwbbe 200 is a newly discovered b[e ] supergiant star , and represents the fifth such object to be identified in the smc . |
aside from being the major constituent of the outer crust of the earth , oxides are of considerable interest for their wide applications in various technological fields such as electrochemistry , catalysis or microelectronics.@xcite .
they present very diverse electronic ground states which range from superconducting , metallic , semi conducting ( intrinsic , charge density or spin density like ) to insulating .
> from a theoretical point of view , a proper description of their electronic properties is still an active research subject . on the one hand ,
several transition metal oxides require an inclusion of many - body effects beyond the available effective one - electron approaches , such as the density functional theory ( dft ) or the _ ab initio _ hartree - fock ( hf ) method . on the other hand , even for insulating charge transfer oxides , such as mgo and sro , whose electronic ground state is well accounted for in a one - electron picture , the calculation of quasiparticle ( qp ) energies within the dft or the hf method gives rather poor results .
indeed , one has to face the so - called energy band gap problem @xcite . indeed , when using the kohn - sham eigenvalues of the highest occupied and the lowest empty states , the resulting gap usually underestimates the experimental data .
on the other hand , the corresponding difference between the hf eigenvalues overestimates the gap severely .
alternative ways to calculate the band gap in infinite systems have been found , in the framework of hf or dft .
one of them is based on total energy differences between ground and charge transfer states @xcite .
the other one relies on the value of the gap found in the one electron spectrum of the ionized or hole - doped system @xcite . however , such methods are unable to provide the full qp spectrum . in the last decade , due to the implementation of angular resolved photoemission experiments able to cope with charging effects on insulating oxides , detailed qp spectra of mgo@xcite , srtio@xmath0@xcite or tio@xmath2@xcite among other oxides ,
have become available .
the experimental achievements demand systematic improvements in the theoretical description of the electronic excitations of simple oxides . in semiconducting and ionic crystalline compounds , corrections to the dft eigenvalues , as obtained from the kohn - sham equations ,
may be estimated by first - order many - body perturbation theory @xcite , with respect to the difference between the exchange - correlation ( xc ) self - energy and the corresponding xc potential of the dft , for which various approximations are available . within the same approach ,
the self - energy operator is treated in the gw approximation proposed by hedin@xcite .
such calculations have been carried out with success in several semiconducting and insulating crystals @xcite , permitting the prediction of the quasiparticle energies and the calculation of optical properties of real solids .
one of the crucial points in such calculations is the evaluation of the self - energy operator , which is numerically very expensive , because of the required knowledge of the full inverse dielectric matrix of the system under study .
this is a bottleneck for the application of the method to systems with a large number of atoms in the unit cell , such as surfaces @xcite and defective or low symmetry solids . in order to reduce the remarkable numerical effort and
be able to predict electronic excited states in a wider range of real systems , one may use efficient gw methods , in which a model dielectric function is used to mimic the screening properties of the system under study .
pioneering work in this direction has been performed by bechstedt and del sole in a tight - binding scheme @xcite and by gygi and baldereschi @xcite within the dft in the local density approximation ( lda ) for the exchange and correlation energy .
the key of their work relies _ ( i ) _ on the use of a model dielectric function , which mimics the main screening properties of semiconductors , and _
( ii ) _ on an approximated self - energy operator in which both local fields and dynamical screening giving rise to opposite contributions to band gaps are neglected .
@xcite for a large family of semiconductors , the band gaps obtained through these simplified methods show good accordance both with experiments and full gw results , in which the screening properties of the system are calculated from first principles .
more recently , an efficient gw method that approximately includes dynamical - screening and local field effects , without increasing the computational effort , has been formulated @xcite and proved to reproduce quasiparticle energies and band gaps for either bulk zincblende solids ( ranging from si , gaas , alas to the more ionic znse)@xcite or non cubic systems ( bn@xcite , sic polytypes@xcite ) satisfactorily .
the method is based on a linear expansion of the dynamical contribution to the exchange - correlation self - energy , on the use of a model dielectric function , and on a local ansatz for the treatment of the intrinsically non local screened interaction .
depending on the system , the speedup of the calculation may result of two order of magnitude with respect to corresponding full gw ones .
moreover , this kind of gw calculations is feasible on a standard workstation .
the aim of the present paper is to study the ground state properties of cubic magnesium oxide , strontium oxide and strontium titanate , in the dft - lda scheme , as well as their quasiparticle band structures within the efficient gw scheme . on one hand , no previous gw calculation of sro or srtio@xmath0 exists , to our knowledge . on the other hand , these charge transfer oxides are characterized by different fundamental gaps , ranging from @xmath3 3 ev ( srtio@xmath0 ) to @xmath3 7.7 ev ( mgo ) , as given by experimental data .
their bandstructures may substantially differ from those of the semiconductors cited above , due to the presence of strongly localized electronic states .
thus , such a study may also represent a test for the approximate gw scheme , originally proposed for small - gap semiconductors , to show its capabilities relatively to a different class of physical systems .
the paper is organized as follows .
the theoretical framework is depicted in section ii and it is followed , in section iii , by a description of the computational ingredients used for the description of structural and electronic properties of mgo , sro and srtio@xmath0 .
the results obtained for these systems are presented separately in sections iv , v and vi , respectively .
finally , in section vii we present our conclusions and the perspectives opened by our work .
the energy @xmath4 of a quasiparticle in a band @xmath5 at a given @xmath6 point in the brillouin zone can be obtained through the equation : @xcite @xmath7 \phi_{n \vec k}(\vec r ) \ , + \ , \int d^3\vec r^\prime \
, \sigma(\vec r , \vec{r^\prime } ; e_{n \vec k } ) \phi_{n \vec k}(\vec{r^\prime } ) \ , = \ , e_{n \vec k } \phi_{n
\vec k}(\vec r)\ ] ] which is written in atomic units ( @xmath8=@xmath9 = @xmath10=1 ) .
@xmath11 is the external potential originated by the ions , @xmath12 is the hartree potential and @xmath13 is the mass operator , usually called the self energy operator , which is in general non local , non - hermitian and energy dependent . in principle
it contains the exchange and correlation effects and is state ( @xmath14 ) dependent .
one can compare the above equation to that of kohn and sham @xcite : @xmath15 \psi_{n
\vec k}(\vec r ) = \ , e_{n \vec k}^{(0 ) } \psi_{n \vec k}(\vec r)\ ] ] where @xmath16 is the exchange correlation potential of the dft ( for which several approximation are available ) , and the superscript @xmath17 will denote the quantities computed at the dft level from now on .
@xmath18 is usually considered as a first approximation to the quasiparticle energy .
the failure of this scheme for evaluating transition energies @xmath19 between different quasiparticle states in solids has been evidenced by many authors @xcite .
one can correct the dft eigenvalues @xmath20 , in a first - order perturbation theory with respect to @xmath21 , and estimate the qp energies as : @xmath22 the evaluation of @xmath23 is in general a very difficult task .
a possible approximation is the gw one , @xcite in which a perturbation expansion for the self - energy can be constructed and stopped at the first order : @xmath24 where @xmath25 is the one - particle green function , @xmath26 is the screened coulomb interaction and @xmath27 .
following bechstedt and coworkers,@xcite the real part of the self - energy operator ( which is relevant for the corrections to the qp bandstructure ) may be separated in a dynamic and a static contribution , and written as : @xmath28 @xmath29 and @xmath30 are the static screened exchange ( sex ) and the coulomb
hole ( coh ) terms , respectively , and @xmath31 the energy dependent contribution .
although the previous equation is purely formal , it is the starting point of simplified gw schemes , since one can work out approximated expressions for each contribution separately .
the most expensive step towards the evaluation of gw corrections to @xmath18 is the calculation of the energy - dependent screened interaction @xmath32 .
although many improvements have been proposed for an efficient calculation of the dielectric response function from first principles , @xcite still this stage is both computer time and memory very demanding when a large basis set is used for the description of the electronic structure . in particular
, this is the case of oxides in a plane - wave pseudopotential approach . on the other hand , the improvement of dft lda band structures obtained by using schemes originally formulated on phenomenological basis , such as the scissor operator or the slater s @xmath33 potential
, @xcite showed that the essential physics can be described through the use of effective or state - dependent potentials . from a more fundamental point of view
, one can work out a model for the screening properties of the system , and look to what extent the computed qp spectra account for the observed experimental values .
possibly , the model should also be simple enough to be generalized to a large number of different systems .
following previous studies , @xcite we use a model dielectric function to estimate both the static and the dynamical contributions to the self - energy ( eq.[eq:5 ] ) , since it is the key ingredient entering the efficient gw approximations . for the diagonal part of the screening function , we use an analytical model originally proposed by bechstedt and del sole
: @xcite @xmath34 ^{-1 } \label{eq:6}\ ] ] where the fermi vector @xmath35 and the thomas - fermi wave - vector @xmath36 depend on the average electron density @xmath37 .
this expression mimics the free electron gas behavior at high @xmath38 . at @xmath39 , eq.[eq:6 ] gives the value of the optical dielectric constant @xmath40 of the oxide .
moreover we use a local _ ansatz _ for the static coulomb screened interaction , analogously to that of hybertsen and louie @xcite : @xmath41 \label{eq:7}\ ] ] where @xmath42 is the screened coulomb interaction of a virtual homogeneous system characterized by a finite optical dielectric constant .
@xcite quasiparticle energies determined from the above _ ansatz _ result in good accordance with full gw calculated ones.@xcite it permits , together with eq.(6 ) , to obtain a simple analytical expression for the static part of the coulomb hole self energy @xmath30 .
the explicit formula has been given elsewhere .
@xcite for the calculation of @xmath29 in eq.[eq:5 ] , the fourier transform of @xmath43 is needed , namely @xmath44 . in order to reduce sharply the time needed to compute this term
, we proceed in two steps .
firstly , we retain only the diagonal terms in the screened coulomb interaction .
secondly , we account for local field effects ( _ i.e. _ non diagonal terms with @xmath45 in @xmath46 ) by using suitable state - dependent densities to compute @xmath35 and @xmath36 . given a state @xmath47
, we define the effective electron density @xmath48 , which determines the typical screening lengths @xmath49 and @xmath50 , relevant for the estimation of @xmath51 ( eq.[eq:3 ] ) as the expectation value of the ground state electronic density @xmath52 on the actual state : @xmath53 since the relevant screening lengths @xmath50 and @xmath49 are increasing functions of @xmath54 , the latter is intended to give an approximate description of the local fields in @xmath55 through an enhanced screening for the electronic states which contribute effectively to the total electron density . in the oxides considered here ,
the effective densities @xmath54 range from few tenths of the mean valence electron density @xmath56 for conduction states up to 45 @xmath37 for deep valence states . as a consequence ,
the lengths @xmath50 and @xmath49 are shorter for the screening of valence band states than for conduction band states .
this difference is enhanced with respect to more covalent semiconductors , such as si and gaas @xcite , in which the valence electron distribution is more delocalized and less dissimilar from that of the conduction states than in oxides such as mgo .
this approximate scheme for treating local field effects was analyzed in detail and its reliability was shown through a comparison with full gw calculations , in the case of both small and wide gap semiconductors @xcite . in ref .
, it was also shown that the use of a non - diagonal model dielectric function does not clearly improve band gaps while bringing additional computational costs
. the diagonal matrix element of the dynamical part of the self energy ( the third operator on the right side in eq.[eq:5 ] ) on the state @xmath57 : @xmath58 can be linearly expanded in energy around the dft value @xmath59 , with linear coefficient : @xmath60 this approximation for the energy dependence of the self - energy operator turns out to be valid in a wide energy range , as confirmed by the results of full gw calculations on bulk si.@xcite solving the dyson equation within the first - order perturbation theory with respect to @xmath61 , and using eqs.[eq:3 ] and [ eq:5 ] , the qp shifts for the @xmath62 state hence read : @xmath63 where @xmath64 and @xmath65 are the expectation values of the static coh and sex contributions to the self energy , respectively .
> from the gw expression of @xmath13 , replacing g by g@xmath17 , computed at the dft level @xcite : @xmath66 @xmath67 ^ 2 } \label{eq:13}\ ] ] where @xmath68 the model @xmath69 is diagonal in @xmath38 space and the local field effects are accounted for through the use of the state averaged density @xmath54 ( eq.[eq:8 ] ) .
the @xmath70-dependent dielectric function is obtained from the static one by means of the plasmon - pole approximation ( ppa)@xcite . the ppa has been widely used to compute semiconductor and insulator quasiparticle bandstructures , even in the energy range of semicore states @xcite . in order to account for the differences @xmath71 in eqs.[eq:12 ] and [ eq:13 ] , replacing them with the state independent form @xmath72 as in ref . , and using the sum rule @xmath73 , one obtains : @xmath74 @xmath75 therefore , once the spectrum and eigenfunctions obtained in the framework of the dft are known , the calculation of @xmath76 and @xmath77 according to eqs.[eq:15 ] and [ eq:16 ] needs no additional external parameters apart from the model dielectric function , and results a simple task that can be carried out on a standard workstation . we have made extensive checks to test the reliability of the approximations used to obtain eqs.[eq:15 ] and [ eq:16 ] , in the case of silicon .
some transitions between high symmetry points of some valence and conduction band states in si are compared in tables i to those obtained through full gw .
the energy of the transitions between high symmetry points , as well as the fundamental gap in si , are in very close agreement with those obtained from full gw calculations and with experiments for both methods .
as far as the qp eigenvalues in si are concerned , the top of the valence band @xmath78 is shifted downwards with respect to the dft values of 0.28 ev .
this can be compared with the upward shift of 0.07 ev , as found by godby _
et al._@xcite . .
computed vertical transitions and minimal valence
conduction band gap for silicon ( in ev ) .
@xmath79 : ref .. @xmath80 : ref .. the experimental values are quoted in ref .. [ cols="<,^,^,^,^",options="header " , ]
we have calculated the gound state properties of cubic oxides mgo , sro and srtio@xmath0 within the dft - lda , and applied a perturbative gw method , based on a dielectric model which includes approximately local field and dynamical effects , to determine their quasiparticle energies .
we have tested the dependence of the calculated spectra on the parameters entering the calculation , such as the optical dielectric constant @xmath81 and the lattice parameter @xmath1 .
in particular , we have shown that quasiparticle energies in alkaline - earth oxides are sensitive functions of the lattice parameter , even at the lda level . as a consequence ,
the use of either the experimental or the theoretical values for @xmath1 can influence the quality of the results whenever @xmath82 differs from @xmath83 . through a careful comparison with the available experimental data and previous ab
initio gw calculations , we have showed that the simplified gw method works reasonably well for systems , such as mgo and sro , in which the electronic states of the valence band and the bottom of the conduction band mainly consist of @xmath84 and @xmath85like states . moreover , in the three oxides , gw corrections have been found to be crucial to account for the energies of semicore electron states , such as o @xmath86 , ti @xmath87 , ti @xmath88 , sr @xmath89 and sr @xmath90 , which agree well with the peak positions recorded in photoemission spectra .
when localized @xmath91like states contribute to bands around the fermi level , our method produces an overestimation of the self - energy corrections .
this is the case of srtio@xmath0 , for which it is not clear at this stage whether a higher order perturbative approach , in which the lda wavefunctions are renormalized , is needed , or whether this failure calls for a more refined treatment of local field effects .
we thank lucia reining for fruitful discussions and careful reading of the manuscript , tristan albaret for his help in generating the pseudopotentials , and eric shirley for kindly providing us unpublished results on self - energy corrections for mgo .
computing resources were granted by the idris / cnrs computational centre in orsay , under project 980864 .
g.c . acknowledges support from the commission of the european communities ( framework @xmath92 ) programme under the networks scheme . in sro
, the bottom of the conduction band @xmath93 has mainly sr @xmath94 character , while at @xmath95 the sr @xmath96 character prevails . in any case
, a noticeable degree of second neighbour hybridization is present . | the structural properties and the band structures of the charge
transfer insulating oxides sro , mgo and srtio@xmath0 are computed both within density functional theory in the local density approximation ( lda ) and in the hedin s gw scheme for self - energy corrections , by using a model dielectric function , which approximately includes local field and dynamical effects .
the deep valence states are shifted to higher binding energies by the gw method , in very good agreement with photoemission spectra . since in all these oxides
the direct gaps at high - symmetry points of the brillouin zone may be strongly sensitive to the actual value of the lattice parameter @xmath1 , already at the lda level , self - energy corrections are computed both at the theoretical and the experimental @xmath1 . for mgo and sro , the values of the transition energies between the valence and the conduction bands is improved by gw corrections , while for srtio@xmath0 they are overestimated .
the results are discussed in relation to the importance of local field effects and to the nature of the electronic states in these insulating oxides . |
for an analytical investigation the motion of the vortex is commonly described employing the thiele equation.@xcite this equation is exact for the steady state motion of a non - deformable magnetization pattern . however
, this assumption holds true only for the small vortex core . due to the spacial restriction the vortex has to deform while the core is moving .
this yields a small modification of the thiele equation that is especially important for current - driven vortex motion in view of the non - adiabatic spin torque . with the magnetization @xmath59 and the magnetic field
@xmath60 a general version of the thiele equation reads@xcite @xmath61 ( \vec h \cdot \vec m)\\ & - \frac{m_s \mu_0}{\gamma } \int dv \ ,
\sin(\theta ) ( \vec \nabla \theta \times \vec \nabla \phi ) \times ( \vec v + b_j \vec j)\\ & - \frac{m_s \mu_0}{\gamma } \int dv \ , ( \vec \nabla \theta \vec \nabla \theta + \sin^2(\theta ) \vec \nabla \phi \vec \nabla \phi ) ( \alpha \vec v + \xi b_j \vec j ) \end{split } \label{eq_thile_start}\ ] ] with the saturation magnetization @xmath62 , the gyromagnetic ratio @xmath30 , the current density @xmath63 , the gilbert damping @xmath9 , the non - adiabaticity parameter @xmath10 , and the coupling constant @xmath64 between current and magnetization . @xmath65 and @xmath66 are the out - of - plane and in - plane angle of the magnetization , respectively .
the velocity @xmath67 of the magnetization pattern may depend on the position .
assuming that the magnetization pattern does not deform the velocity is independent of the position . then eq .
( [ eq_thile_start ] ) can be written in its well known form@xcite @xmath68 with the velocity @xmath69 of the vortex core . here
@xmath70 ( \vec h \cdot \vec m)\ ] ] denotes the force on the magnetization pattern .
@xmath71 is the gyrovector and @xmath72 is the diagonal dissipation tensor with @xmath73 and @xmath74 .
the term @xmath75 describes the dissipation of energy due to the changing magnetization .
the integrand in the gyrovector is non - zero only in the small vortex core where the out - of - plane angle @xmath65 varies while the integrand in the dissipation tensor is also non - zero outside the core .
close to the boundaries of the sample the magnetization pattern moves slower compared to the center as it can be seen in fig .
[ model ] .
thus the velocity in the third term of eq .
( [ eq_thile_start ] ) depends on space .
aiming a similar form as in eq .
( [ eq_thile ] ) we replace the spatially dependent velocity @xmath76 in the third term of eq .
( [ eq_thile_start ] ) by an effective value @xmath77 which is independent of the position .
this effective velocity occurs only in the third term as the second term is located at the vortex core . for a homogeneous current flow
@xmath78 is constant over the sample .
thus we do not replace the current by an effective value .
the equation then reads @xmath79 the effective velocity @xmath77 depends on the core position @xmath21 and the core velocity @xmath69 . for small deflections of the vortex core , i.e. , small deformations of the vortex , @xmath77 can be expanded in @xmath80 and @xmath69 . for @xmath81
the magnetization is static and @xmath82 .
thus the first non - vanishing term in the expansion is proportional to @xmath69 .
here and hereafter we write @xmath83 since the effective velocity @xmath77 is always smaller than the velocity @xmath69 of the vortex core , the new constant @xmath37 is smaller than @xmath13 . inserting eq .
( [ eq_diss_gamma ] ) in eq .
( [ eq_thile2 ] ) yields a modified thiele equation @xmath84 employing the same conversions as used for the original thiele equation @xcite we find an expression for the velocity of the vortex core @xmath85 we investigate a square thin - film element with a current in @xmath18 and a field in @xmath20 direction . with the stray - field energy for small deflections@xcite @xmath86 and the total zeeman energy@xcite @xmath87
we get a force of @xmath88 here @xmath23 and @xmath24 are the lateral extension and the thickness of the square , respectively . as for the original thiele equation in the absence of current and field
the excited vortex performs an exponentially damped spiral rotation around its equilibrium position .
the free frequency @xmath89 and the damping constant @xmath90 are slightly changed compared to their values derived from the homogeneous thiele equation . in the following we express @xmath91 by the frequency and the damping constant .
the velocity of the vortex then reads the values of @xmath37 and @xmath13 can be determined by micromagnetic simulations . for these simulations we used our extended version of the object oriented micromagnetic framework ( oommf ) that includes the adiabatic and non - adiabatic spin torque.@xcite the position of the vortex core
was defined as the point with the maximum out - of - plane magnetization . to determine this maximum
, the simulation cell with maximum out - of - plane magnetization and its next neighbors are interpolated with a polynomial of second order . for the simulations the material parameters of permalloy , i.e. , a saturation magnetization of @xmath95 a / m and an exchange constant of @xmath41 j / m where used .
for the determination of @xmath37 the vortex was excited by a magnetic field pulse .
the subsequent oscillation was then fitted with the first two terms in eq .
( [ eq_solution_sup ] ) .
@xmath37 can then be determined from eq .
( [ eq_omega_free_gamma ] ) . finally the value of @xmath13 was determined by fitting an excitation with a direct current .
the results are shown in fig .
[ d_gamma ] for different edge lengths @xmath23 and different thicknesses of the sample .
it can be clearly seen that @xmath37 is smaller than @xmath13 . using the original thiele equation for the description of the vortex motion @xmath37 and @xmath13
are assumed to be equal .
figures [ fit1 ] and [ fit2 ] show an example for the fit of a numerically calculated vortex - core trajectory using both theories .
the theory based on the modified thiele equation shows better accordance than the theory based on the original thiele equation .
it can be seen that the thiele equation has to be modified for a sufficient description of the dynamics of current - driven magnetic vortices in the presence of a non - adiabatic spin torque .
this modification takes the deformation of the outer part of the vortex into account . | the strength of the non - adiabatic spin torque is currently under strong debate , as its value differs by orders of magnitude as well in theoretical predictions as in measurements . here
, a measurement scheme is presented that allows to determine the strength of the non - adiabatic spin torque accurately and directly .
analytical and numerical calculations show that the scheme allows to separate the displacement due to the oersted field and is robust against uncertainties of the exact current direction .
a spin - polarized current flowing through a ferromagnetic sample interacts with the magnetization and exerts a torque on the local magnetic moments . for conduction
electron spins that follow the local magnetization adiabatically it has been shown that the interaction via spin transfer can be described by adding a current - dependent term to the landau - lifshitz - gilbert equation.@xcite this equation has been extended by an additional term that takes the non - adiabatic influence of the itinerant spins into account.@xcite theoretically , several mechanisms have been proposed as the origin of the non - adiabatic spin torque , leading to different orders of magnitude for its strength.@xcite thus a precise measurement of the non - adiabatic spin torque is necessary to give insight into its microscopic origin .
a determination of its strength is further important for a reliable prediction of the current - driven domain - wall velocity.@xcite currently measured values of the non - adiabatic spin torque for permalloy differ by one order of magnitude,@xcite thus the strength of the non - adiabatic spin torque is under strong debate . in these experiments the observed motion of a domain wall
was compared with micromagnetic simulations to determine the non - adiabatic spin torque .
this analysis is highly susceptible to surface roughness and oersted fields . due to its high symmetry and spacial confinement
a vortex in a micro- or nanostructured magnetic thin - film element is a promising system for the investigation of the spin - torque effect.@xcite vortices are formed when the in - plane magnetization curls around a center region . in this few - nanometer - large center region , called the vortex core , the magnetization turns out of plane to minimize the exchange energy .
there are four different ground states of a vortex .
these states are labeled by the direction of the out - of - plane magnetization , called polarization @xmath0 , and the sense of rotation of the in - plane magnetization , called chirality @xmath1 .
polarizations of @xmath2 and @xmath3 denote a core that points parallel or antiparallel to the @xmath4 axis , respectively .
a chirality of @xmath5 denotes a counterclockwise curling of the in - plane magnetization while @xmath6 denotes a clockwise curling .
it is known that vortices are displaced from their equilibrium position when excited by spin - polarized electric current pulses.@xcite the spatial confinement of the vortex core within the film element yields an especially accessible system for measurements with scanning probe techniques , such as soft x - ray microscopy , x - ray photoemission electron microscopy , or scanning electron microscopy with polarization analysis .
an analytical solution of the extended landau - lifshitz - gilbert equation shows that for a current - driven vortex the forces due to the adiabatic and the non - adiabatic spin torque are perpendicular to each other.@xcite in this work we present a scheme which allows to measure the contributions due to the adiabatic spin torque , the non - adiabatic spin torque , and the oersted field separately .
it bases upon analytical calculations @xcite and overcomes the two main difficulties that occur in an experiment .
the first problem arises from an additional vortex displacement due to the oersted field accompanying the current flow.@xcite this displacement is comparable in size to the displacement due to the non - adiabatic spin torque and both displacements point in the same direction.@xcite thus , the unknown contribution of the oersted field has to be separated from the measured signal .
the second problem is the exact determination of the displacement angle .
since the displacement due to the adiabatic spin torque is about one order of magnitude larger than the displacement due to the non - adiabatic spin torque , a small uncertainty in the direction of the current through the sample would cause large errors in the determination of the non - adiabatic spin torque parameter . to test the applicability of our analytical findings they are applied to vortex displacements obtained from three - dimensional micromagnetic simulations . for the analytical calculations we start from a modified version of the thiele equation @xcite
@xmath7 that takes deformation of the vortex into account.@xcite here @xmath8 is the velocity of the vortex core , @xmath9 is the gilbert damping , @xmath10 the strength of the non - adiabaticity , @xmath11 the force on the vortex , @xmath12 the @xmath4 component of the gyrovector , and @xmath13 the diagonal element of the dissipation tensor . the coupling constant @xmath14 between the current and the magnetization depends on the saturation magnetization @xmath15 and the spin polarization @xmath16 .
the assumption of a magnetization pattern which rigidly gyrates holds true only for the small vortex core . due to the spatial confinement the remaining part of the vortex has to deform while the core is moving .
@xmath17 is a phenomenological parameter that takes into account a reduced dissipation due to this deformation.@xcite we will investigate a square thin - film element with a homogeneous current flowing in @xmath18 direction .
the accompanying oersted field is accounted for by a homogeneous field @xmath19 in @xmath20 direction . for small displacements of the vortex core from its equilibrium position
the demagnetization energy can be expanded up to second order in the core displacement @xmath21 .
the force on the vortex is then given by @xcite @xmath22 with the lateral extension @xmath23 , and thickness @xmath24 of the system .
the factor @xmath25 parameterizes the confining potential.@xcite this equation can be solved for harmonic excitations of the form @xmath26 and @xmath27 .
after all transient states are damped out , the position of the vortex core is given by @xmath28 with @xmath29 , the gyromagnetic ratio @xmath30 , @xmath31 , the free oscillation frequency of the vortex @xmath32 , and its damping @xmath33.@xcite from eq .
( [ eq_solution ] ) it is obvious that an oersted field has the same influence on the vortex as the non - adiabatic spin torque .
thus the presence of an oersted field can disturb the measurement of the non - adiabatic spin torque . in experiments
the coordinate system is given by the sample axis .
a small uncertainty of the direction of the current flow , e.g. due to a rotation or imperfections of the sample , yields a mixing of the displacement components , resulting from the adiabatic spin torque and the smaller non - adiabatic spin torque , relative to the sample axis .
this mixing causes a large error in the measurement of the displacement originating from the non - adiabatic spin torque .
an excitation with a direct current causes a displacement of the vortex core to a new steady - state position.@xcite a benefit is that a direct current allows for a measurement with a non - time - resolving technique .
the displacement of a vortex driven by a direct current is given by @xmath34 the sign of the displacement induced by the oersted field depends on the chirality of the vortex , while the displacement due to the adiabatic spin torque is determined by the polarization.@xcite the non - adiabatic spin torque causes a displacement that is independent of the vortex properties @xmath0 and @xmath1 .
vortices with different @xmath0 and @xmath1 values can be achieved by remagnetizing the sample . comparing the displacement of three vortices with different polarizations and chiralities
it is therefore possible to separate the contributions of all three forces to the displacement of the vortex . from eq .
( [ eq_solution_dc ] ) we find @xmath35 these equations are schematically illustrated in fig .
[ diff ] . from eqs .
( [ eq_non - ad ] ) and ( [ eq_ad ] ) it is possible to determine the non - adiabaticity parameter as @xmath36 since this equation is independent of the strength of the oersted field , the angle of the sample , the frequency @xmath32 , the damping @xmath33 , and the parameter @xmath37 , it yields the sought measurement scheme . with this scheme
a direct determination of @xmath10 is accessible .
only one micromagnetic simulation for the determination of @xmath38 is necessary since @xmath38 is independent of @xmath10 and @xmath39 .
micromagnetic simulations of the experimental setup allow to determine the positions of the vortex core with a precise knowledge of the micromagnetic parameters of the system .
the simulations therefore allow to test the analytical results in eqs .
( [ eq_solution_dc ] ) and ( [ eq_xi ] ) . for the simulations the material parameters of permalloy ,
i.e. , a saturation magnetization of @xmath40 a / m and an exchange constant of @xmath41 j / m , are used . since we are interested only in the steady final position of the vortex we used a gilbert damping of @xmath42 to ensure a fast damping of the transient states to reduce computation time .
as a sample system we considered a square thin - film element of length @xmath43 nm and thickness @xmath44 nm with a cell size of 2 nm in the lateral directions and 10 nm perpendicular to the film .
this system allows for a reasonable computation time . for the simulations we used our extended version of the object oriented micromagnetic framework.@xcite figure [ displacement ]
shows the displacement of the vortex core in simulations without oersted field . as predicted by eq .
( [ eq_solution_dc ] ) the displacement in the direction of the current flow is proportional to @xmath10 and the displacement perpendicular to the current flow is independent of @xmath10 . from these simulations the value @xmath45 can be determined . in experimental samples
we are faced with an oersted field and an uncertainty of the direction of the current flow . to mimic the oersted field in the simulations we applied an in - plane field perpendicular to the current .
the strength of the field is given by @xmath46 nm .
the uncertainty of the direction of the current flow was taken into account by rotating the sample by 5 degrees .
figure [ positions](a ) shows the positions of the vortex core for both simulations .
it becomes visible that the oersted field and the rotation of the sample strongly shift the core positions , complicating the determination of @xmath10 . to test the analytical model we compared the non - adiabatic spin - torque parameter @xmath47 that was inserted into the simulations with the value @xmath48 that was calculated from eq .
( [ eq_xi ] ) using the core positions . here
it is worth noting that the value of the oersted field and the angle of the sample are not needed for the calculation of @xmath48 .
the results are shown in fig .
[ positions](b ) .
it can be seen that all the perturbations that are inserted in the simulations can be effectively excluded by the analytical calculations . in experimental samples
we are also faced with the anisotropic magnetoresistance ( amr ) effect that leads to inhomogeneous current paths , i.e. , a higher current density in the vortex core .
simulations including these inhomogeneous current paths yield a small shift to lower values of @xmath48 .
this shift is up to 2 % for an amr ratio of 10 % .
in the remaining part we will discuss the experimental accuracy in the determination of @xmath10 that can be achieved with the presented scheme . in experiments direct currents of densities up to @xmath49
a / m@xmath50 have been realized in permalloy on a diamond substrate.@xcite assuming a spin polarization of 0.5 we get a spin - polarized current density of @xmath51 a / m@xmath50 , i.e. , the maximum shown in fig .
[ displacement ] .
this yields values of up to @xmath52 m / s .
the displacements of the vortex in the numerically investigated samples are small compared to the experimental resolutions available .
a larger displacement of the vortex can be achieved by increasing the lateral size of the structure .
for example simulations of a square thin - film element of length @xmath53 nm and thickness @xmath44 nm yielded values of @xmath54 and @xmath55 s. with these values eq .
( [ eq_ad ] ) yields @xmath561100 nm . we assume that the core position can be measured with a resolution of @xmath57 nm . equation ( [ eq_xi ] ) then yields that @xmath58 can be realized .
this resolution ranges from 5 % to 50 % depending on the value of @xmath10.@xcite the resolution can be further increased by using thin - film elements with still larger lateral sizes .
in conclusion we present a robust and direct measurement scheme for the non - adiabatic spin torque using the displacement of magnetic vortices .
this scheme allows to distinguish between the displacements of the vortex core due to the non - adiabatic spin torque , the adiabatic spin torque , and the oersted field , independently of the direction of the current flow .
we also showed that an inhomogeneous current due to the amr effect can be neglected .
the scheme thus allows a precise measurement of the non - adiabatic spin - torque parameter @xmath10 . financial support by the deutsche forschungsgemeinschaft via sfb 668 `` magnetismus vom einzelatom zur nanostruktur '' and via graduiertenkolleg 1286 `` functional metal - semiconductor hybrid systems ''
is gratefully acknowledged .
29 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , * * , ( ) . ,
* * , ( ) .
, , , * * , ( ) . , , , , * * , ( ) . , , , , * * , ( ) . , , , , , * * , ( ) . , , , , , , ,
* * , ( ) . , , , , , , * * , ( ) .
, , , , , , , , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , , , , , , , ( ) , . , , , , , , , , , , , * * , ( ) . , , , , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , , ,
* * , ( ) . , , , , ,
, * * , ( ) . , , , ,
, * * , ( ) . , , , ,
, * * , ( ) . ,
* * , ( ) .
, , , , , * * , ( ) . ,
* * , ( ) .
, , , , * * , ( ) . . , , , , , * * , ( ) . , , , , , * * , ( ) . , , , , , , * * , ( ) . ,
* * , ( ) . |
the history of the star formation rate in galaxies is one of the main ingredients required to understand their formation and evolution . whilst the roberts time scale for gas consumption by star formation gives a rough idea of the evolution of the stellar and gas content , more elaborate models ( as summarised , e.g. by sandage , 1986 )
can reproduce a wide variety of observables , from the evolution of the bulge - to - disc ratio in the hubble sequence to the integrated colours as a function of redshift . however , in sharp contrast with these global views where the star formation rate sfr(t ) is a monotone function of time , current bottom - up hierarchical scenarios of structure formation predict far more complicated histories , where star formation episodes are related to , if not driven by , halo mergers .
the star formation history is , in this context , a more or less faithful reproduction of the merger history , rich in both strong and minor events at redshifts @xmath0 or so .
while it is possible to understand some of the observed properties of high redshift galaxies in this framework , a far more direct test can be achieved with the fossil record of the star formation history itself : the colour - magnitude diagrams ( cmds ) of the resolved stellar populations in nearby galaxies .
progress in this field has been steady , driven both by ground - based wide - field surveys and by deep , high - resolution , narrow - field hst studies , reaching well below the main sequence turn off point in many satellites of the local group ( see aparicio 1998 , mateo 1998 for recent reviews ) .
the unprecedented quality of the current cmds requires powerful tools to invert them in order to derive the star formation history ( sfh ) which gave rise to the observed distribution of stars in these diagrams .
the methods used so far are based on comparisons between the observed cmd and synthetic cmds computed assuming a given sfh , and then looking for the best matching sfh .
for instance tolstoy & saha ( 1996 ) use a bayesian likelihood technique , mighell ( 1997 ) a classical @xmath1 optimisation , ng ( 1998 ) a poisson merit function , while gallart et al .
( 1999 ) and hurley - keller et al .
( 1999 ) minimise a counts in cells statistic for well - defined boxes in the cmd .
alternatively , dolphin ( 1997 ) makes a linear decomposition in terms of fiducial cmds and solves for the best matching final cmd , in terms of a @xmath1 statistic .
one of the problems with most of these techniques is that there is no guarantee that the actual sfh belongs to the set of ( usually parametric ) sfhs explored within a family of functions defined a priori .
for instance , in the case of the carina dsph , hurley - keller et al .
( 1999 ) find the best fitting 3-burst solution , leading to results which are at variance with those from mighell ( 1997 ) , who uses a non - parametric approach .
in addition , it is unclear to what extent these best matching solutions are actually good fits to the observed cmds .
we have therefore developed an entirely new method based on a combination of bayesian statistics with variational calculus which does not suffer from the limitations listed above .
full details of the method are given in hernandez et al .
( 1999 , paper i ) , and will not be repeated here .
very briefly , the method makes 3 key assumptions : ( 1 ) the metallicity of the ensemble of stars is known and has a small dispersion ; ( 2 ) the initial mass function is given and there are no unresolved binary systems ; and ( 3 ) both distance and colour excess are known to within some uncertainties .
the method then takes as inputs the positions of @xmath2 stars in a colour magnitude diagram , each having a colour @xmath3 and luminosity @xmath4 , with ( in this example , uncorrelated ) associated errors @xmath5 and @xmath6 respectively . using the likelihood technique , we first construct the probability that the @xmath2 observed stars resulted from some function @xmath7
this is given by @xmath8 where @xmath9 in this expression @xmath10 is the density of stars of mass @xmath11 along the isochrone of age @xmath12 , and only depends on the assumed imf and the set of stellar tracks ( and in particular the durations of the different evolutionary phases ) .
the @xmath13 factors are the differences in luminosity and colour of the observed star @xmath14 with respect to the luminosity and colour of a star of mass @xmath11 at time @xmath12 .
we refer to @xmath15 as the likelihood matrix , since each element represents the probability that a given star @xmath14 was actually formed at time @xmath12 with any mass .
following the discussion of paper i ,
we may write the condition that the likelihood has an extremal as the variation @xmath16 , allowing a full variational calculus analysis to be used . developing first the product over @xmath14 using the chain rule , and dividing the resulting sum by @xmath17 , one obtains @xmath18 introducing the new variable @xmath19 defined as @xmath20 into equation [ eq : sum ] we can develop the euler equation to yield @xmath21 where @xmath22 is an integral constraint . we have now transformed what was an optimisation problem , finding the function that maximises the product of integrals defined by equation [ eq : likelihood ] , into an integro - differential equation with a boundary condition ( at either @xmath23 or @xmath24 ) which can be solved by iteration to get , non - parametrically , the function @xmath7 .
further details on the numerical aspects of the procedure are available in paper i. it is important to point out that our method has distinctive advantages over other techniques : ( 1 ) the variational calculus allows a fully non - parametric reconstruction , free of any astrophysical preconceptions ; ( 2 ) there is no time - consuming comparisons between cmds , since the function ( not the parameters ) @xmath7 that maximises the likelihood if solved for directly ; ( 3 ) the cpu scales linearly with the time resolution required in the reconstruction .
paper i also presents a detailed study on the influence of the assumptions that were made : the effect of changing the imf or the presence of unresolved binaries is essentially a normalisation problem , which does not change the overall shape or localisation of a burst of star formation , while a wrong estimate of the metallicity has drastic effects on the position of a burst , a result of the age - metallicity degeneracy .
see also paper i for the effect of photometric uncertainties in the effective time resolution of the reconstructed sfh .
note that since the imf and the fraction of binaries ( and the distribution function of their mass ratios ) are unlikely to be measured , an absolute normalisation of the star formation rate can not be achieved .
[ fig_umi ]
given the main assumption made so far in order to derive a fully non - parametric variational reconstruction of the sfh , namely a determination of the metallicity of the ensemble of stars , only a few galaxies fulfill this requirement .
in addition , in order to minimise any possible systematic effects between different data sets and reduction procedures , as well as crowding / blending effects , we selected from the hst archive an homogeneous sample of five dsph galaxies ( leo i , leo ii , draco , carina and ursa minor ) of the local group which were reduced with the same procedures .
it is only such an internally consistent data set which allows robust comparisons between different galaxies , with the proviso that the volumes sampled by the wfpc2 field may not be representative of the entire galaxy .
full details of the sfh reconstructions for these galaxies are presented in hernandez et al .
( 2000 , paper ii ) .
the dsph umi can be used as an example of our procedure .
figure [ fig_umi ] gives the wfpc2 cmd we used as input .
note the increase of the 2@xmath25 error bars with decreasing luminosity , and the removal of stars on the horizontal branch ( incompatible with the evolutionary phases included in our isochrones ) .
since incompleteness may be important , and isochrones are degenerate at this level , we also remove stars below @xmath26 , so we are left with @xmath27 stars in total . for the fiducial values of distance , metallicity and colour
excess indicated ( slightly different but consistent with the more recent determinations by mighell & burke , 1999 ) , the variational method gives the reconstructed sfh indicated with the black line on the right panel of fig .
[ fig_umi ] .
this function maximises the likelihood , as defined in eq .
[ eq : likelihood ] , for the given data set . to assess the robustness of this reconstruction , many functions were reconstructed by changing the observational parameters within their error bounds indicated on fig .
[ fig_umi ] .
the _ envelope _ of such solutions is given as the shaded area on the figure , and represents the uncertainties in the reconstruction given by the uncertainties in the observations , so that _ any _ function that will be contained within these bounds will maximise @xmath17 , given these errors .
the bulk of the stellar population in umi was therefore formed more than 12 gyr ago , with a peak at 14 gyr .
some residual activity may have been present at younger ages , but it is caused by stars bluer than the main turn off point or by blue stragglers . given the uncertainties in the photometry , we can not resolve any episodes in the star formation rate within the main burst at 14 gyr , but we can conclude that the burst lasted less than 2 gyr ( fwhm ) .
having found the best solution ( in terms of maximising the likelihood ) does not guarantee that it is also a good solution .
the classical way to assess the goodness of fit , via fisher s information matrix , does not apply here ( there are no parameters ) , so we apply again bayesian techniques to check whether the observed cmd could result from the cmds produced by the best reconstruction of the sfh . to do this
, we do a counts in cells analysis and apply saha s w statistic ( saha , 1998 ) , basically @xmath28 with @xmath29 number of cells the cmd is divided into , and @xmath30 and @xmath31 the number of stars in cell @xmath14 in two cmds . to compute the distribution function of @xmath32 , a series of model - model comparisons are made , that is , synthetic cmds resulting from realisations of the reconstructed sfh are compared pairwise . to check whether the observed cmd could arise from this distribution , the @xmath32 statistic resulting from comparisons between the observed cmd and a series of cmd realisations are made .
if the distribution functions are compatible , at some statistical level , the hypothesis that the observed cmd can be produced by the reconstructed sfh is accepted . in the case of umi ,
the mean and dispersion of the model - model @xmath32 distribution is 47 @xmath33 5 , while the data - model distribution gives 44 @xmath33 4 , so we can deduce that the synthetic cmds are compatible at better than 1@xmath25 with the observed cmd .
the best solution is also a good solution .
[ fig_leo2 ] the dsph leo ii ( see fig .
[ fig_leo2 ] ) presents a similar case , whereby the bulk of stars was formed in a burst slowly rising at 12 gyr with a maximum around 8 gyr and an abrupt decrease at 6 gyr .
however in this case , the very sensitive @xmath32 statistic says that the observed cmd is not compatible ( at the 2@xmath25 level ) with the synthetic diagrams resulting from this best solution .
the reason is likely to be the relatively large spread in metallicity , around 0.3 dex , showing the limitation of the method .
[ fig_carina ] the case of the carina dsph ( fig .
[ fig_carina ] ) is more interesting .
there are clearly two sub - giant branches , identified with two epochs of star formation at 8 and 3 gyr .
given the 2@xmath25 errors indicated in fig .
[ fig_carina ] and simulated cmds with these errors , we can say that the bursting episodes in carina lasted more than a 1 gyr , and that there was no epoch in the 28 gyr interval without star formation activity .
the @xmath32 statistic indicates a good solution .
it is however puzzling that such an extended star formation history , over 6 gyr , resulted in such a small metallicity , with a dispersion of 0.2 dex or less .
either these values are only representative of the tip of the giant branch , or the enrichment in metals had a non standard history .
[ fig_leo1 ]
leo i presents a similar , extended , history of star formation ( see fig .
[ fig_leo1 ] ) .
some activity was present around 1013 gyr ago , compatible with the presence of horizontal branch stars recently detected at the ntt ( held et al .
2000 ) but not present in our hst data .
the bulk of the observed stellar populations was formed continuously , with peaks of activity at 8 and 4 gyr , in agreement with an independent analysis by gallart et al .
( 1999 ) . in this case
however , the @xmath32 statistic indicates that the maximum likelihood solution is not a good solution at the 2@xmath25 level .
again , a likely explanation is that the dispersion in metallicity ( around 0.3 dex ) is too large for the method to work
. a full bi - variate non - parametric reconstruction of both @xmath7 and @xmath34 is required .
if the local group is a representative volume of the universe , the star formation activity at high redshift was dominated by these dsph that dominate the luminosity function .
a properly averaged sfh does not show any significant epoch in the comoving density of sfr ( tolstoy 1999 , paper ii ) .
there seems to be no link either between the epoch of the bursts and perigalacticon passages of these satellites .
the evolution of the star formation rate in the galactic disc is another basic function required to understand the formation of the disc , its chemical evolution and , more generally , the luminosity evolution of spiral galaxies .
previous attempts at determining the sfh in the galaxy have relied on indirect methods , basically relations between age and some astrophysical property such as chromospheric activity or metallicity and then correct with evolutionary models for the stars which have disappeared from the sample .
a good example of the complexities of these techniques is provided by rocha - pinto et al .
( 2000b ) . here
we use the method outlined in [ sec_theory ] since the solar neighbourhood has a small dispersion in metallicity centred on the solar value .
further details will be found in hernandez et al .
( 2000b , paper iii ) .
the first step is to define a volume - limited sample of well - measured stars , and the ideal catalogue for this task is obviously the hipparcos catalogue .
this provides a direct way to infer the sfh of the solar neighbourhood , as opposed to more indirect methods ( e.g. rocha - pinto et al .
although the completeness of hipparcos varies both with spectral type and galactic latitude , a cut at @xmath35 for the sample with parallax errors smaller than 20% provides a reasonable sub - sample , once binaries and variable stars are removed .
a typical cut at @xmath36 can produce an absolute - magnitude limited sample complete in volume with a well understood error distribution .
the absolute magnitude limit , say at @xmath37 , implies that only stars younger than about 3 gyr enter the sample , but the kinematical and geometrical corrections are minimised . with such samples , and their error distributions , we simulated synthetic cmds with two different sfhs : a 3 burst scenario and a constant sfr over 1.5 gyr .
figure [ fig_hip_tests ] shows that our maximum likelihood variational method reconstructs correctly such star formation histories , even though the number of stars is very small .
note also that the cmds look very similar , yet they are produced by very different sfhs .
[ fig_hip_tests ] the inversion procedure applied to the actual hipparcos diagram for the solar neighbourhood is shown on fig .
[ fig_hip_sfh ] .
the superb quality of the data allows us to reconstruct its sfh with the unprecedented resolution of 50 myr , and a clear pattern emerges : over a roughly low - level constant sfr , there are several distinct peaks , separated by about 0.5 gyr . as before , the envelope gives the area where any function will maximise the likelihood , when different @xmath38 cuts are applied .
note that the envelope increases with age , a reflection of larger uncertainty .
this best solution is also a good solution ( see paper iii ) .
[ fig_hip_sfh ] a possible interpretation of the quasi - periodic episodes in the star formation rate of the solar neighbourhood may be given in terms of interactions with either spiral arms or the galactic bar .
as the pattern speed and the circular velocity are in general different , the solar neighbourhood periodically crosses an arm region , where the increased local gravitational potential might possibly trigger an episode of star formation . in this case , the time interval @xmath39 between encounters with an arm at the solar neighbourhood is @xmath40 where @xmath11 is the number of arms in the spiral pattern .
recent determinations tend to point to large values for the pattern speed ( e.g. mishurov et al .
1979 , avedisova 1989 , amaral and lpine 1997 ) close to @xmath41 km s@xmath42 kpc@xmath42 , which would imply that the regularity present in the reconstructed @xmath7 would be consistent with a scenario where the interaction of the solar neighbourhood with a two - armed spiral pattern would have induced the star formation episodes we detect .
this is consistent with the current knowledge on the nearby spiral structure ( valle 1995 ) , and is reminiscent of the explanations put forward to account for the inhomogeneities observed in the hipparcos velocity distribution function , where well - defined branches associated with moving groups of different ages ( chereul et al .
1999 , skuljan et al . 1999 , asiain et al . 1999 ) could perhaps be also associated with an interaction with spiral arm(s ) , although in this case the time scales are much smaller .
of course , other explanations are possible ; for example the cloud formation , collision and stellar feedback models of vazquez & scalo ( 1989 ) predict a phase of oscillatory star formation rate behaviour as a result of a self - regulated star formation rgime .
close encounters with the magellanic clouds have also been suggested to explain the intermittent nature of the star formation rate , though on longer time scales ( rocha - pinto et al .
we put forward a maximum likelihood technique , coupled to a variational calculus , which allows the robust , non - parametric reconstruction of the evolution of the star formation rate from the information contained in colour - magnitude diagrams .
a full bayesian analysis is also applied to assess whether the best solutions found are also good fits to the data .
its main limitation , at the moment , is the prior knowledge of the metallicity of the ensemble of stars in a cmd .
for the first time , an objective reconstruction of star formation histories without any a priori or model - dependent information is applied to an homogeneous sample of dwarf spheroidals of the local group .
we find a wide variety of sfhs , with bursts of activity uncorrelated to any special epoch or event , like perigalacticon passages . among many other things ,
this also implies that late accretion was not important in the formation of the galactic halo ( gilmore et al .
2000 ) . in the solar neighbourhood observed by hipparcos
, we infer with an unprecedented resolution of 50 myr its star formation history over the past 3 gyr , finding a surprising regularity of star formation episodes separated by some 0.5 gyr .
a possible explanation is that the solar neighbourhood interacted with two spiral arms or the galactic bar , triggering star formation at each interaction .
these bursts are likely to induce the formation of massive star clusters which slowly dissolve into the galactic disc .
amaral l.h . , lpine j.r.d . , 1997 , mnras , 286
, 885 aparicio a. , 1998 , in iau symp 192 , the stellar content of local group galaxies , ed .
p. whitelock & r. cannon ( san francisco : asp ) , 20 asiain r. , figueras f. , torra j. , 1999 , a&a , 350 , 434 avedisova v.s .
, 1989 , astrophys .
30 , 83 chereul e. , crz m. , bienaym o. , 1999 , a&a suppl .
135 , 5 dolphin , a. , 1997 , new ast . , 2 , 397 gallart c. , freedman w. , aparicio a. , bertelli g. , chiosi c. , 1999 , aj , 118 , 2245 gilmore g. , hernandez x. , valls - gabaud d. , 2000 , the galactic halo : from globular clusters to field stars , 35th lige conf . , ed .
a. noels et al . in press ( astro - ph/9910409 )
held , e.v .
et al . , 2000 , apj , 530 , l85 hernandez x. , valls - gabaud d. , gilmore g. , 1999 , mnras , 304 , 705 ( paper i ) hernandez x. , gilmore g. , valls - gabaud d. , 2000a , mnras , 317 , 83 ( paper ii ) hernandez x. , valls - gabaud d. , gilmore g. , 2000b , mnras , 316 , 605 ( paper iii ) hurley - keller d. , mateo m. , nemec j. , 1998 , aj , 115 , 1840 mateo m. , 1998 , ara&a , 36 , 435 mighell k.j . , 1997 , aj , 114 , 1458 mighell k.j . ,
burke , c.j .
, 1999 , aj , 118 , 366 mishurov y.n . ,
pavloskaya e.d .
, suchkov a.a . , 1979 ,
azh , 56 , 268 ng , y.k . , 1998 ,
a&as , 132 , 133 rocha - pinto h.j . , scalo j. , maciel w.j . ,
flynn c. , 2000a , apj , 531 , l115 rocha - pinto h.j . ,
maciel , w.j .
, scalo j. , flynn c. , 2000b , a&a , submitted ( astro - ph/0001383 ) saha p. , 1998 , aj , 115 , 1206 sandage , a. 1986 , a&a , 161 , 89 skuljan , j. , hearnshaw , j.b . , cottrell , p.l .
, 1999 , mnras , 308 , 731 tolstoy e. , 1999 , in dwarf galaxies and cosmology , ed .
thuan , c. balkowski , v. cayatte , j. tran thanh van ( paris : editions frontires ) tolstoy e. , saha a. , 1996 , apj , 462 , 672 valle , j.p . , 1995 , apj , 454 , 119 vazquez e.c .
, scalo j.m . , 1989 ,
apj , 343 , 644 | the unprecedented quality of recent colour - magnitude diagrams of resolved stellar populations in nearby galaxies requires state - of - the - art techniques to infer the star formation histories which gave rise to the observed distributions .
we have developed a maximum likelihood technique , which coupled to a variational calculus allows us to make a robust , non - parametric reconstruction of the evolution of the star formation rate .
a full bayesian analysis is also applied to assess whether the best solutions found are also good fits to the data . applying this new method to wfpc2 observations of five dsph galaxies of the local group
, we find a wide variety of star formation histories , with no particular epoch being dominant . in the case of the solar neighbourhood observed by hipparcos
we infer , with an unprecedented resolution of 50 myr , its star formation history over the past 3 gyr .
the surprising regularity of star formation episodes separated by some 0.5 gyr could be interpreted as the result of interactions with two spiral arms or the galactic bar .
these bursts possibly trigger the formation of massive star clusters which slowly dissolve into the galactic field . |
lattice qcd has established the existence of a transition from hadron gas to a new state of strongly interacting matter where quarks and gluons are no longer confined inside hadrons and which is usually called the quark gluon plasma @xcite .
the nature of this transition depends on the quark content and quark masses . for infinite or very large quark masses
the transition is a 1st order deconfining transition . in the opposite case of zero quark masses
one may have a 2nd order chiral phase transition for 2 flavors or a 1st order chiral phase transition for 3 flavors . for intermediate masses
the transition is just a rapid crossover , meaning that thermodynamic quantities change very rapidly in a narrow temperature interval .
the boundary of the 1st order transition region of 3 flavor qcd as a function of mass has been studied using improved ( p4 ) @xcite and standard staggered actions @xcite .
there is a significant discrepancy regarding the value of the quark masses , or equivalently the value of the pseudo - scalar meson masses where the transition changes from crossover to 1st order . with an improved action
it was found that the 1st order transition ends for a pseudo - scalar meson mass of about @xmath4 mev , while with the standard action it ends for pseudo - scalar meson masses of about @xmath5 mev @xcite or larger @xcite .
it has been observed that the pressure and energy density normalized by its ideal gas value shows almost the same behavior as a function of @xmath6 for su(3 ) pure gauge theory , 2 flavor , 2 + 1 flavor and 3 flavor qcd @xcite .
thus flavor and quark mass dependence of these quantities in the first approximation , is determined by flavor and quark mass dependence of transition temperature @xmath7 .
therefore it is very instructive to study the flavor dependence of the transition temperature .
such a study has been performed in ref .
@xcite on lattices with temporal extent @xmath0 using the improved staggered p4 action . in the past most simulations with 2 and 3 flavors of staggered fermions
have been done using the hybrid molecular dynamics r ( hmdr ) algorithm @xcite , often called simply the r - algorithm .
it has finite step - size errors of @xmath8 where a step - size @xmath9 is used in the molecular dynamics evolution .
recently the rational hybrid monte carlo ( rhmc ) algorithm has been invented which allows simulations of theories with fractional powers of the fermion determinant , for example 2 and 3 flavors of staggered fermions , without finite step - size errors @xcite .
therefore the most recent thermodynamics studies use the rhmc algorithm @xcite .
it has been observed in ref .
@xcite that the use of the exact rhmc algorithm reduces the value of the critical quark mass where the transition turns to 1st order by @xmath10 in the case of the standard staggered action .
the purpose of this paper is twofold .
first we would like to study the transition in 3 flavor qcd , extending the previous studies to smaller quark masses and smaller lattice spacings using the improved p4 staggered fermion action .
second we would like to compare the r - algorithm with the new rhmc algorithm .
the rest of the paper is organized as follows . in section
ii we discuss the calculational set - up . in section iii
we show our results for finite temperature calculations .
section iv discusses the zero temperature simulations needed to set the scale and the transition temperature in physical units . in section v we present a comparison of the r and rhmc algorithms .
finally , section vi contains our conclusions . a technical discussion of different fat link actions is given in appendix a. in appendix
b we discuss the eigenvalues and eigenvectors of the p4 action in the free field limit .
most of the simulations discussed in this paper were done using the p4fat3 action , which is also simply called the p4 action @xcite . to improve the rotational symmetry , which is violated on the lattice , bent 3-link terms are added to the 1-link term of the standard staggered action @xcite .
properly chosen coefficients of the 1-link and the 3-link terms can eliminate , at tree level , the @xmath11 errors in the dispersion relation for staggered fermions @xcite .
the violation of flavor symmetry which is present in the staggered fermion formulation can be significantly reduced by replacing the normal link in the 1-link term by a fat link which is a sum a of normal link and 3-link staples @xcite .
this 3-link fattening is the origin of the name p4fat3 .
though this type of fat link action is a big improvement over the standard staggered fermion action , further improvement of the flavor symmetry can be obtained by adding 5- and 7-link staples @xcite .
in fact one can eliminate the effect of flavor symmetry breaking at order @xmath12 using a suitable combination of 3- , 5- and 7- link staples leading to what is called the fat7 action @xcite .
we have also done calculations with the p4 action with fat7 fat links .
unfortunately it turns out that on the coarse lattices used in our study of 3 flavor qcd thermodynamics this action has some undesirable features .
it leads to the occurrence of a bulk transition , which we will discuss in more detail in appendix a. to study staggered fermions with less than four flavors , we use the rooting procedure , i.e. each fermion flavor is represented by @xmath13 , where @xmath14 is the staggered fermion dirac operator .
for a recent discussion of this procedure see ref .
most of our simulations have been done using the standard r - algorithm @xcite . as in ref .
@xcite the step - size of the molecular dynamics evolution was set to @xmath15 for the staggered quark mass @xmath16 .
we also performed calculations with the rhmc algorithm . in this algorithm
an optimal rational approximation is used to evaluate the fractional power of the determinant @xcite .
more precisely one finds the optimal approximation for @xmath17 where @xmath18 is the usual staggered fermion matrix and for the three flavors @xmath19 .
using sufficiently high order polynomials the rational approximation can be made arbitrarily precise for the given spectral range of the fermion operator .
for the standard staggered fermion action the spectral range of @xmath20 is well known , the smallest eigenvalue of this matrix is @xmath21 .
the largest eigenvalue can be estimated in the free field limit to be @xmath22 .
for the case of the p4 staggered action the smallest eigenvalue is the same , while the largest eigenvalue in the free case is @xmath23 . in appendix
b we give the derivation of this result .
it turns out that in all our simulations the largest eigenvalue was smaller than 5.0 , so we choose @xmath24 as the upper limit on it . for the range of the quark masses studied by us , which include quark masses as light as 1/20th of the strange quark mass , it is sufficient to use polynomials of degree 12 to achieve machine precision with the rational approximation
. therefore in the metropolis accept / reject step we used polynomials of degree 12 . in the molecular dynamics evolution
we used a less stringent approximation of the determinant since any errors in the evolution , including the @xmath25 step - size errors , are eliminated by the accept / reject step .
it has been found that in most cases it is sufficient to use polynomials of degree 5 - 6 without compromising the acceptance rate .
furthermore , the stopping criteria for the conjugate gradient inversions can be relaxed to @xmath26 in the molecular dynamics evolution without significant effect on the acceptance rate . in the monte - carlo
accept / reject step we typically use @xmath27 for the conjugate gradient stopping criteria .
the length of the trajectory was @xmath28 in units of molecular dynamics time .
the gauge fields and quarks contribute different amounts to the force in the molecular dynamics evolution .
the contribution of the gauge fields is larger than that of quarks . on the other hand
the cost of the evaluation of the force coming from the gauge fields is small .
the opposite is true for the part of the force coming from the fermions .
therefore it is reasonable to integrate the gauge force on finer molecular dynamics time scale than the fermion force @xcite . in our simulations
we typically used 10 gauge field updates per fermion update .
the 3 flavors of staggered fermions are simulated as 1 + 1 + 1 , i.e. we used a factor @xmath29 for each fermion flavor .
although this increases the number of inversions , the reduced force allows for a larger step - size @xmath30 for the same acceptance rate .
the step - size @xmath9 was chosen such that the acceptance is about @xmath31 . to achieve this ,
the stepsize was typically of the order of the strange quark mass used in our 2 + 1 flavor study @xcite . .
parameters of the numerical simulations [ cols="^,>,^,^,^",options="header " , ]
in this paper we have studied the phase transition in 3 flavor qcd at finite temperature using @xmath0 and @xmath32 lattices . for the quark mass corresponding to the second order end - point
we find the critical temperature to be @xmath2 .
the transition temperature in the chiral limit is about @xmath33 smaller than the above value . for a given pseudo - scalar meson mass
the difference between the transition temperature in 3 flavor and 2 + 1 flavor case is less than @xmath34 .
we also find that the cut - off dependence of the transition temperature in 3 and 2 + 1 flavor qcd is very similar .
furthermore , we find that finite step - size errors present in the r algorithm are negligible , at least for the p4fat3 action at the quark masses studied .
this work has been supported in part by contracts de - ac02 - 98ch1 - 886 and de - fg02 - 92er40699 with the u.s .
department of energy , the helmholtz gesellschaft under grant vi - vh-041 and the deutsche forschungsgemeinschaft under grant grk 881 .
the work of c.s . has been supported through ldrd funds from brookhaven national laboratory .
the majority of the calculations reported here were carried out using the qcdoc supercomputers of the riken - bnl research center and the u.s .
in addition some of the work was done using the ape1000 supercomputer at bielefeld university
in this appendix we are going to discuss the properties of the finite temperature transition in the case of the p4fat7 action and compare it to the p4fat3 case .
in general the gauge transporter in the 1-link term of the p4 action can be replaced by combination of the link variable and different staples , called the fat link , without changing the naive continuum limit .
this is true provided the coefficient of different terms in the fat link satisfy appropriate normalization conditions .
for example in the case of the fat link with the three link staple only , this condition reads @xmath35 . introducing five and seven link staples in addition to the three link staples
give the so - called fat7 link @xcite . in this case
the normalization condition reads : @xmath36 .
it is possible to eliminate the leading order coupling to the high momentum gluons with momenta @xmath37 , @xmath38 and @xmath39 , i.e. to suppress the flavor changing interaction at order @xmath40 if the coefficients are chosen as @xcite @xmath41 this gives then the value @xmath42 for the coefficient of the 1- link term to get the naive continuum limit . in fig .
[ fig : pbpfta7 ] we show the chiral condensate calculated on the @xmath0 and @xmath32 lattices .
we can see that for small quark masses the transition becomes strongly first order .
the value of the chiral condensate in the low temperature phase is also much larger than for the p4fat3 action discussed in the main text . for the smallest quark mass
the discontinuity in the chiral condensate is about the same for @xmath0 and @xmath32 lattice , although we would expect it to decrease by roughly a factor of @xmath43 when going from @xmath0 to @xmath32 .
this could mean that we are dealing with a bulk transition . in fig .
[ fig : betac ] we also compare the pseudo - critical couplings for the p4fat3 and p4fat7 actions .
we see that for large quark masses the @xmath44 dependence of the pseudo - critical coupling is similar , though their values are significantly different . for small quark masses
the pseudo - critical couplings calculated for @xmath0 and @xmath1 come very close together , again suggesting that the transition may be a bulk transition .
we did calculations also on @xmath45 lattice at @xmath46 . in fig .
[ fig : betac ] we compare the chiral condensate calculated on @xmath47 , @xmath48 and @xmath45 lattices .
we see a sharp drop in the value of the chiral condensate , which occurs at the same @xmath49 for @xmath32 and @xmath50 .
this again indicates a bulk transition .
( left ) and @xmath32 ( right ) lattices for p4fat7 action.,title="fig:",width=264 ] ( left ) and @xmath32 ( right ) lattices for p4fat7 action.,title="fig:",width=264 ] one may wonder which feature of the p4fat7 action is responsible for the bulk transition .
the main difference of the p4fat7 action compared to p4fat3 action as well as to other fat link action ( e.g. asqtad ) is the negative sign of the one link term .
close to the continuum limit the normalization condition @xmath51 should insure that the combination of 1- , 3- , 5- and 7-link terms will describe a conventionally normalized , positive dirac kinetic energy . however , at stronger coupling where the gauge field are more disordered , the staples with many links are expected to give a significantly smaller contribution and the 1-link term may dominate , resulting in an effective kinetic energy term with a possibly negative sign . while this would simply correspond to non - standard sign and normalization conventions for the dirac kinetic energy , it raises the possibility that this effective kinetic energy term will change sign as one passes from strong to weak coupling .
such a sign change could induce a bulk transition .
in addition , the change in magnitude of the coefficient in the effective kinetic energy ( small for strong coupling and large for weak coupling ) would appear reversed in the chiral condensate ( large for strong coupling and small for weak coupling ) consistent with the observed behavior . to verify this we did calculations with p4fat7 action but with different coefficients which we call the p4fat7 action .
the coefficients were chosen to be @xmath52 for this action we found no evidence for a strong first order transition but only a crossover .
this can be seen for example in the behavior of the chiral condensate shown in fig .
[ fig : pbpfat7p ] . both the value of the chiral condensate and the location of the transition point is very similar to that of the p4fat3 action . , @xmath48 and @xmath45 lattices at @xmath46 ( right ) . ,
title="fig:",width=264 ] , @xmath48 and @xmath45 lattices at @xmath46 ( right ) . ,
title="fig:",width=264 ] lattice at @xmath46 .
we also show the chiral condensate calculated with p4fat7 and p4fat3 actions at the same quark mass .
the data points for p4fat3 action have been shifted horizontally by @xmath53 for better visibility .
the band in the right figure corresponds to ferrenberg swendsen re - weighting.,title="fig:",width=264 ] lattice at @xmath46 .
we also show the chiral condensate calculated with p4fat7 and p4fat3 actions at the same quark mass .
the data points for p4fat3 action have been shifted horizontally by @xmath53 for better visibility .
the band in the right figure corresponds to ferrenberg swendsen re - weighting.,title="fig:",width=264 ] we also calculated the eigenvalues @xmath54 of the p4fat7 dirac operator .
the normalized distribution of the lowest 50 eigenvalues @xmath55 is shown in figs .
[ fig : eig_m01 ] and [ fig : eig_m001 ] for the quark masses @xmath56 and @xmath46 , respectively .
we have used 100 configurations for @xmath46 , and 200 configurations for @xmath56 . given the above definition of @xmath55 the breaking of the chiral symmetry manifests itself in a non - zero density at @xmath57 . in fig . [
fig : eig_m01 ] we show the eigenvalue distribution for the larger quark mass @xmath56 .
it shows the expected features : large density of eigenvalues near @xmath58 in the low temperature phase ( @xmath59 ) , significant drop of eigenvalue density around zero at the transition ( @xmath60 ) and zero density of eigenvalues at the origin for the deconfined phase ( @xmath61 ) .
the situation is different for the smallest quark mass @xmath46 , where we see non - zero density of eigenvalues at @xmath58 even in the deconfined phase ( @xmath62 ) .
the large decrease in the density of eigenvalues at the origin when going from the confined phase ( @xmath63 ) to the deconfined explains the large drop in the value of the chiral condensate . calculated on @xmath47 lattice in the confined phase ( @xmath59 ) , at the transition ( @xmath64 ) and in the deconfined phase ( @xmath61 ) . , title="fig:",width=207 ]
calculated on @xmath47 lattice in the confined phase ( @xmath59 ) , at the transition ( @xmath64 ) and in the deconfined phase ( @xmath61 ) .
, title="fig:",width=207 ] calculated on @xmath47 lattice in the confined phase ( @xmath59 ) , at the transition ( @xmath64 ) and in the deconfined phase ( @xmath61 ) .
, title="fig:",width=207 ] calculated on @xmath47 lattice in the confined phase ( @xmath63 ) , at the transition ( @xmath65 ) and in the deconfined phase ( @xmath62 ) .
, title="fig:",width=207 ] calculated on @xmath47 lattice in the confined phase ( @xmath63 ) , at the transition ( @xmath65 ) and in the deconfined phase ( @xmath62 ) .
, title="fig:",width=207 ] calculated on @xmath47 lattice in the confined phase ( @xmath63 ) , at the transition ( @xmath65 ) and in the deconfined phase ( @xmath62 ) .
, title="fig:",width=207 ]
in this appendix we discuss the calculations of the largest and the smallest eigenvalue of the staggered dirac operator in the free field limit .
let us start our discussion with case of the standard staggered fermions .
the free - field staggered dirac operator acting on a single - component fermion field is given by @xmath66 where @xmath67 are the normal staggered phases .
consider @xmath14 acting on a momentum eigenstate @xmath68 + 2 m \,\chi(p ) \,\mathrm{e}^{\imath p \cdot x},\end{aligned}\ ] ] where @xmath69 thus , the staggered dirac operator has non - diagonal terms that couple together states at different corners of the brillouin zone . in momentum space , we can write the fermion matrix as @xmath70 then , we have for @xmath71 @xmath72 noticing that @xmath73 ( mod 2@xmath74 ) , and interchanging @xmath75 and @xmath76 labels in the second piece of eq .
( [ eq : intmunu ] ) , we see that all the off - diagonal pieces of @xmath71 cancel , and we are left only with the diagonal piece , @xmath77 as expected , we have a hard lower bound on the eigenvalue spectrum of @xmath78 . the upper bound @xmath79 is realized when @xmath80 .
the situation for free - field p4 fermions is similar , but a little bit more complicated . here
, the p4 dirac operator is given by @xmath81 where @xmath82 and @xmath83 .
again , we can examine how @xmath14 acts on the momentum eigenstate in eq .
( [ chifield ] ) . as before
, we see that we have off - diagonal pieces that come about as a direct result of the presence of staggered phases .
@xmath84 or , in matrix form , @xmath85 calculating @xmath71 , we see that the off - diagonal pieces are eliminated in the same way as in the naive staggered case .
thus , we are left with only diagonal terms , @xmath86 we see that @xmath87 .
finding @xmath88 requires us to maximize the function @xmath89 . doing this
, we find the maximum when two of the components of @xmath90 are equal to @xmath91 and the other two components are equal to @xmath92 .
for example , @xmath93 .
this yields @xmath94 for the p4 action .
a similar calculation for the naik action yields @xmath95 . for completeness
, we also quote the eigenvectors of the free p4 dirac operator . using a slightly different method
we find , @xmath96 u^0_{\rho ' } \ ; \mathrm{e}^{\imath 2 p \cdot x},\ ] ] where we implicitly sum over @xmath97 .
here we have used hypercube coordinates @xmath98 and @xmath99 , where @xmath98 labels the hypercube and @xmath99 is the offset within the hypercube , @xmath100 with @xmath101 .
furthermore , @xmath102 is defined by @xmath103 \mathrm{e}^{\imath p \ : ( \rho-\rho')}$ ] and @xmath104 is a constant vector depending only on @xmath99 .
finally , the eigenvalue of the free p4 dirac operator is @xmath105 f. karsch , et al .
suppl . * 129 * , 614 ( 2004 ) c. schmidt , c. r. allton , s. ejiri , s. j. hands , o. kaczmarek , f. karsch and e. laermann , nucl .
suppl . * 119 * , 517 ( 2003 ) [ arxiv : hep - lat/0209009 ] .
f. karsch , e. laermann and c. schmidt , phys .
b * 520 * , 41 ( 2001 ) | we study the finite temperature transition in qcd with three flavors of equal masses using the r and rhmc algorithm on lattices with temporal extent @xmath0 and @xmath1 . for the transition temperature in the continuum limit
we find @xmath2 for the light pseudo - scalar mass corresponding to the end point of the 1st order transition region . when comparing the results obtained with the r and rhmc algorithms for p4fat3 action we see no significant step - size errors down to a lightest pseudo - scalar mass of @xmath3 . |
four dimensional @xmath0 gauged supergravity coupled to @xmath1-vector multiplets admits regular , static , bps black holes which are asympotic to @xmath3 . for very particular choices of prepotential ,
gauging parameters and background charges , there exists a remarkable analytic solution @xcite due to cacciatori and klemm ( ck ) .
these solutions can have spherical , flat or hyperbolic horizon geometries - hypermultiplets and there is also a rich solution space @xcite but in the current work we will restrict our focus to vector - multiplets .
there has been some futher study of these ck solutions as well as related non - bps @xmath3 black holes @xcite ] and can be interpreted in m - theory as m2-branes wrapped on riemann surfaces @xcite .
to understand the structure of these new black holes we find it quite useful to compare with the much better understood , asymptotically flat , single - center , bps black holes in @xmath0 ungauged supergravity @xcite .
these asymptotically flat black holes preserve eight supercharges at infinity , four along the bulk of the solution and eight again at the horizon where the usual enhancement occurs . while the scalar fields are unfixed at infinity , they are typically fixed at the horizon giving rise to the moniker _ attractor mechanism_. for the black hole solutions of @xmath0 gauged supergravity relevant to the current work , the asymptotic @xmath3 region will preserve all eight supercharges , while along the bulk of the solution just two will be preserved and this is enhanced to four supercharges at the horizon .
the vector - multiplet scalars will tend to be fixed at infinity as well as at the horizon , albeit to different values ; giving rise to a varying effective cosmological constant . whereas ungauged @xmath0 supergravity is _ invariant _ under an action of @xmath4
, the gauged theory should be _ covariant _ under this action of the symplectic group . to enforce this covariance , one should work in a formalism which allows for magnetic gaugings .
this is a somewhat complicated issue , the basic idea is that one can introduce magnetic gauge fields along with auxiliary tensor fields @xcite but having said that , the explicit component lagrangian of @xmath0 gauged supergravity with both magnetic and electric gaugings has not been written out .
it should nonetheless be possible to extract the relevant formulae from @xcite . in this work
we follow the pragmatic approach of @xcite and work with a theory which is a straightforward symplectic completion of @xcite . using this symplectically covariant formalism
we solve the algebraic bps equations for horizon geometries of the form @xmath5 with arbitrary electric and magnetic charges within a theory constructed from a general cubic prepotential .
schematically , the horizon equations express the charges and gauging parameters in terms of the scalar fields and metric components , a solution to this system constitutes inverting these and solving for the scalar fields and metric components in terms of the charges and gaugings . in ungauged supergravity , black hole horizons in theories with a general cubic prepotential were first systematically studied by shmakova @xcite where an implicit solution was provided up to set of @xmath1 real quadratic equations for @xmath1 variables .
in addition , when @xmath6 , an explicit solution was obtained .
this has been extended in @xcite where the explicit solution was found for situations where the special khler manifold is a symmetric space , the final answer being s_bh= where @xmath7 is the quartic invariant .
the general solution obtained in the current work for black hole horizons in gauged @xmath0 supergravity is implicit up to a set of @xmath1-holomorphic quadratic equations in @xmath1-complex variables . due to this holomorphicity ,
the solutions space is non - empty and at least zero - dimensional .
these equations can also be explicitly solved when the special khler manifold is symmetric , giving rise to an analytic solution for the scalar fields and the entropy .
we find that the entropy is related to , but not equal to , the quartic invariant @xmath8 .
our solution for the entropy depends on both the charges and the gaugings .
we also find an explicit solution when both one magnetic charge and one magnetic gauging parameter vanish p^0=0 , p^0=0 .
it is of significant interest to show generally that these horizon solutions can be continued to the uv and connected to a bps @xmath3 vacuum .
there is currently no general analytic solution like the ungauged case , nonetheless we construct numerically a particular example of interest , that which corresponds to an internally spinning m2-brane wrapped on @xmath9 in a calabi - yau fivefold .
this paper is organised as follows . in section 2 .
we introduce the ansatz and review the bps equations for horizon solutions .
we derive a form of the equations which is particularly amenable to solution and expand these equations for a general cubic prepotential . in section 3
we solve the bps equations and consider in more detail the solution for symmetric special khler manifiolds .
we present the canonical stu mode example which can be embedded in m - theory .
we also numerically compute a particular dyonic black hole which corresponds to a spinning m2 brane wrapped on a riemann surface in a calabi - yau fivefold .
we largely follow the conventions of @xcite but here we repeat the four dimensional black hole ansatz : [ ansatz ] ds_4 ^ 2&= & e^2u dt^2- e^-2u dr^2- e^2(v - u ) ( d^2+h()^2 d^2 ) [ metansatz ] + a^&= & ^(r ) dt- p^(r ) f ( ) d , with h()=\ { : & s^2 ( = 1 ) + : & ^2 ( = -1 ) . and the scalar fields are radially dependant @xmath10 .
this ansatz is general enough to describe asymptotically flat or @xmath3 black holes however this work we will primarily be interested in the horizon geometries where the radial dependance is fixed . as explained in @xcite one can easily include flat horizons .
the electric and magnetic charges are p^&= & _
s^2 f^[elinv ] , + q _ & & _ s^2
g_= -e^2(v - u ) _ ^+ _
p^ , [ maginv ] where @xmath11 is the symplectic - dual gauge field strength g _
= r _ f^- _ * f^ .
this ansatz works in theories with just electric gaugings , we develop below the generalization needed to include magnetic gaugings .
the horizon of a static bps black hole is of the form @xmath5 .
this requires the metric functions to take the form e^u= , e^v= and all scalar fields are constants z^i(r)=z_0^i .
when the theory has just electric gaugings , the horizon equations are @xcite p^q_&= & 1 + e^-i l^q_&= & + & = & e^i + p^&= & - ( e^-i l^ ) - r_2 ^ 2 ^ q _
+ q_&= & - ( e^-i _ ) - r_2 ^ 2 _
^ q_. we have introduced a certain phase @xmath12 of the supersymmetry parameter _
e^u/2 e^i _
0a where @xmath13 is an @xmath14 doublet of constant spinosr satsifying the two projections _ 0a&= & i_ab ^0 _ 0^b , + _ 0a&= & - ( ^3)_a^ b ^01 _ 0b .
working with magnetic and electric gaugings requires introducing a fair amount of data from special geometry and we have included the necessary background in appendix [ app : special ] .
a key result is that we can expand any symplectic vector in terms of @xmath15 and their conjugates .
for example the background charges @xmath16 and gaugings what is typically denoted @xmath17 .
furthermore this is different still from the notation in @xcite where the gauging parameters were denoted @xmath18 and @xmath19 .
we hope the interested reader will persevere regardless .
] = p^ + q _ , = p^
+ q_can be expanded as & = & i - i + i ^ _ - i ^i u_i + & = & i - i + i ^ _ - i ^i u_i so that = , , = , , _
i=,u_i , _ i=,u_i .
key objects for our analysis are the symplectic matrices @xmath20 and @xmath21 ( see appendix [ app : special ] ) which together satisfy = -i , u_i =
i u_i . using these one arrives at the symplectically covariant horizon equations @xcite - r_2 ^ 2 & = & -4 ( ) [ symphor1 ] + & = & e^i [ symphor2 ] + , & = & 1.[pq1 ] since the gravitino is charged , dirac quantization implies , while the bps condition selects this integer to be unity .
eq is itself a symplectic vector thus we can extract its components by contracting with @xmath22 and @xmath23 resulting in & = & , [ wz1 ] + _ i&= & _ i.[wz2 ] it is worth noting that - constitute @xmath24 real equations for the @xmath25 variables @xmath26 so generically we expect there to be two constraints on the solution space which is parameterized by @xmath27 . in the limit @xmath28 of vanishing gauging parameters ,
these equations reduce to the attractor equations of ungauged supergravity , apart from which should not be enforced in that limit .
one qualitative difference in the gauged case is that the lhs of depends on the scalar fields through @xmath29 term which makes it somewhat more complicated to disentangle these equations .
the central goal of this paper is to decouple these equations such that the scalar fields and radii are solved for in terms of the charges @xmath16 and the gaugings @xmath30 . using and we get ir_2 ^ 2 = i+i - i^ _
-i^i u_i and so - are equivalent to .
the derivation of is a key step in this analysis as we have eliminated the dependance of the scalar fields on the lhs in , nonetheless there remains a dependance on @xmath31 .
defining the holomorphic quantities ^&= & p^+ i r_2 ^ 2 p^ , [ pdef ] + _ & = & q_+ i r_2 ^ 2 q_[qdef ] and using details from appendix [ app : special ] .
we can expand explicitly for an arbitrary cubic prepotential : ^0&= & 2i e^k/2 + 12 i e^k d_y , i^i[fpfq1 ] + ^i & = & ^0 z^i -2i e^k/2 ^i [ fpfq2 ] + _ 0 & = & -^0 d_z + 6i e^k/2 ^i d_z , i [ fpfq3 ] + _ i & = & 3 ^j-6ie^k/2 ^j d_z , ij [ fpfq4 ] .
this system is @xmath32 holomorphic equations for the @xmath33 complex variables @xmath34 .
this must be supplemented with and in addition we must solve for @xmath31 . like - this generically this gives a system with 2 additional real constraints on @xmath35 .
it is interesting that the variables which we formulate the equations in terms of , namely @xmath36 are almost objects which should be _ specified _ before solving any given background . in particular
@xmath37 define the theory under consideration while @xmath27 define the vacuum selection sector , the glitch is of course that @xmath36 also involve @xmath31 .
our strategy is to solve for all variables in terms of @xmath36 and then use our final equation to solve for @xmath31 in terms of the gaugings and the charges .
we now solve the horizon equations - for models derived from a general cubic prepotential f = d_ijk and with arbitrary charges @xmath16 and gaugings @xmath30 . to this end
, we construct using - a set of @xmath1 holomorphic , quadratic equations : d_ijk^j ^k & = & ^0 _
i -4 e^k d_ijk^j ^k [ zizj ] .
we define _ i&= & d_ijk ^j ^k - ^0 _ i , [ pidef ] + ^i&=&2i e^k/2 ^i so that becomes [ pieq ] .
when @xmath39 we can explicitly solve : e^k/2^i= and then from we solve for the scalar fields @xmath40 z^i= ( d_^-1)^ij _ j where ( d_^-1)^ijd_,jk=^i_k . we then obtain @xmath41 from : = 6 e^k/2 d_y , i ^i .
finally we have the relation 0=2 _ 0 + ( d_^-1)^ij _ i _ j from which one can obtain @xmath42 in terms of @xmath43 as well as one constraint on the charges .
assuming more generally that @xmath44 we can then solve for @xmath40 z^i= .
[ zisol2 ] then can be used to solve for @xmath45 and thus : e^i & = & -6 e^k/2 d_y , i ^i [ zsol ] from we can extract the solution for one radius @xmath46 and the phase of the spinor @xmath47 as a function of the the charges , gaugings and the second radius ( which governs the entropy ) @xmath48 . in principle once the solution for @xmath49 has been obtained one immediately obtains @xmath40 from . the final step to obtain @xmath31 is to solve , which one can show is equivalent to ^0 _ 2 ( , ) & = & ^i _ i [ q0eq ] where we have introduced the quadratic symplectic invariant _ 2(,)= - ^_. [ quadinv ] this gives equations from both the real and imaginary parts , one of which can be used to solve for @xmath31 and the other gives a constraint on the charges . to extract a completely explicit solution to , we will restrict to the case of @xmath50 being a symmetric space .
finally one must also impose the dirac quantization condition .
it is important to check regularity of any resulting solution .
there is an important constraint on the scalar fields that the metric @xmath51 be finite and one must have @xmath52 so that the spacetime metric is well defined . from
we see that @xmath46 can always be taken real since any phase is absorbed into @xmath12 .
when @xmath50 is a symmetric space , we can solve explicitly .
the key object is the contravariant three - tensor @xcite ^ijk = which for symmetric spaces is constant and satisfies ^ijkd_j(lmd_np)k&=&^i_(ld_mnp ) , + d_ijk^j(lm^np)k&=&_i^(l^mnp ) .
the solution to is then ^i&= & [ zisol ] where we have defined _ & & ^ijk _ i _ j _ k + & = & ( ^0)^2 _ 4(,)+4 _ 2(,)^2and we have introduced the familiar symplectic invariant _ 4 ( , ) & = & - ^_^2- ^0 ^ijk _ i _ j _ k+4_0 d_ijk ^i ^j ^k + d_ijk ^ilm ^j ^k _ l _ m . [ quartic ]
we are now in a position to unpack and we get 2 _ 2(,)= , the only regular solution takes the @xmath53-sign in : 0=_4 ( , ) .
[ quarticzero ] recalling the definitions and , then separating into real and imaginary parts gives two polynomials in @xmath54 0&= & a_0 + a_4 r_2 ^ 4 + a_8 r_2 ^ 8 [ r2eq1 ] + 0&= & a_2 r_2 ^ 2 + a_6 r_2 ^ 6 [ r2eq2 ] where the @xmath55 are quartic invariants of the diagonal action on the charges and gaugings .
the explicit expressions for the @xmath55 can easily be obtained from but do not appear to be particularly enlightening for our purposes .
having said that , we do have a_0=_4(p , q ) , a_8=_4(p , q ) .
if either @xmath56 or @xmath57 then for regularity , both must vanish . in this case , we have r_2 ^ 4=. [ r2quad ] and no extra constraint .
regularity implies that @xmath58 which gives a bound on the space of charges . otherwise , if both @xmath59 and @xmath60 we have r_2 ^ 4=- [ r2sola2a6 ] and the constraint 0=a_0a_6
^ 2-a_2a_4a_6+a_2 ^ 2 a_8 .
[ chargeconstraint ] again we have the bound @xmath58 .
one can also use this explicit solution to obtain an explicit expression for @xmath45 and thus @xmath46 .
it is of some interest to express this in a compact form in terms of the explicit invariants of the diagonal symplectic action on @xmath61 @xcite . while it is interesting to consider the limit @xmath28 of the above solution to connect with bps black hole horizons in ungauged supergravity , this limit is not regular .
nonetheless we note that the key equations in @xcite are identical in form to but real instead of holomorphic . in that work
the entropy is given by s&= & where _ i&= & 3 d_ijk p^j p^k -p^0 q_i is essentially identical to @xmath62 in and ^i = e^k/2 || y^i solve the set of @xmath1 real quadratic equations _
i= d_ijk ^j^k . [ deltaeq ]
we note that similarly as above , when @xmath50 is a symmetric space one can solve with ^i= from which we find s=. this provides an alternative derivation from that in @xcite for the explicit entropy of bps black holes when the special khler manifold is a symmetric space .
the model with @xmath63 and prepotential given by f&= & - is commonly refered to as the stu - model . in gauged supergravity this model has considerable complexity due to the gaugings @xmath37 and the spectrum of solutions within this model varies discontinuously with these gauging parameters . for a particular set of gaugings , there is a known embedding of this gauged supergravity theory into m - theory on @xmath64 @xcite .
this model is given by p^&= & -(0,g , g , g ) , q_=(g,0,0,0 ) . if we rotate to another duality frame using & = & a & b + c & d , a = d= \{1,0,0,0 } , b =- c= \{0,1,1,1 } we get f&= & -2i + p^&=&0 , q_= g. and so the gaugings in this frame are purely electric
. there exists a remarkable analytic solution @xcite in this duality frame for black holes which are _
purely magnetic_. the horizon geometries which arise as the ir of these magnetic black holes are specified by three independant charges ; there is four charges and just one constraint , @xmath65 and the radius @xmath31 comes from .
more generally , from the computations above , we see that in this model the general space of solutions is parameterized by _
six _ charges ; the eight charges have two constraints , one from and the extra constraint .
so for these dyonic configurations the entropy comes from .
one of course needs to check that @xmath66 but one finds that this can generically be arranged .
while there is a nice analytic solution for the entire magnetic black hole in this duality frame , we have been unable to find an analytic solution for the black hole with additional electric charges . with non - trivial electric charges ,
the real part of the vector multiplet scalars @xmath67 is turned on , which complicates significantly the formulae .
nonetheless we have computed numerically the entire black hole solution for a particular choice of electric and magnetic charges and @xmath68 .
the plots are shown in fig .
[ fig : flow3vecsmetric ] and [ fig : flow3vecsscalars ] for the radial variable @xmath69 .
the charges have been chosen to be p^=(5,-1,-1,-2 ) , q_=(0,1,-1,0 ) with these choices of charges , we find that @xmath70 , @xmath71 and @xmath72 along the whole flow although one does not need to restrict to such a simple dyonic configuration . and
@xmath73 , @xmath74 and @xmath75 ] these black holes have a simple interpretation in terms of wrapped branes @xcite .
consider the local calabi - yau five fold @xmath76 which is the product of four line - bundles over a riemann surface @xmath9 : the four magnetic charges of the gauged supergravity are related to the chern - numbers of these line bundles and the bps condition is tantamount to @xmath77 .
the additional electric charges which we have found in this work correspond to spin of these m2-branes along the particular @xmath78 isometries of @xmath76 .
we have solved the horizon equations for bps black hole in @xmath0 gauged supergravity with @xmath1-vector multiplets by employing a generalization of the types of special geometry techniques that have proved useful in the past for the study of bps black holes in ungauged supergravity .
the level of detail of our solution is essentially equivalent to the state of the art for ungauged supergravity but it is curious that our set of quadratic equations are holomorphic whereas the similar equations in the ungauged case are real .
this holomorphicity guarantees a non - vanishing solution space but one must still check regularity .
it would seem that a reasonable strategy to understand black holes in gauged supergravity would be to continue to utilize the many tools which have been put to good use in the ungauged theory , much as we have done in this paper .
the study of higher derivative corrections in the gauged case presents a formidable challenge . by the general principles of holography , it is desirable to embed these black holes into a uv complete theory of gravity such as string or m - theory , but currently the only known embedding is for the particular stu - model discussed in section [ sec : mtheory ] . while it would certainly be a welcome development to derive quite general models of gauged supergravity with vector multiplets from m - theory , in lieu of that
, it is natural to include hypermultiplets where there is a richer set of known embeddings ( one can find numerous explicit examples in @xcite ) . studying m - theory enbeddings might help understand the physical meaning behind the combinations @xmath36 .
these are a combination of charges , which specify the vacuum selection sector and gauging parameters which appear in the action and define the theory .
the pairing does not appear to be so natural , it would be nice to have a physical understanding of why the bps equations have a neat form when this combination are used . in this work we have largely focussed on the horizon geometries but it would be quite helpful to have a general proof that such horizons can be uv completed to an @xmath0 , @xmath3 solution . in the case of the ck black holes where an analytic solution is at hand , one can easily check that the range of magnetic charges for which a regular black hole exists is identical to the range of charges for which a horizon geometry exists .
one can thus conclude that the ck solution space exhausts the space of magnetic black holes in that particular gauged supergravity theory .
it would be a significant development to have analytic solutions for more general dyonic black holes much like the general solution is known in ungauged supergravity .
we hope to return to these and other issues in the immediate future .
* acknowledgements : * it is a pleasure to thank mohab abou - zeid , davide cassani , atish dabholkar , sheer el - showk , boris pioline , martin ro@xmath79ek and brookie phipps williams for useful conversations and communcation .
i would also like to acknowledge the collaboration with michela petrini and alberto zaffaroni on related topics .
in this appendix we collect various formula for the special geometry of @xmath50 the vector - multiplet scalar manifold , essentially to establish a consistent notation . all this material
is well known but there are numerous different conventions in the literature .
we have the cubic prepotential f = d_ijk , [ cubicprepot ] where @xmath80 and will often work with special co - ordinates x^&= & ( 1,z^i ) , + z^i&= & x^i + i y^i , where @xmath81 .
we use the notation d_z=
d_ijk z^i z^j z^k , d_z , i = d_ijk z^j z^k , d_z , ij = d_ijk z^k and similarly for @xmath82 etc .
the metric on @xmath50 can be explicitly given in terms of @xmath83 : g_ij&= & + & = & -+ + g^ij&= & -d_y(d_y^-1)^ij+2 y^i y^j where @xmath84 is the khler potential and e^-k=-8 d_y .
the complex scalar fields @xmath40 of the vector multiplets are co - ordinates on @xmath50 but special geometry requires consideration of certain sections of an @xmath4 bundle over @xmath50 .
these sections are denoted @xmath85 however the more natural objects are the rescaled sections : = l^ + m_=e^k/2 x^ + f_which satisfy m_&= & _ l^ + _ & = & _ + i _ .
one should note that @xmath86 and @xmath87 are the gauge kinetic and topological terms in the @xmath0 lagrangian .
an explicit expression for @xmath88 is _ = _
+ 2i where f_= _ _ f .
we also use the covariant derivative of the sections which are defined to be u_i d_i = _ i + ( _ i k ) . the components of @xmath23 are u_i = f^i _ + h_i and satisfy h_i = _ f_i^. of particular importance is the symplectic inner product @xmath89 a , b= b^a_- a^b _ under which we have , & = & -i , u_i,_= ig_i .
this allows us to expand a symplectic section in terms of @xmath15 and their conjugates .
for example the background charges are given by & = & i - i + i ^ _ - i ^i u_i so that = , , _ i=,u_i .
using we can compute the components of @xmath22 : = l^ + m _
= e^k/2 x^ + f _ where \
{ x^0&=&1 + x^i & = & z^i + f_0&=&-d_z + f_i & = & 3 d_z , i .
the components of @xmath23 are ( u_i)^0 & = & d_i l^0 = -12 i e^k d_y , ie^k/2 + ( u_i)^j & = & d_i l^j= e^k/2 ( ^j_i -12i e^k d_y , i z^j ) + ( u_i)_0 & = & d_i m_0= e^k/2 - 3d_z , i+12i e^k d_y , id_z + ( u_i)_j & = & d_i m_j = e^k/26d_z , ij-36i e^k d_y , id_z , j further objects we use in the main computation are & = & 1 & - + 0 & 1 & 0 + 0 & ^-1 1 & 0 + -&1 = a & b + c & d where a&= & + ^-1 + d&= & ^-1 + b&= & c^t= -^-1 combined with the symplectic form = 0 & -11 + 11 & 0 . using these explicit expressions one can observe the property = -i , u_i = iu_i .
m. m. caldarelli and d. klemm , `` supersymmetry of anti - de sitter black holes , '' _
_ * b545 * ( 1999 ) 434460 , http://arxiv.org/abs/hep-th/9808097[hep-th/9808097 ] .
n. halmagyi , m. petrini , and a. zaffaroni , `` bps black holes in ads4 from m - theory , '' http://arxiv.org/abs/1305.0730[1305.0730 ] .
s. bellucci , s. ferrara , a. marrani , and a. yeranyan , `` d=4 black hole attractors in n=2 supergravity with fayet - iliopoulos terms , '' _
_ * d77 * ( 2008 ) 085027 , http://arxiv.org/abs/0802.0141[0802.0141 ] .
t. kimura , `` non - supersymmetric extremal rn - ads black holes in n=2 gauged supergravity , '' _ jhep _ * 1009 * ( 2010 ) 061 , http://arxiv.org/abs/1005.4607[1005.4607 ] .
g. dallagata and a. gnecchi , `` flow equations and attractors for black holes in n = 2 u(1 ) gauged supergravity , '' _ jhep _ * 1103 * ( 2011 ) 037 , http://arxiv.org/abs/1012.3756[1012.3756 ]
. k. hristov and s. vandoren , `` static supersymmetric black holes in @xmath3 with spherical symmetry , '' _ jhep _ * 1104 * ( 2011 ) 047 , http://arxiv.org/abs/1012.4314[1012.4314 ] .
d. klemm and o. vaughan , `` nonextremal black holes in gauged supergravity and the real formulation of special geometry , '' _ jhep _ * 1301 * ( 2013 ) 053 , http://arxiv.org/abs/1207.2679[1207.2679 ] .
k. hristov , c. toldo , and s. vandoren , `` phase transitions of magnetic ads4 black holes with scalar hair , '' http://arxiv.org/abs/1304.5187[1304.5187 ] .
a. gnecchi and c. toldo , `` on the non - bps first order flow in n=2 u(1)-gauged supergravity , '' _ jhep _ * 1303 * ( 2013 ) 088 , http://arxiv.org/abs/1211.1966[1211.1966 ] .
s. barisch , g. lopes cardoso , m. haack , s. nampuri , and n. a. obers , `` nernst branes in gauged supergravity , '' _ jhep _ * 1111 * ( 2011 ) 090 , http://arxiv.org/abs/1108.0296[1108.0296 ] .
a. donos , j. p. gauntlett , and c. pantelidou , `` magnetic and electric ads solutions in string- and m - theory , '' _ class.quant.grav .
_ * 29 * ( 2012 ) 194006 , http://arxiv.org/abs/1112.4195[1112.4195 ] .
a. donos , j. p. gauntlett , and c. pantelidou , `` semi - local quantum criticality in string / m - theory , '' _ jhep _ * 1303 * ( 2013 ) 103 , http://arxiv.org/abs/1212.1462[1212.1462 ]
. j. m. maldacena and c. nunez , `` supergravity description of field theories on curved manifolds and a no go theorem , '' _ int .
* a16 * ( 2001 ) 822855 , http://arxiv.org/abs/hep-th/0007018[hep-th/0007018 ] .
j. p. gauntlett , n. kim , s. pakis , and d. waldram , `` membranes wrapped on holomorphic curves , '' _ phys.rev . _
* d65 * ( 2002 ) 026003 , http://arxiv.org/abs/hep-th/0105250[hep-th/0105250 ] .
s. ferrara , r. kallosh , and a. strominger , `` n=2 extremal black holes , '' _ phys .
* d52 * ( 1995 ) 54125416 , http://arxiv.org/abs/hep-th/9508072[hep-th/9508072 ] .
k. behrndt , d. lust , and w. a. sabra , `` stationary solutions of n=2 supergravity , '' _ nucl.phys . _
* b510 * ( 1998 ) 264288 , http://arxiv.org/abs/hep-th/9705169[hep-th/9705169 ] .
b. de wit , h. samtleben , and m. trigiante , `` magnetic charges in local field theory , '' _ jhep _ * 09 * ( 2005 ) 016 , http://arxiv.org/abs/hep-th/0507289[hep-th/0507289 ] .
b. de wit and m. van zalk , `` electric and magnetic charges in n=2 conformal supergravity theories , '' _ jhep _ * 1110 * ( 2011 ) 050 , http://arxiv.org/abs/1107.3305[1107.3305 ] .
l. andrianopoli _ et al .
_ , `` general matter coupled n=2 supergravity , '' _ nucl .
* b476 * ( 1996 ) 397417 , http://arxiv.org/abs/hep-th/9603004[hep-th/9603004 ] .
m. shmakova , `` calabi - yau black holes , '' _ phys.rev . _ * d56 * ( 1997 ) 540544 , http://arxiv.org/abs/hep-th/9612076[hep-th/9612076 ] .
s. ferrara and r. kallosh , `` universality of supersymmetric attractors , '' _ phys .
rev . _ * d54 * ( 1996 ) 15251534 , http://arxiv.org/abs/hep-th/9603090[hep-th/9603090 ] .
s. ferrara and m. gunaydin , `` orbits of exceptional groups , duality and bps states in string theory , '' _
_ * a13 * ( 1998 ) 20752088 , http://arxiv.org/abs/hep-th/9708025[hep-th/9708025 ] .
b. pioline , `` lectures on on black holes , topological strings and quantum attractors , '' _ class .
_ * 23 * ( 2006 ) s981 , http://arxiv.org/abs/hep-th/0607227[hep-th/0607227 ] .
m. gunaydin , g. sierra , and p. k. townsend , `` the geometry of n=2 maxwell - einstein supergravity and jordan algebras , '' _ nucl .
_ * b242 * ( 1984 ) 244 . b. de wit , f. vanderseypen , and a. van proeyen , `` symmetry structure of special geometries , '' _ nucl .
_ * b400 * ( 1993 ) 463524 , http://arxiv.org/abs/hep-th/9210068[hep-th/9210068 ] .
n. halmagyi , `` work in progress , '' .
m. j. duff and j. t. liu , `` anti - de sitter black holes in gauged n = 8 supergravity , '' _ nucl .
_ * b554 * ( 1999 ) 237253 , http://arxiv.org/abs/hep-th/9901149[hep-th/9901149 ] .
m. cvetic , m. duff , p. hoxha , j. t. liu , h. lu , _ et al .
_ , `` embedding ads black holes in ten - dimensions and eleven - dimensions , ''
_ nucl.phys .
_ * b558 * ( 1999 ) 96126 , http://arxiv.org/abs/hep-th/9903214[hep-th/9903214 ] .
j. p. gauntlett , s. kim , o. varela , and d. waldram , `` consistent supersymmetric kaluza klein truncations with massive modes , '' _ jhep _ * 04 * ( 2009 ) 102 , http://arxiv.org/abs/arxiv:0901.0676[arxiv:0901.0676 ] .
a. donos , j. p. gauntlett , n. kim , and o. varela , `` wrapped m5-branes , consistent truncations and ads / cmt , '' _ jhep _ * 1012 * ( 2010 ) 003 , http://arxiv.org/abs/1009.3805[1009.3805 ] . | we study static bps black hole horizons in four dimensional @xmath0 gauged supergravity coupled to @xmath1-vector multiplets and with an arbitrary cubic prepotential .
we work in a symplectically covariant formalism which allows for both electric and magnetic gauging parameters as well as dyonic background charges and obtain the general solution to the bps equations for horizons of the form @xmath2 . in particular
this means we solve for the scalar fields as well as the metric of these black holes as a function of the gauging parameters and background charges . when the special khler manifold is a symmetric space , our solution is completely explicit and the entropy is related to the familiar quartic invariant . for more general models
our solution is implicit up to a set of holomorphic quadratic equations .
for particular models which have known embeddings in m - theory , we derive new horizon geometries with dyonic charges and numerically construct black hole solutions . these correspond to m2-branes wrapped on a riemann surface in a local calabi - yau five - fold with internal spin . nick halmagyi + 0.5 cm laboratoire de physique thorique et hautes energies , + universit pierre et marie curie , cnrs umr 7589 , + f-75252 paris cedex 05 , france + 0.5 cm [email protected] + _ dedicated to the loving memory of sue halmagyi _ |
for 32 edge - on spiral galaxies , spanning hubble type and mass , we fit the edge - on infinite disk model by @xcite on the irac mocaics : @xmath0 .
we applied stellar dominated 3.6 and 4.5 @xmath1 m channels and the pah emission at 8.0 @xmath1 m , with the stellar contribution subtracted ( using the prescription from * ? ? ? * ) .
the data is from our dedicated go program ( go 20268 ) and the _ spitzer _ archive .
our program aims to populate the lower range of disk masses . from the fit
we obtain the scale - length ( @xmath2 ) , the scale - height ( @xmath3 ) , and the face - on central surface brightness ( @xmath4 ) .
the edge - on disk model fits the stellar channels very well but the pah channel less so because of star - formation structures , such as hii regions , in the pah maps .
the disk s total luminosity is inferred from the fitted model : @xmath5 , with @xmath2 the scale - length and @xmath4 the face - on central surface brightness ( see * ? ? ?
combined with the rotational velocities from hyperleda , we can construct the tully - fisher relation for our galaxy disks and a plot of disk oblateness as a function of rotational velocity ( and hence dynamical mass ) .
figure [ f : tf ] shows the inferred tully - fischer relation for these disks in one of the stellar channels , 4.5 @xmath1 m .
notably , the slope ( @xmath6 ) is 3.5 , similar to what @xcite found but contrary for the increasing trend of slope with redder filters .
the effects of age and metallicity of the stellar population become independent and opposite effects on the color - m / l relation in nir ( see * ? ? ?
* figure 2d ) .
hence , the shallower slope in the irac stellar channels , could well be this metallicity effect starting to dominate .
the tully - fisher relation for disks in our sample .
notably , the slope is very similar to the one found by @xcite for the sings sample .
open circles are the poorer disk fits ( @xmath7).,scaledwidth=80.0% ]
figure [ f : hz ] shows the relation between the disk oblateness the ratio of scale - length and -height , @xmath8 relative to the oblateness in 4.5 @xmath1 m .
the 3.6 and 4.5 @xmath1 m channels both trace the stellar population and their oblateness measurements agree well .
the pah mosaic ( 8 @xmath1 m ) , shows flattening compared to stellar disk : @xmath9 .
the pah disk s flattening is substantial in the case of the more massive disks ( @xmath10 ) .
this appears to corroborate the reasoning of @xcite that massive disks are vertically unstable and hence form the characteristic dust lanes . the oblateness ( @xmath8 ) of the 3.6 @xmath1 m fit ( open triangles ) and 8.0 @xmath1 m pah map ( filled squares ) in terms of the 4.5 @xmath1 m . a similar oblateness as the 4.5 @xmath1 m .
the oblateness of the stellar filters , 3.6 and 4.5 @xmath1 m is very similar .
the pah disk is flatter than the stellar , especially in massive disks ( @xmath10 , vertical dotted line).,scaledwidth=80.0% ]
hence , from our disk models of irac data , we conclude : * the tully - fisher relation has a shallower slope ( @xmath11 ) than naively expected from trends with filter ( figure [ f : tf ] ) , which may be a metallicity effect of the stellar population s
m / l . * the pah disk is flatter than the stellar one , especially in more massive spirals ( figure [ f : hz ] ) .
the effect is not as pronounced as found by @xcite and @xcite based on dust lanes .
the unique perspective of _ spitzer _ on edge - on spirals allows us to quantify the vertical structure in spiral disks .
the pah maps are the product of both the ism density and star - formation .
the stellar channels are effectively unaffected by extinction and can therefore serve as a reference to quantify dust lanes and smaller dust extinction structures .
the phenomenon of truncation the sudden change in exponential behavior of disks can also be characterized for all irac bands . | edge - on spiral galaxies offer a unique perspective on disks .
one can accurately determine the height distribution of stars and ism and the line - of - sight integration allows for the study of faint structures .
the spitzer irac camera is an ideal instrument to study both the ism and stellar structure in nearby galaxies ; two of its channels trace the old stellar disk with little extinction and the 8 micron channel is dominated by the smallest dust grains ( polycyclic aromatic hydrocarbons , pahs ) . @xcite probed the link between the appearance of dust lanes and the disk stability . in a sample of bulge - less disks
they show how in massive disks the ism collapses into the characteristic thin dust lane .
less massive disks are gravitationally stable and their dust morphology is fractured .
the transition occurs at 120 km / s for bulgeless disks . here
we report on our results of our spitzer / irac survey of nearby edge - on spirals and its first results on the nir tully - fischer relation , and ism disk stability . |
the interaction of oxygen with silver surfaces has been intensively studied because of the key role of silver in many important industrial oxidation reactions such as , for example , partial oxidation of methanol to formaldehyde or ethylene epoxidation.@xcite these reactions occur very efficiently when ag powders are employed as a catalyst .
since the particles which compose these powders are affected by many surface defects of different nature and with different concentrations , their chemistry can be extremely different from that of perfect low - miller - index surfaces .
indeed the presence of various inequivalent adsorption sites suggests that atomic and molecular oxygen can adsorb forming species of different chemical nature .
the microscopic characterization of such adsorbed species is a fundamental issue to understand the catalytic activity of ag powders .
a large amount of experimental and theoretical results clearly indicates that steps , kinks , and other surface imperfections play a crucial role in surface chemistry .
for example , the study of many systems such as n and o adsorbed on ru(0001),@xcite o on pt(111),@xcite or c@xmath0h@xmath1 on ag(100)@xcite has shown that monoatomic steps can bind adsorbates more strongly than terrace sites .
moreover it has been shown that this kind of defects can completely determine the kinetics of dissociation in systems such as , for example , n@xmath0 and no on ru(0001),@xcite o@xmath0 on pt(111),@xcite n@xmath0 on fe(110 ) and fe / ru(0001).@xcite the role of steps in the dynamics of o@xmath0/ag interaction has been recently investigated by rocca and coworkers@xcite who have studied the ag(410 ) and ag(210 ) surfaces .
these surfaces are vicinal of ag(100 ) and are characterized by open ( 110 ) steps and three - atom- and one - atom - row wide ( 100 ) terraces , respectively . using a supersonic molecular beam to dose o@xmath0 onto ag surfaces at selected angles of incidence and different kinetic energies , and characterizing the adsorbed species by vibrational spectroscopy ( hreel ) , these authors have investigated the various oxygen species which form on these surfaces .
it was concluded that o@xmath0 dissociation occurs mainly at the steps and that atomic oxygen adsorbs at different surface sites . in particular , hreel spectra show two peaks in the frequency region of the o / ag stretch modes , at about 32 and 40 mev .
these frequencies are quite similar to those observed in o / ag(100 ) ( 33 mev)@xcite and on the added - row reconstructed o / ag(110 ) surface ( 40 mev).@xcite it was thus proposed to assign the low energy peak to adatoms occupying ( 100 ) terrace sites and the one at higher energy to atomic oxygen adsorbed at the step edges forming o ag o chains similar to the added rows found in reconstructed o / ag(110 ) surfaces.@xcite in o / ag(210 ) an additional peak around 56 mev was found , and it was proposed that it is due to the vibration of an oxygen atom occupying a subsurface site
. motivated by these results we have performed _ ab initio _ calculations aimed at understanding the role of the steps in the adsorption of atomic oxygen on the ag(410 ) and ag(210 ) surfaces . in the present work we limit our investigation to on - surface atomic oxygen , subsurface and molecular adsorption will be addressed in separate papers .
our results show that on these vicinal surfaces the adsorption configuration in which the adatoms decorate the steps in a ( 1@xmath21 ) geometry is particularly stable .
the oxygen adsorption on the two surfaces presents however significant differences related to the availability of adsorption sites on terraces . indeed , on ag(410 ) ,
terrace and step edge sites have comparable chemisorption energies and , in particular , the hollow sites on terraces are slightly more favored than the step edge sites when the adatoms are far apart from each other . at higher coverage , instead , adatoms slightly prefer to decorate the steps . on ag(210 ) , the absence of favorable terrace sites makes the step decoration much more stable than any other adsorption configurations .
the formation of these o ag o chains at step edges will be shown to significantly affect both the geometrical structure and the electronic properties of the surface .
as we have already mentioned , the occurrence of adsorption on both the terraces and steps of vicinal ag(100 ) surfaces was also proposed based on an analysis of the hreel spectra , recently measured for these systems.@xcite our calculation of the vibrational properties of these systems performed within density - functional perturbation theory
provides a further support to this analysis .
the paper is organized as follows . in section [ sec : theo ] we describe the theoretical approach and computational details of our work .
section [ sec : res ] contains our results for the energetics , the geometrical and electronic structures , and for the vibrational properties of the systems we have investigated .
the last section is devoted to our conclusions .
all the calculations have been performed within density functional theory ( dft ) using the generalized gradient approximation ( gga ) with the perdew - burke - ernzerhof ( pbe ) exchange - correlation functional.@xcite vibrational properties have been calculated using density - functional perturbation theory ( dfpt).@xcite we have used the pseudopotential method with ultra - soft pseudopotentials @xcite and plane - wave basis sets up to a kinetic - energy cutoff of 27 ry ( 216 ry for the charge - density cutoff ) .
details about the ag and o pseudopotentials are reported in ref . .
brillouin zone ( bz ) integrations have been performed with the gaussian - spreading special - point technique@xcite with a smearing parameter of 0.03 ry .
all the calculations have been performed using the pwscf and phonon packages,@xcite while molecular graphics has been generated with the xcrysden package.@xcite surfaces are modeled with periodic super - cells . for the ag(410 ) and ag(210 ) surfaces we have used slabs of 20 ( 410 ) layers and 14 ( 210 ) layers , respectively . for the ag(100 )
surface and the ( 2@xmath21 ) added - row reconstructed ag(110 ) surface [ ag(110)p(2@xmath21)o ] we have used slabs of 7 layers .
adjacent slabs are separated by a vacuum region of at least 16 a.u . and
o atoms are adsorbed on both sides of the slabs .
all the structures have been fully relaxed until the hellmann - feynman forces were smaller than @xmath3 ry / a.u .
atomic oxygen adsorption on ag(410 ) and ag(210 ) has been modeled by ( 1@xmath21 ) and ( 2@xmath21 ) surface super - cells . for the ag(410 ) surface the bz integrations
have been performed using a ( 8@xmath24@xmath21 ) and a ( 4@xmath24@xmath21 ) uniform shifted @xmath4-mesh@xcite for the ( 1@xmath21 ) and ( 2@xmath21 ) structures , respectively , while for ag(210 ) ( @xmath5 ) and ( @xmath6 ) meshes have been used for ( 1@xmath21 ) and ( 2@xmath21 ) structures , respectively . for the c(2@xmath22 )
cell on ag(100 ) we have used a ( @xmath7 ) mesh , while for the ag(110)p(2@xmath21)o surface we have used a ( @xmath8 ) mesh .
spin polarization effects have been neglected after checking that the interaction between o and ag substrate results in a negligible spin moment on the o atom .
the chemisorption energies , @xmath9 , are referred to the clean ag(@xmath1010 ) , @xmath11 or @xmath12 , surface and the isolated oxygen molecule : @xmath13 where the total energy of the adsorbate
substrate system , of the clean surface , of the isolated o@xmath0 molecule , and the number of adsorbed oxygen atoms are represented by @xmath14 , @xmath15 , @xmath16 and @xmath17 , respectively . with this definition ,
stable adsorbates have negative chemisorption energies . considering the errors in the structural relaxation and the energy convergence with respect to the plane wave cutoff energy and the k - point density
, we estimate an overall numerical accuracy for the chemisorption energies of about 30 mev .
the pbe gga functional used in the present work slightly overestimates the binding energy of the o@xmath0 molecule : 5.50 ev , as calculated using spin - polarized gga with a 15 - wide cubic box , while the experimental value is 5.23 ev.@xcite molecular overbinding is a common drawback of current implementations of dft .
this fact , of course , does not affect the conclusions of the present work which mainly concerns the _ relative _ stability of different adsorption configurations .
vibrational frequencies are calculated by diagonalizing the @xmath18 dynamical matrices of large slabs containing more that 50 atomic layers .
the common procedure for calculating the dynamical matrices of such large slabs is to patch the dynamical matrices obtained for much smaller slabs ( in fact , the same which were used to determine the structural properties , as described above ) with the interatomic force constants calculated for the bulk metal.@xcite the dynamical matrices of the ag bulk have been calculated on a 4@xmath24@xmath24 grid of * q*-points in the bz and the interatomic force constants have been obtained by a fourier transform.@xcite for the thin slabs , we have computed only the parts of the dynamical matrices which refer to the displacements of the adsorbates and of the uppermost ag layers . diagonalizing the dynamical matrix of the extended slab we have calculated the vibrational modes of the systems and we have identified the adsorbate - substrate modes from the displacements eigenvectors .
the estimated numerical accuracy of the vibrational frequencies so obtained is of about 1 mev .
in fig . [ fig : surf ] we display the four adsorption sites that we have considered in the present work : site a with o just above the step and coordinated with three ag surface atoms ; sites t@xmath19 and t@xmath0 with the adatom in the hollow sites on the ( 100 ) terrace ; site b where o adatom lies in the hollow site just below the step .
note that no terrace t@xmath19 and t@xmath0 sites are available on the ag(210 ) .
top and bridge sites are not considered here since we have found that they are much less stable than hollow sites , as on ag(100 ) surface.@xcite let us first consider the adsorption of atomic oxygen on the ag(410 ) surface . in table [ tab:1 ]
we present the chemisorption energies corresponding to different adsorption configurations ( see fig . [
fig : configs ] ) . at low coverage ( @xmath20 1/8 ml)@xcite where the adatoms stay far apart from each other , the hollow sites on terraces ( t@xmath19 and t@xmath0 ) and the step - edge site ( a ) are almost degenerate in energy , with terrace sites being only slightly more favored than site a , while site b is much less stable . in these configurations
the distances between the o atoms are quite large ( 8.32 ) and this suggests a negligible interaction between the adatoms , as confirmed by test calculations done at lower coverages in which the o@xmath21o distance is increased . at higher coverage ( @xmath20 1/4 ml )
we have examined various possible geometries which we indicate with the notation ` s@xmath19s@xmath0 ' , where the s s stand for two near a , b , t@xmath19 and t@xmath0 sites being occupied by two o adatoms in a ( 2@xmath21 ) super - cell .
the s@xmath19s@xmath19 arrangements correspond to a ( 1@xmath21 ) periodicity where o adatoms form rows parallel to the steps .
those configurations in which the separation between the adatoms is larger than 5 are not considered since we expect that their adsorption energies should be very similar to those at low coverage .
indeed , as we will show , already for o o distances as small as 4 the interaction between the adsorbates has small effects on the chemisorption energies .
our results show that the s@xmath19s@xmath19 configurations with both the adatoms on the terrace or above the step and the a
t@xmath0 arrangement are the most stable adsorption geometries . as in the case of low coverage
the differences in chemisorption energy between these sites are quite small , but in this case we find that oxygen adatoms slightly prefer to decorate the step edges .
one expects that occupying adjacent sites should be unfavorable because of the electrostatic repulsion between negatively charged oxygen adatoms .
indeed , for example , feibelman and co - workers@xcite have shown that o adatoms on pt(211 ) vicinal surface decorate the steps in a ( 2@xmath21 ) adsorption geometry while the ( 1@xmath21 ) structure where the distance between adjacent oxygens is about 2.8 is significantly less stable .
our results indicate that this effect on ag(410 ) surface is important only when nearest neighbor sites are occupied , i.e. when the distance between the adatoms is less than 4 ( as in a
t@xmath19 , t@xmath19t@xmath0 , t@xmath0b and a b , configurations ) , while it is weak and uneffective in the other geometries .
compare for example the a t@xmath0 and a t@xmath19 configurations which consist of one adatom sitting above the step while the other sits at a terrace site . in the a
t@xmath0 arrangement the distance between the o atoms is 4.23 and their chemisorption energy ( @xmath210.80 ev ) is quite close to the values found for oxygen adsorption at sites a and t@xmath0 at low coverage ( @xmath210.75 ev and @xmath210.83 ev ) , indicating a small adsorbate - adsorbate interaction . in the a
t@xmath19 configuration , instead , the o
o distance is only 3.26 and the adsorption energy of the adatoms ( @xmath210.59 ev ) is strongly reduced with respect to when oxygen is adsorbed at sites a and t@xmath19 at low coverage ( @xmath210.75 ev and @xmath210.80 ev ) .
moreover , it is interesting to observe that in the a
t@xmath19 geometry the electrostatic repulsion strongly affects also the locations of the adatoms .
indeed we find that the adsorbates are slightly displaced from the hollow positions and that their distance , 3.26 , is considerably larger than the separation between two nearest hollow sites of the clean surface ( about 2.9 ) .
this is a clear indication that the two adatoms repel each other .
similar calculations have been performed also for the ag(210 ) surface .
the two surfaces differ in the width of the terrace ( see fig .
[ fig : surf ] ) .
no terrace hollow sites are available on the ag(210 ) surface . in table
[ tab:1b ] we report the chemisorption energies for o adsorbed in various configurations on this surface .
these results display the same trend as obtained for ag(410 ) , although the chemisorption energies are somewhat smaller here . due to the lack of stable adsorption sites on the terraces , the formation of o ag o rows at the step edges ( i.e. the a a configuration )
is by far energetically favored on this surface .
the second most stable configuration at higher coverage ( @xmath22 ml ) is therefore a b , which is 0.24 ev less bound than a a .
in contrast , on ag(410 ) several other configurations that involve terrace sites are more stable than a b and much closer in energy to the a a configuration . comparing the chemisorption energies of various configurations , irrespective of the o coverage ,
we find that the difference between the most favored a a configuration and the second most stable one is 0.03 and 0.12 ev for ag(410 ) and ag(210 ) surfaces , respectively .
one of the most peculiar results of our calculations is the increase of the stability of oxygen adatoms when they decorate the step edge in a ( 1@xmath21 ) geometry .
we believe that the mechanism responsible for this behavior could be related to the arrangement of o adatoms to form o ag o chains at the upper side of the ( 110 ) steps .
indeed , the a
a adsorption configuration is similar to that occurring in the reconstruction of the ag(110 ) surface upon oxygen adsorption , in which o atoms align in added - rows along the [ 001 ] direction , thus forming stable o
o structures@xcite ( the ag(110)p(2@xmath21)o surface is shown in fig . [
fig : added ] ) .
as we have pointed out in the introduction , this similarity between the two structures was first suggested from a comparison of the vibrational properties of the two surfaces.@xcite before analyzing in detail the structural and electronic features related to the formation of these chains , it is interesting to consider the effects of the substrate relaxation on the chemisorption energies .
for this reason we have repeated some calculations for o / ag(410 ) keeping the substrate fixed in the optimized clean surface geometry and allowing the adatoms to relax ( see values in parentheses in table [ tab:1 ] ) .
our results show that at low coverage the t@xmath0 site is more stable than a site even if substrate relaxation is not considered . on the other hand , in the ( 1@xmath21 ) periodicity o adatoms in a a and t@xmath0t@xmath0 configurations have the same chemisorption energy on a fixed substrate .
this result indicates that the enhanced stability of atomic oxygen on the step in the ( 1@xmath21 ) geometry is mainly determined by the energy gain due to surface relaxation allowed by the presence of the steps . in order to illustrate the characteristic features of the o ag o chain formation at the step edges ,
we discuss now the effects of oxygen adsorption on the geometry of the surface . to facilitate the discussion ,
we first define a few labels for silver atoms
ag@xmath23 , ag@xmath24 , and ag@xmath25as in fig .
[ fig : legend].@xcite in table [ tab:2 ] we report the optimized structural parameters describing the oxygen adsorption on ag(410 ) at @xmath20 1/8 ml and at @xmath20 1/4 ml for the s@xmath19s@xmath19 configurations .
first of all it is interesting to observe the dependence of the bond lengths between the adsorbate and the nearest metal atoms on the number of neighboring substrate atoms . as predicted by effective medium theory , @xcite the larger the number of nearest metal atoms the longer the bond length . indeed
the results in table [ tab:2 ] ( observe in particular @xmath26 and @xmath27 ) show that the three - fold coordinated step - edge is the site in which o is closest to the surrounding silver atoms ; o in the two four - fold coordinated terrace sites has similar structural features , while o when sitting in the site just below the step , close to four atoms of the terrace and to two atoms of the step is the furthest from neighboring metal atoms .
we now focus on the structural features of oxygen adsorption at sites a and t@xmath0 . in fig .
[ fig : str1 ] we compare the structure of the first three layers of the substrate for the a and a a geometries . at low coverage
the main substrate distortions due to the presence of the adatom are a shift of 0.18 of the nearest ag terrace atom towards the center of the terrace ( see @xmath28 in table [ tab:3 ] ) and an expansion of 0.16 of the distance between the two ag atoms on the step near oxygen ( see @xmath29 in table [ tab:3 ] ) .
the result is that the adatom sits in a site equidistant ( @xmath30 = 2.21 ) from the three ag surface atom surrounding it . in
the a a configuration the distortion has different features . in this case symmetry
does not allow the two ag step atoms near oxygen to be displaced parallel to the step , so that these atoms move only outward .
this is indicated by a marked expansion of the bond lengths between the ag step atom and the ag atoms on terrace ( by 0.25 ) and below the adatom ( by 0.16 ) . in this case
the oxygen is closer to the ag step atoms ( @xmath31 = 2.10 ) than to the nearest ag atom on terrace ( @xmath32 = 2.24 ) .
note also that the angle , @xmath33 , between the bonds of oxygen with the two ag atoms of the step which directly measures the alignment of the ag o
ag structures is larger at high than at low coverage ( 164@xmath34 versus 156@xmath34 ) .
thus , in the a a geometry oxygen and silver atoms on the step tend to align forming an almost straight o ag o chain which is more stable than other adsorption configurations . as we have already pointed out ,
the o ag o chains at the step edges have a similar structure as the added rows in the reconstructed o / ag(110 ) surface , so it is interesting to compare the geometries of the two systems . in table
[ tab:2 ] we report some predicted structural quantities for the ag(110)p(2@xmath21)o surface .
our results are in good agreement with experiments@xcite and previous dft calculations.@xcite as we can observe , the alignment of o and ag atoms in the added rows on ag(110 ) is more pronounced than in the a a configuration ( compare , for example , the angle @xmath33 in the two geometries ) , but the similarity between the two structures is clear .
[ fig : str2 ] shows the structures of t@xmath0 and t@xmath0t@xmath0 configurations . in these systems
the height of the adatom with respect to the terrace is larger than when the oxygen is chemisorbed on the step .
also in this case in the ( 1@xmath21 ) geometry the silver atoms in the row of oxygens move up towards adatoms , but here the alignment between o and ag atoms is less pronounced than when oxygen atoms are on the step . in table
[ tab:2b ] we report the optimized structural parameters describing the oxygen adsorption on ag(210 ) at @xmath20 1/4 ml and at @xmath20 1/2 ml for the s@xmath19s@xmath19 configurations , while fig .
[ fig : a - oag210 ] shows the side views of the corresponding optimized o / ag(210 ) structures .
these data and plots reveal the same trend as for ag(410 ) surface .
the most apparent feature is that the height of the o adatoms above the surface is much smaller for a and a a sites than for the b and b b sites .
note also the difference between the pattern of the substrate reconstruction for a ( low coverage ) and a a ( high coverage ) configurations .
like on the ag(410 ) surface , the adsorption of o adatoms induces an outward relaxation of step ag atoms for both the a and a a configurations . in the case of the a configuration
this relaxation is not large enough to compensate the inward relaxation of the clean substrate ( see the inset of top - left panel of fig . [
fig : a - oag210 ] ) , therefore ag atoms at the step appear to be pushed somewhat inward , like in the case of the clean ag(210 ) substrate .
on the contrary , this relaxation is quite significant in the a a configuration , resulting in a marked expansion of the bond length between the ag step and terrace atoms ( @xmath35 ) , and between the ag step atom and ag atom located below the o adatom ( @xmath36 ) .
note that the step ag atoms appear to be pushed outwards .
contrary to the a a geometry , in the a configuration the step ag atoms also relax laterally along the step - edge direction away from o adatom the distance between the two step ag atoms that are bound to the same o adatom being increased by 0.17 . in order to better characterize the different behavior of oxygen on the step in a and a a configurations we have analyzed the corresponding electronic structure . here
we present the analysis of the electronic structure of o / ag(410 ) only the conclusions for o / ag(210 ) being similar .
the interaction between oxygen and silver removes electronic charge from the silver atoms neighboring the o adatom .
the electron deficit regions around the silver atoms are shown in fig .
[ fig : charge ] ( blue isosurfaces ) .
it is interesting to observe that while at low coverage the deficit regions around the three ag atoms are very similar , thus indicating a similar donation of charge , in the a a configuration the flow of charge to oxygen adatoms is mainly due to silver atoms on the step .
this suggests that the o ag o chains that form in the ( 1@xmath21 ) geometry have the structure of electrostatically stable @xmath37 strings .
the analysis of the density of states reveals that the bond between oxygen and silver is not purely ionic . in fig .
[ fig : dos ] we display the density of states projected ( pdos ) onto the oxygen adatom and onto the nearest silver atoms for both the a and a a geometries .
the pdos on o atoms shows that the interaction of the o 2@xmath38 state with the 4@xmath39 metal band results in the formation of two regions of high density of states , which correspond to bonding and antibonding states .
the bonding states are located at the bottom of the 4@xmath39 ag band while the antibonding ones are mainly below the fermi level .
note that the o 2@xmath40 level lies below the bottom of the valence band ( at about @xmath2117 ev with respect to the fermi level ) and it is not shown in the figures .
the main difference between the pdos in the two geometries is that at high coverage additional features appear 67 ev below the fermi level .
the figure clearly shows that these states are due the hybridization of oxygen 2@xmath38 orbitals and 4@xmath39 orbitals of the silver atoms on the step ( ag@xmath41 ) . by inspection of the components of these orbitals we find that the mixing is mainly between o 2@xmath42 orbital parallel to the step - edge and ag 4@xmath43 orbital parallel to the terrace plane
this hybridization is very efficient because of the good alignment between o and ag atoms on the step .
the spatial distribution of these states is illustrated in fig .
[ fig : ildos ] , where we display the integrated local density of states ( ildos)@xcite in the energy window ( @xmath217.0 , @xmath216.0 ) ev comprising the two small peaks in the pdos .
the figure shows quite clearly the bond between oxygen and silver atoms on the step .
it is worth noting that in the other s@xmath19s@xmath19 configurations there are no states in this energy region even if we find an increase of the density of states around @xmath44 ev , especially in the t@xmath0t@xmath0 geometry .
the study of the electronic structure of the added - row reconstructed ag(110 ) surface confirms that the presence of bonding states 67 ev below the fermi level is related to the formation of aligned o ag o
ag structures .
indeed we find similar spectral features also in the dos of ag(110)p(2@xmath21)o and their spatial distribution closely resembles that shown in fig .
[ fig : ildos ] .
note that the results of experimental studies@xcite show a quite weak oxygen - induced feature in this energy range . in table [
tab : osc ] we present the frequencies of the adsorbate modes and the directions of the corresponding displacement eigenvectors ( the angles are defined in fig .
[ fig : ref - osc ] ) for different o / ag systems . before analyzing the vibrational modes of oxygen on ag(410 ) and ag(210 ) , it is useful to discuss the vibrational properties of the o / ag(100 ) system with oxygen adsorbed in hollow sites and of the added - row reconstructed ag(110 ) surface
. data for the ag(100)c(2@xmath22)o structure which had already been calculated in ref . in the local density approximation ( lda)have been recalculated within the present gga approach , giving quite similar results ( not unusually , gga frequencies result to be slightly red - shifted with respect to lda ) .
the @xmath45 mode perpendicular to the surface vibrates at 30 mev in agreement with the peak at 30 - 32 mev observed in the hreel spectra of the non - reconstructed o / ag(100 ) surface.@xcite the modes @xmath46 and @xmath47 at 50 mev are two in - plane degenerate modes . for the ag(110)p(2@xmath21)o structure
we find a vibrational mode , @xmath46 , perpendicular to the surface at 38 mev .
this value has to be compared with the feature around 41 mev observed in hreel spectra.@xcite besides this dipole - active mode two other adsorbate modes exist , @xmath45 at 28 mev and @xmath47 at 73 mev , in which the adatom vibrates parallel to the surface plane .
the polarization of the softer mode is perpendicular to the added row while the other one is parallel to it .
it is interesting to observe that the @xmath47 mode is even harder than the in - plane modes ( @xmath46 and @xmath47 ) of ag(100)c(2@xmath22)o because the oxygen atom sits closer to the nearest silver atoms in the added row reconstructed ag(110 ) surface ( @xmath48 = 0.10 ) than in the ag(100)c(2@xmath22)o structure ( @xmath48 = 0.71 )@xcite .
we now focus on the vibrational properties of the a - a adsorption geometry on ag(410 ) and ag(210 ) .
the mode at higher frequency ( @xmath47 at 68 mev and 71 mev for ag(410 ) and ag(210 ) respectively ) in which the adatom vibrates parallel to the ag o ag chain at the step is very similar to the @xmath47 mode of the ag(110)p(2@xmath21)o structure .
the softer modes are due to vibrations of the adsorbate in a plane perpendicular to the step edge ( these two modes for the ag(410 ) surface are shown in the upper panel of fig .
[ fig : modi ] ) .
the @xmath45 mode at 24 mev is nearly parallel to the ( 110 ) plane of the step while the @xmath46 mode at 37 mev is almost perpendicular to it .
these modes resemble very closely the @xmath45 and @xmath46 modes of the ag(110)p(2@xmath21)o surface .
the displacement pattern of these two modes suggests that they could both be dipole active .
in particular the mode @xmath46 could be responsible of the spectral feature around 40 mev observed in the hreel spectra of these surfaces . for the ag(210 ) surface
we have also investigated the a configuration to understand the effect of the ag o
ag chain formation at the step on the vibrational properties . in this case both the frequency and the displacement eigenvectors of the softer modes , @xmath45 and @xmath46 , change very little with respect to the a a arrangement .
the @xmath47 mode is polarized parallel to the step and is significantly softer than in the a a structure ( 46 versus 71 mev ) mainly because in this case the height of the adatom with respect to the step ag atoms is larger than in the a a geometry . since the hollow terrace sites on ag(410 )
compete in energy with the step edge sites , we have also investigated the vibrational properties of a very stable adsorption configuration which involves terrace sites , namely the t@xmath19t@xmath19 geometry . in the lower panel of fig .
[ fig : modi ] we show the @xmath45 and @xmath46 modes in the plane perpendicular to the step .
the @xmath45 mode at 29 mev is due to the vibration of o nearly perpendicular to the ( 100 ) terrace plane while the @xmath46 mode at 32 mev is due to o vibrating almost parallel to this plane .
the harder mode , @xmath47 , is at 58 mev and it corresponds to the oxygen vibration parallel to the step edge .
our calculations show clearly that the @xmath45 mode is very similar to the dipole active @xmath45 mode of the c(2@xmath22 ) o adsorption geometry on ag(100 ) .
it seems thus reasonable to expect that the spectral feature around 32 mev observed in hreel spectra on ag(410 ) could be due to this stretching mode .
finally we have calculated the vibrational modes for the b b adsorption geometry on ag(210 ) .
an adatom at site b is significantly less stable than at the step edge , but it is interesting to investigate its vibrational properties since this is the only terrace site available on ag(210 ) and the hreel spectra of this surface show , in addition to the 40 mev peak , other spectral features around 30 mev and 56 mev . as it is shown in table [
tab : osc ] , the higher frequency mode , @xmath47 at 37 mev , is parallel to the step edge , while the other two modes are in the plane perpendicular to it . in particular
the @xmath45 mode at 27 mev is nearly perpendicular the ( 110 ) step plane , while the other , @xmath46 at 34 mev , is almost parallel to it .
two important results emerge from this analysis .
the @xmath45 mode in the t@xmath19-t@xmath19 configuration of o / ag(410 ) at 29 mev lies very close to the @xmath45 mode of o / ag(100 ) ( 30 mev ) , in correspondence with the spectral feature around 32 mev observed in hreel spectra on ag(410).@xcite the @xmath46 modes of the a - a configuration in o / ag(210 ) and o / ag(410 ) ( both at 37 mev ) lie close to the @xmath46 frequency in the added - row o / ag(110 ) surface [ ag(110)p(2@xmath21)o ] at 38 mev , in correspondence with the spectral feature around 40 mev observed in the hreel spectra of o / ag(410 ) and o / ag(210 ) surfaces.@xcite we believe that this correspondence between infrared - active vibrational modes , calculated for specific configurations , and the features found in the experimental hreel spectra supports the proposed assignment of those spectral features to atomic species adsorbed on terraces and at step edges .
problems still remain in the interpretation of the hreel spectra of o / ag(210 ) . on one hand ,
the feature experimentally found at 32 mev could be due to the vibration of oxygen occupying the only available terrace sites ( site b ) , but this assignment remains uncertain since the adsorption on these sites is energetically unfavored . on the other hand ,
no vibration is predicted near 56 mev , where a feature in the experimental spectrum is also found .
this feature could be the signature of other o species , possibly adsorbed subsurface , which we have not investigated in the present work .
for sure , these issues call for further theoretical and , possibly , experimental work .
in this paper we have investigated the adsorption of oxygen on ag(410 ) and ag(210 ) surfaces .
our results show that the step edge decoration is particularly stable on both surfaces .
while on ag(410 ) adsorption on terrace and at the step edge compete in energy and no clear preference is predictable , on ag(210 ) the step decoration is significantly favored with respect to other adsorption configurations .
the formation of this stable o ag
o rows at the step edges strongly affects both the structural and the electronic properties of the surface , and leaves a clear fingerprint in the hreel spectrum of these systems at 40 mev .
the situation is less clear for the hreel peak at 32 mev . on ag(410 )
it is reasonable to assign this spectral feature to adatoms occupying terrace sites . on ag(210 ) instead this o ag stretch mode could be due to o atoms adsorbed on terraces at the step foot , which however are predicted to be metastable adsorption sites . among the adsorption configurations that we have investigated ,
no candidate was found to support the vibrational mode experimentally found at 56 mev .
this suggests that other o species , possibly adsorbed subsurface , would be responsible for this spectral feature .
this work has been supported by infm ( _ iniziativa trasversale calcolo parallelo , sezioni f e g _ , and _ pais chemde _ ) and by the italian miur through prin .
all numerical calculations were performed at the cineca national supercomputing center in bologna ( italy ) .
99 r. a. van santen and h. p. c. e. kuipers , adv . catal . * 35 * ( 1987 ) 265
. b. hammer , surf .
sci . * 459 * , 323 ( 2000 ) .
p. j. feibelman , s. esch , and t. michely , phys .
lett . * 77 * , 2257 ( 1996 ) .
a. kokalj , a. dal corso , s. de gironcoli , and s. baroni , j. phys .
b * 106 * , 9839 ( 2002 ) .
s. dahl , a. logadottir , r. c. egeberg , j. h. larsen , i. chorkendorff , e. trnqvist , and j. k. nrskov , phys .
lett . * 83 * , 1814 ( 1999 ) .
t. zambelli , j. wintterlin , j. trost , and g , ertl , science * 273 * , 1688 ( 1996 ) .
p. gambardella , .
ljivananin , b. hammer , m. blanc , k. kuhnke , and k. kern , phys . rev
. lett . * 87 * , 056103 ( 2001 ) .
r. c. egeberg , s. dahl , a. logadottir , j. h. larsen , j. k. nrskov , and i. chorkendorff , surf .
sci . * 491 * , 183 ( 2001 ) .
l. savio , l. vattuone , and m. rocca , phys .
* 87 * , 276101 ( 2001 ) .
l. savio , l. vattuone , and m. rocca , j. phys .
matter * 14 * , 6065 ( 2002 ) .
l. vattuone , l. savio , and m. rocca , phys .
lett . * 90 * , 228302 ( 2003 ) .
f. bautier de mongeot , a. cupolillo , u. valbusa , and m. rocca , chem .
lett . * 302 * , 302 ( 1999 ) .
l. vattuone , u. valbusa , and m. rocca , surf .
lett . * 317 * , l1120 ( 1994 ) .
r. courths , s. hfner , p. kemkes , and g. wiesen , surf .
sci . * 376 * , 43 ( 1997 ) .
t. schimizu and m. tsukada , surf .
lett . * 295 * , l1017 ( 1993 ) .
j. p. perdew , k. burke , and m. ernzerhof , phys .
* 77 * , 3865 ( 1996 ) .
s. baroni , p. giannozzi , and a. testa , phys .
lett . * 58 * , 1861 ( 1987 ) .
p. giannozzi , s. de gironcoli , p. pavone , and s. baroni , phys .
b * 43 * , 7231 ( 1991 )
. a. dal corso , a. pasquarello , and a. baldereschi , phys .
b * 56 * , r11369 ( 1997 ) ; a. dal corso , phys . rev .
b * 64 * , 235118 ( 2001 ) .
s. baroni , s. de gironcoli , a. dal corso , and p. giannozzi , rev .
phys . * 73 * , 515 ( 2001 ) .
d. vanderbilt , phys .
b * 41 * , r7892 ( 1990 ) . g. cipriani , d. loffreda , a. dal corso , s. de gironcoli , and s. baroni , surf .
sci . * 501 * , 182 ( 2002 ) .
h. j. monkhorst and j. d. pack , phys .
b * 13 * , 5188 ( 1976 ) .
m. methfessel and a. t. paxton , phys .
b * 40 * , 3616 ( 1989 ) .
, , , and , pwscf and phonon : plane - wave pseudo - potential codes , ( ) . , * * , ( ) . , and , ( ) .
the bz integrations have been performed using the uniformly spaced and shifted @xmath4-meshes .
such a mesh is generated by the following recipe : ( i ) define the sequence of numbers @xmath49 ; and ( ii ) generate the following @xmath50 distinct @xmath51 points : @xmath52 , where @xmath53 , @xmath54 , and @xmath55 are reciprocal lattice vectors .
it can be shown , taking into account the tranlational symmetry , that in case of even mesh ( i.e. even set of @xmath56 values ) such mesh coincides with the monkhorst - pack mesh .
@xcite our value for the bond energy of o@xmath0 is smaller than other dft - pbe results . for example , hammer _
et al._@xcite obtained 5.81 ev and 5.99 ev with pseudopotential calculations , while perdew _
et al._@xcite obtained 6.24 ev with an all - electron calculation . in agreement with the results reported in ref .
, we have found that the bond energy of o@xmath0 is quite sensitive to the details of the pseudopotential .
however the conclusions of the present work are not affected by the pseudopotential approximation , as confirmed by test calculations performed with an harder pseudopotential for o which gives a o@xmath0 bond energy closer to the values reported by hammer.@xcite note that in our calculation the nonspherical density for the oxygen atom is taken into account .
b. hammer , l. b. hansen , and j. k. nrskov , phys . rev .
b * 59 * , 7413 ( 1999 ) .
the coverage is defined as the number of oxygen atoms divided by the number of ag surface atoms in the unit - cell .
although the ag(@xmath1010 ) substrate possesses the mirror - plane symmetry , which is parallel to [ @xmath57 crystal axis and perpendicular to [ 001 ] axis ( i.e. the plane passes through the ag@xmath24 and ag@xmath58 atoms in fig . [
fig : legend ] ) , we have not imposed this symmetry in our calculations . therefore the two r atoms in fig .
[ fig : legend ] are located in slightly asymmetric positions , and in tabs .
vi the distances labeled as @xmath59 , @xmath60 , @xmath61 , and @xmath62 are averaged distances .
b. hammer and j. k. nrskov , in _ chemisorption and reactivity on supported cluster and thin films _
, edited by r. m. lambert and g. pacchioni ( kluwer academic , the netherlands , 1997 ) , pp .
. m. canepa , p. cantini , f. fossa , l. mattera , and s. terreni , phys .
b * 47 * , 15823 ( 1993 ) .
h. katagiri , t. uda , and k. terakura , surf .
* 424 * , 322 ( 1999 ) .
d. sekiba , h. nakamizo , r. ozawa , y. gunji , and h. fukutani , surf .
sci . * 449 * , 111 ( 2000 ) .
f. stietz , a. pantfrder , j. a. schaefer , g. meister , and a. goldmann , surf .
* 318 * , l1201 ( 1994 ) .
d. loffreda , a. dal corso , s. baroni , l. savio , l. vattuone , and m. rocca , surf .
sci . * 530 * , 26 ( 2003 ) .
.chemisorption energies , @xmath9 , and o o distances , @xmath63 , for various chemisorption sites on ag(410 ) surface .
values in parentheses are obtained keeping the substrate fixed .
[ cols="<,<,^,<",options="header " , ] ] , for a and a a configurations ( top views ) .
top panels : isosurfaces at @xmath210.015 e / a@xmath64 ( blue ) .
bottom panels : charge density contours in a plane passing through o , ag@xmath41 and ag@xmath24 atoms .
contours are drawn in linear scale from @xmath65 to @xmath66 e / a@xmath64 , with the increment of @xmath67 e / a@xmath64 . grey ( red ) balls represent silver ( oxygen ) atoms .
, width=340 ] 7.0 , @xmath216.0)ev for the a a configuration ( top view ) .
the isosurface is at 0.015 e / a@xmath64 ( yellow ) and the 10 contours are drawn in linear scale from 0.015 e / a@xmath64 to 0.15 e / a@xmath64.,width=302 ] and @xmath68 of the displacement eigenvector @xmath69 .
observe that the @xmath70 axis of the coordinate system is perpendicular to the terrace ( 100 ) plane while the @xmath71 axis is parallel to the step edge.,width=302 ] and @xmath46 for the a - a and the t@xmath19-t@xmath19 configurations on ag(410 ) .
for the a - a ( t@xmath19-t@xmath19 ) geometry the mode @xmath45 vibrates at 24 mev ( 29 mev ) while the mode @xmath46 at 37 mev ( 32 mev).,width=302 ] | the structure and dynamics of atomic oxygen adsorbed on ag(410 ) and ag(210 ) surfaces has been investigated using density functional theory .
our results show that the adsorption configuration in which o adatoms decorate the upper side of the ( 110 ) steps forming o ag o rows is particularly stable for both surfaces . on ag(210 ) , this arrangement is more stable than other configurations at all the investigated coverages . on ag(410 ) , adsorption on the terrace and at the step edge are almost degenerate , the former being slightly preferred at low coverage while the latter is stabilized by increasing the coverage .
these findings are substantiated by a comparison between the vibrational modes , calculated within density - functional perturbation theory , and the hreel spectrum which has been recently measured in these systems . |
for a family @xmath8 of @xmath9-graphs , let @xmath10 denote the maximum number of edges in an @xmath3-vertex @xmath9-graph which contains no member of @xmath8 . determining @xmath10 is a fundamental question in graph theory which becomes extremely difficult when @xmath11 .
let @xmath12 and call this value the _ turn density _ of @xmath8 ( as has been pointed out many times , it is easy to show that this limit exists ) . when @xmath8 consists of a single graph @xmath1 , we write @xmath13 for @xmath14 .
let @xmath15 denote the complete @xmath0-graph on four vertices .
over 70 years ago , turn famously conjectured that @xmath16 , but this conjecture is still unproved @xcite .
in fact , when @xmath11 there are very few @xmath9-graphs for which the turn density is known ( see @xcite for a detailed account ) . despite this general difficulty , there is a special @xmath0-graph called the _ fano plane _ for which much is known .
the fano plane , denoted @xmath17 , is the projective geometry of dimension @xmath18 over the field with @xmath18 elements ; alternatively , @xmath17 is the @xmath0-graph on seven vertices @xmath19 with the seven edges @xmath5 .
let @xmath20 denote the balanced complete bipartite @xmath0-graph , which is obtained by partitioning a set of @xmath3 vertices into parts of size @xmath21 and @xmath22 and taking as edges all the triples intersecting both parts .
since @xmath20 is @xmath18-colorable and it is easy to see that @xmath4 is not , @xmath20 contains no copy of @xmath4 .
therefore , @xmath23 .
ss @xcite conjectured that this lower bound is asymptotically best possible and hence @xmath24 .
a few decades later , de caen and fredi @xcite proved ss conjecture via a clever use of so - called link graphs .
a few years later , keevash and sudakov @xcite and independently fredi and simonovits @xcite proved the exact counterpart of this result ; that is , @xmath25 for sufficiently large @xmath3 .
let @xmath26 be a @xmath9-graph with vertex set @xmath27
. given any subset @xmath28 , @xmath29 , the _ degree _ of @xmath30 , denoted by @xmath31 , is the number of edges of @xmath26 that contain @xmath30 . for simplicity , when @xmath30 consists of one vertex @xmath32 or two vertices @xmath32 and @xmath33 , we write @xmath34 , and @xmath35 instead of @xmath36 and @xmath37 , respectively . when @xmath38 , we call @xmath35 the _ co - degree _ of @xmath32 and @xmath33 , while the set of vertices @xmath39 such that @xmath40 is called the _ co - neighborhood _ of @xmath41 and will be denoted by @xmath42 . for each integer @xmath43 ,
let @xmath44 .
we call @xmath45 the minimum
_ @xmath46-degree _ of @xmath26 . for a family @xmath8 of @xmath9-graphs ,
let @xmath47 denote the maximum value of @xmath48 in an @xmath3 vertex @xmath9-graph @xmath26 which contains no member of @xmath8 and let @xmath49 .
mubayi and zhao @xcite prove that this limit exists in the case @xmath50 and lo and markstrm @xcite prove that this limit exists for all @xmath51 ( a fact previously sketched by keevash @xcite ) . note that the case @xmath52 just reduces to @xmath14 .
when @xmath38 , we call @xmath53 the _ co - degree density _ of @xmath8 . for general @xmath9-graphs , a simple averaging argument shows that @xmath54 when @xmath55 ( see @xcite section 13.2 ) .
it is also pointed out in @xcite section 13.2 that for any graph @xmath1 , @xmath56 .
the same argument applies to any finite family @xmath57 as well .
[ min - degree ] for a finite family @xmath8 of @xmath9-graphs , @xmath58 .
( sketch ) .
let @xmath59 .
let @xmath60 be any small positive real .
let @xmath3 be sufficiently large as a function of @xmath61 .
let @xmath26 be a @xmath9-graph with @xmath62 . by @xcite proposition 4.2
, @xmath26 contains a subgraph @xmath63 on @xmath64 vertices with @xmath65 . since @xmath66 and @xmath67 as @xmath68 ,
when @xmath3 is large enough , we have @xmath69 .
so @xmath63 contains a member of @xmath57 and therefore @xmath26 contains a member of @xmath57 .
so the minimum degree problem is essentially the same as the turn problem .
the minimum co - degree problem however is drastically different .
for instance , there are @xmath0-graphs @xmath1 with @xmath13 arbitrarily close to @xmath70 and yet @xmath71 ( see @xcite ) . in general
, there has not been a very good understanding of the relationship between @xmath13 and @xmath72 ( see @xcite and @xcite for detailed discussions ) .
similar to the situation with the turn density , not much is known about @xmath72 even for small graphs @xmath1 such as @xmath15 ( in this case czygrinow and nagle @xcite conjectured that @xmath73 ) .
mubayi @xcite initiated the study of @xmath74 , where @xmath4 is the fano plane .
as pointed out earlier , @xmath20 contains no copy of @xmath4 .
so , @xmath75 .
mubayi @xcite proved an asymptotically matching upper bound thus establishing @xmath76 .
he further conjectured that @xmath77 , for sufficiently large @xmath3 .
this was later proved by keevash @xcite using a very sophisticated argument involving hypergraph regularity , quasi - randomness , and stability ( we should mention that keevash proves the stronger statement that the extremal example is stable " . also , the scope of keevash s paper is not limited to the problem of determining the co - degree threshold for the fano plane . ) . in this paper
, we give a simple proof of mubayi s conjecture which is in the same spirit as mubayi s original proof of @xmath76 .
our main result is [ main ] there exists @xmath78 such that if @xmath79 , then @xmath80 .
since we are giving a new proof of an old result , it is worth mentioning that we only need @xmath78 to be large enough so that supersaturation " holds ( see section [ lemmas ] ) .
while we do not make an attempt to compute the value of @xmath78 , it is considerably smaller than the value of @xmath78 needed for the use of regularity in @xcite .
the paper is organized as follows . in section [ lemmas ]
we give some lemmas and introduce a family of @xmath0-graphs called _ rings_. in section [ fano ] we prove theorem [ main ] by making use of the family of rings . in section [ turan ]
we determine the turn density of the family of rings . finally , in section [ remarks
] , we conclude with some remarks and open problems .
for any @xmath9-graph @xmath26 , the _ @xmath81-blowup _ of @xmath26 , denoted @xmath82 , is the graph obtained from @xmath26 by cloning each vertex @xmath81 times . for a family of @xmath9-graphs @xmath8 , let @xmath83 .
erds @xcite used supersaturation to show [ erdos - blowup]@xcite for any finite family of @xmath9-graphs @xmath57 and any positive integer @xmath81 , @xmath84 .
keevash and zhao @xcite proved an analogous result for the co - degree density .
[ codegreeblowup][kz ] for any finite family of @xmath9-graphs @xmath57 and any positive integer @xmath81 , @xmath85 . the same supersaturation argument in fact gives [ blowup ] for any finite family of @xmath9-graphs @xmath57 and any positive integer @xmath81 , and any @xmath86 , @xmath87 , @xmath88 .
we also make the following trivial observation based on the definitions .
[ subgraph ] let @xmath57 and @xmath89 be two families of @xmath9-graphs .
let @xmath90 .
suppose that for every member @xmath91 , some subgraph of @xmath26 belongs to @xmath57 .
then @xmath92 .
so , in particular , @xmath93 .
we now define a family of @xmath0-graphs , called _ rings _ , which will play a central role in our proof of theorem [ main ] .
let @xmath94 and let @xmath27 be a set of at most @xmath95 vertices surjectively labeled with @xmath96 .
let @xmath97 be the family of @xmath0-graphs on @xmath27 with edge set @xmath98 , where addition is defined modulo @xmath99 .
let @xmath100 be the ( unique ) member of @xmath97 which has exactly @xmath95 vertices and call @xmath100 a ring on @xmath95 vertices .
let @xmath101 and @xmath102 .
[ familyblowup ] for all positive integers @xmath94 and @xmath103 , we have @xmath104 and @xmath105 . since @xmath106
, we have @xmath107 . on the other hand , for every @xmath108
, @xmath109 clearly contains a copy of @xmath110 , since in any member of @xmath109 there are @xmath99 distinct copies of @xmath111 . by proposition [ subgraph ] and lemma [ codegreeblowup ] , @xmath112 .
thus , @xmath113 . by a similar argument
, we have @xmath114 . [ lmproperty ] a hypergraph @xmath1 on @xmath115 vertices
is said to have the _
@xmath116-property _ if every subset of @xmath117 vertices contains at least one edge of @xmath1 .
mubayi and rdl @xcite recursively constructed for every @xmath94 a family @xmath118 of @xmath0-graphs with the @xmath119-property .
they showed that @xmath120 for each fixed @xmath94 and used this to establish an upper bound on the turn density of @xmath121 ( sometimes referred to as the @xmath0-book with @xmath0 pages ) .
this family @xmath118 also played a key role in mubayi s proof of @xmath122 . here , we observe that for every @xmath99 the graph @xmath100 has the @xmath123-property and we will also show that @xmath124 is small .
then , by using @xmath125 instead of @xmath126 we are able to establish @xmath127 .
[ edge ] @xmath100 has the @xmath123-property .
clearly @xmath100 has @xmath95 vertices .
let @xmath128 be any set of vertices in @xmath100 that contains no edge .
we show that @xmath129 . for each @xmath130 , if @xmath131 then @xmath132 ( addition modulo @xmath99 ) otherwise we would have an edge .
this implies @xmath129 .
next , we show that @xmath124 is small by using an auxiliary directed graph .
first we recall some old results concerning short directed cycles in directed graphs . as usual , for a directed graph @xmath133 , let @xmath134 and @xmath135 denote the minimum out - degree and in - degree of @xmath133 respectively .
caccetta and hggkvist @xcite conjectured that if @xmath133 is a directed graph on @xmath3 vertices with @xmath136 , then @xmath133 contains a cycle of length at most @xmath137 .
while their conjecture remains open , chvtal and szemerdi @xcite gave a simple proof of a slightly weaker statement .
[ chsz ] let @xmath133 be a directed graph on @xmath3 vertices . if @xmath136 ( or @xmath138 ) , then @xmath133 contains a directed cycle of length at most @xmath139 .
there have been improvements on this result
. however , theorem [ chsz ] suffices for our purposes .
[ ring - codegree ] for all @xmath94 we have @xmath140 . by lemma [ familyblowup
] , it suffices to prove that @xmath141 .
let @xmath142 .
let @xmath143 be a small positive real and let @xmath144 .
let @xmath3 be sufficiently large as a function of @xmath143 .
let @xmath26 be a @xmath0-graph on @xmath3 vertices with @xmath145 .
let @xmath133 be an auxiliary digraph with vertex set @xmath146 such that @xmath147 is an edge of @xmath133 if and only if @xmath148 and @xmath149 are edges of @xmath26 ( in other words , if and only if @xmath150 ) .
let @xmath151 .
then @xmath133 has @xmath152 vertices . for any @xmath153
, its out - neighbors in @xmath133 are precisely all the @xmath18-subsets of @xmath154 and thus ( using @xmath3 being sufficiently large ) @xmath155 by theorem [ chsz ] , @xmath133 contains a directed cycle @xmath156 of length at most @xmath157 .
the subgraph of @xmath26 corresponding to @xmath156 is a member of @xmath158 .
let @xmath159 be the @xmath0-graph obtained from the complete @xmath0-partite @xmath0-graph with vertex set @xmath160 by adding the vertex @xmath161 and the three edges @xmath162 . notice that @xmath163 .
we obtain theorem [ main ] as a corollary of the following more general theorem .
[ f * ] for sufficiently large @xmath3 , @xmath164 for all @xmath165 . in the introduction we pointed out that @xmath20 gives the lower bound @xmath166 ,
thus it suffices to prove @xmath167 . by theorem [ ring - codegree ] and lemma [ codegreeblowup ] , @xmath168 .
let @xmath3 be large enough such that @xmath169 .
let @xmath26 be a graph on @xmath3 vertices with @xmath170 .
then @xmath26 contains a copy @xmath1 of @xmath171 for some @xmath172 . for each vertex @xmath173 in @xmath100
, let @xmath174 denote the clone of @xmath173 in @xmath171 .
for all @xmath175 , let @xmath176 .
summing over all @xmath175 and using the exact condition @xmath170 ( the only place where the exact condition is needed ) , gives @xmath177 this implies that there exists some @xmath178 which is contained in more than @xmath99 different sets @xmath179 .
therefore , by lemma [ edge ] , there are vertices @xmath180 such that @xmath181 is an edge in @xmath100 with @xmath182 , @xmath183 and @xmath184 .
so in @xmath171 , @xmath185 induces a complete @xmath0-partite @xmath0-graph and thus @xmath186 ( see figure [ fplus ] ) .
let @xmath187 denote the family @xmath188 .
the fact that @xmath100 has the @xmath189-property and the family @xmath125 has small co - degree density was key to our short proof of mubayi s conjecture .
conceivably , the family @xmath125 can be useful elsewhere in the study of the turn problem for @xmath0-graphs .
for instance , if @xmath125 also has relatively small turn density , then it could potentially be used in bounding the turn densities of other @xmath0-graphs , just like how @xmath126 was used by mubayi and rdl @xcite . in this section
, we show that similar to @xmath126 the family @xmath187 also has turn density at most @xmath190 .
in fact , we will show that the turn density of @xmath191 is exactly @xmath190 .
the family @xmath187 does , however , have some advantages over @xmath126 .
one , it has the @xmath123-property versus @xmath126 having the @xmath192-property .
two , the structure of @xmath100 is simple and explicit , while in forcing a member of @xmath126 , we do not quite know which particular structure that member has .
next , we show that @xmath187 has turn density at least @xmath190 via a construction inspired by the `` half - graph '' constructions from bandwidth problems .
[ ring - lower - construction ] let @xmath193 and @xmath194 .
let @xmath195 be a @xmath0-graph on @xmath196 whose edges are all the triples of the form @xmath197 and @xmath198 where @xmath199 .
it is easy to check that @xmath200 .
[ ring - lower ] for all @xmath3 the graph @xmath195 given in example [ ring - lower - construction ] contains no member of @xmath191 and hence @xmath201 .
observe first that , based on the definition of @xmath195 , for any @xmath202 with @xmath203 , the pair @xmath204 has no co - neighbor in @xmath205 and the pair @xmath206 has no co - neighbor in @xmath207 .
suppose for a contradiction that @xmath26 contains a copy @xmath1 of @xmath100 , for some @xmath99 .
suppose @xmath208 @xmath209 and @xmath210 .
for each @xmath173 in @xmath100 , let @xmath174 denote its image in @xmath195 under a fixed isomorphism from @xmath100 to @xmath1 .
for any @xmath211 in @xmath212 ( or @xmath213 ) , let @xmath214 denote its subscript in @xmath212 ( or @xmath213 ) . in other words , if @xmath215 , then @xmath216 .
there are two cases to consider .
_ for some @xmath217 , @xmath218 and @xmath219 are in the same set . without loss of generality , we may assume that @xmath220 and that @xmath221 are both in @xmath212
. then @xmath222 must both be in @xmath213 .
furthermore , by the observation we made at the beginning of this proof , @xmath223 . by repeating this argument
, we get @xmath224 , which is a contradiction .
_ for all @xmath217 , @xmath218 and @xmath219 are in different sets . by the symmetry of @xmath100 , we may assume that all the @xmath218 s are in @xmath212 and all the @xmath219 s are in @xmath213 .
based on the observation we made at the beginning of the proof , we now must have @xmath225 for all @xmath226 ( with addition defined modulo @xmath99 ) .
this leads to a contradiction like in case 1 .
we now prove the main result of this section .
this follows immediately from the following lemma .
given a @xmath0-graph @xmath26 and a vertex @xmath32 , the _ link graph _
@xmath227 of @xmath32 is a @xmath18-graph whose edges are all the pairs @xmath228 such that @xmath229 .
[ ring - upper ] @xmath230 . by proposition [ min - degree ] and lemma [ familyblowup ]
, it suffices to prove @xmath231 .
let @xmath3 be sufficiently large as a function of @xmath99 .
let @xmath26 be a @xmath0-graph on @xmath3 vertices with @xmath232 .
we prove that @xmath26 contains a member of @xmath158 .
create an auxiliary digraph with vertex set @xmath146 ( all @xmath18-subsets of @xmath233 ) where @xmath147 is an edge of @xmath133 if and only if @xmath148 and @xmath149 are edges of @xmath26 ( in other words , if and only if @xmath234 is in the link graph of both @xmath235 and @xmath174 ) .
let @xmath151 .
let @xmath236 be a vertex in @xmath133 .
since @xmath237 , the link graph of @xmath161 has at least @xmath238 edges and the link graph of @xmath173 has at least @xmath238 edges .
therefore there are at least @xmath239 edges in the intersection of their link graphs , which implies @xmath240 .
so we can apply theorem [ chsz ] to the directed graph @xmath133 to obtain a directed cycle @xmath156 of length at most @xmath241 .
notice that the directed cycle @xmath156 corresponds to a subgraph of @xmath26 which is a member of @xmath158 .
proposition [ ring - lower ] and lemma [ ring - upper ] now yield @xmath242 .
we have now determined the turn density of the entire family of rings .
however , computing its value for any single member @xmath100 appears to be difficult .
after all , @xmath243 is just @xmath15 and determining @xmath244 has been notoriously difficult . a quick observation that one can make is [ evenupper ] for any positive integers @xmath245 we have @xmath246 .
thus , for all even @xmath99 , we have @xmath247 .
since @xmath248 is contained in the @xmath249-blowup of @xmath250 , we have @xmath251 .
now suppose @xmath99 is even .
since @xmath252 , we have @xmath253 .
recall that the conjectured value for @xmath244 is @xmath254 .
for the lower bound , turn s construction @xmath255 is obtained by partitioning @xmath3 vertices as equally as possible into three sets @xmath256 and including as edges all triples of the form , @xmath257 , @xmath258 , @xmath259 , @xmath260 for all @xmath261 ( see figure [ turanexample ] ) .
it is straightforward to check that if @xmath255 contains @xmath100 for some @xmath99 , then @xmath99 must be divisible by @xmath0 .
hence , @xmath255 contains no @xmath100 when @xmath262 .
so we have the following .
[ evenlower ] for @xmath263 , @xmath264 .
so by propositions [ evenupper ] and [ evenlower ] , if turn s conjecture is true , then we would have @xmath265 for every even @xmath99 with @xmath262 .
[ bothexamples ] finally , for odd @xmath99 , the following construction shows that @xmath266 is larger than @xmath267 .
let @xmath268 be a @xmath0-graph on @xmath3 vertices where the vertices are partitioned into three sets @xmath256 with sizes @xmath269 , @xmath270 whose edges are all triples of the form , @xmath271 , @xmath257 for all @xmath261 and @xmath272 ( see figure [ oddexample ] ) .
it is easy to check that @xmath273 and that if @xmath268 contains @xmath100 then @xmath99 must be even .
furthermore , we can iterate this construction inside @xmath274 and @xmath275 to push the density above @xmath267 while maintaining the fact that there are no odd rings .
suppose @xmath276 , then the density of @xmath268 before iterating is @xmath277 ; this gives an optimal value of @xmath278 when @xmath279 .
after iterating , the density becomes @xmath280 ; numerical methods give an approximate optimal value of @xmath281 when @xmath282 .
thus we have the following . [ oddlower ] for odd @xmath99 , @xmath283 .
let @xmath284 be obtained by adding the edges @xmath285 to @xmath286 .
note the following simple observation .
[ q3 ] for odd @xmath99 at least @xmath287 , @xmath100 is contained in the blow - up of @xmath284 .
the final results in this section are obtained by using razborov s flag algebra calculus . since the upper bounds are not tight ( and we do nt intend to formally publish the bounds obtained from these calculations ) , we refer the reader to @xcite , @xcite , @xcite for an explanation of the method and its applications .
[ uppers ] 1 . for even @xmath99 , @xmath288 .
2 . for odd @xmath289 ,
@xmath290 3 . for @xmath99 an odd multiple of @xmath0 , @xmath291 .
4 . for @xmath99 an even multiple of @xmath0 , @xmath292 . in each case we use lemma [ codegreeblowup ] to transfer a statement about the blow - up of a graph to a statement about @xmath100 . 1 .
proposition [ evenupper ] shows @xmath100 is contained in the blow - up of @xmath15 and flag algebra calculations give @xmath293 ( see @xcite ) .
2 . proposition [ q3 ] shows @xmath100 is contained in the blow - up of @xmath284 and flag algebra calculations give @xmath294 .
3 . @xmath100 is contained in the blow - up of @xmath286 and flag algebra calculations give @xmath295 4 .
@xmath100 is contained in the blow - up of @xmath243 and @xmath286 and flag algebra calculations give @xmath296 the results of this section are summarized below , with the lower bounds coming from propositions [ ring - lower ] , [ evenlower ] , [ oddlower ] and the upper bounds coming from proposition [ uppers ] .
perhaps the most interesting thing about the upper bounds for rings is that it is possible to get nearly tight results for every value of @xmath99 using only flag algebra calculations for small @xmath0-graphs and lemma [ codegreeblowup ] .
given the results of this section and theorem [ ring - codegree ] , it would be interesting to solve the following problem .
determine @xmath266 or @xmath298 for each fixed value of @xmath99 .
let @xmath249 be a prime power and let @xmath299 be @xmath300-graph with vertex set equal to the one dimensional subspaces of @xmath301 and edges corresponding to the two - dimensional subspaces of @xmath301 .
we call @xmath299 the _ projective geometry _ of dimension @xmath18 over @xmath302 ; note that @xmath303 is the fano plane . in @xcite , keevash also proved the following more general theorem about projective geometries furthermore , there is a nearly matching lower bound when @xmath249 is an odd prime power ( see @xcite and @xcite ) .
the proof we present in section [ fano ] relies on the fact that there is a family of @xmath0-graphs @xmath305 , such that each member @xmath306 has the @xmath307-property and @xmath308 .
in fact , our same proof could be used to give a simple proof of theorem [ keevashpg ] if there was an affirmative answer to the following question .
[ lm ] let @xmath309 .
does there exist a finite family @xmath310 of @xmath9-graphs such that for each @xmath311 there exists a positive integer @xmath99 such that @xmath312 has the @xmath189-property and @xmath313 .
it seems conceivable that obtaining a @xmath9-graph with the @xmath116 property for some other values of @xmath115 and @xmath117 might give us the same benefit and be easier to obtain ; however , this is not the case . on one hand
, we must have @xmath314 so that equation holds . on the other hand
, when @xmath315 the complete balanced bipartite @xmath9-graph @xmath316 , which has @xmath317 , does not contain any subgraph with the @xmath116-property ( any subgraph of @xmath316 with @xmath115 vertices must contain an independent set of size @xmath318 ) .
so we must have @xmath319 , which implies that @xmath115 is even and @xmath320 .
a different problem is the following .
instead of determining the co - degree threshold for a single copy of @xmath1 in @xmath26 , one can ask about the co - degree threshold for @xmath321 vertex disjoint copies of @xmath1 in @xmath26 ( assuming @xmath322 divides @xmath323 ) .
this has been referred to as the _ tiling _ or _ factoring _ problem and received much attention lately .
interestingly , the co - degree threshold for tiling with @xmath15 and @xmath324 have been determined ( see @xcite , @xcite ) , but the co - degree threshold for a single copy of @xmath15 or @xmath324 is still unknown and appears to be difficult . since the co - degree threshold for a single copy of @xmath4 is known and seems to be much easier than @xmath15 or @xmath324 , it would be interesting to determine the co - degree threshold for tiling with @xmath4 .
let @xmath3 be divisible by @xmath325 and let @xmath26 be a @xmath0-graph on @xmath3 vertices . determine the minimum value @xmath326 such that @xmath327 implies that @xmath26 contains @xmath328 vertex disjoint copies of @xmath4 .
the relationship between the edge density of a hypergraph and its subgraphs with large co - degree is also very intriguing .
even the following simple questions do not seem to have an easy answer .
a @xmath9-graph @xmath1 is said to _ cover pairs _ if @xmath1 has at least @xmath329 vertices and every pair of vertices lies in some edge , i.e. @xmath330 . l. caccetta and r. hggkvist , on minimal digraphs with given girth , proceedings of the ninth southeastern conference on combinatorics , graph theory , and computing ( florida atlantic univ . ,
boca raton , fla . , 1978 ) , pp .
181187 , congress .
xxi , utilitas math . ,
winnipeg , man . , 1978 .
v. ss , remarks on the connection of graph theory , finite geometry and block designs , colloquio internazionale sulle teorie combinatorie ( roma , 1973 ) , tomo ii , pp . 223233 .
atti dei convegni lincei , no .
17 , accad .
lincei , rome , 1976 . | given a @xmath0-graph @xmath1 , let @xmath2 denote the maximum value of the minimum co - degree of a @xmath0-graph on @xmath3 vertices which does not contain a copy of @xmath1 .
let @xmath4 denote the fano plane , which is the @xmath0-graph @xmath5 .
mubayi @xcite proved that @xmath6 and conjectured that @xmath7 for sufficiently large @xmath3 . using a very sophisticated quasi - randomness argument , keevash @xcite proved mubayi s conjecture .
here we give a simple proof of mubayi s conjecture by using a class of @xmath0-graphs that we call rings .
we also determine the turn density of the family of rings . |
Subsets and Splits